CR-10S Pro, Micro Swiss Hot End, Thermal Runaway (temperature drops fluctuations)

[quarky42]

EDIT: This thread has been very helpful. There is definitely a hardware problem here and Creality has been SUPER slow about and "round about" in regards to providing help. Long story short: When my heated bed comes on, the DC voltage drops from 23.4 volts down to 11 volts. When the printer is printing, 11V is not enough for the hot end and the hot end loses temperature until the heated bed shuts off. Once the heated bed shuts off, then the hot end recovers to its set point.

RE: "Before filing an issue be sure to test the 1.1 and/or 2.0 "bugfix" branches to see whether the issue is already addressed." -- I was directed to use the Creality_DWINTest branch because of my BLTouch V3 (March build date). So, I'm not sure that I should or could use the bugfix branch without a lot of modifications. I am willing to try that branch but I would need my BLTouch V3 (March) and my Micro Swiss Hot End working on it.

My Printer: New CR-10S Pro (Control Board "v2.4 2018-08-13" printed on it) using Creality_DWINTest Branch pulled down 5/21/19 8:23pm MST, customized and loaded to the printer with the Arduino IDE Mods: HotEndStock enabled, ABL_BLTouch enabled (V3 / March), AllMetalHotEnd enabled (MicroSwiss HotEnd Installed, using the stock cooling fans currently.)

One of the more recent M303 Auto PID tunes was : M301 P25.99 I2.50 D67.61 U1 (PID tuning was ran with the nozzle 6mm off the bed. Fan was commanded "on" through the touch screen. I couldn't detect a difference in the fan using the M106 command or using M107 to turn it off.... not sure what was going on there. I tried specifying P0 and P1....neither controlled the part cooling fan.)
I am rerunning the auto tune again right now. I'll edit back when I get a chance.

I measured my thermister to be 102 kOhms both at the break out board connector and at the control board connector verifying the whole path. Both boards solder joints look okay. I do not see any signs of a cold solder joint.

I measured 24.3V on each of the legs of the power supply.

I measured 1.4V at the extruder, x, and y stepper pots. Also, 1.7V at the z stepper pot. Based on what little I was able to find, I believe these are correct.

Description

Edited this section heavily: After the fact I see now that when my heated bed comes on the nozzle temperature drops anywhere from 15 to 30 degrees depending on a couple of other factors. So, it looks like a hardware problem at this point, but this thread has been VERY helpful, so if it ends up being purely hardware, I hope that the thread will live on in some form because I didn't find anything else online about the nozzle temperature dropping when the print started on a CR-10S Pro. This may help others in their troubleshooting.

Steps to Reproduce

1. On my CR-10S Pro, Load the Creality-DWINTest firmware and custom firmware settings described above. Perform a factory reset. Calibrate the extruder and save the newly calculated extruder steps value. Run the PID Auto Tune and save the results to the EEPROM.
2. Slice a simple tiny model printing at 205C degrees hot end, 60mm/s, with 60C bed temperature.
3. Print and watch it fail.

I've discovered that if the heated bed is off, then the nozzle temperature is very stable like it should be. When the heated bed is on and is heating up during its thermal cycling the nozzle temperature drops.

Expected behavior: I expect it to hold to within 1 or 2 degrees of 205 C instead of apparently turning off the heater for 30 seconds. I have seen other Marlin firmwares on other printers "wait" near the setpoint temperature for 5 or 10 seconds before starting a print so that it can stabilize. This seems to start printing very shortly after reaching the set point.... There should be a 15 second (maybe more) delay before printing is started to let the temperature stabilize. (I don't think that is the cause of the problem, just an observation.)

Actual behavior: I think I already described this: The temperature starts to drop and loses 20 to even 30 degrees 30 seconds after the first plastic hits the bed. The system halts during the print with a thermal runaway condition because the temperature is too low in most cases. I've only had 2 prints out of ~30 reach 100% without the error coming up, but they absolutely dropped temperature multiple times during each print.

Here's a video I shot: https://www.youtube.com/watch?v=ZvkEPcxjZHU I may or may not have flipped the printer the bird several times in the video. That is aimed squarely at the printer that I've wasted the most part of 3 days screwing around with this thing just trying to get it to print a simple cylinder piece.

