Erratic Heatbed Temperatures Gen 7 1.3.1

Hi,

I've been using my Gen 7 1.3.1 electronics seamlessly for ages now (brilliant electronics) though the other night I noticed my PCB Heatbed was registering at >140C! I immediately shut down the printer, restarted and monitored my heatbed temperatures. I found that the preset temperature was correctly set at 115C which the system powered up to, then killed power until the temperature dropped to 100C before kicking in again. During this off power phase though the temperature reading for my Hotend jumped up by 20C even though unpowered! This seems to show every time the heatbed powers down!

After the heatbed power kicks back in the temperatures are very erratic! One minute the bed will heat to 115C, cool down to 80C (massive range!), then upto >140C and beyond. Then it will work fine again for a while, then powered constantly until turned off! All the time though the temperature shows as set to 115C!

I've tried reloading the firmware, tried different control software, restarted, the works! I believe the problem lies within the Gen 7 electronics! Any ideas or pointers?

Obviously not software, nor thermistors and the Mosfet seems to be switching, though randomly!

Any help or advice would be appreciated.

Obviously power supply goes to the knees when the heated bed is turned on. No neccessarily the power supply it's self, it might be a problem of how the juice is distributed on the board.

Gen7 v1.2 had more trouble with this and the description for fixing v1.2 might get you on the right track to enhance your v1.3.1: [reprap.org] As v1.3.1 has an additional PSU connector, the fix should require only one cut and one new wire.

v1.4 has the above fix "built in".

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Thanks for the fast reply and the links, unfortunately I don't think the modifications have helped!

I believe I have followed the instructions towards the modification correctly by braking the tracks, as such isolating the high powered section and re-linking in the ATmega as shown in the attached photos.

As can be seen in the temperature graph, the light blue line shows the normal fluctuations in the heatbed temperature compared to the preset temperature (purple line). After one cycle the temperature just keeps increasing, even once the power is cut! It seems as though the off signal from the ATmega is not transmitting.

Any further advice would be greatly appreciated.

From Printed Parts

From Printed Parts

From Printed Parts

Nice temperature graphs, indieflow. How do you make them?

The GND fix looks right. To get rid of the rectangle in the left half even better, it might help to do something similar for the 5V reference voltage. This should be connected to AVCC of the ATemga only, which is pin 30.

On the right half I don't see an electronics problem. It's the firmware apparently not recognizing the bed is over temperature. Odd, as it works before the extruder kicks in and it works for the extruder, too. Am I assuming correctly you use Teacup firmware? Then heater_tick() in heater.c needs debugging.

If you don't feel like debugging, I'd turn on BANG_BANG in config.h. This doesn't use all the PID stuff, much simpler and likely more reliable code.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

The temperature graph is produced via the Repetier Host software that I'm using, really nice software with great visualisation features!

Also I am currently using the Repetier Firmware, which has not been altered at all since well before the problem started! I don't think the problem lies with the software/firmware! The heatbed only seems to work for the initial cycle (even then, over a massive range) and then just climbs! The extruder turning on and off has no effect what so ever (I turned it on just to show how it works perfectly) as the heatbed was climbing well before the extruder was turned on and off!

Would I be right in thinking the problem (electronics wise) could only be the ATmega (not signalling), resistors (connecting to the MOSFET) or the MOSFET? I will have to investigate the tracks with my multi meter.

Also am I right in thinking you are suggesting I make a direct connection from the +5V to pin 30 and cut the track that feeds the ATmega?

Quote

Also I am currently using the Repetier Firmware, which has not been altered at all since well before the problem started!

Oh, so it once worked better? Now, that's an important news (for me). Because then it can't be a wrong design of the electronics either. Apparently, something has become broken.

Quote

I don't think the problem lies with the software/firmware!

Well, it very much looks like a software problem. Because setting this jump in the extruder temperature readout aside, the firmware simply doesn't react the way it should. It reads 140 °C and doesn't stop the heater. Do you remember wether this jump was there earlier?

Quote

Would I be right in thinking the problem (electronics wise) could only be the ATmega (not signalling), resistors (connecting to the MOSFET) or the MOSFET? I will have to investigate the tracks with my multi meter.

What counts is what the firmware reads out from the analog input. You can measure the voltage in the analog input, of course, but this is exactly what the ATmega does.

Quote

Also am I right in thinking you are suggesting I make a direct connection from the +5V to pin 30 and cut the track that feeds the ATmega?

Yes. This should improve the stability of the analog temperature voltage and also enables temperature readouts on standby voltage. Standard feature of v1.4 now.

Now things get tricky. Is it perhaps possible the firmware switches off the heater, but the heater continues heating anyways? I'd put in a debug message right next to the "heater off" code to see wether Repetier acts the way it should. You can also measure the signal voltage of the MOSFETs on R11/R12.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Right problem found and now all fixed

After you suggested measuring the signal voltage of the MOSFETs on R11/R12 I found the problem!

The signal was coming from the ATmega to kill the heat though the MOSFET wasn't cutting off! Believing the problem was the MOSFET, I started to change it out for a new one. When I de-soldered the legs, I found one of the legs had broken off near the board! I now believe once the MOSFET heated up, it slightly lifted off the board creating the break from signals!

As you can now see from my new temperature readings I am getting a really nice even heat range! Thanks for the advice and amazing electronics!

From Printed Parts

Now, that was a really tricky one. Congratulations!

And yes, it's amazing how long it takes for a MOSFET to drain it's self. Charge the gate with 5V, remove the signal source and it takes several minutes until it goes off. If it doesn't die from overheating in between.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY