Effects of static electricity on 3D printer Hot Ends

[ChicagoKeri]

In my various researching of CANbus, I might have found a solution to a problem I was having.

It seems that when I built my Voron 2.4 with a CANbus toolhead, a BTT SB2209, I kept destroying Hotend Thermistors.  Typically, I would complete a 4 or 5 hour long print, and upon the next print I would get thermal runaway shutdown and an erratic Thermistor reading.  Replacing the Thermistor would return the printer to service but again, after a long print, the new one would fail.  My collection of failed thermistors included a couple of E3D 3mm, generic 3mm, an E3D Revo (white wires), a Revo Knockoff, (Red Lizard?) and even a Creality 3mm (worth a shot). Halfway through this pile of busted Thermistors, I swapped out the SB2209 to no effect.   Finally, I installed a MellowFly SB2040 which has not yet eaten a Thermistor.

But why?  BTT might not be the best ever but the boards themselves didn't fail.   The Thermistor connector, P13, has only 2 wires, a well-protected sense wire to the MCU and a ground.  The sense wire has nearly 5k of resistance between it and the MCU with protection in the form of a diode and a capacitor to ground and a diode to +3.3v.  Nothing here should be capable of killing even the most fragile of Thermistors. 

So, today I was looking into CANbus alternatives and on Duet3D's website I found a reference to grounding the metal parts of the Hotend to reduce static buildup!  Ah-Hah!

Supposition: pulling plastic through a reverse bowden and pushing it through an extruder might be building up a static charge on the completely isolated metal parts of the Hotend, which is bolted into a Stealthburner made of plastic.  The only connections to the Hotend are the plastic filament path, the Heater wiring and the Thermistor wiring, which likely has weaker insulation than the heater.  Over time, the charge could build up enough to flash through the insulation of the Thermistor and ground itself to either the Ground at P13-1 or though the protection circuitry of P13-2. If the charge goes through the Thermistor element on its way to ground, Poof!

So, What to do about this? 

Well, the direct answer is to take Duet3D's advice and ground the metal parts of the hotend with a 10k - 10MΩ resistor.  This could be done at a heatsink bolt and a mounting bolt of the SB2209.  This would be a bit messy.

So here's my plan:

Print the Stealthburner parts in conductive ABS that is meant to be static dissipative.  In my case, this would not only provide a path to the CAN Toolboard mounting bolts but also a path to the metal Mellow TAP which in turn is bolted to a metal X Carriage, which is bolted to the Metal Linear Rail Carriage. 

Another option is paint the relevant parts with conductive spray paint, such as graphite containing paint. 

Fortunately, I have access to a High voltage insulation breakdown tester capable of 3,000 volts.  I'm going to zap a thermistor or two and see what they're good for.  Also, see if the conductive ABS I just ordered conducts at elevated voltages.

Any thoughts or ideas?

 

 

 

 

[mvdveer]

20 minutes ago, ChicagoKeri said:

Any thoughts or ideas?

Way above my pay grade ! Thanks for the info - learned something new

[VoronManiac]

IMHO, if it was me, I'd decompose the problem into two separate parts.  Run experiments on the exact existing setup to show that ESD was the cause of the problem. And then come up with a mitigation strategy.  Printing a new set of parts would be an interesting data point, but it potentially introduces new unknown variables and doesn't determine root cause.

If it is indeed ESD, we'd have to ask the question why is this not more commonly observed?  What is potentially unique about this particular printer or printing conditions?  If it was me, I'd start by kludging a ground to the hot end.  It doesn't have to look pretty. 

[ChicagoKeri]

1 hour ago, VoronManiac said:

IMHO, if it was me, I'd decompose the problem into two separate parts.  Run experiments on the exact existing setup to show that ESD was the cause of the problem. And then come up with a mitigation strategy.  Printing a new set of parts would be an interesting data point, but it potentially introduces new unknown variables and doesn't determine root cause.