Additional Information

• Included a ZIP file containing your Configuration.h and Configuration_adv.h files: Marlin-Creality_DWINTest_ConfigurationFiles.zip HeadPost_plainWhite.zip

I don't think the Configuration_adv.h was changed from original. It has a new modified date, but I do believe that was an accidental save with no actual change being made. I had a typo in that file and I undid it and hit save. That's it.

I've gone back and cleaned up this original post to try and make it more clear. Hindsight being 20/20 and all.

[InsanityAutomation]

Sounds like the part cooling fan is overpowering the hotend heater. Have you tried with the part cooling fan off until quite a bit higher? Or running a PID tune with the cooling fan at 100%?

[quarky42]

Yes, I have tuned with the part cooling fan on at 50%, 75%, 100%.... makes no difference. On the latest version of my gcode, the part cooling fan is on at 35% at the start of the print and held at 35% the whole time. The skirt and the first couple layers are printed at 205 degrees. The temperature only starts to fall 30 seconds after the extrusion begins.

I can put turn the cooling fan on, preheat the printer to 205, and extrude 300mm of plastic with no loss of temperature. It is only losing temperature during an actual print job. I cannot "make it" lose temperature outside of a print job.

[quarky42]

Also, please note: If the thermal runaway isn't triggered, then the temperature recovers. It goes back up to 205 degrees and sits there for 30 seconds. Then the temperature dives again. The cooling fan hasn't changed in that time. As far as I can tell the heater is turning off for 30 seconds at a time. If I am printing slow enough or get "lucky", it will recover. The print is still ruined with that huge of a temperature swing, but if it does recover, it will go back up to 205 degrees and once it reaches 205 degrees it will sit there for about 30 seconds before it starts to dive again. Something is causing the heater to cycle in this strange way while printing.

[quarky42]

Short of having a voltage probe hooked up to the heater line during the print, I am 95% certain that the heater is turning off for 30 seconds at a time. It is not cycling the way I would expect it to cycle.

[InsanityAutomation]

Very strange indeed... Are you able to get a serial log when this is occurring? Might be some useful data in the temp reports. One thing to possibly try is disabling autotemp on line 143 in config_adv. It may be having an issue offsetting heater power based on extrusion.

[quarky42]

I will do both tonight and report back. If you think of anything else you'd like me to try / check, please let me know and I'll add it to the to-do list. :)

I can capture the serial output. I had been using putty to connect. I finally figured out how to resize the Pronterface window. The first few times I tried to find the "seam" to resize it, the cursor wouldn't light up. It wasn't usable until I could stretch it out much wider, finally.

[InsanityAutomation]

Putty is awesome! Used it for years :)

[quarky42]

I loaded the B6_250K hex file with factory default PID settings. I have the same issues, but now the temperature drops are showing up on preheat as well (though I am doing something a bit different now...I got OctoPrint set back up so I've been using that to control the printer. I will need to So I have graphs, logs, see what the behavior is using the touch screen.)

This latest print was done with the fan completely off. Didn't help.

I haven't had a chance to try loading the firmware from Ardiuno IDE after turning off autotemp in configuration_adv.h yet, but I will try and get to that tomorrow. I wanted to post the log, screenshots, and a timelapse.

Thermal drop (about 20 degrees) during print. This screenshot is after loading BLTouch B6 250k Hex and loading the factory default M301 PIDs: Same thing as above just after the print was over I turned the hot end back on and just let it sit. This is showing that same dipping behavior. I will need to try this from the touch screen instead and see what I get. Before, the touchscreen commanded preheat temperatures was pretty stable. Serial Connection Log File from OctoPrint: serial.log Time Lapse Video: headpo_1_20190528232154.zip

[quarky42]

Should the heated bed be receiving 12 volts or 24 volts? I took a voltage reading earlier on the DC relay for the heated bed and was reading ~12.3v out of the relay. Is that right?

As far as this next graph goes, I think I found my culprit: The heated bed. I turned off the heated bed and just the hot end. The hot end is sitting stable at 205 degrees.

So my theory at this point is that every time the heated bed is coming on my hot end dips. Otherwise I'm at a loss for why turning off the heated bed caused the hot end to behave itself.

It held temperature steady for 10 minutes with the heated bed turned off. I'm assuming at this point that this isn't firmware related, but since you have a lot of experience I am hoping you might steer me in the right direction?

