Heatbed can't keep up with extruder fan [thermal runaway]

Hello,

I'm experiencing issues with my printer (it's a prusa derivative, which I have modified) with a new extruder fan I added. The new fan is great and makes complex models print a lot better than the previous one, but whenever the fan turns on, the bed has serious issues keeping up with the temperature and ends up in a thermal runaway most often than not (specially with small objects).

I'm printing at about 60C degrees, I don't think the bed should have issues keeping that temperature, but maybe I'm missing something. From my internet searches I've deduced I a couple of options I've got, (1) lower the fan speed and (2) change the thermal runaway protection values in the firmware.

I'd rather avoid lowering the fan speed as that would make switching the fan kinda pointless, so I'm calling that plan B. I also read that other people are using this same fan without a problem, so I think there is a problem somewhere in my setup. Atm my thermal runaway settings are:

#define TEMP_BED_RESIDENCY_TIME 10 
#define TEMP_BED_HYSTERESIS 4 
#define TEMP_BED_WINDOW     1

My question is, what would be reasonable values for this?

I have also tried checking if there is something wrong with my setup by checking the voltage from the bed contacts and I read 11.3v at first, so I changed the PSU to give a bit of extra juice (from the PSU I was measuring 11.9V before switching it up) and is currently on about 11.7V on the bed itself. Is there anything else I should check?

Post a picture of the machine.

A 40mm square fan can't possibly move enough air to cool the bed by much. Maybe noise from the fan wires is coupling into the thermistor wires and causing problems for the control code. Don't share wires withe the thermistor, and keep the fan wires away from the thermistor wires. It is best to twist the fan wires and twist the thermistor wires (separately, no together!).

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
Post a picture of the machine.

A 40mm square fan can't possibly move enough air to cool the bed by much. Maybe noise from the fan wires is coupling into the thermistor wires and causing problems for the control code. Don't share wires withe the thermistor, and keep the fan wires away from the thermistor wires. It is best to twist the fan wires and twist the thermistor wires (separately, no together!).

Anything specific I should make sure is visible in the photo?

In this case, the thermistor wires (for the bed) and fan wires go in completely different ways, I do however have the fan wires and the extruder wires (among others) in the same "tube". The wires for everything that is related to the extruder (+autolevel sensor) go over the top of the machine and the wires for the bed (thermistor & heater) go through the back of the printer, and they don't cross paths anywhere I can see (including the RAMPS).

Considering this, if it were coupling I'd have expected the issue to be the extruder's thermistor, but the error I get says it's the bed's thermistor (and I can see that the temperature drops for a few minutes and it seems incapable of keeping it up before the thermal runaway).

EDIT: I just realized I haven't pointed out why I think it's caused by the fan - in case anyone is wondering, the temperature always drops when the extruder fan is running, before that it can hold the temperature fine, and for prints that aren't using the fan the bed temperature rarely if ever goes 1C below target. For large prints (like a large object or multiple pieces spread across the bed) the temperature of the bed does drop, but it can usually hold it close enough to target that it doesn't trigger a thermal runaway .

Edited 2 time(s). Last edit at 01/27/2017 09:17AM by uorbe001.

The bed should not vary by 1C. Turn on PID control or rerun the PID autotuning. You shouldn't see more than 0.2C or so variation in bed temperature unless the heater is underpowered and you're operating at its limit, in which case you should get a higher powered bed heater (and then a bigger power supply, and new controller that can handle the increased current, etc.- there's a reason that printer was cheap...).

How much power is the bed heater capable of delivering? How big is the bed? No matter how large it is, it certainly has a relatively large thermal mass and temperature changes at the bed should take a while - a few seconds at the very least. If the bed temperature reading drops as soon as the tiny fan turns on, there's a problem with the wiring.

pictures of everything....

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

If it's a pcb bed heater, it's normal to have problems holding 110C+ when using part cooling fans, but not being able to hold 60C suggests the heater isn't working properly. If it's a dual 12/24V bed, make sure it's correctly wired for 12V, two of the pads have to be shorted together. If it's correctly wired, measure the bed resistance to check it's in the normal range of about 1 to 1.5 Ohms.

Quote
the_digital_dentist
The bed should not vary by 1C. Turn on PID control or rerun the PID autotuning. You shouldn't see more than 0.2C or so variation in bed temperature unless the heater is underpowered and you're operating at its limit, in which case you should get a higher powered bed heater (and then a bigger power supply, and new controller that can handle the increased current, etc.- there's a reason that printer was cheap...).
.

