Cubex Trio Autopsy

[jwk...@gmail.com]

Why do they drink warm beer in England? ............ Because they use CubeX Refrigerators.

Bad joke, but seems consistent with the experience of many users.

I'm a Electrical Engineer with extensive experience designing high-end, specialized CNC machines. This is my first experience with 3D printers.

I was presented with a killer deal to purchase a near new Cubex Trio with 7 cartridges and other supplies from a local start-up for pennies on the dollar. They were in a time crunch to get their product developed and quickly moved to a professional machine after getting unsatisfactory results from the CubeX. I read the negative reviews on Amazon, but felt confident that I could develop solutions to the stated problems. They did do a fantastic job in the aesthetics. The CubeX is one of the nicest looking machines out there.

The personnel(3 people)at the seller's company were not technical. They stated their issues with this machine as poor print quality, a fan blade came off and Extruder 3 wasn't working.

The posts in this community have helped considerably in getting me up to speed.

Here are my findings and impressions to date.

1. The broken fan blade was apparently caused by someone resting the machine on the corner of a bench. There is no protection to stop this since the fans aren't protected on the bottom and there is no way for a person to know that the whole bottom isn't solid.

2. As I began to replace the 80mm fans (Used fans with LEDs), I was confronted with their goofy mounting system using push-on, rubber “noodles”. It was impossible to salvage these things intact during removal.

3. While replacing the fans, I was amazed by their cooling strategy that blows cooling air on the back side of the circuit boards, instead of the actual components. I plan to install some additional fans to keep the components cooler.

4. The power system is cheaply done. The main power input to the board is about 15V. This passes through a 7812 regulator to obtain 12V. The output of the 7812 then cascades to a 7805 regulator to develop 5V. The regulators do not have heat sinks and do tend to get hot. The main supply is held to the bottom with two tie wraps. Seems to work OK, but it just seems cheap for a $4000 machine. Also, no On/Off switch.

5. Working on the interior components, I noticed that the print table could be lifted with no resistance. The NEMA 23, Z axis motor was not bolted to the frame. It was only held in place by gravity. I am the second owner of this machine, the original owners could have done this, but there were no tool marks around the motor's mounting holes. Sure seems like it came that way from the factory.

6. After resolving these issues, I felt confident to power the machine up. Got KiSSlicer and CubeIt and started doing some test prints. Found it interesting that the cartridge work-around solution was in the project G-Code and not in the system firmware. Want to give thanks to the people who figured this out and made it available to the community. Sucks that Cube encrypted the output of their software so users cannot view the print code before loading it into the machine.

7. I read many articles about people using heated beds for ABS. Didn't fully appreciate the problem until the flat bottom of a small gear I was printing slowly turned into a parabola. Their Magic Glue needs a little extra Fairy Dust. Seems like a heated bed should be included in a $4000 machine. I have one planned.

8. They use a $4 reed switch to do precision position alignment. The placement of this device is in the middle of what appears to be a very noisy section of the machine. I plan to do a relocated sensor using a Hall-Effect switch which is noise tolerant and has superior repeatability.

9. At this point, I decided to tackle the problem with the third extruder. I was able to successfully print from this extruder using bulk filament. The problem was most likely in the cartridge reporting electronics. I removed the cartridge bracket for the third position and discovered that the circuit board underneath contained a DS28E01 chip, the same one that is used in the cartridges. Anyone know why this chip is included here? Position identification?

10. Here's where everything went to Hell. After reassembling the cartridge holder, I inserted a new cartridge into it with the power on. The whole machine went dark. Double checked everything I had done and could not find anything I had done wrong. Could not get it to come back to life.

11. Took the machine apart and started examining the main board. All voltages were OK. Only one of the two red LEDs would light when power was applied. Contacted the company for a replacement board. Their ship date was six weeks out.(warm beer for six weeks) Given the symptoms, it appears that the most likely component is the controller chip. I ordered one, $9, should be here in a few days. Will report back regarding the success or failure of the chip replacement.

12. It appears that a good portion of the problems encountered with this machine are due to the company's strategy to force their customers into using their filament.

[Gary Muir (GRM Products)]

Nice write up! The main board issue is well known to be very sensitive. Rodney Wells has done extensive work on these boards and offers a repair service but with your knowledge I'm sure he can point you in the right direction as you speak the same language. He's in Australia so he'll likely chime in later today. 

[Peter Gregory]

I agree with you about the design choices made building this product.

It's pretty, but the electronics need help.

The main board design reminds me of several home projects I made in the 80's.

Why use linear power regulators when switching regulators are so much better - more amps, less heat.

Dropping 15v -> 12v -> 5v = lots of dissipated heat.

You will notice a world of difference with a heated bed.

Any 3D printer I buy in the future must have one.

You should check out Gary's company.  He has a great heated bed system for the Cubex.  It re-locates the Z-Gap sensor too.

I bought one about a year ago and it's been working great!

[Rodney Wells]

Thanks Gary. Did you get the circuits i sent you.  I sent you my Z axis opto switch up grade I know you have a magnetic relocation upgrade but This is an improvement. 

I would be very interested in your friends assessment of the plugin board.      

[Rodney Wells]

Make pain your friend. lol 

Here we are again.

You are a professional electronics person.  So straight to the point.  The best way to make this into a good reliable printer even a great printer  with out much fuss and the  Marlin open source is not the easy way. 

Re flash the MB controller with the 5.4.2 firmware and make a few rewiring changes you can then print away until your hearts content.  

I have spent almost a year now re engineering there main board The only thing they did when they released  the CubX was to interlock it with there cartridge system. They made no attempt to make it a better printer.

Everything you said in your write up is spot on. I still think the over all deign of the machine is the best its the software they wrote to prioritise it. That  #### it up.     
Please look at some of my post if that can help. I am sure you have your own ideas on how to improve this beast we would love to hear them.

 
Oh the Z stepper is bolted down.  

[Gary]

Hi, I did get the shipment but haven't had a chance to work with it.

 

              Gary Muir

 

              GRM Products

              Hudson NH (USA)

              1-978-293-6741

              http://www.GRMProducts.com

              Heated Bed Install Video link  http://youtu.be/O54pmUfaIvk

--
You received this message because you are subscribed to a topic in the Google Groups "CubeX 3D Printing" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cubex-3d-printing/QIELNShv0cs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cubex-3d-print...@googlegroups.com.
Visit this group at http://groups.google.com/group/cubex-3d-printing.
For more options, visit https://groups.google.com/d/optout.

[Joe Kuter]

Hey Rodney

I changed the controller chip. Initial indications were good in that both LEDs on the controller were lit and the LED on the display panel was lit.  They were dark before the swap.

Only connected the display and USB connectors.

Followed your instructions and downloaded the 5.4.2 firmware suite.  Connected the USB cable and and started the BFB Firmware loader.

The computer doesn't detect a device on the USB Port and the firmware loader indicates "device not detected" 

Re-examined my solder connections to the chip.  They look OK.

The controller specs indicate a dedicated DMA channel for USB, but don't know if some activation procedure is needed for a new chip.

Did I miss anything?

[Rodney Wells]

If you have changed the processor you are starting with a new chip and its complexly blank. there is no boot loader. The USB on the go is not present. 

With out the boot loader you cannot program it via the usb port.  Can you do an  In-Circuit Serial Programming™ (ICSP™)?

They have a security kind of thing going on as the new system needs to be present at the correct location in memory.  all I can do to help is send you a programmed chip.  

Its not fun programming a  64-Pin TQFP device when its not on the board. but if that's what you need it can be done. Or get a ICSP. 

[Matt Kusak]

On Wednesday, September 10, 2014 8:34:07 PM UTC-4, Joe Kuter wrote:

9.  At this point, I decided to tackle the problem with the third extruder. I was able to successfully print from this extruder using bulk filament.  The problem was most likely in the cartridge reporting electronics.  I removed the cartridge bracket for the third position and discovered that the circuit board underneath contained a DS28E01 chip, the same one that is used in the cartridges. Anyone know why this chip is included here? Position identification?