If it is indeed ESD, we'd have to ask the question why is this not more commonly observed?  What is potentially unique about this particular printer or printing conditions?  If it was me, I'd start by kludging a ground to the hot end.  It doesn't have to look pretty. 

Yes, exactly, that would be the best course of action.  As the weather is dry now the static effect should be stronger .   However, I don’t want to put my 2nd Revo at risk by reinstalling the EBB SB2209. Unless I take the Dragon out of the V0.2 and reinstall a regular thermistor 

there are tools for measuring static charge but I don’t have one. Of course any regular meter’s input impedance will bleed the charge off before it could be measured.

what I do have are tools for measuring dielectric strength and also very low and high resistances which can tell me just how isolated the Hotend actually is… and also just how well various ideas might work.  I am intending to make more posts about it as I proceed.

[YaaJ]

Did you notice the statics when separating the part from the buildplate ?
No idea what the adhesion is. Hydrogen bindings, Van der Waals ? etc. Electrostatic ? Forgot everything and didn't search...
Charges are moving. Could they contribute ?

[ChicagoKeri]

2 hours ago, YaaJ said:

Did you notice the statics when separating the part from the buildplate ?
No idea what the adhesion is. Hydrogen bindings, Van der Waals ? etc. Electrostatic ? Forgot everything and didn't search...
Charges are moving. Could they contribute ?

No, there was never any perceptible static when removing completed prints.  The static  charge is likely to build up on the metal Hotend parts as it is at the end of plastic filament that was just pulled fairly quickly through about a meter of reverse bowden tube, made of PTFE.  Static can easily build up by rubbing insulators together.  If the charge builds up to a Voltage high enough to cross the electrical insulation of the Thermistor cartridge or wiring, it can damage the sensitive Thermistor or control board. 
 

measuring static charge is tricky as most voltmeters would bleed the charge off faster than it could be measured.

[VoronManiac]

Keri,  what was the failed resistance of all the thermistors?

A  potential failure mode to rule out is that thermistor died because of the thermal run away.  We'd need to show canbus communication and FW wasn't  locked up and the over temp wasn't caused by the temperature not being registered correctly.  Canbus is canbus and has to be suspect at the hobby level as compared to industrial and automotive.  Perhaps need to capture a recording of reported extruder power and temperature and sanity check.  There's probably a way to do this with Klipper (I'll let the expert chime in here).  And Klipper should have heuristics to detect a run away if it is receiving the correct data. A sudden ESD spike blowing out the thermistor should trigger safeties withing Klipper.  Can just leave the machine on all the time and at some point something should go poof in the log, while printing or not.  Also if the thermistor measurement noise suddenly disappeared in the log, it could indicate a comm failure or canbus controller latch up.

[YaaJ]

30 minutes ago, VoronManiac said:

sudden ESD spike blowing out the thermistor should trigger safeties withing Klipper

And probaly kill the GPIO input, despite the diodes :

[ChicagoKeri]

OK, off we go on an adventure in Static Hunting!

Let's meet Keri's little helpers...

First, here is a plain old Fluke 112 multimeter. It is a quality 600v rated meter capable of measuring Resistance from 0.1 ohm to 40M (Mega) ohms.

 

Second, here is a Carlson Lab Leakage tester. It is primarily used to check Capacitor Leakage (parallel resistance). It can measure resistance up to 1.6G (Giga) ohms.

      

Notice that here it is showing a resistance reading with the sense lead resting lightly on the vinyl clip cover of the return lead.  The vinyl clip cover is intended to be a GOOD insulator!  This should make an EXCELLENT test for extremely high resistance.

 

Here is a real bruiser... a recently maintained Biddle 305 Hi-Pot insulation tester.  This can deliver 3,000 volts at 5mA, which can easily destroy electronics on a 3D printer.  This can be used to test parts to destruction, to see what the breakdown Voltage is.  However, static electricity can easily achieve much higher voltages so it might be of limited value for these tests.

 

Here is a graphite -bearing spray paint that excels at dry lubrication, SlipPlate.  On the sheet of paper, the dry sample checks at 10k (kilo) ohms per 100mm.  This would make an excellent static dissipative coating for Stealthburner parts.

 

My first test is completed, but that's for a different post....

 

Carlson2.heic

[ChicagoKeri]

Test #1

Test the isolation of the Revo Hotend mounted in Voron 2.4.  After all, if the Revo has any connection to Ground at all, no static buildup is possible and my hypothesis about thermistor failure due to static buildup would be wrong..

Test #1 Results:

The metal parts of the Revo have no measurable continuity to Ground.  My Hypothesis is not DOA! 

The Voron is grounded at the Frame and the Heated Bed. The Fluke reads infinite ohms and the Carlson Leakage Tester read no leakage at its most sensitive setting and highest testing voltage (mica setting, 27v) between the Revo and either the Frame or the Heated Bed.

As the Carlson Leakage Tester has an ability to detect about a 1.6G (Giga ohm) resistance, this indicates an extremely high amount of electrical Isolation.  This test was conducted with ABS filament loaded.

Testing will continue!

Edited February 16 by ChicagoKeri
oops

[DangerBit]

I am very curious to see the outcome here!

[Nurgelrot]

For what its worth I ground all my tool heads that are mounted solely to plastic. Usually I do this to the GND line of toolboard using a 10k. If you've ever felt the static build up on a large print bed after a long print its easy to see why its a good idea.  

Edited February 16 by Nurgelrot

[YaaJ]

5 hours ago, Nurgelrot said:

ever felt the static build up on a large print bed

Yes. Not only on a large bed : the v0.2. But only after removing the part from the bed. Everything tends to stick to the bed or the part. Small pieces of plastic (brim or skirt), paper, dust, etc. Means thousands of volts.

Also have a similar problem with another printer, but it is my fault, and it is not related ; huge mistake : mains powered bed, with the SSR on the neutral instead of the live wire (my bad !). 200V, nothing related to static. Capacitive coupling, not dangerous.

Can't wait for the reults !

[EDIT] a few weeks back, I asked on Rerap forums about "earthing ground". No answers...
https://reprap.org/forum/read.php?1,894130

Edited February 16 by YaaJ

[Nurgelrot]

Right. I'm a RRF guy so use a good deal of Duet3d Hardware. We can argue the merits of firmware all day but nobody disputes the qualityand care taken by Duet3d on their hardware. On all the latest versions of the Duet3d toolboards there is a path to ground provided inat least one of the mounting holes to GND via 100k resistor. That's enough evidence for me... Of course if you have little or no metal in most of your toolhead you need to provide some connective wiring. Usually a mounting screw for the heatsink works well. 

[ChicagoKeri]

Test #2 and #3

Can the mounting bolts of common CAN Toolboards be used as a grounding / earthing connection?

BTT EBB2209 - No. There is no measured continuity between the through-plated mounting holes and Power Negative (-) or for that matter to Power Positive (+).

BTT EBB36 - No. There is no measured continuity between the through-plated mounting holes and Power Negative(-).

Mellow Fly SB2040 - No. There is no measured continuity between the through-plated mounting holes and Power Negative. 

Duet3D 1LC - Yes. All four through-plated mounting holes measure 100k ohms to Power Negative.  This should make an excellent grounding path for the metal Hotend Parts... IF Power Negative has any continuity to Earth ground .

As at least one board has continuity from Power Negative to the board's mounting holes, the next question to test is:

Test #3

Does Power Negative from the system main Power Supply have continuity to Earth ground?

Voron 0.2 - No. Power Negative appears completely isolated from Ground with Power Off.

Voron 2.4 - No, Power Negative appears completely isolated from Ground with Power Off.

Both Vorons shows several Megaohms of continuity between Power Negative and Earth Ground with Power On through their Meanwell Power Supplies There is probably an internal capacitor, but no direct link.

