search

DoESLiverpool / somebody-should

A place to document practices on the wiki and collect issues/suggestions/to-do items for the physical space at DoES Liverpool
31 stars 11 forks source link

Creality CR10-S printer is failing with "thermal runaway" #1163

Closed ajlennon closed 5 years ago

ajlennon commented 5 years ago

The bed is heating and it seems to be reading a temperature increase.

But after a while the system halts with a "thermal runwaway" error message.

Also the heating bed connector at the rear of the main PSU box is getting very hot.

Others have apparently seen this problem

https://www.reddit.com/r/CR10/comments/8a6s05/is_it_normal_for_the_bed_cable_connector_to_get/

I've contacted the vendor for a response

Thingomy commented 5 years ago

Hot bed connector was found to be underrated (either 5A or 1A) when it was taking 10A. It turns out the connector was getting hot enough to melt it's solder, and was being held together by heatshrink. Connector has been replaced with bullet connectors.

Box was reassembled and machine was fired up, whereupon a clunk was heard from inside the box, this turned out to be the hotbed mosfet falling off due to overheating melting it's solder. It was soldered back on on the offchance it was a one off; it wasn't a one off. I'm 95% sure there is no short in the hotbed.

Machine is currently in pieces awaiting either somone that knows thier way around mosfets to prod it, or me to bring in a bunch of additional tools.

zarino commented 5 years ago

When I hear the word "mosfet" I immediately think @magman2112 – pinging him here in case he has any expertise to lend!

ajlennon commented 5 years ago

Hehe. Yes @magman2112 was involved with @Thingomy in the investigation.

I've had a response from Ristar, whom I bought the printer from via Amazon

Dear Customer,

Thanks for your contact. We apologize for the case you meet.

Could you please provide us more specific information about the problem? you can also take some photos or a short to show it, then our technical will specify the problem. And we will according to the reasons do the appropriate treatment. We appreciate your cooperation.

Looking forward to your reply.

Kind regards

My response:

Hi,

Thanks for your response. We've done some more investigation here and it appears as though you are >putting more current through the heating bed connector than its specification allows.

Notes:

"Hot bed connector was found to be underrated (either 5A or 1A) when it was taking 10A. It turns out >the connector was getting hot enough to melt it's solder, and was being held together by heatshrink. ... >Machine was fired up, whereupon a clunk was heard from inside the box, this turned out to be the >hotbed mosfet falling off due to overheating melting it's solder. It was soldered back on on the >offchance it was a one off; it wasn't a one off. I'm 95% sure there is no short in the hotbed."

I am not in the office today to provide pictures but the issue appears the same as on this page. So it >looks like others are seeing this problem too.

https://www.reddit.com/r/CR10/comments/8a6s05/is_it_normal_for_the_bed_cable_connector_to_get/

I like this printer a lot and have been recommending it to colleagues in my workspace, but I am >concerned about the issues above. Please help to resolve them so I can continue to recommend you >guys!

Thanks,

Alex Lennon

image

Will keep this issue updated with their response

DocTrucker commented 5 years ago

I'd mentioned about the hazards the Anet printers on the DoES Liverpool google mailing lists last december and the Creality machines and most at the same price break are very similar. Even the much more expensive Up Printer has very small CSA wire for the heaters.

The creality machines at least used to ship with a version of Marlin with thermal protections disabled, thankfully that has either changed or you caught it.

It's note worthy that the Ultimakers are one of the only suppliers to say their machine can run unattended on the CE certification.

All that said I'm working on the safeties on my 'small holding' at home to save any potential grumbles from the insurance company. Passive saftey (ie heaters at 100% duty never reach a dangerous temperature) is first, but is generally frustrating. Firmware safteys are the next option, and finally thermal fuses switches and fuses mounted on the hot zones can save your building if it all goes belly up.

As with most electronics it's the negative leg that is modulated for heater control so you can feed the posative side of the heaters seperately to the rest of the board (tied negative poles) and interlock that PSU on faults. Most boards have a PS_ON pin which can be used to drop relays. If you go down that route remember SSRs and MOSFETs are not safety devices.

jrp commented 5 years ago

I'm planning to be in tonight with the right tools to to do some more diagnostics on this.