Hi Joe,
We had issues with the chip reader contacts and had to get a replacement..  We tried to swap one into another position (just to keep operational) and controller did not allow it to work - We were told that they are position specific..  So I am assuming that the chip is used to identify the cartridge position..

the hardware when working provides very good output - but as you had seen the electronics are it's biggest weakness.  Our Z Gap calibration has never been consistent.

look forward to any fixes you can offer.  Our second controller board now appears to have failed..

[Peter Gregory]

Three things you can do to greatly increase the quality / repeatability of prints:

1) Get a heated bed.  I saw a world of difference from the stock platform and a heated bed platform.  Much less hassle and no glue needed printing to kapton tape.

2) Relocate the Z-gap sensor (using a kit) to another location.  Usually to the left rear column.  This moves the z-gap from the electronic noisy print assembly.

3) Use the home twice before print / z-gap setting.  This moves the bed to a known location before the print / z-gap procedure begins.  This gives repeatable bed placement.

[C.Scheers]

On Wednesday, September 10, 2014 7:34:07 PM UTC-5, Joe Kuter wrote:

10.  Here's where everything went to Hell.  After reassembling the cartridge holder, I inserted a new cartridge into it with the power on.  The whole machine went dark.  Double checked everything I had done and could not find anything I had done wrong.  Could not get it to come back to life.

 
Based on multiple comments similar to this in this forum, and my personal experience with a CubeX Duo with a bad mainboard, I have made a personal decision that I will NEVER insert or remove a filament cartridge with the power on.  If I am going to have a cartridge in the cartridge slot, I always turn the power off first.

Perhaps related to this (I strongly suspect that it is), on the cartridge slot circuit board there is a Zener diode (ZD1).  I've noticed that on both my cartridge slot circuit boards the Zener diode appears to be blown open.

My hardware skills are not world class, so it is possible I am doing something incorrectly, but I measure a hard open across ZD1 in both directions with my meter.  From what I have read, if ZD1 was working at all, I should measure some resistance across it.  If I am incorrect, please correct me.

Again, please correct me if I am wrong, but my understanding is that one use of a Zener diode is to shunt spikes to ground.  Assuming that ZD1 exists in this circuit to shunt spikes that occur when a cartridge is inserted/removed, I hypothesize that eventually ZD1 gets destroyed by a spike and then a spike gets sent down the line to the controller board, destroying that.

Comments?

[Rodney Wells]

Its a voltage regulator, Very basic and going to fail. What you get is full power supply on the main board 

See photo.

[C.Scheers]

On Tuesday, September 16, 2014 3:50:55 AM UTC-5, Rodney Wells wrote:

Its a voltage regulator, Very basic and going to fail. What you get is full power supply on the main board 

See photo.

So what is the effect when it fails?

Should I replace it?  Is there something better that could be put there?

Has anyone else noticed this failed on their machines?
 

[Jetguy]

Simply if that board fails, it feeds the 15V down the 3.3 volt logi pins and blows the mainboard.

Why they didn't use 3.3v power to the circuit in the first place is beyond me.

 

So yes, this is the easy explanation for all of the failures. The circuit is a time bomd waiting for the right trigger to fail and kill your mainboard. Stock firmware will not run with the sockets unplugged so you are kind of stuck.

 

Now that I know this, I would definitely mod the filament spool socket boards and either put a proper 3.3 v regulator or even more smartly, borrow 3.3 volts at the mainboard and power from there.

[Jetguy]

I'd go further than that even, and remove the board physically from the metal socket housing and tuck away undereneath sealed in a shrinkwrap tube.

The idea here is most folks have migrated away from using chipped cartridges and using the Kisslicer and CubeIt which then doesn't need the socket plugged into a cartridge. The mainboard must still see the socket board but put it away from static discharge and other user perils.

 

FYI, what they did is put secure eeprom ICs on these boards to try and prevent the use of chipped filament under the premise of helping users by using the secure eeproms to tell the machine plastic type and temperatures.

