Fixing Rep2X Extruder Heat Soak Issues

[Michael Menefee]

I continue to have issues with my Rep2X extruders (both left and right) over-heating and giving me air prints on longer print runs.  I do not believe this is a result of clogged nozzels or anything else, as there is a definite correlation between print time and air prints. Anything over an hour or so generally results in an air print.

As such, I've been thinking through options for modification of the Rep2X extruder assembly to correct these issues.  However, I know that I'm new to the world of MakerBots, so I thought I'd ask for a bit of feedback before pursuing any of these options.  That, and I'm sure at least a couple of these things have been tried before.

So, if you could take a look at the attached PDF and let me know what you think, I would appreciate it.

My basic thinking is to both remove heat (energy) flowing into the aluminum mounting bar, and more efficiently remove heat from it.  I have read several theories on these boards that the thermal barrier tube itself is causing too much energy to be transferred to the filament prior to it entering the nozzle, and causing air prints as a result of the pressure loss on the filament.  I can think of a few mitigation options, but I'm not yet sure if they are sufficent, so looking for a little feedback.  So far:

- Decrease heat load on support bar by insulating stepper motors from bar
- Increase conductivity between heat sinks and support bar by using heat transfer compound
- Insulate filament guide assembly from radiative heat from back of heat sink through use of thermal blanket

- Increase conductivity between thermal barrier and support bar through use of heat transfer compound

Anyway, appreciate any feedback.  Thanks.

[MBuser]

I don't have a printer, but is there also a path upwards from the heater to the support bar to the stepper wheel enclosure?

[Jetguy]

Sorry, but it does NOTHING for the problem.

The problem is the mounting block to thermal barrier interface. The heat is NOT being conducted from the thermal barrier tube to the aluminum mounting part because of reduced contact patch and the clamp section has a vary small cross section.

 

Your mods won't change anything fro the hot end.

Sorry, I'm being honest. The best and most effective thing is to just replace the bar mount minus the cut away sections and even better, is to use a threaded thermal barrier tube to maximize thermal contact between the tube and the bar.

 

Trust me, I build my own extruders complete from bar stock. I know what I'm talking about.

[furicks]

I was having a similar problem to this when I first got my 2x. Especially with dual extrusion prints where 1 color would sit for a bit. 

I contacted Maker support and they send me a new thermocouple for my right side extruder and that has largely fixed the problem. 

If you've ever seen an error about "heater reads failing" then this is likely your problem. 

Consider that having your extruder stop heating (say if a heater turned itself off because it couldn't read the temp) would look almost identical to having the filament over heat. 

In both cases the drive wheel would strip out the filament causing air print. 

If you're positive it's heat soak:

I disagree that your suggestions will do "NOTHING" but jetguy may be right. The surface area to transfer into the aluminum block is quite low.

The surface area to transfer into the actual heat sink is also quite low, so thermal transfer paste would probably help the whole block stay cooler. 

I also added a few adhesive copper heat sinks to the back of my stepper motors to help heat escape that way instead of through the heat sinks which are quite far.

http://www.amazon.com/gp/product/B00637X42A/ref=oh_details_o00_s00_i01?ie=UTF8&psc=1

Take a temp reading of your stepper motors, or even better the aluminum block, to see if the heat really is getting excessively high. 

If the aluminum block isn't getting very hot (it's being cooled correctly) then I'd have to assume the heat isn't getting out of the nozzel and into the block like jetguy suggested. 

On Monday, June 24, 2013 5:21:21 PM UTC-4, Michael Menefee wrote:

[Michael Menefee]

No need to apologize.  That's exactly the sort of feedback I'm looking for.

Now, NOTHING might be a bit extreme, but I see your point.  I'd agree that the optimal way to solve this is to produce a re-designed support bar and barrier tube assembly, especially given an excessively small contact patch between the tube and the support bar..    I was thinking of adding some thermal compound between the parts, but that's probably still insufficient to overcome the contact patch issue.

Actually, even more than that, I'd like to see some sort of ceramic or other insulation between the heater, the barrier tube, and the rest of the assembly, but that's probably wishful thinking at this point.