[InsanityAutomation]

Where were youre leads when you measured 12v? Only one power supply in the system and its 24v... Power supply may have an issue and is failing under load, but only the hotend shows it as the bed is much slower to respond and change.

[quarky42]

The load side. It only shows voltage on those connectors when the heated bed is on. Doing some more reading I think the heated bed is 12V based?

The other side of the relay showed 24.3v constant. (I did not measure that side under load... Sounds like I need to measure the hot end temperature when the heated bed is on, but based on the behavior I am pretty sure the hot end is losing power while the bed is heating. I am uploading a video that shows the nozzle reach temperature with the bed. When the nozzle starts to drop the bed drops 1 degree and the nozzle only recovers after the bed has recovered.

[InsanityAutomation]

Id try to get a replacement power supply. Keep in mind load side of the SSR is +/+ not +/- think of it like a light switch on that side.

[juliandroid]

It would be interesting if the PSU is throttling (overheating). Inside you have one small fan and if it does not work (if you reverse the polarity it won't run at all)?

[quarky42]

@juliandroid Thank you for taking the time to comment. I don't think I got a chance to add this, but yes I already checked the fan last night and it is indeed running. The fan was pointing into the power supply instead of pointing out of the power supply. I have it pointing out as an exhaust fan and it has made no difference in the problem at all.

I felt the cover of the power supply after the bed and hot end had been on for a while and it was warm, but not uncomfortable at all. (So somewhere between 100 degrees and 110 degrees without getting out my IR thermometer. I had already made sure that all power connections on the power supply and on the control board / dc relay were tight. They were. I didn't find any loose connections when I was initially going through the printer to do the quality control that Creality is known to not always do well.

If there is something I should check inside the power supply, I am willing to carefully do so. At this point I think @InsanityAutomation is spot on... it is likely a weak / bad power supply.

[quarky42]

What's weird is that I can heat up both the heated bed and the nozzle at the same time. It takes about ~7 minute to heat up if I heat up both at the same time. The nozzle heats up a lot faster with the heated bed pre-heated or not running at all. I do not see the same major drops when the heated bed and the nozzle are going at the same time.

@InsanityAutomation Is the bed heater "all or nothing" or is it varied by PWM in any way? I haven't run any PID calibration on the heated bed itself. Is there any value in that? I'm just using the defaults. I did load the 250k hex version of the firmware last night and had no change in the behavior. Since others had confirmed that they used the hex version without problems I figured I would try a "known good" firmware that way to make sure it wasn't something weird in the Arduino IDE project. I did redownload the github version and I made the change that you suggested a few posts back. I did not have time to load it up and try it. I am still willing to if you think I should. I just ran out of time last night.

Here are two videos showing as clearly as I can the temperature fluctuations with the heated bed on and the lack of temperature fluctuations right after turning the heated bed off. There is only about a second or two between the end of the first video and when the second video starts. I stopped and started recording so that I would have two separate files: https://www.youtube.com/watch?v=DnFcbnXpij0 (Preheating takes about 7 minutes in this one. The first temperature dip is a little bit after 8 minutes in. The second temperature dip is at about 8:24 into the video. The third is another 2 minutes and change after that. So a period of a little over 2 minutes between dips)

https://www.youtube.com/watch?v=EZt28lBNxpI I stopped recording the first video and started this one and turned off the bed. You can see that for 6 minutes straight the nozzle temperature is stable. I saw a 1 degree variation on the touch screen at one point.

[quarky42]

I didn't see a 220 / 110 V switch on the power supply. Is there one that I missed somewhere? If so, can someone direct me to it so I can make sure that the power supply isn't on 220V by mistake?

[juliandroid]

If I remember correctly, somewhere on the Creality site was written that the PSU is automatically selecting the right voltage. (Edit: I confirm, on their website it is written on the picture banner, not as a text)

Did you try setting the hotend to something like 240C to see whether time interval between heating and first drop is consistent with lower (205C) or it shorten and more often? I assume that if you have issue with the PSU it will take less time and happen more often if you draw consistently more power on that rail.

Regarding the PSU - I would inspect visually for inflated caps (turned off and not touching anything - for long time there might be 110/220V charges):

Edit: And also check the power transistors located on the heatsink for any sign of overheating.