DocTrucker commented 5 years ago

Expect I may be stating the obvious to you but have a good look over the pluggable heater contacts for signs of arking and check gauge of wire against the expected current draw. If the creality is the machine I'm thinking of it's got (or previous models had) an aluminium heatbed. It's worth ditching the connector and soldering wire (maybe new) directly to the bed and fully stress shield the joint. If you do this preheating the aluminium plate on a gas stove helped me. I needed a gas powered soldering iron to get enough heat to work with high temp (silicone insulated) 14awg, but that is potentially overkill for a standard heatbed. Check thermistors are secure too where they are removable.

Thingomy commented 5 years ago

I had a look tonight at the mosfet arrangment to switch the heated bed, it is an off the shelf board that effectively emulates a relay. It is officialy goosed. That tends to happen when parts get hot enough to melt off. I replaced the failed mosfet with a much lower rated one for the purpose of testing whether the board worked, and it still does nothing.

So we need a replacement for this board.

As for how it died, I'm puzzled. The Mosfet is actually well over rated, (98A is mentioned in the datasheet, which is about what I'd spec for driving a 10A load), and the connectors and wires in the area look up to spec too (other than the one that started all this). It's possible I did something to it during the other fix, but I can't work out what. I even confirmed I never reversed it's power supply, which would have had this effect.

I'm now unavailable for a few weeks, so I'm going to have to hand this one off to the next interested party.

PS. I left a bag of mosfets in the area, could someone either keep them with the machine, or stick them in my box please?

PPS. I got as far as to work out how the mosfet overheated so fast. It failed mostly conductive, at 0.18ohms (including some of the wiring) so as soon as the PSU came on, it imediately outputs a measured 7.8amps to the hotbed; and so runs at 13ish watts, and the heatsink fails to disapate it. It's possible that during it's failing process it's resistance was higher, producing more heat.

DocTrucker commented 5 years ago

Failed closed circuit is a common failure mode of MOSFETs & SSRs. This is why Omron phorbid the use of their devices in safety critical circuits where there is no backup to stop the flow of current in the event of failure.

While they are great for high frequency pwm without having to worry about inductive kick back from the coil, or electrical noise from contact arking, they are crap at dealing with transient voltages that mechanical relays hardly notice.

In automotive use their circuits tend to be built with TVS (transient voltage suppression) diodes and other over voltage and crowbar like protection devices. Designing the protection isn't straight forward as you have to have some idea of the spike that you're protecting from. In the DoES case it could have been created on the device, in it's power supply, come from the mains, lightening within a mile or so, or someone else in DoES recent creation maybe accidentally very electrically noisey. ...or just a poor quality MOSFET board.

Long story short buy your SSRs or MOSFETs boards from reputable sources, ensure there is an additional mechanical relay to kill power to the heaters (on dc or ac side of heater psu), and consider a thermal fuses.

On Wed, 3 Jul 2019, 23:20 Thingomy, notifications@github.com wrote:

I had a look tonight at the mosfet arrangment to switch the heated bed, it is an off the shelf board that effectively emulates a relay. It is officialy goosed. That tends to happen when parts get hot enough to melt off. I replaced the failed mosfet with a much lower rated one for the purpose of testing whether the board worked, and it still does nothing.

So we need a replacement for this board.

As for how it died, I'm puzzled. The Mosfet is actually well over rated, (98A is mentioned in the datasheet, which is about what I'd spec for driving a 10A load), and the connectors and wires in the area look up to spec too (other than the one that started all this). It's possible I did something to it during the other fix, but I can't work out what. I even confirmed I never reversed it's power supply, which would have had this effect.

I'm now unavailable for a few weeks, so I'm going to have to hand this one off to the next interested party.

PS. I left a bag of mosfets in the area, could someone either keep them with the machine, or stick them in my box please?