So given that I need to replace/redesign the barrier tube/heater assembly, do you have any suggestions for where to buy a redesigned support bar and barrier tube/heater assembly?  Or can I retrofit parts from the original Replicator or Replicator 2 extruders?  I have the skill to build what your're talking about out of bar stock, but I currently lack access to the tools I would need to make these parts.  (I miss having a shop full of tools.  A new CNC mill and lathe are on the list, but not for next month....)

[Michael Menefee]

I have never seen any thermocouple related errors.  I have watched the readout temps on the bot during builds and air prints, and it claims that the extruder is running at the commanded temp.  Actually, the temp does not appear to change between "normal" extruding and air printing, which is what drives me to the heat soak conclusion in the first place.  I don't think the 2X is capable of measuring the heat rise in the other parts of the extruder assembly.  I considered trying to electronically log the thermocouple output, but given the LCD does not indicate a change, I think that would be of limited utility, especially since the same thermocouple is probably being used to drive the heater PID loop...

I don't have a reliable way to measure absolute temp of the components right now, but relative temp wise, the support bar, stepper, and heat sink increase by about 20 degrees between "normal" and "air print" modes, again making me suspect the issue is heat soak.  Based on that (highly unreliable reading made with a bare temp probe), back of the envelope calculations suggest that by thermally isolating the various components in the extruder, I could delay heat soak by an hour or more, which might be just enough to get me through the prints I'm trying to make.

I'll agree that what I'm proposing is marginal at best, but it has the advantage of not requiring any machining work, which is a bit difficult to accomplish at the moment.

So, appreciate any suggestions of what to do to solve this.  I'm willing to spend money on parts if someone can point me in the right direction...

[Jetguy]

Yes, I know, harsh. But I sometimes need to make it sink in. I have spent a lot of time making custom extruders over the last 8 months and have been following several other extruder projects.

Again, I build these things mostly from scratch or maybe just a few stock components.

Most recently, I went through a rather frustrating experience of trying to Mod a SeeMeCNC hot end. It had the exact same kinds of problems with PLA. Most users were using ABS fine.

There are many hot ends that don't work well with PLA. They share the same problem, heat soak up in the filament path above normal hot "zone".

 

So, on to the point of how to fix it.

If you download the STLs from the original Replicator, they include a printable version of the dual extruder mounting bar.

If you print that, then you can use it as a drill guide and sizing tool to get the right chuck of aluminum to make the bar from.

Unfortunately, and I'm not saying this to be a jerk, the Replicator 2X is not open source, thus there is no side by side STL you could print and compare the two. You do have the real part thought being a 2X owner.

To make I own bars, I just got my hands on some aluminum bar stock that roughly matched the printed parts at a local machine shop. Shouldn't cost much at all.

I then drill it and tap it with bits and taps from Harbor Freight.

What you do need to switch is unscrew the stock 2X thermal barriers and install these ones from QU-BD

http://store.qu-bd.com/product.php?id_product=25 $8 each.

Or you could just buy the whole dual extruder with no motors, 24 volt heaters, and rig up a mounting plate on the carriage (use the universal mounting plate, not the MakerBot compatible one, that is for a T-O-M)

http://store.qu-bd.com/product.php?id_product=19

Some notes here too folks. QU-BS uses 5mm bore heater cartridges. Makerbot uses 1/4 inch or 6.35mm cartridges. Either heater block still uses an M6 thread for the thermal barrier tube and nozzle. QU-BD uses thermistors by default but can easily be used with a thermocouple.

 

Or, Mbot3D has a copy of the Rep-1 bar

http://www.mbot3d.com/collections/accessories/products/alluminiam-bar-mount $18

And then you need 2 of these http://www.mbot3d.com/collections/accessories/products/stainless-barrel-part $15

So the QU-BD is not exactly a bad deal considering all that you get for about $70.

 

I'm NOT saying you have to make this mod. I'm just answering the question of how you would, if you wanted to. (Sorry, have to throw that disclaimer out there).

 

 

J

[MBuser]

Another quick fix to see if temp is the issue, is to gob on thermal grease to bridge large gaps. Could fill it with copper powder too. If it works, a more permanent fix would be copper filled thermal adhesive. 