[juliandroid]

[quarky42]

How obvious would the signs of heating be on the power transistors? carbon? severe discoloration? I didn't notice anything, but I'll look again.

I'll check the caps. (Yes, I'll make sure not to touch the metallic bits where the angry pixies can leap out at ya! Still a good warning, though. ;) I didn't notice any bulging but I can look for it specifically to be sure.

I have PLA loaded, so I am not going above about 220 degrees on the hot end. 220 just long enough doing the PID tuning causes some of the filament to crystallize. I suspect the same effect you are talking about would happen if I drove the heated bed temperature up to 80 degrees and/or had something blowing on the bed to cool it down. It's tied to the bed heater cycling. My next plan was to watch the hot end voltages when the bed is heating to see what I'm getting when the power supply is under load.

[juliandroid]

I see at least 2 resettable thermal fuses, but I have no idea why if those kick-in your entire printer is not shutting down. From the datasheet ( https://www.meanwell-web.com/content/files/pdfs/productPdfs/MW/RSP-500/RSP-500-spec.pdf ) it seems it has single rail (output), so everything connected to it should shutdown as well. Both over power and over temperature protections have automatic recovery, but there is no way for the printer to keep working for 30 seconds.

The alternative is that only one of the power mosfets is overheating (internal temperature above 150/175C), it shutdowns and leave to the PSU with less power than expected, so the hotend is not completely turned off, but just don't have enough power to keep the temperature steady.

You said you have loaded PLA filament. What about reducing the printing speed significantly, so it becomes quite easy to maintain the hotend temperatures even when the current is reduced due to failing mosfet? I guess you'll see less temp drop than before until the mosfet recovers.

[quarky42]

Yes, printing at slower speed reduces but does not eliminate the temperature drop. One of the two prints I've had complete were at 50 percent speed and the temperature drop was still severe but not as bad... Instead of 30 degrees, it dropped 15. With PID tuning I might be able to get that to 8 degrees... Still not acceptable but it fits in with what you are saying about the mosfets.

With that much information about what might be causing this, I can watch the power supply in FLIR and specifically look okay the mosfets while it's printing and see them over heat in real time. Your analysis of this situation is ginormously appreciated!

Also, I have a killawatt meter. I can watch the power going into the printer with the heated bed, the hot end, and then both the heated bed and hot end. I also looked up the PSU data sheet and with that efficiency rating I can tell if the power supply is drawing more or less power than it should be pulling. (Which would help rule out or indicate a short or a component pulling way more power than it should.)

We've pretty strongly narrowed this down to hardware and the fact that you're still willing to share your expertise and effort on this is really appreciated. I may just end up buying a replacement power supply while I wait for Creality to dick around. Does Tiny Machines sell this particular model? (I'll check later when I'm back to a computer.)

Maybe I should open the power supply and just sit the whole thing on top of a box fan. ;). Lol

[juliandroid]

Btw, I don't recall what was the default fan direction, but from my picture it seems to be an exhaust fan, so you had running your PSU fan in wrong direction and might caused failure. I'm using Noctua fan and I'm not even sure it has enough static pressure and flow compared to the stock, so that could be a potential problem for me (and others performed the mod too), but at least I had it running in the right direction.

I Amazon I saw a lot of "MEAN WELL RSP-500-24", so if you don't find that at Tiny Machines.

Please keep us posted with your findings here.

[InsanityAutomation]

Forget the box fan, run it submerged in liquid nitrogen! Maybe resting on some dry ice! Meanwell is usually a decent brand so it shouldnt be an issue sticking with them.

[quarky42]

@insanityautomation this guy gets it! Go big or go home! Lol.

I used my wattage meter last night. The system pulls between 10 and 14 watts idle. During bed heating only it pulls between 320 and 400 watts. (Since resistance goes up with heat and it takes more power to heat up to higher temperatures this makes sense to me.). During nozzle heating only it pulls about 50 watts (which makes sense for a 40 watt nominal heater cartridge).

Here is where it got interesting: With both the bed and nozzle heating it still only could pull 400 watts. After running for a little while, it could only pull 360 watts max (thermal limiting). So when the bed was heating, it took the whole 360 watts. When the bed stopped heating the nozzle could catch up and it was pulling 50 watts during that time period.

So yeah, pretty sure a bad power supply.