PPS. I got as far as to work out how the mosfet overheated so fast. It failed mostly conductive, at 0.18ohms (including some of the wiring) so as soon as the PSU came on, it imediately outputs a measured 7.8amps to the hotbed; and so runs at 13ish watts, and the heatsink fails to disapate it. It's possible that during it's failing process it's resistance was higher, producing more heat.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=AHL52BBPWNZLQ5ULCGEZI4DP5UQ4HA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODZF2MMI#issuecomment-508274225, or mute the thread https://github.com/notifications/unsubscribe-auth/AHL52BFPWZONJGRWMM3UZHLP5UQ4HANCNFSM4H4MIPHQ .

ajlennon commented 5 years ago

Ristar haven't come back to me.

Have written them a bad review and requested a refund. Will see what happens...

Sean-anotherone commented 5 years ago

https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20190729115336&SearchText=creality+board

ajlennon commented 5 years ago

Yeah we might end up going that way @Sean-anotherone. Can you recommend a board for the Creality CR-10S ?

DocTrucker commented 5 years ago

https://forum.duet3d.com/topic/10542/converting-creality-cr-10-s5-to-duet-wifi-24v

On Tue, 30 Jul 2019, 09:21 Alex Lennon, notifications@github.com wrote:

Yeah we might end up going that way @Sean-anotherone https://github.com/Sean-anotherone. Can you recommend a board for the Creality CR-10S ?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=AHL52BBMDZQGBJUO7DPSJZTQB72YNA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3DGCHQ#issuecomment-516317470, or mute the thread https://github.com/notifications/unsubscribe-auth/AHL52BFAMBVPQSQN6WB4CDTQB72YNANCNFSM4H4MIPHQ .

ajlennon commented 5 years ago

Still no response from Ristar so escalated to Amazon. They have agreed a full refund on return of the Creality printer so it needs packing up for a pickup

ajlennon commented 5 years ago

Say goodbye image

ajlennon commented 5 years ago

The failed unit should be picked up today. I've ordered a replacement (same model) which should be with us by end of week.

ajlennon commented 5 years ago

OK the replacement Creality CR10-S has arrived @jackie1050

I have

I'm just doing a test print but feel free to have a go at bed levelling and some prints yourself.

I'll be in tomorrow with a bit of luck to chase up why Amazon have still not picked up the other unit and to move things around to take this one off the workspace desk...

ajlennon commented 5 years ago

Had a quick go at printing. Seemed to be going well then my test cube came unstuck. Base probably just needs a good clean

JackiePease commented 5 years ago

Sounds good - I won't be in until tomorrow about 1pm but looking forward to trying it out.

Get Outlook for Androidhttps://aka.ms/ghei36

From: Alex Lennon notifications@github.com Sent: Friday, August 16, 2019 4:11:41 PM To: DoESLiverpool/somebody-should somebody-should@noreply.github.com Cc: Jackie Pease jackiepease@hotmail.com; Mention mention@noreply.github.com Subject: Re: [DoESLiverpool/somebody-should] Creality CR10-S printer is failing with "thermal runaway" (#1163)

Had a quick go at printing. Seemed to be going well then my test cube came unstuck. Base probably just needs a good clean

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=ADBVRUB5ICFXNYMR3J7JZCDQE27S3A5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4O4GAA#issuecomment-522044160, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADBVRUHDCCLKPBMVJZ45VKDQE27S3ANCNFSM4H4MIPHQ.

RussCoty commented 5 years ago

Looks good. Is it going to live on that worktop where the other one was? Or is there room near the ultimakers?

JackiePease commented 5 years ago

The space near the Ultimakers is really a desk - we're aiming to return it to that use once we've finished the large filament reel that won't fit in the usual space (and once Felicity's fixed). So probably it will go where the other one was - unless there's a better plan?