I will try rerunning the pid tuning in a while (I have a print running atm), I ran it earlier this week at about 60C, but I will try again. Related to this, is the PID tunning for the heatbed supposed to run with the extruder fan running? I think I ran it with the fan off.

Quote
the_digital_dentist
How much power is the bed heater capable of delivering? How big is the bed? No matter how large it is, it certainly has a relatively large thermal mass and temperature changes at the bed should take a while - a few seconds at the very least. If the bed temperature reading drops as soon as the tiny fan turns on, there's a problem with the wiring.
.

The bed is an MK3 12/24V dual power 200x200, pretty much identical to this one although mine is not from that specific seller. I have it setup for 12V, in case that makes a difference.

I just noticed this on the description of that url for the heatbed:

(5)With 15V on the 12V terminals will maintain approx. 115 degree Celsius without regulation.

Should I change the PSU so that it feeds some extra voltage to the printer?

Quote
the_digital_dentist
pictures of everything....

I uploaded a bunch of pictures here, let me know if I should take some more.

PCB heaters are universally awful, but that isn't the problem here. 60C should be well within its capability. Something else is wrong. A 40mm fan, even directed at the bed plate, can't move enough air to change the bed temperature by any amount that the controller and heater can't handle.

The problem started when you added the extruder fan (print cooling fan? or extruder cooling fan?). Check the wiring. An extruder cooling fan should be wired directly to the power supply. A print cooling fan should be wired to a MOSFET on the controller board.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Are you printing in your house ?...or outside in a cold garage?

Dan

Quote
JamesK
If it's a pcb bed heater, it's normal to have problems holding 110C+ when using part cooling fans, but not being able to hold 60C suggests the heater isn't working properly. If it's a dual 12/24V bed, make sure it's correctly wired for 12V, two of the pads have to be shorted together. If it's correctly wired, measure the bed resistance to check it's in the normal range of about 1 to 1.5 Ohms.

You were onto something. I measured the bed resistance and I got about 3ohms, which is way off according to the spec for this bed. I checked the soldering and one of the pads didn't seem to be completely in contact so I re-soldered that pad and I now get 1.6ohms, which is within the 'normal' range for it.

I have also PID tuned and I managed to do the PID tuning at 90C, which is a lot better than I used to be able to, up until now trying to do it at 70C caused timeouts every time I've done it. After that, I manually set the hotbed at 65C and turned on the fan sat 20% speed and climbed up to 100% without major temperature shifts. I tried the same at 80C, and about 70% fan speed it struggled to hold the temperature, so it'd appear my bed is still a bit sucky, but at least better than before.

Quote
the_digital_dentist
PCB heaters are universally awful, but that isn't the problem here. 60C should be well within its capability. Something else is wrong. A 40mm fan, even directed at the bed plate, can't move enough air to change the bed temperature by any amount that the controller and heater can't handle.

The problem started when you added the extruder fan (print cooling fan? or extruder cooling fan?). Check the wiring. An extruder cooling fan should be wired directly to the power supply. A print cooling fan should be wired to a MOSFET on the controller board.

The fan is a print cooling fan, and it is wired to the D9 mosfet in the RAMPS (bed is to D10). I also have an extruder cooling fan connected to one of the power terminals, which I asumed would be the same as wiring it directly to the PSU, but maybe that's a bad idea?

Quote
techristian
Are you printing in your house ?...or outside in a cold garage?

Dan

Indoors, room temperature is about 20C.

Ok, I just tried a couple of prints after the soldering I mentioned up there. When I print a relatively small thing like the stl I attached (precision-block.stl) at 65C, it can print it now, but the temperature did drop to 63C for a second when the extruder fan turned on (and went back to 64 pretty quickly, until it eventually stabilized at 65 for the rest of the print).

I also tried printing this tiny thing and that sadly gave me a thermal runaway error within a minute of the fan starting.

I call this progress, but it feels like there is still something not quite right. It seems like it's struggling especially when the fan starts blowing and then the heater can keep up with it after a while. If this is the case, would changing the thermal protection values in the firmware to give it some extra time make sense?

In case this is a bed limitation, what kind of beds are recommended?

Yes, if the bed is basically working but the firmware is being a bit over sensitive it would make sense to back off the limits a little. Which firmware are you using? When I switched from marlin to repetier firmware I got the impression that repetier's thermal management is a little better than Marlin's.

The other thing to look carefully at is the design of the fan duct for the part cooling fan. You want the air concentrated on the recently printed filament, and if your current fan duct isn't doing that you may be getting a lot more bed cooling than you need to put up with.

If you are having a hardware issue changing safety procedures in the firmware may not be the greatest idea...

