Replicator 2X right extruder won't reach 230 heating issue.

[Scott D]

I read these post about the 2X being a work horse and then I buy one and we have not been able to get the right extruder tool 0 working since I've had it. I think some if not many changes have happened since it was fist released.
At first the tech said they are aware of it and working on a solution. Basically after asking question it was either sending parts or possible firmware tweak.
I called back after a week and the new tech had me remove the extruder stepper motor. For sure after looking at how kinked and pinched the coupler wire was I thought we had identified the issue. It reached 226 and maybe 227 but never 228 to allow the filament loading script to complete or even be able to print. So they sent me a new coupler. After replacing it I had the same problem. I have a 2X that is basically a R2 for ABS.
Then on Friday they decided to replace the whole told bar. But it bugged me that no real trouble shooting was going on. So over the weekend I swapped the cable harness from tool 1 that worked to tool 0 and and even put the heating element from tool 1 to tool 0. For sure this would tell me what was wrong but to my surprise it had the same issue. So new coupler, different heating elements, new wire harnes and still not working.
So then I thought it could be the board, maybe control A had an issue on the board which control A was controlling the right tool 0. So I moved the wire harnes to control B and moved the coupler from A to B. Now for sure this has to work. As expected the firmware had to be told to heat up the left extruder but was realy hooked up to the right......and still the same issue. So now I thought it had to be something with that heater block and the parts in the mail would fix it.

So today I received the new parts. The heating elements, nozzels, filament tube ceramic and couplers already put togetheron the block. I thought great, happy B-day to me (and yes it's realy my B-Day) so I removed the old and on with the new. I thought for sure this would fix the issue especially after all the troubleshooting I did over the weekend. Boy was I wrong. Not only was I wrong but I have never been so baffled as a engineer in my life. This whole thing makes no sence but even after replacing the whole left and right parts with block it still doesn't work.

What more annoying is they don't want to admit or tell me any more information about the part they suspect it bad that the first tech mentioned. They don't deny anything, but just change the subject with "let's replace this part next, I'll send it out overnight".

The only two parts left are the PSU and the motherboard. And now days the boards are clipped in and not held with screws so I don't have the tool to remove it even if they sent me that part next.

I'm now in that window of the last 3 days to return it to the store I bought it. I think I have been more than patient but also do t want to go back to researching machines before I pick the next one I want. I spent almost a year researching and reading up on these machine be fore makerbot got bought out.

So do I return it for an exchange? Do I try the new FF Creator X since I prefer the metal frame, or do I go with the ultimaker 2 that is out of stock every time I check. These were my top three picks. The store I bight it from demos the R2X for 10 to 12 hours a day and has for the last three months with hundreds of hours on it. But this was one of the first ones and not the ones made recently. Mine had a December 7th 2013 date on the box.

Any ideas and suggestions are welcome.

[Dan Newman]

Can you tell from your motherboard if it is a Rev G or Rev H ?
I've never actually seen a Rev H board myself. I know that changes
were made for FCC compliance certification. And further there was
a change made which required the SD card data read rate be dropped
from 8 down to 1 MHz. I cannot imagine what on the board could make
this happen as you've swapped everything between A and B leaving just
the heater block and side of the heat spreader bar the same. And now
it sounds like you swapped out the heater block as well. Or did
they send you an entire, complete new two nozzle set? Complete
as in nozzles, heater blocks, barrier tubes, heat spreader bar
holding both the barrier tubes, etc.

Dan

[Dave B]

Hi Dan,
You know this situation and trouble shooting steps I have taken more than anyone since we have been emailing directly. It's a Rev H and I have even sent you pictures since it had so many filters on it. I still don't know if it received FCC approval but I have not taken the time to research this.
I tried to post a picture of exactly what was replaced but now it's waiting for approval. I will just email you the picture directly. Since I have posted this topic I have reset default, reset mother board completely, and re flashed back to MBI firmware away from sailfish. I'm never thought it was sailfish but was just making sure all was back to default and then represent again after moving back to MBI firmware.

In fact since I have my most expensive electronic plugged into APC surge and battery back ups I even tried plugging directly into the wall outlet on a different breaker just to rule out a noisy line causing interference but the issue was still happening. It's a old school move, but I had to be sure.

[Scottbee]

Just an FYI, it doesn't take any special tools to remove the MightyBoard. Those are "tangerine pins" and the board will pop right off.

Makerbot sent me a new board when the right extruder fan FET died.

[Scottbee]

MBI has an interesting policy....

When I needed a replacement ceramic insulator for my right extruder they sent me the entire shebang.... Spreader bar, both thermal tubes, both heater blocks, heating elements, thermocouples, nozzles.... the entire assembly. At that time they told me that this was the minimum FRU (field replaceable unit).

I'm not complaining...... not at all!

[Scott D]

Hi Dan, it figures the one who is the most familiar with my situation would reply to my post first. After all we've been emailing directly about this for the last two weeks. In fact I even emailed you pictures of the board because I was shocked at how many filters were on the system from the rev H design. And yes thus is a rev H.

Since I posted this I've done a reset to default, reset motherboard, remote saifish and went back to MBI firmware, and reset to default after moving back to MBI firmware. . Not that I've ever suspected sailish to be the cause but I had to rule out any possibility and make sure that EPROM WAS completely reset and cleared.

I just happen to have all my expensive electronics on APC surge and battery backup 1500 protectors. Just to rule out the chance of the dirty line I've removed it from the APC and plug it directly into the wall, I've even moved it to the other side my house ensuring that I was on a separate breaker off my panel. I know this is old-school, but I had to rule that out also.

[Scott D]

Scottbee, that's good to know. I guess my biggest fear at this time is TIME.
I have maybe another 3 to 5 days before my 30 return runs up and don't want to be stuck with a $2,800 machine that on lye does half of what I bought it for.

But on the other hand MBI has been great to work with. I can't say night about their customer service and to me that says a lot about a company and has my respect.

If it's an issue with Rev H I hope they work it out. If I just got a bad one I hope they might continue to work with me and maybe get a new board and PSU to me before I am screwed out of the option of returning it. But it's a lot of money riding on hopes.

Oh, and I realy wanted this to make my birth day a good day by these parts working so some disappointment I guess.

[Scott D]

Yes, that also surprised me and was above and beyond the normal customer service I would expect. They explained that if I had just replaced the heat block it had a chance of leaking if it was not put back together tight enought so they decided to replace the whole thing.
Even more of a shock was when I received an email later and she went on to explain they are sending both sides tool 0&1 with the block and couplers all assembled.
But like you, I was not complaining.

[Dan Newman]

If you only saw this with MBI's firmware, then I would suspect an
"odd" feature they have in their firmware which is undocumented and
can cause this problem. But since you saw it with Sailfish as well,
that issue is ruled out.

What am I talking about? MBI put in a constant offset which can
be added to the measured temp for a sensor. They called it a
calibration. By default, the value is supposed to be 0 in EEPROM
so adding it in has no effect. But if you had a bad value in EEPROM,
then it would throw the reading for that sensor. It's a bit silly
to use just an additive constant for type K thermocouples. They
are quite linear in the range employed here so you would want both
an additive constant AND a multiplier. Then you'd measure at two
reference temps (e.g., ice water and boiling water) and from that
determine the two calibration values.

When we saw that MBI had put that into their firmware, we did not
follow suit: having just the additive value is pointless. And,
indeed, calibration is pretty pointless here anyway since you do
not need a dead correct value. You just need a proxy value: when
the sensor reads X then that's the right temp for my extruder.

Dan

Ref: Lines 111, 142, and 303 of Heater.cc,

https://github.com/makerbot/MightyBoardFirmware/blob/master/firmware/src/MightyBoard/shared/Heater.cc

[Dan Newman]

Not also that "reset to default" does not completely clear the EEPROM.
To do that, you need to "reset motherboard completely". At least that's
what it's call in RepG. I don't know that MakerWare has that function.
When you use that, it wipes all the EEPROM. The reset to factory defaults
doesn't touch some values -- offets and calibration values.

So, you can jot down the home positions/offsets and toolhead offsets since
they will get nuked. (Well, take a screen snap rather than writing them
down.) Then do a "reset motherboard completely". Then see what happens.

Dan

[Ryan Carlyle]

On Tuesday, February 25, 2014 8:09:31 PM UTC-6, Scott D wrote:

I read these post about the 2X being a work horse and then I buy one and we have not been able to get the right extruder tool 0 working since I've had it. I think some if not many changes have happened since it was fist released. 
At first the tech said they are aware of it and working on a solution. Basically after asking question it was either sending parts or possible firmware tweak. 
I called back after a week and the new tech had me remove the extruder stepper motor. For sure after looking at how kinked and pinched the coupler wire was I thought we had identified the issue. It reached 226 and maybe 227 but never 228 to allow the filament loading script to complete or even be able to print. So they sent me a new coupler. After replacing it I had the same problem. I have a 2X that is basically a R2 for ABS. 