[Andy Soukup]

i've got a 2x and i'm seeing the same issues. it seems that things work really well if your print is extruding at a decent pace, but if it slows down too much to do fine detail or you pause the print for any reason you'll strip the filament and get air prints.

i've came up with these tips so far:

- if you need to pause, use the "load filament" option. it'll lower your bed all the way so you get access to your part and then you can extrude a little before resuming to ensure things are flowing.

- if you strip the filament, unload the filament and gently blow canned air on the hobbled gear when it is turning to remove the debris. this way you don't have to disassemble everything. Be sure not to shoot super cold air (liquid) on it! 

jetguy, do you think it could help things out if you were to flow air across the filament between the hot end and pinch wheel? i know it is mechanically crowded in there, but it may be achievable with a custom bracket? i'm also still using the stock extruder. do you thing the increased range of the printed one will help with this?

[Jetguy]

Well, there are two parts to the extruder.

The pinch wheel feeder system and then the hot end.

 

The big problem you you folks is the entire system has been changed. It's good to change one thing at a time and evaluate the effects rather than change everything as MakerBot has done here.

 

So, one thing is that the plunger system was replaced and now with a completely different and all new molded version that is radically different from anyone's designs for upgrades. In doing so, you are in the UNKNOWN portion of the spectrum. The theory is, this system is supposed to be the best one on the market right? I mean it is the flagship model and the most expensive version and ONLY this bot has it.

 

But, studying the design from a mechanical standpoint, it uses a small short spring compressed to nearly it's max. If you know anything about springs, they have limits and lifetimes. Compressing a small short spring to max "coil crush" depth is a sure fire way to make it lose tension or strength.

Given the leverage inside the system and some other geometric considerations, I don't see how this design could be better than say the Replicator 2 upgrades that are very popular. In other words, the new design isn't even as good as a printable upgrade for the previous bot. So it's 2 things really just here in the filament drive. An unproven and new design, that by all analysis doesn't provide as much pressure, nor does it have as much compliance distance (ability to adjust for filament size and maintain constant pressure) as even a cheap upgrade. Then the same design is causing severe duty cycle on a smaller spring and expecting it to last. I HIGHLY suspect had a person measured the spring on the day they got it, and now when they are having problems, they find the spring has both compressed and gotten shorter, and has lost strength. Being a lever system, that then could equate to an even more radical reduction of pressure against the pinch wheel and result in filament slipage.

 

Then, to add insult to injury, they changed the hot end and mounting bar too resulting in what many feel are less than desirable charateristics for printing in PLA.

 

So you are fighting multiple issue here and they all could affect PLA printing. Extruder aging (not working as good as it did when new) seems highly likey

as I go through this design. I attribute that to the spring, and the fact multiple load, unload cycles, folks have complained of the springs failing. That fully backs up my explanations and further is just one more thing that could contribute to issues with PLA.

 

IMO, you really need to fix both, and that is backed up by the fact folks already have designed upgrades for the filament drive.  (sorry, was going to link some but thingiverse is being hammered today?) The problem is, the filament drive is easy to fix, the hot end is a lot more complicated as no printed parts can just fix it, if you know what I mean here.

 

MBuser and other keep insisting you just add some paste for something and I'm betting that just isn't going to cut it and some users tried the basic thermal paste. Again, YMMV and maybe there is the cheap magical fix?

[MBuser]

Isn't the nozzle supposed to be temp-regulated? 

[Jetguy]

Yes, the nozzle is temperature regulated because the heater block has a heater cartridge (heat source) and the thermocouple(heat sensor) both connected to the mainbard and controlled by the firmware. The nozzle is THEN attached to that heater block.

So NO, the nozzle itself is not directly temp controlled, the heater block it screws into is.

That block and nozzle then must be attached b y something, and that is the thermal barrier tube.

Sorry,this pic below is a custom hot end prototype of a MK7ish design using copper for the heater block but is the same basic dimensions, The thermal barrier tube is an experimental one from Makergear they no longer sell. In actual usage, the tube gets cut down so the thinned section is right next to the heater block with minimal threads into the heater block. In theory, there would be a huge difference in temp from one side of the thinned section to the other (provided the top section is properly heasinked!!!).

