Fix to prevent heat bed cutting out using relay switch

Nice, I'm going to try this. Where did you get that flat temperature probe thing? I noticed that my bed temperature is set to 80 but when I use the IR gun it is showing it is 10 degrees higher (90).

 

Also, I have a fan mounted underneath my robo3d to cool the chip but it is powered by a separate computer PSU. Is there a way to splice in/connect the fan to robo3d's power supply?

 

The existing power supply on the R3D has room to add additional wires to it for more things like a fan. I have the LED's for my print head direct wired to my PSU and have a switch mounted so I can turn them on/off if desired, but mostly I use it as a reminder for if the printer is still powered on or not.

 

Ir temperature guns are not accurate for temperature measurements. Use a thermocouple based temperature meter and attach the thermocouple sensor directly to the surface of your heatbed.

 

Thanks for the reply Das Wookie and Cambo3D.
Das Wookie, I'll look for your LED instructions and see if I can change it for my fan. I'd really like to get this bed issue fixed though because although the fan mount works, I sometimes still get a dip in the bed temp.

 

Unless the firmware is adjusted to use Bang Bang style temperature control, I highly recommend NOT USING THIS SOLUTION.

The heat bed is controlled by what is known as PWM. This means that the mechanical relay will be cycling continuously unless the heater is on 100% or completely off. As a result, the relay will fail mechanically after a relatively short life. Unfortunately, the tech specs aren't published for this relay so the MTBF cycles before failure is not published. I think the Marlin firmware uses 10Hz frequency for the heat bed. That means that the relay will be cycling 10 times per second for hours on end.

Running at anything other than full 100% duty cycle will result in 1,000,000 cycles per 1.15 days of operation. Failure could happen anywhere in the low millions of cycles to the hundreds of millions of cycles. With PID control, it is not a matter of if but when the relay fails.

A failed relay may not be a big deal if it fails off but, failing on will result in the heat bed running at full power until you power off the device.

Either change the firmware to use Bang Bang Style by editing this section of configuration.h:
// PID settings:
// Comment the following line to disable PID and enable bang-bang.
#define PIDTEMP

OR

Buy this device will solve the problem permanently and will not burn out after some few days/weeks of operation:
http://www.automationdirect.com/adc...10A_-_75A_(AD-SSR6_Series)/AD-SSR6M40-DC-200D

 

Thanks Leon. I'm not an electrical engineer so I have to rely on the experts here. I think other might be having this problem as well so if someone would be so kind as to write instructions on how to install this, I would be really grateful. Leon, do you have that part installed for your robo3d? (just read your printer post where you did) Would it be better to go with the 5$ radioshack method and change the code or would the $40 relay be better.

 

The Opto-SSR is a technically better solution but the standard relay with bang bang is good enough.

Personally, I can't accept good enough, which is why I've gone to such great lengths to modify my printer. So, if you can accept the good enough solution, there will be no adverse affects on your print quality.

 

Jerry,

Has the configuration.h been modified for the units you're testing.? The behavior you describe is much more like bang bang control than PWM.

If you watch the blink rate on the LEDs when it controls the hot end, that's the PWM frequency. The same thing is going to the bed relay and this is exactly what Tesseract was seeing.

Leon

 

My heat bed also cut out. I found this thread, and figured I would investigate to see what the problem is, and propose an alternate solution. Tipping the printer over I see this:

Based on the other posters, I also think the problem is with the upper fuse in this picture. Hard to tell if it is F1 or F2, but it is the PTC fuse labeled "11A". I brought out the positive side of the heater element lead, and attached it to a jumper so I could use a current meter. I also used a voltmeter so I could see that the voltage goes to zero, indicating the fuse has blown. Here is my test setup:

I could not replicate the problem when I was watching. I am going to blame it on my house being much colder at night than in the afternoon. I will re-calibrate my heights and try to print again tomorrow. However, I do have some notes.