If I understand your issue, my R2x does the same thing, but it doesn't actually keep me from printing. It usually just takes 2-3x longer to preheat the right extruder than the left. It hangs at ~227C for a long time (sometimes 10 minutes) before reaching 230C. It takes longer for the right extruder to reach temp if the left extruder is already hot, so I think it's a problem with the power draw management code in the firmware. That's the only function I can think of that would cause interaction between the two heater circuits.

It's annoying when I try to change filaments on a cold machine. But otherwise I always preheat for a while before starting a print (to warm up the chamber) and both extruders always gets to temp eventually.

Things to try:

• Let it preheat for up to 30 minutes before you give up on it. In theory, the PID control will notice the steady temperature discrepancy and slowly ramp up the heater output over time to correct the offset. I haven't put much effort into interpreting the default PID tuning parameters, but it should take several minutes at minimum for the integral correction to become noticable. 
• Lower your preheat temp (in firmware via LCD display) and build temp (in custom profiles) to 227 to see if it can reach that temp easier. This will indicate if there's a specific physical issue that keeps it from hitting 230C, or if it's a control issue.

Oddly, my right extruder also cools down significantly faster when the heaters are cut off. Which would suggest a physical difference between the two hot ends. But they look identical as far as I can see. And both thermocouples are spot-on against a reference TC. Not sure what that means -- maybe a small machining or assembly difference between the two sides.

[Dan Newman]

On 25/02/2014, 6:44 PM, Scott D wrote:

> Hi Dan, it figures the one who is the most familiar with my situation would reply to my post first. After all we've been emailing directly about this for the last two weeks. In fact I even emailed you pictures of the board because I was shocked at how many filters were on the system from the rev H design. And yes thus is a rev H.

Oh, and on the rev G mightyboard, if you left the USB cable plugged in to the bot but
not attached to a computer, then you could get spurious a/d reads for the thermocouples.
I've seen that as well as others. It's a shared ground between the USB and the rest
of the motherboard and so that antenna -- usb cable -- can be dumping noise onto the
a/d ref. ground and throw readings. But, this should have been ruled out when you
did the A/B swap.

Dan

[Scott D]

Agreed, why would they add this in the firmware? Even if they have started calibrating before they leave the factory I would have no way to change that value in the EPROM. But I would hope a reset mother board (not to be confused with defaults) would have cleared that value if they are doing an offset from zero. Or maybe not, and if that's the case I have no way of confirming or changing this with MBI firmware. But it's null anyway since it was the same issue with sailfish.

[Ryan Carlyle]

Dan,

For the offset "correction" factor, would it display a non-adjusted number on the LCD while using the adjusted number in the heater control? Seems like a spurious offset value in the EEPROM would just induce temp error, not a failure to reach the setpoint.

[Scott D]

Yup, I did the rest mother board, not just the default with RepG after taking my offsets and doing an EPROM backup since I would have planned on going back to sailfish once it's fixed.

[Dan Newman]

On 25/02/2014, 7:22 PM, AL M wrote:
>
>
> <https://lh3.googleusercontent.com/-rX8Kzky85_M/Uw1djeekn5I/AAAAAAAAAIo/NOnQaUVajJY/s1600/DSC_8558.JPG>
> This is a Rev.H from a 2X if you want more pics i would be glad to take more

Thanks, but I've seen pics. When I write that I haven't seen a rev H, I mean I haven't
had one to hold in my hands and study all the components and traces. BTW, even better
than snapping a picture is tossing the board on a scanner. But don't post such pics --
they are too good in most cases and serve a good input to software used in reverse engineering
boards. When MBI sent me a Rep 2 (#18) and I mentioned it on this list I received within a day an
e-mail from someone I'd never heard of and whose gmail address didn't turn up in
searches. They politely asked if I could scan the board and send them pictures since
they wanted to compare it to their Rep 1 and see if it was worth them upgrading to a
Rep 2. Yeah, right. I never responded. (And if you're still out there, now you know
why: I didn't believe you and even if I did, I wouldn't have sent you scans or photos.)

Dan

[Scott D]

Ryan,
I tied adjusting the preheat setting, it again get hung 3 degrees before what ever I set it to.
The longest I have let it sit was 20 minutes. But I also ip understand that after 15 it has code that if it does not reach temp it will cut off to protect the machine from further damage. So if this code activates would this not also interfeard with the PID offset ramping up?

And yes, sounds like you have a recent machine? My right also cools down way faster. I didn't interpret this as the right and being a different hardware, but I kind of you to it as the right never really reaching 227 as it stated she cooldown so much quicker.

Now you gave me an idea since they replaced the whole bar. I can take the left off the old bar and put this on the right with the new bar installed. If it shows the same issue then the left a right are identicle and the problem lis something else.

[Dan Newman]

> If I understand your issue, my R2x does the same thing, but it doesn't
> actually keep me from printing. It usually just takes 2-3x longer to
> preheat the right extruder than the left. It hangs at ~227C for a long time
> (sometimes 10 minutes) before reaching 230C. It takes longer for the right
> extruder to reach temp if the left extruder is already hot, so I think it's
> a problem with the power draw management code in the firmware.

What there is is minimal: if either or both extruders are heating up
and a command is then received to heat the build platform, then "pause"
the extruders and allow the platform to heat up. Once the platform
hits its target, then "unpause" the extruders and allow them to heat
up. "Pausing" means seeing what the current temp is, then telling that
heater to hold there -- use the current temp as a temporary set point.
Unpausing is then restoring the prior set point.

There's no other power management in the firmware. The gcode does a little
by dropping the stepper motor currents and later raising them back up.

So, if one heater takes a lot longer, I would suspect an issue with
that heater core; i.e., it has higher resistance than its sibling.

But if this were the case, then the A/B swap should have seen it.
(Not stated is that the issue stayed with the errant extruder POSITION
even when the heater cores were swapped. Thus it seemed to be an issue
with something not swapped such as the heater block itself.)

> Oddly, my right extruder also cools down significantly faster when the
> heaters are cut off. Which would suggest a physical difference between the
> two hot ends. But they look identical as far as I can see. And both
> thermocouples are spot-on against a reference TC. Not sure what that means
> -- maybe a small machining or assembly difference between the two sides.

That's quite interesting. I wonder if the heater blocks are machined
in such a way that there's significant differences in the depth of the
tapped hole for mounting the t/c in its little brass holder? Or something
similar such that over time they each stabilize to the same temp,
but one takes longer to do so and then drops faster as well? (Such as
if it's partially air gapped while the other one isn't.)

Dan

[Dan Newman]

On 25/02/2014, 7:26 PM, Ryan Carlyle wrote:
> Dan,
>
> For the offset "correction" factor, would it display a non-adjusted number
> on the LCD while using the adjusted number in the heater control?

Yes, that was one of the odd things about how it was implemented. It's
only used by the heater management: gets temp, then adds offset, then
does management. The LCD display just gets temp and then displays it.
No use of the offset.

> Seems like a spurious offset value in the EEPROM would just induce temp error,
> not a failure to reach the setpoint.

I don't know how high of a temp a R2X extruder can reach. But if there
was a spurious value making it try to hit, say, 295C then that wouldn't
trigger an error. There is code which tries to detect if the heater
just isn't heating up enough and then throw an error. But I don't know
that that code actually works. Sailfish has the same code that MBI has -- it
inherited it -- but it's always looked a little suspect to me and I've
never actually tested it myself.

Dan

[Joseph Adams]

I had this problem with my bot after upgrading my filament guide tubes so that I could print PLA reliably. My problem seemed to be that I was getting too much heat transfer through the guide tubes given the factory PID settings. I adjusted the PID and now it's close to normal again (takes about 2.5 minutes to reach 230, my rep2 takes about 1.5 minutes). The trade off however is that I didn't spend the time to optimize the settings, so I get a little bit of oscillation now of about 2-3 degrees. It's acceptable for me for now. The bot is at my office so I can't tell you the exact values I used, I will post in them in the morning. In the meantime as a workaround, since you are using sailfish, you can set the target temp to overshoot by 5 degrees (either in the slicer or your preheat settings) and then when the bot reaches your intended temp, manually lower the target temp. This will allow the build to start. For example, if you want to build at 230, set slicer target temp to 235, now once the bot gets to 230 manually lower the target temp to 230 through the machine's menu and the build will start. Alternatively you could just preheat the machine to 240 and by the time the build starts it will already be at 230. I know these are not solutions but it is a way to get builds started.