@juliandroid: Having the fan blow in would not cause a failure in a good power supply. Air flow is air flow and this power supply has thermal protection built in. So it would protect itself. It might not run optimally with a wrong fan direction, but this one doesn't run any differently with the fan pointing the other way. I also read the same data sheet that @InsanityAutomation read. It shows the fan direction on there. It is definitely in the exhaust direction.

One concern I have: The heated bed is taking almost twice the wattage it is rated for. The bed is a 12v 18A (220 watts) rated. With just the bed heating up only, the system is pulling up to 400 watts. If I replace the power supply, how much will the bed pull, then? Will it burn out? Is it pulling too much?

Maybe the bed has a failure that has shortened the power supply's life.

[InsanityAutomation]

That power draw sounds about right, but remember it is 24v not 12v, if you measure from either load wire on the ssr to the power supply - you should have 24v. Can measure resistance of the bed itself and I can compare it to mine tonight.

[juliandroid]

Yes, the SSR switches 24V, but after initial load it just maintain the set temperature. One thing that is strange - you said, you have power meter and the power rating for input is 100-240V AC 5.9A. So, for my power country that is 230V5.9A ~ 1360W and the DC output is only 24V21A (~500W). Not sure where that difference is coming, but shouldn't be from efficiency. Anyway, the main point here is that on the outlet side I would expect to see significantly more than those 400W on the AC side, reported by you when you heat both. I'll try myself and see what wattage is reported :)

[quarky42]

At first all I came across was 12v heated beds for the CR-10S. Now I am starting to see some 24V ones.

This one shows 24v and 220 watts: https://m.banggood.com/Creality-3D-24V-220W-235235mm-Aluminum-Heated-Bed-Hot-Bed-Kit-For-Ender-3-3D-Printer-p-1335217.html

So I was mistaken about the voltage. (I wondered why they would have a 24v PSU and a 12v bed, but that's all I found at first). But I still think the wattage is way too high on the heated bed: https://m.banggood.com/Creality-3D-24V-220W-235235mm-Aluminum-Heated-Bed-Hot-Bed-Kit-For-Ender-3-3D-Printer-p-1335217.html

Is there any way to limit the heated bed to 220 watts from the controller side? I doubt it. I figure the bed just gets 24V applied to it and that's it. However if that's the case then why does the heated bed pull more power the hotter it gets? (If resistance is going up as it heats up and it is getting the same 24V, then the current and wattage should be going down? I'm setting guessing my thoughts on that from before.)

[juliandroid]

I've just checked behaviour of my PSU. I've setup the hotend to 205C and bed to 60C. The power consumption (on the AC side) maxout at 505W. It then drop slightly to around 500W and when the temperature became closer and closer to the target it gradually dropped to around 450W (heating according to the PID). So, most of the time it was in the range of 470-500W.

Edit: No limitation from the controller side, just the SSR and how you are going to cooling it down. Some recommends mounting the SSR on the aluminium frame few centimetres on the side, because even currently it heats up to around 70C, while mounting it on the frame drops to 45-50C. So, I would definitely move the SSR to protect it.

What I'm going to do some day is to mount 220V bed heater (needs proper grounding on the bed itself), so you are offloading the PSU dramatically.

[quarky42]

Regarding expecting to see more power being drawn: exactly!

New CR-10S Pro's come with a big aluminum heat sink on the DC Relay now. So, mine already has that to help it shed heat.

Back to the wattage:. Exactly! I would expect it to be able to pull up to 500 watts for a good power supply. This one is supposed to be 89 percent efficient. So that's a little under 450 watts usable power. Figure about 15 watts overhead for the controller, fans, etc (based on what I saw). That is about 435 watts for all things (motors, bed heater, nozzle heater).

Since I'm getting up to 395 watts maximum, draw, that is 365 watts usable (after conversion loss). I've seen the bed take a lot more power than I think it is rated for. The bed is starving the nozzle heater. The videos show it pretty clearly, too.

Definitely not getting the full power out of the power supply. When the bed heater is on by itself the printer pulls between 360 and 395 watts. When both the bed heater and the nozzle heater are on, it still only pulls 360 watts (likely because the power supply has heated up and has even lower capacity at that point). So seeing the hot end temperature drop while the bed heater is active is making more sense.