First: The relay is not an optimal solution for a number of reasons as pointed out by Leon. I will add that I do not know the operating time of the radioshack relay, but I measured the PWM pulse width to be 1ms at about 7.5Hz. I highly doubt that a $5 relay is going to be able to operate in 1ms. With this solution we might be under "bang bang" control after all, since the relay probably won't even wiggle in 1ms. I guess I would have to hear it to know. I also don't want to have to go in and modify software files and fix a bunch of wiring. Much easier to replace one PCB component.


Second: I'm not entirely sure this is the temp sensor, so I could be wrong- but what I think is the temp sensor is taped to the heater element with capton. That's probably not going to last. If this is the sensor element, it probably should be glued on or something. Is this the temp sensor? Is there a recommendation for something better than tape?


Second: There is a wide range of values where a fuse might blow- they are pretty much only useful for protecting against short circuits that would otherwise cause serious damage or a fire. The power flowing to my heater element is 9.25A at 11.68V. That voltage is down from 12.25 when unloaded (~.6V lost in wiring and connections is to be expected and doesn't concern me). The RAMPS board schematic says that this fuse is used:
http://www.bourns.com/data/global/pdfs/mfr.pdf
In the datasheet (page 3), it says that the fuses derate with temperature. At 40C, the hold current for the MF-R1100 is 9.4A. If this fuse is exposed to between 40 and 50C ambient temperature (it IS sitting below a 108W heater!), it might blow after a few minutes of operation.

Of course, my house isn't 40C. But I have a theory for how it might get there. The PCB is oriented with the fuse facing downwards in normal operation (trapping rising hot air). The fuse is boxed in on 3 sides by the PTC for the extruder, the power connector, and the heater connector, while also being in a semi closed environment under a heater. Imagine air heating off the large flat sides of the PTC, rising up towards the board and getting trapped by walls on three sides. If the PTC is a significant source of heat, then it is possible that it could heat up significantly above ambient. Here is a picture of how I see this:


The max resistance in this fuse is .01Ohms, so at 9.25A it would dissipate .86W- not a tremendous amount, but possibly enough to trip the fuse when it is also warm during the day-time (+20C above ambient?). It blew when my house was around 20-25C, but now that my house is 15-20C, it has been working quite well. I am probably right on the cusp of working/not-working. This is probably a rare problem to have, but I bet we see more of this when the weather warms up next year.

The highest fuse Bourns offers in the MF-R line is the MF-R1100. So instead, I am going to order the Littelfuse 16R1200G ($.94) and 16R1400G ($1.09) from digikey:
http://www.digikey.com/product-search/en?vendor=0&keywords=16R1200G
http://www.digikey.com/product-search/en?vendor=0&keywords=16R1400G
It seems that I am very close to the working boundary, so the 12A fuse will most likely work. People in very hot climates might need the 14A fuse.

 

My apologies, I should have put more attention to my language and not used the word fuse. Everywhere I said fuse, I really meant "PTC Resettable Fuse" or "Polyswitch". I should have clarified that.

The point I was attempting to make was that the *Polyswitch* on the control board appears to be just slightly undersized for the task at hand, given its load, location, and orientation. Having acheived commercial success with a shipping product and revenue stream, R3D should consider a custom build of the RAMPS board (it is open source) that has a different *Polyswitch* on it. R3D must have placed an order for at least 1,000 RAMPS boards, so why would they not ask the supplier to replace the one part that is slightly deficient for our application?

To me, that seems more intuitive and efficient because it requires less manufacturing labor and lower BOM cost than downgrading the temperature control technology from SOLID STATE PWM (21st century) to MECHANICAL ON-OFF (20th century).

But that still leaves me with a question: If the proposed solution is to bypass the *Polyswitch*, why bother with a relay? Why not just replace the polyswitch with a piece of wire?

 

My fuse reads: X16 CF1100 or GF100.
The X looks more like a logo than a character.

 

Where are you seeing that. I can't read anything on mine it is just a yellow square....

 

Hard to read, orange on orange background It should be on the inside, facing to F1.

 

I can't read that without bending them. My F1 says something like R500 1335 though