[Ryan Carlyle]

I've strapped a TC to the heater block where the nozzle snugs up to it (up inside the ceramic insulation) and the LCD temps were spot-on to the real temp. So I don't think it's something like an air gap. The offset parameter causing the PID control to think 228=230 would actually explain everything except the faster cooldown time.

If I remember correctly, my Kill-A-Watt said both heater cores pulled the same wattage (when on) while I was screwing around with it. That would rule out different resistance. I wasn't paying much attention though.

[Ryan Carlyle]

This is smart. For stock firmware, you would just set your preheat setting 3-5C higher than your intended build setting. Instead of preheating up to temp, it should cool down to temp and then start. But there is a risk that it will pause to let something else heat up / cool down, and the troublesome extruder will cool past the desired setpoint.

[Dan Newman]

On 25/02/2014, 8:40 PM, Ryan Carlyle wrote:
> I've strapped a TC to the heater block where the nozzle snugs up to it (up
> inside the ceramic insulation) and the LCD temps were spot-on to the real
> temp. So I don't think it's something like an air gap. The offset parameter
> causing the PID control to think 228=230 would actually explain everything
> except the faster cooldown time.

Faster cooldown indicates that the right is dissipating heat faster.
The same cause could also be causing it to take longer to heat up. But,
backed into the far right corner after a print is done would, all other
things equal, cause it too take a tad longer to cooldown than the left
extruder, what being in closer proximity to the warm X stepper and,
post homing, closer to the toasty Y stepper as well.

Any chance the extruder heatsink fans are somehow creating an air
wash that impacts the uninsulated bottom face of the right extruder's
heater block? (The early Rep 2's at least followed the design of
the Rep 1's and have the heater blocks insulated on all six sides.
Took longer to assemble that way and hence the move to the slide-on
insulation which doesn't insulate all faces of the heater block.)

Or are there some asymmetries that would cause the right to cool faster?
I don't have an R2X myself. My Rep 1 DUAL doesn't exhibit such behavior,
but it's heat spreader bar and heater block insulation is very different
from the R2X.

Dan

[Ryan Carlyle]

When I was testing the cooldown rate, I had the carriage centered over a cool build plate. The system is designed for minimal air wash for ABS, so I can't imagine that's an issue.

I can't think of any obvious asymmetries. The only area that comes to mind is the non-threaded thermal barrier interface... Maybe the cooling block inside radius is different or something. That would significantly affect the contact area while not being obvious.

I have a diagnostic test in mind, but it's time for bed so I can't do it tonight. If the problem is asymmetrical cooling and the right extruder is unable to overcome the heat loss, raising the PID Kp and/or lowering Ki should allow it to reach temperature by upping the heater duty cycle. If the problem is caused by a phantom firmware TC offset, adjusting the PID constants will have no significant effect because the firmware thinks it's already on target.

[Scott D]

Hello Ryan and all,
I am grateful of all the knowledge and thinking outside the box style of diagnostics and suggestions provided in this thread. As an engineer myself sometime I know I need to walk away and let another set of eyes look at the issue since sometime you may miss the smallest thing. 3D printing being new to me is one of those times and I am surprised at all the great advice and fast replies I received.
But I'm not sure if anyone should spin any more cycles on this issue. I am currently on the phone again with MBI and they have admitted to me that this is now a know recognized intermittent issue with only some of the R2X units. I am being told the software/firmware team is working on a fix but current no ETA.

I am left with two options, send back the mighty board and extruders so they can repair what ever is casing the issue, or return the unit. I did ask if they had a Rev G board I stock and after checking he confirmed they dont.

My only concerns are from when I swapped the wire harnes and coupler to control B on the board and it still didn't work. That would tell me it's not a board issue and has to be something with the extruder. So I'm not sure how the firmware team plans to correct this.

And as far as I know sailfish is it's own firmware, so why did the issue still happen with sailfish unless the main hardware code is the same between both firmwares?

Well three weeks in, a lot of time troubleshooting and replacing parts with no guarantee of a fix or at least an ETA. It sounds like it started 3 weeks ago when they started seeing a patern. Now they need to be sure on identifying the issue and if firmware can fix it. Again this is not happening with all the new R2X machines and is intermittent. But no identifiable way of knowing what machine is affected by serial number. Extruders are not serialized, I believe they have confirmed it's with only the RevH board so far on the know cases reported.

It does still baffle me, but if think I'm burnt out on the issue and it's a lot of money for and uncertain issue so it might be best I take it back. I may even hold off on buying another until they confirm it's corrected and order it from them directly. It's a shame I will be loosing out on the $400 off sale that I got this one for. Even if I exchange it that won't reset my purchase date and I may get another affected unit, or not. Still a gamble in my eyes.

I can say for sure that I would not normally take apart a new item that costs so much in the first year of owning it unless for maintenance. But dealing with this issue has given me more insight/experience hands on learning for being my first 3D printer. So I did get somthing from this and learned that makerbot customer service is top-notch above most others I have delt with. The apple/Subaru or who ever you like of 3D printers.

If you have the same issue I'm sure they will figure it out and correct it.

[Joseph Adams]

Here are my PID settings I mentioned from an earlier post.  Sorry if you’ve already tried changing this, I haven’t been following closely enough to know.  Best of luck!

 

--

You received this message because you are subscribed to a topic in the Google Groups "MakerBot Operators" group.

To unsubscribe from this topic, visit https://groups.google.com/d/topic/makerbot/8YbqmwbqzK8/unsubscribe.

To unsubscribe from this group and all its topics, send an email to makerbot+u...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.

[Joseph Chiu]

I have wrestled with the 2X in a somewhat similar fashion.  I'm not so sure that it is a firmware problem, per se, as the same firmware manages the thermal regulation of the Replicator and the R2.  

My own feeling is that there is a problem with the way the thermal flows work in the design of the hot end and its mounting bar -- and that it can't reliably close the loop on regulating the actual temperature in the hot-block and the nozzle.  I don't have enough data to prove it, but that's how it feels to me right now.

Perhaps it's fixable in firmware -- but given the code space constraint that is well discussed around here, it may be hard to implement any degree of sophisticated feed-forward correction to the hot end...

--
You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.
To unsubscribe from this group and stop receiving emails from it, send an email to makerbot+u...@googlegroups.com.

[Ryan Carlyle]

On Tuesday, February 25, 2014 11:14:46 PM UTC-6, dnewman wrote:

Faster cooldown indicates that the right is dissipating heat faster.
The same cause could also be causing it to take longer to heat up. 

Upon controlled re-testing, I was wrong about the faster cooldown. I stripped down to just the hot ends and cooling bar sitting on the carriage, and they cool down at exactly the same rate. So I modeled the cooldown as a simple heat transfer function: 

dTemp/dt = k * (extruderTemp - airTemp)

Note the excellent model fit:

So that gives me a function that I can use to calculate heat loss for any arbitrary temperature. That lets us model the theoretical heat-up based on the heater control code. Looking through the heater.cc code Dan linked earlier, the extruder heaters are always run full-blast until the extruder temp is within 15C of the setpoint. So the heat-up from ambient to (setpoint-15) has constant input power. Constant input, minus our function for heat loss, generates a heatup model. This model curve also matches the measured data very well:

By adjusting the assumed input power, the model can be made to fit the right extruder or left extruder up to the PID activation point. 

The right heatup curve implies only ~90% the input power of the left heatup curve. Which makes me very confident that my right extruder is receiving ~10% less heating power than the left extruder, for reasons unrelated to the PID controls.

So, I checked the heater resistance. The left heater reads 13.6 ohms (42 watts @ 24v) and the right heater reads 15.9 ohms (36 watts @ 24v). So the heat-up speed difference is completely explained (for my R2x stripped to the hot ends) by heater variation.

I'm going to keep experimenting to see if I can get it to maintain a 2 degree offset like it often does.

[Ryan Carlyle]

What. The. Shit.

I re-assembled everything and redid the heat-up and cooldown tests. The right extruder gets MORE cooling fully assembled, and the left extruder gets LESS cooling fully assembled.

This is kind of baffling. But it explains a lot -- My right hot end has less power and more cooling, so it just struggles to get on target.

[Scott D]

Ryan, I thank you for continuing to look into this but I also understand that your machine is affected in a similar way that mine was.
When I read your first post of the heater differences I thought to my self that I had already swapped the left and right heater, I even swapped the control plug on the board that provides the power to the heater and I had no luck, then I read this one with "What. The. Shit." and I dont mean to laugh but since I am just as baffled and have tried so many things to get it to work I just had a laugh a little. But at least you machine does eventually get on target, mine would not reach target after 15 or 20 minutes.

