Low Heated Bed Resistance - Solution

I have a heated bed PCB with a resistance of about 0.6 ohms. I had read in other posts that these are essentially defective but I decided to try it out anyway with a RAMPS 1.4 configuration.
It worked for about 3 prints at 60 degrees however when I tried to raise the temperature to 110 degrees the fuse on the RAMPS blew and since this time I was not able to even get to 60 degrees without the fuse blowing. I thought that I was stuck here, unable to efficiently reduce the current, and complained to the supplier of the PCB. Then a thought hit me, why not duty cycle the PCB.
So, for all of those stuck with a heated bed with low resistance, I have made some modifications to Marlin which duty cycles the heated bed. The modifications are in the Dabble63 fork and I have opened a pull request. The parameter is HEATER_BED_DUTY_CYCLE_DIVIDER in the configuration.h file.
With these modifications I am able to raise the bed temperature to a maximum of 82 degrees, not enough for ABS but plenty for PLA.

After glancing over your commit I think what you do here is software PWM, at a fixed rate of 50%.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

You are correct that it is a software PWM, it is crude but appears to be effective. The rate seems to be very slow.
The logic is that bed_is_on is incremented from 0 to HEATER_BED_DUTY_CYCLE_DIVIDER repeatedly. When bed_is_on is 0 i.e. for 1 out of HEATER_BED_DUTY_CYCLE_DIVIDER times, the heat bed is turned on.
I found that a value of 4 i.e. turn on for 25% of the time, worked for me.

The current with half the resistance is doubled, so the heating in the MOSFET and the fuse will be quadrupled, hence why you need 25% PWM, but then you only get half the power.

The Prusa bed has four sections in parallel, so you could rewire them as two pairs in series to get the normal resistance.

Also the resistance increases about 40% with temperature, so you could increase the PWM with temperature to get a higher final temp.

---

[www.hydraraptor.blogspot.com]

Unfortunately the while the power to the bed is halved, the power to the fuse is not....
I tried increasing the power to 3/8 after 75 degrees but this blew the fuse. The resistance at 70 degrees is about 0.8 ohms, my multimeter is not very accurate at these values, so while I am sure that the resistance increased, it is hard to say by how much.
I like your idea of rewiring the PCB. I did not realize that it has 4 banks in parallel.

If the resistance of each bank is R then the current resistance is R/4. With 2 banks in series I would have 2R/2=R whereas the desired value is R/2.
In other words, 2.4 ohms resistance and once again half the power although this time no problems with the fuse .

The only idea I have is to disconnect two of the banks e.g. the outside ones leaving a smaller print area. If you have any other idea please let me know.

You can wire the positive bed connection direct to the PSU to avoid the problem with the fuse but you will need a very good MOSFET to handle twice the current and most PC PSUs will probably die if you take 20A from them despite what they say on the label.

Can you get your money back from the supplier? Presumably they made the mistake of getting the board plated. I get mine from here: [blog.think3dprint3d.com]

---

[www.hydraraptor.blogspot.com]

Been thinking about what I will do with the bed. Before I had the idea of PWMing it I ordered a new one from think3dprint3d, by chance the one you recommended, and it is on its way. I have not yet had a response from snipermand which was where I ordered the first board from.
My power supply has a 21A and an 18A 12V supply so it should be able to handle the current. The MOSFET will probably fry however. I also like having a fuse in the supply line. For the moment I will probably keep the board for printing in PLA or smaller items in ABS - if I cut two tracks it should heat up the middle part very quickly. The alternative is to build an external switched circuit which can handle the current, as there is no need to switch quickly, perhaps a relay will do the trick.
Thanks for the ideas.

I had exactly the same problem, although my solution was fairly long winded!