Get Outlook for Androidhttps://aka.ms/ghei36

From: RussCoty notifications@github.com Sent: Friday, August 16, 2019 8:25:16 PM To: DoESLiverpool/somebody-should somebody-should@noreply.github.com Cc: Jackie Pease jackiepease@hotmail.com; Mention mention@noreply.github.com Subject: Re: [DoESLiverpool/somebody-should] Creality CR10-S printer is failing with "thermal runaway" (#1163)

Looks good. Is it going to live on that worktop where the other one was? Or is there room near the ultimakers?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=ADBVRUF5HCJZDVYIJ5JSP53QE35JZA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4PPUQI#issuecomment-522123841, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADBVRUE62Y6LILTGOLRSRR3QE35JZANCNFSM4H4MIPHQ.

ajlennon commented 5 years ago

Good plan. The setup went better than I expected. I’m quite annoyed Amazon haven’t picked up the original one yet. I’ve been onto them a number of times. Will pursue them again tomorrow and move the pick up package somewhere, then move the new one to where the original one was and we can get back to 3D printing :)

JackiePease commented 5 years ago

If it's staying at DoES you should give it a name

Get Outlook for Androidhttps://aka.ms/ghei36

From: Alex Lennon notifications@github.com Sent: Friday, August 16, 2019 8:50:00 PM To: DoESLiverpool/somebody-should somebody-should@noreply.github.com Cc: Jackie Pease jackiepease@hotmail.com; Mention mention@noreply.github.com Subject: Re: [DoESLiverpool/somebody-should] Creality CR10-S printer is failing with "thermal runaway" (#1163)

Good plan. The setup went better than I expected. I’m quite annoyed Amazon haven’t picked up the original one yet. I’ve been onto them a number of times. Will pursue them again tomorrow and move the pick up package somewhere, then move the new one to where the original one was and we can get back to 3D printing :)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=ADBVRUG3TSTUDG47PQYCTHLQE4AGRA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4PRJVA#issuecomment-522130644, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADBVRUG56VE7J3A6P7Z2G53QE4AGRANCNFSM4H4MIPHQ.

ajlennon commented 5 years ago

Oooh yes. Well hopefully this won’t blow up. I meant to say let’s keep checking the heat on the connectors.

So what ideas do people have for the new printer? Throwing it open...

ajlennon commented 5 years ago

CAD66DAF-67A6-4BEB-BBF7-6B3B935ED92D

ajlennon commented 5 years ago

Checked the back connectors and they’re not hot at all

jrp commented 5 years ago

Could I bother you for a picture of the offending connectors on the new machine? I'm intregued as to how they have redesigned it.

JR

ajlennon commented 5 years ago

Looks same as before. Nice and cool at the moment

735FC47D-1C44-41A1-BB03-FBCBBA2B1934

ajlennon commented 5 years ago

So a bit more of an update from yesterday's attempts to print.

I had a lot of failures. Mainly the plastic wasn't sticking to the glass surface. I tried cleaning it and I tried a few different types of glue.

I wonder if the head is dragging against the plastic at the top.

I did get one "boaty" print which was really really good so it's possible but it's just a bit temperatmental at the moment.

Perhaps those who are more expert in these matters @Sean-anotherone or @jrp could cast their eye over it...?

DocTrucker commented 5 years ago

Manually level the bed as best you can (system doesn't need to be on) to give the probe based bed level correction a fighting chance.

A clean glass bed (IPA, Ethanol, or Acetone should be fine) should be all you need for PLA, but basic, cheap hair strong hair spray may give you a.little more of a fighting chance.

Heat the bed upto temperture if required and leave for half an hoir or so. Run a bed probe and store the correction mesh. Not sure how to view the bed level correction results on Marlin, but check to ensure it doesn't show any evidence of twist about the y axis. Ideally you'd be best off manually levelling the bed to the point where bed probing isn't required, but the machine may not be stable enough to do that.

CAD the largest x and y dimension cube you can. Slice it with a a grid fill density low enough to give 10-20mm spacing of the hatch lines. Print only the first layer and look at the width of the lines over the bed. Slight variations or errors will bee shown up quite clearly by the lines becoming thicker or thinner. If all consistant and your wall thickness for a single wall cube is on target then it's something a little more in depth.

On Sun, 18 Aug 2019, 14:04 Alex Lennon, notifications@github.com wrote:

So a bit more of an update from yesterday's attempts to print.