So I just got back from my local electronics store and the sale on the R2X is still on, I think it ends this weekend. But even worse is they only had one left in stock and they are not ordering any for now.  I decided to make the exchange after being assured that the 30 day return clock starts over. I almost had to return it before I left, I guess since makerbot wont take any returns on units with problems from the store they now have to try to fix it to resell it as a open box discount. So it was pointed out to me after the exchange that the store is no longer taking returns on makerbot printers unless its un-opened. I guess I got in just under the wire since the POS was not updated yet and my receipt says 30 day return unless stated differently above, and I have nothing written about. They checked with the manager and said they will honer that as long as I have the receipt.

Still I dont want to screw the store. My intention is not to return this one also but its a gamble on the one I just picked up is affected or not. Or is it really intermittent or do some people think its normal to wait 30 minutes before being able to printing and not everyone is reporting this issue. So for now the box stays un-opened as I give myself another day to ponder my options. Maybe I should go over to the other thread about "what printer would you buy today"? Na, I know my other two options but dont know if I want to wait.

Well if I open it I will let you know if its really intermittent for me, or if this unit heats up correctly with the right extruder. I was hoping they had one in stock marked as a November build since the December 7th build tag is the one I was seeing the issue with. I cant conform when in November or before December 7th the change was made, but I know the demo 2X machine they run is from sometime in November that was marked on the box. It would be a shame but I think my local store may stop selling this brand and keep selling the other six brands of 3d printer they have in stock.  

[Dan Newman]

> So, I checked the heater resistance. The left heater reads 13.6 ohms (42
> watts @ 24v) and the right heater reads 15.9 ohms (36 watts @ 24v). So the
> heat-up speed difference is completely explained (for my R2x stripped to
> the hot ends) by heater variation.

Those heater cores are spec'd to be 40 W heaters, but I've never been
clear on whether that's when heated up or when cold. I'm also unsure
of how much the resistance goes up on those cores when heated up as
I've never bothered to measure one when it's hot.

But at 40W and 24V, that's a target resistance of 14.4 Ohms. That makes
your measured 15.9 Ohms high by a tad over 10%. So, allowing for the
error in your measurement, it sounds like 10% too high. I wonder what
the accuracy specs are on those cores.

Dan

[Ryan Carlyle]

I SOLVED IT. It's an incredibly easy fix. Just take off your fans / heat sinks and put a piece of Kapton over the gap in the carriage, then re-assemble. The problem is the way the fans and heat sinks circulate air -- something about the air flow through the carriage hits the right extruder a lot harder than the left. The heater can't keep up with the heat loss from air blowing directly on the hot end.

Give me a minute to update my charts and write up what I did with pics.

[Scott D]

Ok, I have to wait to see this, I was just thinking of a kapton tape solution. But I was going to put some kapton tape around the filament tube to see if I could reduce the transfer of heat to the block a little, in turn allowing the hot end to hold more heat? but then again that would mean I open the new one. Now depending on what you post I just may have to open the new box. If this one has the same issue I may just use your fix.

So how thick is the kapton tape you are using?
If I remember correctly the fans and heat sinks are identical and I even swapped the fans in one of my tests. But this was also switched on the plugs. So if the, um .. never mind. I just debunked my own thought. My dexcom is yelling at me to I need to go check my sugar. Ill be back in 15 once I can think again.

[Dan Newman]

On 26/02/2014, 7:51 PM, Ryan Carlyle wrote:
> *I SOLVED IT.* It's an incredibly easy fix. Just take off your fans / heat

> sinks and put a piece of Kapton over the gap in the carriage, then
> re-assemble. The problem is the way the fans and heat sinks circulate air
> -- something about the air flow through the carriage hits the right
> extruder a lot harder than the left. The heater can't keep up with the heat
> loss from air blowing directly on the hot end.

Congrats on finding AND resolving it! My wild speculation about the extruder
heatsink fans creating an air wash hitting the right extruder wasn't too
far off. (But it was completely wild and left base as I've never seen an
actual R2X.)

Dan

[Dan Newman]

On 26/02/2014, 8:04 PM, Scott D wrote:
> Ok, I have to wait to see this, I was just thinking of a kapton tape
> solution. But I was going to put some kapton tape around the filament tube
> to see if I could reduce the transfer of heat to the block a little, in
> turn allowing the hot end to hold more heat? but then again that would mean
> I open the new one. Now depending on what you post I just may have to open
> the new box. If this one has the same issue I may just use your fix.
>
> So how thick is the kapton tape you are using?

Shouldn't have to be thick at all to block the air circulation. Main
feature is that it be able to survive the temps. And since Kapton is
meant for that, it's a good choice.

Dan

[AL M]

I have ran into a similar after upgrading the carriage i could not get the right side print to stick to the HBP i formed kapton tape i a L shape under the fans to block the air and no problem since didnt realize it also affected the temp good deal   

[Ryan Carlyle]

This is REALLY good. Or bad, since it's an airflow issue that breaks some people's R2x. Let's say "interesting."

Aside from tolerance variation in the heater core power output, the left and right hot ends are perfectly symmetrical. When you take off the fans and heat sinks, the two sides heat-up/cool-down along identical curves. You can swap the connections to the board and it works exactly the same. This is what you would expect.

When you attach the fans and heat sinks, the RIGHT extruder struggles to reach temperature. While the LEFT extruder heats FASTER than it does in free air. This was a serious "what the fuck" moment. Heat/power transfer from the cooler extruder to the hotter extruder shouldn't happen. Thermodynamics, dude!

As it turns out, the fans and heat sinks are NOT symmetrical -- they are identical. The fans blow air with some swirl (because they are spinning) and that swirling matters a lot. The fans blow air through a gap in the heat sink and carriage directly onto the right extruder! That heated air then blankets the left extruder (or something like that) thus reducing its exposure to cool ambient air. So you're sucking heat out of the right extruder, and protecting the left extruder from ambient cooling. Super weird air flow profile.

Here's the gap:

Here's where right fan air blows large amounts of air directly through a mysterious gap in the heat sink fins:

Here's where you put Kapton tape (any thickness) over the gap to block the air flow (but don't put it directly between the heat sink and the cooling bar):

I put the tape over the entire gap. I just didn't feel like taking the left fan off again for the pic. Tired of those stupid bolts.

And here's my data comparing the heating/cooling performance for different configurations:

Note that adding the Kapton tape to fill the gap produces the FASTEST HEATING and SLOWEST COOLDOWN of all configurations I tried. Both extruders easily hit the setpoint once you stop blowing air directly onto them.

[Ryan Carlyle]

FYI, I used 1" wide, 2 mil tape from McMaster-Carr. You could just as easily cut a piece off the build plate replacement sheets that come with the machine. (They're too thin for build plate use anyway -- you want a 5 mil sheet for that.)

[AL M]

[AL M]

I trimmed mine real close to the heat sinks dont think you would want it under them in my case i could not believe at first that was all that was keeping the print sticking and never even thought to look at the temp

[Dan Newman]

On 26/02/2014, 8:45 PM, Ryan Carlyle wrote:
> This is REALLY good. Or bad, since it's an airflow issue that breaks some
> people's R2x. Let's say "interesting."

Thanks for the followup. This has been an interesting mystery to follow.

Also worth noting that MBI's assemblers are inconsistent about the fan
airflow orientation. Some folks get their bots with the airflow directed
towards the heat sink, others with it pulling away from the heatsinks, and
others with a mix of the two. Clearly, the assembly directions have not
included looking at the airflow arrow molded into the side of the
fan. This may, in part, explain why some folks with this rev of
the heat spreader bar may not experience this.

Dan

[Scott D]

Congrats Ryan, And thank you, great work. After 2 weeks of trouble shooting and testing all electronics this was really starting to bug me.
So do you think this may increase any heat creep while printing, I guess you wont know until you build something over a couple of hours.
I also wonder if the mysterious gap on the heat sink was an attempt to get the X2 to print PLA?

I think the thing I question the most is "What is MBI planning on adjusting in the firmware to compensate for this?" I only wonder since if you slow down the RPM of one fan to redirect the flow speed of air so the right reaches temp, this may work for X2's having this problem but might cause issues with prints for people that didn't have this problem.

Well this seem like your fix will work or MBI can replace affected users with new heat sinks. A much cheaper part to replace than the whole right side like they sent me.
 Again thanks and great work.

[Jetguy]

Hate to be the naysayer but if cooling air isn't flowing as designed, heat soak of the plastic carriage, especially in the belt clamp area sounds ripe for disaster.

Again, common sense here. Thermal insulation or not, you still have infrared radiation from the 230C HEAT block right next to plastic. With no convection at all, the plastic will get hot on long prints.

For every action, there is an equal and opposite reaction. You think you solved one problem and may have caused another one.