 

That tube connects the hot end(the nozzle, heater block, heater core, thermocouple, and insulation to the rest of the extruder).

 

So one side of that metal tube is 230C since it screws halfway into the heater block and the nozzle meets and seals to it making a sealed path for the filament to hit the nozzle on the inside, be melted by the controlled heat and comes out the 0.4mm hole on the bottom side.

 

The point here is the thermal barrier tube has 5 very important jobs:

It connects the hot parts to the cold parts

It seals the filament path and nozzle in the hot zone so plastic doesn't leak out everywhere at the 2k+ PSI or more.

It attempts to limit the heat conduction from hot end upwards to the mounting block and motor (material choice and overall cross section, along with contact area at both ends determines how much heat is conducted)

Provides a smooth narrow guide path for the filament until it hits the melt zone. The filament is under compression load. The motor is pushing it down and 1.75mm filament can obviously bend sideways. If it does, it then rubs the walls and creates friction or resistance to actually reaching the melt zone.

 

So back to the assembled MK7 picture, the upper aluminum block is the one attached to the motor and must be cool and is colled by the heatsink and fan that bolts to the front of it, and even the motor body itself can be a cooling effect if the motor is cooler than the mounting block is being heated to by the hot end, via the thermal barrier.

 

And that's the real problem, We really do not care how warm the motor is, nor do we care about the block that attaches to the motor. What we want to cool is the filament path of the thermal barrier tube. It's being heated at one end to 230C. That heat wants to heat the entire thing to 230C just like the heater block and nozzle. The only way to cool that much heat is to conduct it to a larger chuck of aluminum (AKA the mounting block). If you don't have good conduction between the tube and the heat sink, it will be hot, that's just basic laws of physics. Heat wants to go somewhere, you must give it a path. Otherwise, it sits somwhere and that somewhere is the thermal barrier tube.

[TaErog]

Depending on your problem there may be a very KISS fix.

 

First 2 questions:

Are you printing ABS? (it seems like you are - just making sure)

Are you Printing with the top enclosure? (if not? I would find it strange I have done long full build plate prints with no problem regularly and know others that also do it with the 2x and have no problems at all without the top enclosure, and would be quite surprised that you would need to do anything drastic to fix this, But as others and I do not have this problem with the printer I can't know exactly what is going on there) what is your ambient temp like?

 

If you ARE printing with the top enclosure/dome, then I may have a answer for you..

The top enclosure holds in too much heat.  So, on my first long print (with the top enclosure on it came rather late) I saw it start failing . . I opened the top to look closer and it was VERY warm, and a blast of hot air was released . . un-strangely enough the print started extruding again correctly in short order (still messed the print)

So, I just put a few 1" tape roles in the corners and lifted the enclosure dome enough to let some air out . . .this reduced the heat build up greatly and allowed a new print to finish (had no ill effect on the print).  I also noticed the unused head temp was a good way to gauge the chamber/head ambient temp overall (now it is not accurate but gives a great and rather consistent reading between prints to see if there is going to be a overheating problem.  I keep mine under 58c but your range might be different.

 

In the end I just decided to cut a whole in the top of the dome and place a book on the hole to control the air . . . (later I will make/get a louver to make it look better and have settings)  I thought of a low RPM PC case fan with thermistor but decided it was too much complexity for such a simple function.

 

I hope that may help . .

 

 

 

[MBuser]

That's a great idea: use the unused extruder temp as a proxy for the chamber temp.

[Andy Soukup]

it may be wort trying to control build area temp with a cheap controller like this one. it could open a valve or turn on a fan. you'll have to be careful where you measure the temperature at. maybe attach it to the stepper motor?

http://www.ebay.com/itm/Digital-All-Purpose-PID-Temperature-Control-Controller-STC-1000-W-Sensor-110V-/370773791107?pt=LH_DefaultDomain_0&hash=item5653d71183

[Michael Menefee]

So, in reply to this and some of JetGuy's thoughts:

It had occurred to me that the ambient temp inside the bot with the "lid" installed could be the cause of part of the issues we're seeing with heat soak.  If the heat sinks and fans are all sized for a certain "normal" ambient, running them into a static air mass insulated from the surrounding environment could certainly be problematic.