I had a lot of failures. Mainly the plastic wasn't sticking to the glass surface. I tried cleaning it and I tried a few different types of glue.

I wonder if the head is dragging against the plastic at the top.

I did get one "boaty" print which was really really good so it's possible but it's just a bit temperatmental at the moment.

Perhaps those who are more expert in these matters @Sean-anotherone https://github.com/Sean-anotherone or @jrp https://github.com/jrp could cast their eye over it...?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=AHL52BEHCOFNJEKSO6UOHMDQFFCEDA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4Q7UKQ#issuecomment-522320426, or mute the thread https://github.com/notifications/unsubscribe-auth/AHL52BGHO2QNVJSG7JAUCJDQFFCEDANCNFSM4H4MIPHQ .

irenge commented 5 years ago

Thanks!

Screenshot from 2019-08-19 12-30-18

ajlennon commented 5 years ago

(oops - bit of confusion - that was me :) )

ajlennon commented 5 years ago

ok with a lot of help from JC I think we're getting there.

I'll start building up a page on the Wiki to document this stuff.

For now I'll close down this issue...

Thanks for all the help everybody!

DoESsean commented 5 years ago

I would like to suggest the name Ramon.

There will be a small prize if anyone can figure out the convoluted way I arrived at that suggestion by the 27th of August.

ajlennon commented 5 years ago

Ramen?

On 20/08/2019 10:16, Seán Gleeson wrote:

I would like to suggest the name Ramon.

There will be a small prize if anyone can figure out the convoluted way I arrived at that suggestion by the 27th of August.

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=AALXOKQZRND74OSI4XX7PCTQFOY6RA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4VUG7A#issuecomment-522929020, or mute the thread https://github.com/notifications/unsubscribe-auth/AALXOKX2HYU74G6ES7XM6JLQFOY6RANCNFSM4H4MIPHQ.

DocTrucker commented 5 years ago

Quite clearly because you wish it had more RAM on it... ;-)

On Tue, 20 Aug 2019, 10:16 Seán Gleeson, notifications@github.com wrote:

I would like to suggest the name Ramon.

There will be a small prize if anyone can figure out the convoluted way I arrived at that suggestion by the 27th of August.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=AHL52BBVTMKYG2L5AZKPLNLQFOY6RA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4VUG7A#issuecomment-522929020, or mute the thread https://github.com/notifications/unsubscribe-auth/AHL52BBNJIBXAWQBBYZJ52LQFOY6RANCNFSM4H4MIPHQ .

DoESsean commented 5 years ago

Those are both one step names, nowhere near convoluted enough. This is a three step name.

On Tue, Aug 20, 2019 at 10:51 AM DocTrucker notifications@github.com wrote:

Quite clearly because you wish it had more RAM on it... ;-)

On Tue, 20 Aug 2019, 10:16 Seán Gleeson, notifications@github.com wrote:

I would like to suggest the name Ramon.

There will be a small prize if anyone can figure out the convoluted way I arrived at that suggestion by the 27th of August.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=AHL52BBVTMKYG2L5AZKPLNLQFOY6RA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4VUG7A#issuecomment-522929020 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AHL52BBNJIBXAWQBBYZJ52LQFOY6RANCNFSM4H4MIPHQ

.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/DoESLiverpool/somebody-should/issues/1163?email_source=notifications&email_token=ACFHYMBVDT5AHMBWCLHNRKTQFO5DBA5CNFSM4H4MIPH2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4VXKHI#issuecomment-522941725, or mute the thread https://github.com/notifications/unsubscribe-auth/ACFHYMFADKMI47ECAQYCDCLQFO5DBANCNFSM4H4MIPHQ .

MatthewCroughan commented 5 years ago

Taking over the testing on this now, since the first layer of large prints fail. Will post my results to the wiki, with prefix "Matt:" prior to the statement. We can then concatenate it.

ajlennon commented 5 years ago

Ok can we (you) break this out into a new setup issue as I think we’ve addressed the meat of what this one was about