[Ryan Carlyle]

It might have been deliberate for print cooling, or to "bias" the extruder heating to stay on more for thermal stability. Hard to say what the intent was. Since it only really affects one extruder, I have a feeling they made the cut-out from the carriage to reduce weight and just never identified the issue during testing. It looks more accidental than deliberate but who knows.

A heater core that runs on the high side of the power tolerance range can probably get the right hot end up to temp. I suspect that's why yours never reached temp, and mine eventually did, and some people have no issues at all.

A firmware update could probably "bandaid" fix this problem by re-tuning the PID control parameters for the right hot end. I'm pretty sure raising the proportional constant quite a bit would allow it to hit the set point. You just can't raise it too high or the temp will overshoot and oscillate.

[Dan Newman]

> I also wonder if the mysterious gap on the heat sink was an attempt to get
> the X2 to print PLA?

The heatsink has always had a gap on the left -- ever since it was introduced
for the Mk7. (I just pulled the solidworks files from Thingiverse and checked;
file W40-15.SLDPRT; also visually checked my ToM, my Rep 1, and my Rep 2.)

> I think the thing I question the most is "What is MBI planning on adjusting
> in the firmware to compensate for this?" I only wonder since if you slow
> down the RPM of one fan to redirect the flow speed of air so the right
> reaches temp, this may work for X2's having this problem but might cause
> issues with prints for people that didn't have this problem.

Fans are not on hardware PWM capable pins. They'd have to introduce
yet another interrupt in the firmware to do that. I doubt they would.
Reversing the fans (manually remove them and turn them over) to draw air
away from the heatsinks might resolve the issue but introduce a different
one -- insufficient cooling of the thermal barrier tube. Plenty of folks
have reported that their jams were fixed when they turned the fans to blow
onto the heatsinks. Only conceivable firmware workaround might be to
change the default PID tunings. MBI has been using the same Kp, Ki, Kd values
since the ToM days. (And they seemed to work well until this.)

Dan

[Scott D]

That's what I feared. But back to dans comment of units being assembled differently. Maybe one fan should be blowing out and the other one blowing in? I may have to take a trip to the store again to see how their demo unit is running. But they also only print demos with the right extruder and I'm not sure they ever tried to print with the left.
Still I believe Ryan's research does confirm it's not an issue with some electronic part or differently machined part from the left to the right.

[Dan Newman]

On 26/02/2014, 9:57 PM, Dan Newman wrote:
>
>> I also wonder if the mysterious gap on the heat sink was an attempt to get
>> the X2 to print PLA?
>
> The heatsink has always had a gap on the left -- ever since it was introduced
> for the Mk7. (I just pulled the solidworks files from Thingiverse and checked;
> file W40-15.SLDPRT; also visually checked my ToM, my Rep 1, and my Rep 2.)

Meant to include this picture from the Mk7 assembly directions. It shows the
heatsink.

Dan

[AL M]

Maybe back a few months when i said i had no problem printing PLA on this 2X most people thought i was nuts but i could only print on the right side and when i upgraded to Carls the right side would not let the ABS stick till i blocked the air now i have to use external air to print PLA

[Dan Newman]

On 26/02/2014, 9:51 PM, Ryan Carlyle wrote:
> It might have been deliberate for print cooling, or to "bias" the extruder heating to stay on more for thermal stability.
> Hard to say what the intent was. Since it only really affects one extruder, I have a feeling they made the cut-out from the carriage

Ahhh. Cut out from carriage. I saw the other posting with,

I also wonder if the mysterious gap on the heat sink was an attempt to get
the X2 to print PLA?

and thought that folks were considering that asymmetry in the heatsink to be
part of a recent design change. (And I can see the cutout in the carriage
from Ryan's photos.)

> A firmware update could probably "bandaid" fix this problem by re-tuning the PID control
> parameters for the right hot end. I'm pretty sure raising the proportional constant quite

FWIW, the PID parameters are in EEPROM and, from RepG at least, can be changed. I
don't recall offhand if they are presented in MakerWare's equivalent or not. (But
I can tell you how to update the .json file with the EEPROM map to add them. You
can update it and make editable all sorts of parameters which MakerWare doesn't
present. E.g., that odd temp offset value.)

Dan

[Joseph Chiu]

IIRC, that section where the heat sink fins are deleted was originally intended for the thermal cutoff switch that were later repositioned to the outside of the hot block.

--
You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.

To unsubscribe from this group and stop receiving emails from it, send an email to makerbot+unsubscribe@googlegroups.com.

[Ryan Carlyle]

MW lets you edit the PID constants and acceleration settings via USB. It looks like most of the firmware settings are in there. It's one of the dropdown menus; I forget what it's called.

The right extruder airflow problem gets a lot worse if you run just the fans without the heatsinks, so I think the main issue is the direction of air swirl coming off the fan blades. The heat sinks block a little of this airflow but not much. Having all the fins would probably help a little more. 

[Joseph Chiu]

Ryan,

Based on your graphs, I think you've done it.  

My initial reaction is that you should wait to declare victory, because I've had some false successes when I was battling with the R2X.  In my case, I had problems with both extruders, but the right was definitely the more difficult side -- so in retrospect, it seems to jive with your findings.

Great job!  My hat's off to you!

Joseph

--

You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.

To unsubscribe from this group and stop receiving emails from it, send an email to makerbot+u...@googlegroups.com.

[Ryan Carlyle]

Joseph,

It's possible that removing that airflow will negatively affect filament cooling, such as bridging and overhangs. That's my main concern. Basically the carriage gap acts like a poor-man's blower fan on the print. 

The longer-term solution might be printing a baffle/diffuser to spread the airflow evenly over both extruders, and perhaps adjusting the PID settings to compensate for the cooling. I'll have to do some print testing now that the gap is blocked.

Or maybe it just needs a just a layer of ceramic tape and kapton around the hot end to keep the air off it...

[AL M]

Ryan it will in my case help with overhangs and prints sticking to BP

[Joseph Chiu]

Well, so one possibility to consider is whether all you need to do is to put some thermal insulation and kapton over the exposed side of the hot block.  I did try fully encasing the hot block, as one of my experiments on the R2X, but that was with an already iffy hot-end, so my experiments there are probably irrelevant.  

On Thu, Feb 27, 2014 at 1:05 PM, Ryan Carlyle <temp...@gmail.com> wrote:

Joseph,

It's possible that removing that airflow will negatively affect filament cooling, such as bridging and overhangs. That's my main concern. Basically the carriage gap acts like a poor-man's blower fan on the print. 

The longer-term solution might be printing a baffle/diffuser to spread the airflow evenly over both extruders, and perhaps adjusting the PID settings to compensate for the cooling. I'll have to do some print testing now that the gap is blocked.

Or maybe it just needs a just a layer of ceramic tape and kapton around the hot end to keep the air off it...

On Thursday, February 27, 2014 2:52:54 PM UTC-6, Joseph Chiu wrote:

Ryan,

Based on your graphs, I think you've done it.  

My initial reaction is that you should wait to declare victory, because I've had some false successes when I was battling with the R2X.  In my case, I had problems with both extruders, but the right was definitely the more difficult side -- so in retrospect, it seems to jive with your findings.

Great job!  My hat's off to you!

Joseph

On Feb 26, 2014 7:51 PM, "Ryan Carlyle" <temp...@gmail.com> wrote:

I SOLVED IT. It's an incredibly easy fix. Just take off your fans / heat sinks and put a piece of Kapton over the gap in the carriage, then re-assemble. The problem is the way the fans and heat sinks circulate air -- something about the air flow through the carriage hits the right extruder a lot harder than the left. The heater can't keep up with the heat loss from air blowing directly on the hot end.

Give me a minute to update my charts and write up what I did with pics.

On Wednesday, February 26, 2014 9:42:11 PM UTC-6, Scott D wrote:

Ryan, I thank you for continuing to look into this but I also understand that your machine is affected in a similar way that mine was..

When I read your first post of the heater differences I thought to my self that I had already swapped the left and right heater, I even swapped the control plug on the board that provides the power to the heater and I had no luck, then I read this one with "What. The. Shit." and I dont mean to laugh but since I am just as baffled and have tried so many things to get it to work I just had a laugh a little. But at least you machine does eventually get on target, mine would not reach target after 15 or 20 minutes.

So I just got back from my local electronics store and the sale on the R2X is still on, I think it ends this weekend. But even worse is they only had one left in stock and they are not ordering any for now.  I decided to make the exchange after being assured that the 30 day return clock starts over. I almost had to return it before I left, I guess since makerbot wont take any returns on units with problems from the store they now have to try to fix it to resell it as a open box discount. So it was pointed out to me after the exchange that the store is no longer taking returns on makerbot printers unless its un-opened. I guess I got in just under the wire since the POS was not updated yet and my receipt says 30 day return unless stated differently above, and I have nothing written about. They checked with the manager and said they will honer that as long as I have the receipt.