Check out the thermistor wiring like the Dentist said. Interference may be telling the thermistor it is colder than it actually is. Perhaps invest in an infra red thermometer to see what happens to the bed when the fan comes on. Is it actually losing temperature??

A pcb heater should heat to 70c with ease. I used to run mine with two 50mm blower fans (pointed at print /nozzle, not heat bed).

Does your software give a temperature graph? Check it for spikes, I had similar problems once and it turned out to be the contacts on the thermistor to board plug.
After re crimping the contacts the errors went away and the temperature curve was way smoother.

Cheers,
Robin.

Quote
Origamib
Check out the thermistor wiring like the Dentist said. Interference may be telling the thermistor it is colder than it actually is. Perhaps invest in an infra red thermometer to see what happens to the bed when the fan comes on. Is it actually losing temperature??

I have tried touching the heatbed when the fan is running (static, I don't want to burn myself!) and it does feel like the area where the fan is blowing is colder to the touch than other areas. I'll see if I can get a better read if I can get my hands on a thermometer.

Also, just wanted to say that I have done a few prints (varying sizes) over the weekend and all of them succeeded, in a few of them the temperature did go down to 59C when the fan started running but it quickly went up to 60C, so my main problem (being unable to print anything with the fan) is solved. There is still that tiny print that gave me a thermal runaway which I want to look into, but I suspect it might be related to the part being placed at like 1cm from where the thermistor is and the fan running at full speed as it hasn't happened with anything else.

Depending on your fan duct and fan, full speed may just be too much for a low power bed to handle. I use 50mm blower fans and rarely use more than 60%. I've just been working on some new fan ducts to bring the air jets closer to the nozzle and it looks like that may lead to a reduction in required fan speed, even though the newer ducts restrict the flow more than the previous ones (but get it closer to where it needs to be).

Edited 1 time(s). Last edit at 01/31/2017 11:19AM by JamesK.

I have the same problem as the original poster, the heatbed doesn't keep the temperature. Also seeking for a cure for that. Some prints come out fine, some not. A guaranteed failure can be reproduced with the print fan.

I am using a (black) MK3 heatbed in the 12V configuration, the gray side (non printed side) is on the upside and I'm using a borosilicate glass plate on top of that. the sensor is inside and measures 85kOhms (shouldn't it be 100kOhms??), underneath the board, I have a radiator insulation layer which reflects the heat towards the board itself. The board is connected to one of two separate MOSFET-Boards (one: extruder, one: bed) and that is connected to a RAMPS 1.4 board and I've soldered out the green connector and soldered in the 12V cables which are coming from the PSU. On the Bed MOSFET board, I'm measuring 11.67V on the incoming side and 11.51V on the outgoing side.

I've tried PID-Tuning and those values really fail, it does not start printing and dies before or sometimes dies at the first 1-2 layers. In fact, the original values are a bit better.

Default:
#define DEFAULT_Kp 22.2
#define DEFAULT_Ki 1.08
#define DEFAULT_Kd 114
#define DEFAULT_bedKp 10.00
#define DEFAULT_bedKi .023
#define DEFAULT_bedKd 305.4

After PID-tuning (coming out worse) :
#define DEFAULT_Kp 25.11
#define DEFAULT_Ki 2.74
#define DEFAULT_Kd 57.59
#define DEFAULT_bedKp 548.09
#define DEFAULT_bedKi 76.03
#define DEFAULT_bedKd 987.80

Marlin 1.1.4

Bed-PID-tuning WITH a fan on fails with timeouts.

What can I do? Any help is appreciated.

Edited 1 time(s). Last edit at 08/29/2017 06:43PM by RepMike.

100k is nominal; (obviously) it varies with temperature so 85k is nothing to worry about.

What is your bed resistance? (Remember to subtract your probe lead resistance (the value your meter shows when you short the probes together))

My bed's resistance is 1.5 Ohms. I got it running by the way by the following procedure: I've done the PID tuning for the bed up to 90°C, a value which it can do after a while (takes long though). Then I took those values and reused them to Heat the bed up to 110° and it works like a charm now.

I just read this had new comments while looking through my post history, and in case anyone lands here I want to say I no longer have this issue. A few months after I posted this, I replaced my PCB bed with a silicone heated bed, and that thing works like a charm. Nothing seems to make the new heated bed even blink (after pid tunning), no matter what I throw at it, and it is way quicker to heat up than the old PCB one. Not gonna say that's a good solution for everyone, but for me, it worked like a charm.

@uorbe001 thanks for posting your solution, I appreciate it! Once I resume my hobby (currently stalled for holiday and other reasons), I will look at that solution.