 

That said, I've seen heat soak issues without the lid.  I actually have yet to receive my top enclosure, so when I do print with an enclosure on top, it is make shift at best.  However, I could certainly believe that the lid is trapping too much heat.  The thought of adding a vent or some other sort of baffle had occurred to me as well.  The idea of the lid is to keep drafts out of the printer as I understand it.  I would certainly believe venting hot air could be done without creating too strong of an air current if the mechanism were set up properly.

 

To answer your questions:

1. I'd like to print with PLA, PVA, and ABS.  I was sold the machine under the impression that "PLA was no problem".  I have only been successful with ABS so far, and then only under certain conditions.

 

2. Have tried with it on and with it off.  Don't have a strong correlation either way. 

 

 

I think I need to do more testing...

[TaErog]

 So what is you internal temps (or ball park?) What do you set your HBP at and did you check of that is accurate? (again ball park)

What is your ambient temp of the room the printing is happening?

Is there any real drafts (as in moving air that is of cooler temp)?

 

If there is no strong drafts I would say print without a top entirely, especially if you have high ambient temps (my ambient temps are always below 21c - note my printer reports 22-25)  I had done so for quite a long time before getting mine.  (I only use it when I open my windows and there is a chance of a real breeze)

 

I would test your temps with an external device.  As I said my printer reports a few deg high consistently and the discrepancy is different with the HBP.

I also check to see of your head heatsinks/fans are attached tightly.  Even with the reduced physical contact with the extrusion tube allot of heat is transferred to the block and then to the HS bad physical connections make a real difference. (I build PC's robots etc and use HS compound all of the time - a very thin layer really helps with metal to metal conduction)

Also HS are much less efficient when there is less of a delta between the air traveling through them and there own temp.

Another mod I was thinking of was a corrugated tube (like the ones you use for cable management without the slit) and run that to the heads from the back where the filament is feeding. this will allow the HS to get much cooler external air and keep the block much cooler with no real problem of drafts (the cool air will be warmed by traveling through the HS so no worry there)  I have yet to try this.    I have a old water cooler that would work great and the block is just about the right size . . but that is crazy overkill . . but fun to think on.

 

What is nice with having a controlled vent on the lid is that you can control the convention of air through the printer and as the cooler air is sucked through the cracks of the enclosure it is warmed by the HBP and internal air quite quickly, this should not effect the print at all. You can always crack the front access for a bit more air.  what we are worried about is temp differentials not overall temperature for curling and cracking/delaminating. 

 

I was always dubious on the "PLA was no problem" as I never really got that connotation from the sales pitch and even if I did the printer was specifically designed to do ABS and doing so can (and seemingly did) compromise the printer on printing another plastic with VERY different characteristics this is not really surprising or unexpected.  It CAN print PLA but not without some real temp control and external fans. BUT some simple mods can fix that, so really not a big deal for any DIY 3d printer people.  I had to modify my TOM to print PLA I fully expected to modify my 2x to also.  PVA and nylon should not be as big a problem looking at there characteristics (but not hard data yet)

 

Good luck!

 

 

Ps

the cheap controller looks interesting, heck it is cheaper then ones that I specked to build.  It would be great to insure a constant temp. though add another thing to power, attach and deal with.  Still a real option. (thanks for sharing that)

 

[Jamesarm97]

Kind of sounds like the same problem we ran into last night on my hot end ;(

I was over at Jetguy's house working on a new printer with a custom hot end and I could feed PLA and extrude all I want but when I clicked print it would jam up by the time the axis homed and bed came up to temp. I would have to reverse the filament out some, then extrude again while pushing down and it would start working but not long enough to print. This was even with a fan blowing across the top of the tube with heat sinks. 

[Sean Trowbridge]

I am following this thread with interest since I am hitting all of these same issues on my 2x, especially when trying to dual-strude.  

(BTW - I did find that leaving the lid off made things a lot better.  Other than that I've tried all kinds of other crazy tweaks as well, like adding a "sacrificial pillar" to my model that prints first on every layer (which helps get the plastic flowing before it gets to the real model every time a toolhead change happens) and changing my local copy of the RepG code's merging logic to go merge layers as LRLRLRLR... instead of LRRLLRRLL...)