Still I dont want to screw the store. My intention is not to return this one also but its a gamble on the one I just picked up is affected or not. Or is it really intermittent or do some people think its normal to wait 30 minutes before being able to printing and not everyone is reporting this issue. So for now the box stays un-opened as I give myself another day to ponder my options. Maybe I should go over to the other thread about "what printer would you buy today"? Na, I know my other two options but dont know if I want to wait.

Well if I open it I will let you know if its really intermittent for me, or if this unit heats up correctly with the right extruder. I was hoping they had one in stock marked as a November build since the December 7th build tag is the one I was seeing the issue with. I cant conform when in November or before December 7th the change was made, but I know the demo 2X machine they run is from sometime in November that was marked on the box. It would be a shame but I think my local store may stop selling this brand and keep selling the other six brands of 3d printer they have in stock.  

On Wednesday, February 26, 2014 10:08:19 PM UTC-5, Ryan Carlyle wrote:

What. The. Shit.

I re-assembled everything and redid the heat-up and cooldown tests. The right extruder gets MORE cooling fully assembled, and the left extruder gets LESS cooling fully assembled.

This is kind of baffling. But it explains a lot -- My right hot end has less power and more cooling, so it just struggles to get on target.

[Ryan Carlyle]

Yeah, I'll try that too. But having more air blowing on the right extruder than the left strikes me as an issue independent of whether the hot block is being cooled too much. If it's hitting the hot block, it's hitting the print area. For consistent results you would want both nozzles to receive the same amount of airflow.

[Dan Newman]

> Or maybe it just needs a just a layer of ceramic tape and kapton around the
> hot end to keep the air off it...

Which is precisely how MBI used to make these. All six sides were covered with
a layer of ceramic "tape" then wrapped with Kapton tape. But that took longer
to assemble and eventually the kapton cooked enough that it needed replacing.
The appearance after 1000 or so hours led to support calls. (ToM and Cupcake
owners who assembled their own bots knew what to do and had spare kapton
tape from their assembly. Not so with Rep 1 owners who would call MBI or post
to this list.)

Dan

[Ryan Carlyle]

Print quality update with carriage gap filled!

I reprinted a 4 hour right-extruder print from earlier in the week, in order to do a print quality comparison test. I went straight from the original .x3g file but this isn't a perfect comparison because I took my Y-stepper cooling unit out and moved the printer to a different room, which means the printer interior is now a bit cooler. It's also possible that the filament diameter changed since somewhat since the print was sliced. YMMV. This is MBI Yellow ABS. 

Parameters were something like this:

2 shells, 30% linear infill

3 floors, 3 roofs

40 mm/s exterior shell, 50 mm/s interior shell

90 mm/s infill

Internal and external spurs on (I think this hurt the overhangs a lot)

230C extruder, 115C HBP w/ ABS slurry

Packing density calibrated before the first print, not the second print

As expected, there is a very slight degradation of bridging, overhangs, and surface finish. But it's incredibly minor. I couldn't tell the difference if I didn't have them side-by-side in front of me. Right is before, left is after:

These are upside down relative to the printed position. All surface curves are spherical or cylindrical. There was a larger "rectilinear" piece on the build plate along with this, which accounted for the majority of the build time. That piece came out indistinguishable between the two prints.

I would say the difference between the two pieces is "below the margin of error." It could be explained by filament diameter change, 5 degrees of build chamber temp, or random chance just as easily as covering up the carriage gap.

[Scott D]

Seems Ryan beat me again :-).
All seriousness I ended up opening the new rep2x and have the same exact issue. This does at least reset my 30 return, but I sat on it for 3 days not opened.

So based off Ryan's research and dans theory, and my concern that jetguy also expressed I decided I'm going to go a slightly different way. Instead of covering the whole gap or he whole right side in he caraige I thought it might be good to only cover part of it. Just enough to reach the wanted set temp. The goal was to get my left and right to act the same way. I wanted to see both hit 230, drop to 229 and go back to 230. I cut multiple sized parts out of a Buisness card with slightly different widths. In my testing I also noticed that the position placement of the card make a big difference.

What I found to be the best was if the left side of the card was lined up with the right side of the left fan. 1mm in width did make a difference of 2 degrees in change.

I'm using card stock after I realized the temp never gets hot enough in this area after running at full temp for 45 minutes. But when I test print I might cover the paper with kapton tape to be safe.

I don't have any fancy equipment. Or software? to graph the heat up or cool down temps, but I did record it on video and might speed it up for a quick view on YouTube if anyone is interested. It only shows the temp reading from each extruder on the LCD. But even that never goes out more than 3 degrease difference during the heat up or cool down.

Ryan, can you share the GCode file to your sample you were using for your testing? I would like to do my testing with the same code so your not the only one waisting filament and time printing while testing this issue.

[Ryan Carlyle]

Yeah, I can post it when I get back to my computer. I didn't use anything fancy for the charts -- I just read temps off the display every 10 seconds and typed them into Excel.

[Ryan Carlyle]

Here's the x3g file I used. Your results may vary if your filament is a different diameter.

[Scott D]

Well I don't want to waist any more of my filament so Im going to use the makerbot ABS natural that comes with the 2X.

I just measured it at 1.75 to 1.77, hopefully that is close to what you are running?

[Ryan Carlyle]

Yeah, that should be fine. It won't be far off.

It's kind of a challenging print -- the contact area of the ball section is very small, and there is a lot of curling on the small overhang. You might be better off using a test print off Thingiverse.

[Scott]

Thanks for the offer on the test print, but naaaaa. I like a challenge. That why I bought a 3D printer. For the challenge of learning something new. Once I get the machine down I'll start learning how to design.   

Scott

--
You received this message because you are subscribed to a topic in the Google Groups "MakerBot Operators" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/makerbot/8YbqmwbqzK8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to makerbot+u...@googlegroups.com.

[Ryan Carlyle]

Ok. I recommend putting some ABS+acetone slurry on the front-center of the build plate so the ball section sticks, but leave the right 1/3rd of the plate clean so you don't get over-adhesion on the big flat section. I positioned the pieces on the plate to avoid some scratched kapton.

[Scott]

Well that would have been a good idea, oh well. I started the print already after some challenges with filament feeding. But then I realized I was trying these new sealed containers for the filament. Turns out it makes it hard for it to spool out correctly. So I guess they go back. 

Well I guess we will see how it does with acetone wipe at 115C on stock makerbot kapton tape. 

I can tell you this heat plate seems more even and even a little hotter on the surface than the last one. And the fact that you layer it out in the corner and edge is even a better test since my last one seemed to have had a cold corner and edges and would always warp in the corners. So it's 20 minutes in, I'll just let it run now since I got the feeding issue resolved. 

Scott

[Scott D]

So I realize not everyone has the option to return a makerbot, but since I did I needed to carefully weigh my options. After some testing and shakey at best inconclusive results I have become growing concerned that a firmware tweet will fix this issue all the way and give me the print quality I'm looking for. Mt original decision to purchase the 2X was because of the ABS for what I plan on printing. One patern I noticed is that the rev G board came out sometime in December.

So what I decided was to give up on dual printing, I dint think I will miss it that much. Exchange the R2X or the R2 made in November. They had one November left. Got home and confirmed it had the updated extruder, rev G board, square heat block with the ceramic tape and kapton tape that has proved reliable at maintaining proper temperature. And with the sale prices on these I got back a little over $400. So I was going to buy bottle works heat bed with one extra glass plate and aluminum arms anyway. After shipping would be about the same $400 and I could print ABS again with the option of PLA for draft prints and a better print smell, faster heat up time and less electric until I need to print the final part in ABS with the HBP returned on.

And I guess the heat bed would help with larger PLA prints set at a lower temp. Same price in the end, not testing anything new and only loosing the dual option of printing but gaining PLA for draft prints. It was a compromise, but one I could settle for.

So now off to lowes to get some plexie cut for the front door and sides, I think I'm going to try so magnet setup for easy removal and placement back on.

[Dan Newman]

On 01/03/2014, 3:25 PM, Scott D wrote:
> So I realize not everyone has the option to return a makerbot, but since I did I needed to carefully weigh my options.
> After some testing and shakey at best inconclusive results I have become growing concerned that a firmware tweet will fix
> this issue all the way and give me the print quality I'm looking for. Mt original decision to purchase the 2X was because of the ABS
> for what I plan on printing. One patern I noticed is that the rev G board came out sometime in December.

You must mean rev H. Rev G came out in October 2012 when the Rep 2 was launched. The Rev H came out around
the same time the Rep 2X launched (but was necessarily on the first Rep 2X's.)