Conclusion:

If one would like a good grounding / earthing path for the metal parts of the Hotend, a separate wire should probably be led up to the extruder assembly, probably with a high value resistor in series.  Another alternative would be to jumper Power Negative to Earth, probably though a resistor, and then ground the the Hotend through the toolhead but that might cause unforeseen problems.

[Dirk]

4 hours ago, ChicagoKeri said:

Mellow Fly SB2040 - No.

So... does this mean, that your theory that Mellow PCB is protects thermistors is not true? or is it the minor difference in the circuits between Mellow and BTT circuits that protect Mellow connected thermistors against all this terrible thermistor-killing-static?

And the next question of course is... what are we going to do so our popular hardware providers (Mellow / BTT/ Fysetc/ MakerPro) are going to add earthing to PCBs like Duet does?

[ChicagoKeri]

3 hours ago, Dirk said:

So... does this mean, that your theory that Mellow PCB is protects thermistors is not true? or is it the minor difference in the circuits between Mellow and BTT circuits that protect Mellow connected thermistors against all this terrible thermistor-killing-static?

And the next question of course is... what are we going to do so our popular hardware providers (Mellow / BTT/ Fysetc/ MakerPro) are going to add earthing to PCBs like Duet does?

 My Theory appears to stand, as the Mellow has yet to eat a thermistor.  As to why this is so, I still suspect that static discharge is the likely cause although of course I could be wrong.  Even though the Mellow does not have an earthing connection to the mounting holes, the Thermistor is connected directly to 24v Negative on one pin and connected through a 100nf capacitor  to the other pin. Even if it did have such a connection to the mounting holes, both it and the Hotend are mounted in a plastic Stealthburner so it wouldn't matter anyway.  My Theory remains that this capacitor reduces the discharge through the thermistor element itself to a survivable level on the Mellow while the added resistance of the BTT board does not. What long-term effect this might have on the toolboard or power supply is beyond my ability to predict.

Of course, it is best to avoid having any static discharges anywhere near a Toolboard, a mainboard or any Thermistors in the first place!  To that end, I am setting about earthing the metal parts of the Hotends in my printers.

An interesting observation:  The BTT Octopus 1.0 /1.1 as well as the BTT SKR Pico all use a Thermistor input circuitry very similar to the BTT SB2209, while the Ultimaker Einsy Rambo and the Duet3D 1LC toolboard both use circuitry much like the Mellow SB2040.

Edited February 19 by ChicagoKeri
oops

[Dirk]

Thank you @ChicagoKeri for explaining it to me so simple-me understands it. 

Also very interesting observation at the end that actually confirms your 'gut feeling'.

I will send an email to BTT tech support today and see if they will read your investigation / research and do something with it.

So if I understand your recommendation at the end of the last message, I can connect a wire, from one of the GND sources on the SB2209 PCB (for example the main GND of the CAN cable) and the other end to a screw on top of the rapido hotend (4 screws are used to fix the rapido heatsink to the plastic SB) with a 100kOhm resistor in between (From @Nurgelrot 's post) ?

Thanks.

Let me write that mail and order a resistor

[ChicagoKeri]

54 minutes ago, Dirk said:

Thank you @ChicagoKeri for explaining it to me so simple-me understands it. 

Also very interesting observation at the end that actually confirms your 'gut feeling'.

I will send an email to BTT tech support today and see if they will read your investigation / research and do something with it.

So if I understand your recommendation at the end of the last message, I can connect a wire, from one of the GND sources on the SB2209 PCB (for example the main GND of the CAN cable) and the other end to a screw on top of the rapido hotend (4 screws are used to fix the rapido heatsink to the plastic SB) with a 100kOhm resistor in between (From @Nurgelrot 's post) ?

Thanks.

Let me write that mail and order a resistor

You're very welcome!

For best effect of grounding / earthing something to DC Negative, in our case 24v Negative, the 24v Negative would have to be connected to chassis / earth ground.

There appears to be a lively debate in other places as to whether or not one should install a jumper between chassis/main Earth and DC Negative.