At any rate, I'd love to fix this problem "the right way" so if there is any way to improve the extruder design I'm all in.  But I'm surprised to hear you recommend switching to the QU-BD thermal barriers.  I tried to switch my old Thing-O-Matic to QU-BD extruders and I never did get them to work right; it seems to me in retrospect that the reason was this same "heat soak" issue, since the clogging behavior seems very similar.  Have you actually seen improvements from using the QU-BD equipment? What about older Makerbot barriers?

It'd be great if this change helps (since I've got the things already sitting in my garage) but I just wanted a bit of clarification.

[Jetguy]

I just built a custom QU-BD extruder for Mark Cohen last week and the only mod to the thermal barrier tube was adding a wider chamfer to the inlet hole at the top to allow for easier filament loading. I had the thermal barriers side by side a genuine MakerBot one and that was the only difference. It's a very minor thing. They do chamfer for the inlet but it's not as wide of a cone as the genuine version. I just used a standard countersink bit in a cordless drill which did a great job of polishing and widening the cone shape. In all a 5 minute operation. You can buy those tubes from MakerBot through support. I think they charged me $10. I don't know that they will sell you the mounting bar from a Replicator 1 since they quit selling that bot. I mean there are sources around, and I don't know what QU-BD made before but I have no problem saying it's not bad for the price now. Is it plug and plat, well nothing is that I have seen, so I look at cost and material properties along with the fact I don't have to contact MakerBot support and play the week long waiting game to buy something that should just be in the store. 3 years of this, over $10k spent with them, and you get "beyond" annoyed.

 

I don't know about previous QU-BD versions but did hear and they are searchable a few folks had minor complaints in the beginning. I know Mark did say to me he had problems with the "gear" that they use stock for filament feed but I am only recommending the hot end and I did my testing with PLA as a worst case scenario.

 

Again, This is a hotly heated debate and I really do not understand the other side of the argument. All I'm trying to say is an alternative that may fit and may work better. I mean I have no problems EVER with my Replicator hot ends with ABS or PLA and switch back and forth all the time. I built extruders for other based largely on the design and it works for them too. I have studied the temp thing and did some tests and will do some more tests. I did find that coating the threads in the mounting block further increased oozing performance (or lack of ooze that is). I know for a fact, the net effect is that the melt zone is being moved closer to the heater block and manually inserting filament and removing shows exactly where the melt begins and ends.

 

That said, with any extruder I have been using ABS printed extruder spring/bearing idler upgrades since the beginning. I don't even touch the plunger systems and I have a small pile of them. Of the 3 used extruders I've gotten all of them showed groves in the plunger, hence why we know that was never a good solution IMO.

SO yes, in general I think we all are at building and modding are own extruders since none of them come perfect off the shelf.

Let me caveat that with, I need to upgrade my recommendation for the Replicator 2 if and when they start shipping them with the new molded upgrades on them. That is finally the one extruder I can say should be up to my personal specs from the factory and shipping to you and should not require user intervention. That is a good thing and I think only by complaining as much as what has been done here in the forum forced MakerBot's hand at great expense to fix that. So while some folks hate that I complain, take that as one HUGE example what we say here can have an influence on MakerBot improving the product for you guys. That is what this forum is about, helping users have the best printer possible. I think some folks lose sight of that.

[Damian Gto]

I do see your point and I do agree some of it.

The hotend on the rep2x is not the best solution, but it do work good.

But I would like them to add a heat protection under the cooling block, like ceramic tape.

That would remove the heat from the nozzle to affect the cooling block and the filament feeder.

That would be an easy and very cheap way to make it even better.

Like you did see on my temperature experiment I only used aluminium tape and I could see a positive result from it.

It also make sense. Heat do travel up. if there is nothing to stop it it will go up to the filament feeder.

That is how it do look now.

I also feel that small change that has low cost if more easy for MBI to change in the production.

Both this solution and the solution I made on the build plate is very low cost change and they could easy do it if they want to, but the result for the user is a lot bigger.

[MBuser]

One advantage I think there would be to a long production print of the same part, is that once you get the heat adjustments dialed in, it hopefully won't  change much from part to part (compared to if you printed a different part each time).