Dan

[Dan Newman]

P.S. I think you'll be quite pleased as long as you don't want to do dual printing.
And if you do, the electronics support it and you can get a dual carriage from
Carl Raffle and assorted parts from QU-BD, Carl, and MBI.

Dan

[Ryan Carlyle]

Dualstrusion is important to me, but I wouldn't recommend an R2x to anyone who doesn't need it. The high carriage mass, heat-warping stock HBP, and overall fidgetiness of ABS are good reasons to get an R2 instead.

[Scott D]

Dan, yes I ment rev H, the same one that was in the 2X. I just realixpzed I miss spoke in my previous comment. But either this one is not affected by the issue or I won't see it because it's the squar heat block and tape, or because it's a single extrusion set up? I also know about carls carriage and that may be a upgrade in the future. For cost savings I was thinking about Carl caraige and FF full nozzle set up they sell for 150' half the cost of MBI and no ceramic trials.

Ryan, dual was just a bonus for me. ABS is important to me. So this way with the R2 I get rid of nit dealing with the wrapped plate and more stable Z axis by updating to the bottle works parts and still get ABS. And a proven board. Dual may be an upgrade down the road for personal use but not required now. But if I do I take the chance of running into this same issue if I have an affected board???

It is nice not having to deal with the middle of the HBP raising and I was able to level the plate and go. I've already printed 4 of the 5 samples on he SD card with out issue. I didn't even clean the plate for a test, no tape either, lol. PLA on this plate sticks harder than ABS on glass with suave hairspray. But of course I don't recommend nit cleaning the plate.

I already found a 24v, 14.6A PSU for about $30, I'm going to do this one like the FF with a mini PSU installed under the bottom. Update the fuse and all to a 15A.

The clear filament it came with is sort of heat also. No color PLA. On my 6th test print now I finally have a little warping. Ok, so I should clean it.

[Dan Newman]

On 01/03/2014, 4:28 PM, Scott D wrote:
> Dan, yes I ment rev H, the same one that was in the 2X. I just realixpzed I miss spoke in my previous comment.
> But either this one is not affected by the issue or I won't see it because it's the squar heat block and tape,
> or because it's a single extrusion set up?

My Rep 2 doesn't have the issue. But it also has the old assembly technique in which the extruder
is wrapped with ceramic tape on the top and underside of the block. It's completely wrapped on
four sides and barely wrapped on the front and back. When I have to rewrap it, I'll cover the
front and much of the back as well. It does not have that quicker-to-assemble slide on insulation
which leaves the top and bottom -- the two largest faces -- exposed.

> It is nice not having to deal with the middle of the HBP raising and I was able to level the plate and go.
> I've already printed 4 of the 5 samples on he SD card with out issue. I didn't even clean the plate for a test,
> no tape either, lol. PLA on this plate sticks harder than ABS on glass with suave hairspray. But of course I
> don't recommend nit cleaning the plate.

Be warned that the acrylic plate may eventually warp. A number of early Rep 2 owners got warped
plates. (I think MBI just had them cut and didn't think about flatness.) MBI replaced those
for folks who had issues. But, if you print models which cover a good portion of the plate, you'll
find it warping after a few months: laying that hot plastic down on the acrylic warms it up enough
for the acrylic to start relieving internal stresses and warp in the process.

> I already found a 24v, 14.6A PSU for about $30, I'm going to do this one like the FF with a mini PSU
> installed under the bottom. Update the fuse and all to a 15A.

That's what a number of us have done. Note that unless they moved to a different power switch, the
one they were using wasn't rated for 15A. The one in my Rep 2 wasn't even rated 10A. I replaced
it when I added a 14.6A PSU.

Dan

[Scott D]

Dan, I suspect that new ceramic is part of the issue after my testing the other night. This is why I mentioned if a go dual with this machine I would buy the parts from FF, nozzles are cheap to change on it. I think the way the heat is pushed off the ceramic with the fan air and not so much off the taped blocks also caused me more issues trying to get the bridges and the tops of the 20mm calibration cube done correctly. It's heating up the top of the print. Ceramic hold heat but also dissipates it faster than I'd it was wrapped in kapton tape.

But in the past for me now that I have the rep 2. So far all test prints have been flawless. We are calling for 12 inches of snow so I will set up for some large prints to run since I'll be home. Something I never felt comfortable doing with the rep 2x design I had.

[Scottbee]

Another data point:

I never noticed this problem with my first 2X (manufactured in October 2013).

Received our new 2X (manufactured January 2014) and I ran a test.  With the cover off and the front lid open the right extruder will not come up to 230C if both fans are running (the left extruder is above 50C).  If I slide in a business card under the right heatsink it will come up to temperature.  Or, if the left extruder drops below 50C and the left fan turns off the right extruder will get to 230C.

I need to take a look at the old 2X to see what the differences are (if any).

[Scott D]

Great, someone that owns both, but sorry to hear you have a machine with that issue.
I was searching for a picture of the carriage and trying to figure out if it was changed at all.
I think I confirmed most every other part as being the same.
If the carriage is the same then it has to be something with the board?
Although I returned mine went to the Replicator too, I'm still very curious on finding out if there's a difference in parts?

[Ryan Carlyle]

Scottbee, can you put an ohmmeter on the heater elements for your two printers and see if your old R2x has a higher-power heater on the right extruder than the new R2x? (Just unplug the heater from inside the carriage wiring bundle and stick probes on the pins.) 

Or perhaps there's some difference in the hot end insulation or fan blade geometry? I could see there being an issue if the fans were changed.

Might also be a good idea to check that the firmware PID settings are identical between the old and new R2x.

[Dan Newman]

On 02/03/2014, 11:22 AM, Ryan Carlyle wrote:
> Scottbee, can you put an ohmmeter on the heater elements for your two
> printers and see if your old R2x has a higher-power heater on the right
> extruder than the new R2x? (Just unplug the heater from inside the carriage
> wiring bundle and stick probes on the pins.)
>
> Or perhaps there's some difference in the hot end insulation or fan blade
> geometry? I could see there being an issue if the fans were changed.
>
> Might also be a good idea to check that the firmware PID settings are
> identical between the old and new R2x.

Mind you, your garden variety Ohmmeter is not going to give a very accurate
reading in this low resistance range. And, the resistance of the leads -- often
about 2 Ohms -- will be in that reading. Some meters let you "zero" that out,
but that also is a big hit on accuracy. Most reliable reading would be
with a meter which supports Kelvin leads (4 wire leads).

[Ryan Carlyle]

If your meter can't tell the difference between 14 ohms and 16 ohms, you need a better meter. I'm using a ~$130 Extech with two probes and the difference in measurements has good agreement with the difference in heat-up rate between the two hot ends. Touching the probes together reads 0.0, so it must've been factory zero'd.

Alternatively, you can block the carriage gap and trend the heat-up curves like I did. That'll show if one heater is under-powered relative to the other.

[Dan Newman]

On 02/03/2014, 11:40 AM, Ryan Carlyle wrote:
> If your meter can't tell the difference between 14 ohms and 16 ohms, you
> need a better meter. I'm using a ~$130 Extech with two probes and the
> difference in measurements has good agreement with the difference in
> heat-up rate between the two hot ends. Touching the probes together reads
> 0.0, so it must've been factory zero'd.

Relative differences cheap meters can do. My point, which wasn't clear,
is that Scottbee should read too much into absolute differences between
his readings and yours which you posted. Different
inexpensive meters will give different absolute results, in a large
part owing to the resistance of the leads themselves. After all 1 - 3 Ohms
resistance variation between meters for the leads alone is upwards of 20% variation.

Dan

[Joseph Chiu]

This is one of those cases where I would take two cheap meters (Free, with coupon at Harbor Freight! -- helps that they're only five minutes from home) and measure current and voltage simultaneously.

--

You received this message because you are subscribed to the Google Groups "MakerBot Operators" group.

To unsubscribe from this group and stop receiving emails from it, send an email to makerbot+unsubscribe@googlegroups.com.

[Scottbee]

Ummm... I own an Engineering lab.  I have NIST traceable milli-Ohm (4-wire) meters at my disposal.  But I think in this specific case my old faithful desk Fluke 179 will probably do....   ;)

On the new machine both left and right heater elements came in at 14.3 Ohms.  We can argue the accuracy of that reading, but the repeatability of the meter is probably excellent..... so I'd say there is little to  no difference between the sides. 

My older machine is in the middle of a print... probably not going to probe it for a while.

[Ryan Carlyle]

I figured you'd have adequate measuring tools for the job :-)

Interesting that your heaters came out matching, since mine were so far apart.