(Fully agree with the earlier estimates that the power supply is faulty... But I wanted to try and back it up with measurements based on the tools I have available.)

I'm thinking if a Mean Well 600W power supply will fit in there, I'll put that in. They are 20mm taller. I'd rather not run a power supply at full load.

[juliandroid]

Sounds reasonable :) Can you please upload an image of the SSR and the aluminum heat sink? I'm interested whether I can further improve my cooling of the SSR due to the planned switch to 220V AC heatbed :)

[quarky42]

Sure, I'll post a picture back when I get a chance. In the meantime I managed to get the video clips uploaded and put into a playlist showing the power draw of my printer at different points:
1> Initial Bed Heatup 2> A little while later, the bed is at 36C and climbing,
3> Bed is at 54 degrees and climbing to 60C (So it is near the end of the heatup cycle) 4> Bed has reached 60C, is at steady state, so then I start heating up the hot end to 205C and we watch that heat up all the way as well as see what happens when the bed cycles.

I really like this little wattage meter. Probably one of the best "a cheap tool to have on hand in case I need it" purchases I've made.

Here's the playlist: https://www.youtube.com/playlist?list=PLBpIws7wZqQGts1Dw50CgzbvVv3JygCIn

[quarky42]

@juliandroid Here is that picture of the SSR with the aluminum heatsink.

I really wish Creality had rotated the heatsinks 90 degrees... both that silver one and the black ones. If they were rotated, then the fans would cool the heatsinks down even better.

[quarky42]

@InsanityAutomation It sounds like the power draw isn't correct. @juliandroid saw 500 to 505 watts being drawn on their system.

I do stand corrected on the voltage of the heated bed, however the 24V heated beds I find online all say 220 Watts. That heated bed is drawing 315 or more watts by itself. In @juliandroid's example of 505 watts that would be:

505 Watts * 0.89 efficiency = 449.5 Usable watts 449.5 - 15 watts (LCD, LEDs, Fans, Control Board) = 434.5 watts 434.45 - 40 watts heater cartridge = 394.5 watts.

So based on some number crunching I would say that the heated bed needs somewhere between 385 and 395 watts by itself. (The heater cartridge is rated for 40 watts and I saw mine use about 40 watts pretty consistently, but it could have been 45 watts or maybe even 50. The bed, by far, uses a lot more power.)

According to the specs of the power supply it should be able to run up to 130% of its capacity, if I'm reading it correctly. Still, running at 500 watts on a 500 watt power supply does seem a little crazy to me.

[quarky42]

Looks like there is around 54mm height available in that base or less. So being able to run a 600 or 650 watt power supply in there is not possible. Damn. Another 10mm or 20mm would be ideal. If Creality made the base 15mm taller, then we'd have the option of switching power supplies easily.

Hmm...make a 10mm aluminum frame shim all around the base for a taller power supply? Hmm...I could do that with extrusions.

I guess if I did the AC bed upgrade using a DC/AC Solid State Relay, that would free up a ton of overhead. I could probably run everything else on 100 watts pretty easily. :/ Hmm.

[juliandroid]

@quarky42 Thanks for the picture! I also see grey fan mount blowing the relay - is that printed by you or also stock now?

Yep, the PSU is rated for 504W, 24V DC max, so with 89% efficiency that makes everything except for the motors that already consumes 89% of the maximum. With all of the motors an extra 25W and you are looking at 95% load :) From the datasheet, the AC input is 2.65A, 230V ~ 609W which related to the rated 504W make rather smaller number, so I guess there is something like extra 5% reserve, but in general it is OK for a reputable PSU to be used on 100% - after all it should work in the rated temperatures/environment and load. Personally I prefer having something like 50% backup... and I'll achieve that by replacing the heating bed with 230V version :)

[chunter1]

@juliandroid Here is that picture of the SSR with the aluminum heatsink.

I really wish Creality had rotated the heatsinks 90 degrees... both that silver one and the black ones. If they were rotated, then the fans would cool the heatsinks down even better.

Thanks for the funny image of the heatsink prooving that nobody at Creality has the slightest clue of thermal management. :) My printer came with the SSR mounted in the air, not touching anything. I am using a thin layer of thermal conductive foil between the SSR and the printer case now. ;)

[quarky42]

A big chunk of aluminum is better than stand offs which is what they were using.