Some power supplies have an internal capacitor that ties DC Negative to Earth for interference suppression but none appear to tie it directly to Chassis Earth, that decision being left to the user. Capacitors have the effect of passing alternating current through but blocking DC. Of course, such a capacitor would probably bleed off the rapid pulses caused by, oh, say, static discharge arcing through Thermistor insulation...

Until I make up my mind this question, I am proceeding to tie the Hotend metal parts to chassis ground / earth through some resistor, bypassing the toolboard.  While I am at it, I will eventually earth the stepper motors because static can easily build up on the drive belts as well, though it would be far less likely to have a bad effect. My reasoning is that avoiding static buildup at the source is more important than ensuring it takes a specific, contested and debated path.  I might at some point tie 24v Negative to Earth if I find a convincing argument to do so.

It's never a bad idea to have some resistors on hand. Assortments are cheap and readily available.  In this case, static dissipation, a high-value resistor of 10k (10,000) ohms through 1M (1,000,000) ohms will reliably drain off static charges before they can accumulate to harmful voltages while yet preventing large current flow in case of accidental contact with, oh, say, mains power.

PS.  I am not claiming that BTT's Thermistor Input circuit design is necessarily defective. It does appear to protect the board itself quite well.  Neither am I claiming that Mellow, Ultimaker and Duet use a different topology specifically to address this issue. I simply don't know why the engineers chose somewhat different circuits.  

 

Edited February 19 by ChicagoKeri
added PS

[YaaJ]

Finally found the explanation for the resistors on the mounting holes (= EMC reduction if the board is mounted to some conductive plane :  Duet cares, chinese manufactureres don't and save some components)
There's also one on the Altium channel.

 

 

Edited February 19 by YaaJ

[Dirk]

13 hours ago, YaaJ said:

chinese manufactureres don't

This video starts with a non-native english speaker where the introduction takes 2.5 minutes (from the 14). Sorry... lost me after 10 seconds. But thanks.

I think saying that Chinese manufacturers do not care, is more discriminative than my way of selection of videos. On the contrary, the Chinese manufacturers that have supported the DIY 3D printer / CNC  community in the last decade, do care a lot. They are very passionate and cooperative with the community. I think without these "Chinese" I would have to buy my Voron with French, German or US parts meaning my printer would have cost 10.000 euros instead of 1000. 

If you do not care as a company and are not passionate, do not care, I do not believe you will be able to serve such a big population from all over the world, with all their different cultures, customs, habits, expectations from companies. 

16 hours ago, ChicagoKeri said:

It's never a bad idea to have some resistors

I did order 100 100k ohm resistors for 2 euros. I hope to have them in 2 weeks  I will try the suggestion you give with my new Rapido 2 HF on the SB2209 PCB.

I did send an email to BTT support with reference to this thread and the other one where @ChicagoKeri shows her experience with the different CAN PCBs. I have requested to send the request to their 'technical team'. I got an answer that it was forwarded to the BTT engineers and they will have a look at it. 

I do think they care. And I am sure they are going to take this serious. I will let you know about their reaction.

 

16 hours ago, ChicagoKeri said:

PS.  I am not claiming that BTT's Thermistor Input circuit design is necessarily defective.

No. But you have explained meticulously and very clearly the single difference between the two boards.  

[ChicagoKeri]

Test # 4

Conductive coating test

"Just connect a wire to the Hotend!" ... so as to provide a conductive path for discharging static to Earth

Has anyone ever tried this on a Stealthburner? There is no place to attach a wire to the Heatsink or hotend, at least not for a Revo Voron. Nor would this be an easy attachment to create.   While the initial testing of the "conductive" ABS appears to indicate no apparent conductivity at lower voltages (further testing later),  I have a promising conductive spray paint and another that looks nearly identical when dry.

You see, one can apply conductive paint to the plastic parts, thus defining a conductive path from the Heatsink mounting bolt area to a handy heatset insert like for the accelerometer mount or just to the metal CNC TAP. 

The wider piece of ABS indicates 13k ohms per 100mm.  The narrow piece indicates infinite ohms per 20mm. 