[Scott D]

I did a cheaters test by swapping the left and right elements and it didn't solve it, in fact the left still worked with the right heating element. The left always worked, but ran hot and made bridging berry difficult.

[Ryan Carlyle]

It's definitely based on position on the carriage, which seems to rule out board issues. The only question is why some people DON'T have the problem.

[Scott]

Well what one says is intermittent may not be.
The pattern seems to be anything after December manufacture or assembly date.
I met a person today that bought one in January with a December date. But he has never tried dual printing. So unless they try dual print they might not know about the issue and it may not get reported causing it to appear intermittent?

Scott

On Mar 2, 2014, at 8:09 PM, Ryan Carlyle <temp...@gmail.com> wrote:

It's definitely based on position on the carriage, which seems to rule out board issues. The only question is why some people DON'T have the problem.

--
You received this message because you are subscribed to a topic in the Google Groups "MakerBot Operators" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/makerbot/8YbqmwbqzK8/unsubscribe.

To unsubscribe from this group and all its topics, send an email to makerbot+u...@googlegroups.com.

[AL M]

oops sry 

On Tuesday, February 25, 2014 9:09:31 PM UTC-5, Scott D wrote:

I read these post about the 2X being a work horse and then I buy one and we have not been able to get the right extruder tool 0 working since I've had it. I think some if not many changes have happened since it was fist released.
At first the tech said they are aware of it and working on a solution. Basically after asking question it was either sending parts or possible firmware tweak.
I called back after a week and the new tech had me remove the extruder stepper motor. For sure after looking at how kinked and pinched the coupler wire was I thought we had identified the issue. It reached 226 and maybe 227 but never 228 to allow the filament loading script to complete or even be able to print. So they sent me a new coupler. After replacing it I had the same problem. I have a 2X that is basically a R2 for ABS.
Then on Friday they decided to replace the whole told bar. But it bugged me that no real trouble shooting was going on. So over the weekend I swapped the cable harness from tool 1 that worked to tool 0 and and even put the heating element from tool 1 to tool 0. For sure this would tell me what was wrong but to my surprise it had the same issue. So new coupler, different heating elements, new wire harnes and still not working.
So then I thought it could be the board, maybe control A had an issue on the board which control A was controlling the right tool 0. So I moved the wire harnes to control B and moved the coupler from A to B. Now for sure this has to work. As expected the firmware had to be told to heat up the left extruder but was realy hooked up to the right......and still the same issue. So now I thought it had to be something with that heater block and the parts in the mail would fix it.

So today I received the new parts. The heating elements, nozzels, filament tube ceramic and couplers already put togetheron the block. I thought great, happy B-day to me (and yes it's realy my B-Day) so I removed the old and on with the new. I thought for sure this would fix the issue especially after all the troubleshooting I did over the weekend. Boy was I wrong. Not only was I wrong but I have never been so baffled as a engineer in my life. This whole thing makes no sence but even after replacing the whole left and right parts with block it still doesn't work.

What more annoying is they don't want to admit or tell me any more information about the part they suspect it bad that the first tech mentioned. They don't deny anything, but just change the subject with "let's replace this part next, I'll send it out overnight".

The only two parts left are the PSU and the motherboard. And now days the boards are clipped in and not held with screws so I don't have the tool to remove it even if they sent me that part next.

I'm now in that window of the last 3 days to return it to the store I bought it. I think I have been more than patient but also do t want to go back to researching machines before I pick the next one I want. I spent almost a year researching and reading up on these machine be fore makerbot got bought out.

So do I return it for an exchange? Do I try the new FF Creator X since I prefer the metal frame, or do I go with the ultimaker 2 that is out of stock every time I check. These were my top three picks. The store I bight it from demos the R2X for 10 to 12 hours a day and has for the last three months with hundreds of hours on it. But this was one of the first ones and not the ones made recently. Mine had a December 7th 2013 date on the box.

Any ideas and suggestions are welcome.

[Scottbee]

I concur.  For "fun" I took my new machine and swapped the extruder cabels (at the MB) and the thermocouple connections.  The "problem" tracked the hardware.

Still doesn't explain why my "old" 2X doesn't exhibit this issue, but I have a suspicion.  To verify it I'd need a guru like Dan to chime-in.  If I remember correctly, the PID algorithms start off with the factory base values... and then at some point do some learned behavior auto-adjustment (at least that's what I think I read somewhere).  If/when it does this I don't believe the base parameters (viewable if you peruse the onboard preferences) get changed.  And I don't know if you can view the current values that the bot is actually using.

But... I may be on crack.  Dunno.

[Dan Newman]

On 03/03/2014, 12:28 PM, Scottbee wrote:
> I concur. For "fun" I took my new machine and swapped the extruder cabels
> (at the MB) and the thermocouple connections. The "problem" tracked the
> hardware.
>
> Still doesn't explain why my "old" 2X doesn't exhibit this issue, but I
> have a suspicion. To verify it I'd need a guru like Dan to chime-in. If I
> remember correctly, the PID algorithms start off with the factory base
> values... and then at some point do some learned behavior auto-adjustment
> (at least that's what I think I read somewhere). If/when it does this I
> don't believe the base parameters (viewable if you peruse the onboard
> preferences) get changed.

The base values are never changed. PID is a feedback control loop with
an evolving error term that feeds back in and is only used for a single
run of that heater. Once the heater is turned off or the temp changed
by more than 10C (up or down), you're back to square one.

Dan

[Ryan Carlyle]

Really, failure to reach the setpoint ought to be entirely fixed by the integral term in the PID control. The longer you're off-target, the more the integral should kick in to correct the offset. This isn't happening for some reason. Either the integral parameter is way off or the PID algorithm isn't actually doing integral control. The integral should correct the offset in very roughly 8x the length of time it takes to heat 63% of the way up to set temp.

I think I'm getting uglier prints since I covered up my carriage gap. Specifically, some corners/overhangs are having severe edge curling and some look great. Which makes me think it's an uneven heat radiation issue now that the carriage fans aren't cooling the print at all. So I'm going to keep playing with it.

[Dan Newman]

On 03/03/2014, 1:45 PM, Ryan Carlyle wrote:
> Really, failure to reach the setpoint ought to be entirely fixed by the integral term in the PID control.
> The longer you're off-target, the more the integral should kick in to correct the offset. This isn't happening for some reason.
> Either the integral parameter is way off or the PID algorithm isn't actually doing integral control.

Third possibility: the rate of heat dissipation (which increases with the fourth power of the
temperature) is beyond what the 40W heater cartridge can sustain. However, others have
commented that they could just use a higher set point so I don't lend much credibility
to this third possibility.

The PID source code is

https://github.com/makerbot/MightyBoardFirmware/blob/master/firmware/src/MightyBoard/shared/PID.cc

Of course, there could be a bug and the source code is misreading its EEPROM and ending up
with a 0 term for Ki.

Dan

[Scottbee]

I know what a PID controller is and how it works.  I must have misread a post somewhere that led me to believe that the PID multiplier values could/would get dynamically tuned over "n" number of cycles.  I guess that information was wrong or I misunderstood the post.

Thanks!

[Dan Newman]

On 03/03/2014, 2:03 PM, Scottbee wrote:
> I know what a PID controller is and how it works. I must have misread a
> post somewhere that led me to believe that the PID multiplier values
> could/would get dynamically tuned over "n" number of cycles. I guess that
> information was wrong or I misunderstood the post.

Some of the RepRap firmwares have some poor man's auto tuning built into
them. You may have heard about that at some point.

Dan

[Ryan Carlyle]

Ok, here's a problem. The maximum possible integral component of heater power is only 66% duty cycle, because the PID code clamps the integral error term at 256. (This algorithm implementation has some issues. I can't make heads or tails of the derivative calculations.) 

Dan, do you know how often the PID code is run? What's the polling time for the heater code? I can't tell how often the integral sum accumulates.

Here's the maximum heater PWM output once the integral term saturates, by temperature error:

0C = 66%

1C = 72%

2C = 77%

3C = 83%

4C = 88%

5C = 94%

6C = 99%

7C and up = 100%

Raising the integral constant from 0.33 to 0.50 should correct the issue, by re-scaling the integral gain to allow 100% heater duty cycle. A better (but less convenient) solution would be to de-cap the accumulated error integral. I don't see any good reason for the cap, since the PID control doesn't even turn on until you're within 15C of the target.

[Ryan Carlyle]

To be clear, the default PID formula for heater duty cycle is:

PWM output = max[ 255 , (7*error + 0.33*integral +36*derivative) * 2] / 255

Where integral is the sum of every error measurement, clamped between +256 and -256. Derivative is some kind of time-averaged function that I haven't figured out yet, but it equals zero when the temp is steady, so you can ignore it for static offset error.