One YouTuber noticed the relay in air and mounted it to the aluminum extrusion rail using a couple T nuts for it. Even that was enough to cool the SSR down a great deal.

So in this case I think thermal mass is enough, but since they could get heat sinks cut and which size and direction, they should have had the fins turned 90 degrees so that the air flow would be better.

The double fan bracket seen there is my design inspired by a single fan bracket I saw online.

[juliandroid]

@quarky42 You can contact Keenovo, they are offering 310x320mm silicone pad for CR-10s Pro (110/220V versions). I don't see currently that offer in Ali, but they offered the same size previously for the price of the "normal" CR-10S :) And yes you need DC/AC SSR. The one in the printer is not appropriate :)

So, originally they didn't provide a fan for the SSR? :)

[quarky42]

@juliandroid yeah I agree about running power supplies at around 50 percent load. They run a lot cooler yet still within their efficiency range at around there. For me it is more like 50 to 60 percent ideal...70 percent is okay. Anything over 80 percent is just asking for the power supply to die. I haven't found a power supply yet that I felt was good for 100 percent load 100 percent of the time. :)

Looking at the heated bed upgrade. That's about $70 plus relay and insulation mat. That's not too bad. I want to get this power supply replaced and maybe then I'll do that next once I make sure what I bought initially works. Once it's fixed, fix it some more until it's broke... Or something like that.

[quarky42]

@juliandroid correct. The SSR in an older version used to be installed on posts, no fan no heatsink.

Recently they added a heatsink turned the wrong direction still no fan. This is the version I got. I added the second fan.

I see the heated beds on Amazon. Not willing to go that far until the power supply situation is settled, but will keep it in mind.

[juliandroid]

BTW, adding just a small fan in front of the heat sink will help a lot, since now the fins are pretty much blocked to dissipate heat :D

Yes, that is discouraging - all the stupid mistakes made by Creality are countless now :) One comforting news is that the peak power consumption is rather temporary, not constant. Also this can be mitigated by Creality by heating the bed first, then the hotend. Also they can disable LCD for saving extra 1W :wink: and even put the MCU in sleep mode :rofl:

[chunter1]

Maybe of interest... Starting at the yellow line, the diagram below shows the power-consumption of my CR-10S PRO printing a calibration cube at 40 mm/s. Extruder is at 200°C, bed at 50°C and turned off at layer 10.

[juliandroid]

That is due to the Eco mode in the printer's menu?

[chunter1]

That is due to the Eco mode in the printer's menu?

No, i am using Simplify3D and control everything from there. By the way... If you don't have it already... i can highly recommend using a Raspberry Pi with Octoprint. It's really increadible convenient :) Gonna integrate the Pi into the Creality case next.

[juliandroid]

Yep, I'm having an older Pi that I want to use it for OctoPrint precisely :) So many things I want to do, so less time to do them :)

[quarky42]

The inside of my power supply and the DC relay when the heated bed has reached 60C:

Edit: Here's the temperature gradient turned off:

[quarky42]

That inductor coil though... 70.8 degrees C. Yeah, it's getting a little too hot in there at 360 to 395 watts of power draw.

[juliandroid]

From temperature side, I don't see nothing imminent that could cause the problem. Maybe it is time for PSU replacement. I cannot imagine that the problem is coming from the heatbed (both to consume less than rated and also stop working periodically), although for example I've seen sensors working perfectly while cold (normal resistance) that goes to "infinite" when heated and that cycles over and over again :)

[quarky42]

Yeah I've reached out to the seller, Creality, and Mean Well with the information. The PSU has a 3 year warranty and Mean Well has a USA based website at least so maybe they can do a warranty exchange in a reasonable amount of time.

I did notice some clicking (rapid cycling on/off) of the solid state relay though. I think it had more to do with the PSU reaching it's faulty overload limit. The SSR relay led and the power led on the PSU both blink rapidly for a few seconds. I'll see about getting that video loaded.

At this point I've just been adding info to the thread because I'm able to do a bit more testing and in the hopes that once I get a replacement PSU that the information here ends up helping someone else that comes across it.

I appreciate the thoughtful discussion that everyone has contributed to this far!

[InsanityAutomation]

Rapid blinking is expected in pwm control as it gets close to temp. PID takes over within 10c of setpoint.