I won't call it evil, they had a plan to make it easier but at the same time, they did things to protect from 3rd parties making cartridges and make is really hard under their system to make a device to go into the socket or replace the socket and fool the mainboard. What they didn't account for is all the pretty plastic of the machine is a problem. The metal cartridge holders are isolated electrically from the frame and ground by the bottom plastic plate. That means it's quite easy to create a huge static charge while inserting a cartridge or touching the machine in general. Eventually, all it takes is that discharge to strike that cartridge socket board and blam it fails and sends 15 volts down to a 3.3volt rated processor.

 

So, if you want to run chipped filament, PLEASE ground the metal cartridge holders to the frame and potentially to an outlet. Powering down the machine while changing a cartridge will NOT eliminate the potential for damage. ONLY grounding the now electrically isolated metal brackets will help with the static issue.

[C.Scheers]

As far as I can tell, the EEPROM on the board does NOT specify anything about the filament.  The EEPROM in the filament cases does that.

When you change the mainboard, the activation code does not change.  This indicates that the serial number is stored in the EEPROM on one of these boards.  I assume it is stored on the slot 1 board.  I also believe that this EEPROM says whether the machine has 1, 2, or 3 heads. Note on the Cubify order page that there are distinct orderable items for a CubeX slot 1 holder, depending on whether it is a CubeX, Duo, or Trio.

I think that the problem seen when a mainboard fails is that the I/O pin that reads the slot 1 EEPROM is blown, so the mainboard can't read the EEPROM to figure out what kind of printer it is, so it refuses to initialize.

I'm kind of intrigued by the suggestion to just remove the board from the holder.  This would work for normal operations, but sometimes it is necessary to print a .CUBEX file.  In particular, when printing the test pattern for the offset between jets.  Does someone know of a comparable alternate test pattern?

As far as grounding the cartridge holder, it may already be grounded to the mainboard through the mounting screws.  I'll have to check that the next time I have one out.

Grounding the cartridge holders to an outlet does NOT help as the mainboard is not grounded.  If you are going to ground the cartridge holders, please make sure that the mainboard is grounded to the same point.


So, to keep the capability to print .CUBEX files, should I replace the blown ZD1s on my cartridge slot boards?

[Rodney Wells]

YES that's it.  Built in failure system.   The processor goes up in smoke.  As I said very basic regulator.  A shunt load system.  its a voltage regulator a child would make.

[Rodney Wells]

Have you changed the chip on the main board?  how are you going to re program the MB? also put 5.4. 2 in.  Give It a try. You will be surprised at what 3D system really did.

[C.Scheers]

I got another mainboard from Cubify.  Are you still looking for bad boards?

What I am wondering is if I should replace the blown ZD1 on my cartridge slot boards?

[Jetguy]

If you plug in a blown cartridge slot board into that brand new mainboard, you will likely kill the new mainboard.

Replacing the ZD1 is not a fix. Understand that if it failed, it already sent 15 volts to the other ICs on the board and thus they are blown.

Again, if you have a situation where it's obvious they are already blown with visible burn marks, it would be downright asking to blow the mainboard to plug those in. And just because the failed ZD1 is all you know about, I wouldn't trust that board again by just replacing ZD1.

On Friday, September 26, 2014 3:22:50 AM UTC-4, C.Scheers wrote:

[Matt Kusak]

Rodney are you still offering the MB repairs / Modifications?

IF so I have a failed MB that I'd like to send to you for you to repair and load with 5.4.2.

[Rodney Wells]

Yes I do . Send it over and it will be back in your machine in about week or so. 

[Rodney Wells]

Hello.

I been on holidays the last  week or so. haven't been on the forum. Back now.  

 

Please send to, 

 

Rodney Wells

11 Stonehaven cls.

Sunbury 3429

Australia.

[rossc...@gmail.com]

Rodney,

Are you still repairing these boards? I've ordered one from Cubeify already but i may as well get this one fixed to.

Thanks

[cu...@tactustechnology.com]

Rodney, It has been a while since you posted anything here. Are you still performing repairs for the Cubex Main Board? I have a nice Cubex Duo with heated bed that has only one led lit on the Main Board, even after disconnecting everything but the first module reader.