 

The culprits:

The "wide" of course was painted with SlipPlate graphite spray paint.  The "narrow",  with  Rustoleum Ceramic.

         

 

So, here is a Stealthburner Revo Voron Mount that has been sprayed in the relevant areas.

Resistance reading from the Heatsink to the uppermost heatset insert for the accelerometer mount is 38k ohms.  This would make a good place to attach an earthing wire.  Alternately, the resistance reading from the Heatsink to the upper attachment points to the metal Mellow CNC TAP is 10k ohms, so possibly attaching the wire to the TAP would be better.

So, for now, I'm going to Earth the Hotend in this fashion.  It's as easy as disassembling the Stealthburner, giving a quick flashing of SlipPlate and re-assembly with a wire running down to Chassis Earth.

 

If nothing else, remember this...

[Dirk]

8 hours ago, ChicagoKeri said:

Has anyone ever tried this on a Stealthburner?

I did

The rapido hotend is fixed in the printhead with four screws.

When you test continuity on the extruder from the heatsink to the tip of the nozzle, there is none. I think because of the coating on the heatsink. But th screws are continuous with all metal parts of the extruder, including the Nozzle. So I hooked a cable to one of the screws on top.

The black wire... The red and yellow/blue wires are thermistor and heater cables.

Then I soldered the other end of the cable to a resistor I had... A 120 ohm... The 100kohm ones aren't here yet. It was to see if it was possible...

And I led it to the SB 2209 and looked for an easy GND port. The biggest one was the one where the can cable plugs in...

Not printed anything yet... Will do so when the 100k ones arrive.

But I like your paint idea better.

[ChicagoKeri]

My Hotend Static Solution (so far) 

OK, I'm pretty sure I've got the Revo well Earthed and unlikely to zap anything even though the Mellow SB2040 has been great in this regard. I feel that it is important to reduce the chance of static buildup as I intend to do some long-ish prints soon, so I did this before further testing.

I went and re-printed all of the Stealthburner parts except for the front cover and applied SlipPlate graphite spray paint to everything I though could possibly stand to be Earthed.   This includes Idler and driven gear bearing holders for the Galileo 2.  I used the "conductive" ABS filament for most parts because it printed exceptionally well with well toleranced parts... I previously had issues with the original printed parts which needed hand fitting.

I'm really glad I did that, because upon disassembly I found a couple of problems with the Galileo2, namely the Carrier Shaft Bearing Retainer (upper right corner of picture above) had cracked, there was insufficient end play of the Carrier Shaft because I had installed the supplied shim without checking the end play and the planetary gears really needed a bit more grease.  

Also, the Mellow TAP had one magnet come unglued, so I glued it back  and the Mellow TAP LDO Linear Rail was all stiff and crunchy so I took it apart, cleaned it out, re-greased it and assembled the carriage in the opposite orientation which seemed  to make it a bit happier.  Naturally, Quad Gantry Leveling seems happy about that.

 

All put back together and now I have a direct path to Earth. 

The 10k ohm resistor is attached to the Upper Accelerometer bolt, which measures 6.2k ohms to the nozzle for a total of about 16k ohms. The thin black Earth wire snakes back up  and attaches to Earth.   I did check, and the 24v Power Supply does appear to have about a 30nf capacitor between 24v Negative and Earth, so one could probably just Earth it to 24v Negative at the Toolhead instead.

 

 

[DanPin]

This is a interesting thread.  What troubles me is that in the orig. post Keri states that the longer prints complete without a problem.

But when she starts the next print, the thermistor is bad.

If it was a static problem from plastic flowing through the extruder, the thermistor should have blown sometime DURING the print.

It seems like the problem happens between the last print and the start of a new print.

If you ever go back to the EBB2209, you should check temperature control after a long print is finished to see what that tells you.

The solution may still be adding a ground path, but the source may well  be elsewhere.

Also, why don't we hear about more thermistor problems in general, maybe something is unique to her printer.

Edited March 7 by DanPin