1) use a sheet of aluminium covered in PET tape to print onto, mount the PCB to the underside of this.
2) Relocate the mosfet for the heated bed to the aluminium sheet, this involved breaking some tracks on the sanguinolou board I was using, I have no experience of RAMPS 1.4. Basically I took the PWM signal and earth from the sanguinolou and used that to switch the mosfet, then powered the bed directly from my PSU.
3) Use a server PSU because they can supply ~40A @ 12V and cost ~£5.

The aluminium sheet acts as a heatsink for the mosfet. This limits my bed temp to 150deg but I haven't felt that an issue when printing in ABS.

I might have guessed it was snipermand. He has a long history of selling things to the wrong spec. Stepsticks that only do x8 for example. I bought some "Makerbot" stepper drivers from him that turned out to be home made PCBs with no voltage regulators or leds and a socketed chip whose datasheet says don't use a socket.

---

[www.hydraraptor.blogspot.com]

Hello everyone!

I, too, bought a heated bed from snipermand only to find myself in the OP's position. I tried nophead's fix of making the four parallel sections into two series sections in parallel, however the resistance came out to be 2.6Ω so I decided to try 2 parallel sections and one series section in parallel with the other two. This gave me a resistance between 1.2-1.4Ω which, correct me if I'm wrong, I think is OK. In order to join up two of the sections into one I simply scraped some of the lacquer off using a hobby knife making sure I take as little copper as possible when doing so and then just put a blob of solder over the two exposed copper connections. Using the same hobby knife I then grinded away through the tracks at the points where the sections join with the voltage buses on the side and voila. I don't yet have my Prusa up and running in order to test this heated bed but I see no reason why it shouldn't work.

If you guys are interested I can take some pictures to show you what I mean with joining the tracks.

It is not clear to me exactly what you have done. It does seem to me that you have different paths for your current with different resistances and hence the current will not be distributed evenly among your heating elements. The result would be uneven heating.
If I am wrong, please show a picture or post a circuit diagram.

Ahhh, I think you may be right with the uneven heating thing. I'll just have to wait and see until my Prusa's all done. The PCB has 4 sections of equal resistance wired in parallel. I now have 2 of those sections untouched and the remaining 2 joined up in series. This new 'series section' is wired in parallel to the other 2 sections. In total I now have 3 sections. I use my PCB with the unetched copper trace upwards and have a 4mm glass mirror on top of it covered in Kapton tape. Hopefully this will help even out the heat around the platform. I'll take a picture tomorrow if you're still interested.

Thanks for the explanation.
Unfortunately you will be delivering 4 times the power to the two sections wired in parallel, compared with the 2 sections wired in series.
I bought a new board from think3dprint3d and I am very happy with it.
My advice, for what it is worth, is to use the board to print smaller items, up to half the size of the board.
The alternative is to build a small external circuit with a relay used to shunt the approx 20 amps of current. If your power supply can handle this current, it will give you a bed that heats up more quickly.

Thinking about this some more....it suddenly occurred to me that the heated bed circuit is independant of the rest of the RAMPS 1.4. The consequence of this is that you could connect the 5V supply line to it and thus limit your current. Unfortunately you will be only transferring half of the power to the heater that a normal heater board would use. The result of this solution would then be similar to the pulsing that I performed in software. This means that the board would be good for PLA but not ABS.

Hi I know this is an old thread but I just wanted some clarification on this...

I have a 24v 10a power supply that I want to use for the bed but I need to increase the resistance in order to use it..Or use this Duty cycle Mod....

I have brielfy tested it and I do get a lower current draw but I'm worried about damaging the PSU and/or the electronics. Should I disregard this Duty cycle Mod and just focus on increasing the resistance of the PCB board, (I have MK2 from reprapworld.com) by rewiring it into 2 series banks in parallel?

Pentatonic: If you still have a photo or circuit diagram that would be a great guide for me..

Cheers,


Edit: Update: Ok I made the mistake of wiring two series banks in parallel which gave me too much resistance (should be 8ohm, but measured 6ohm). So then I wired it as two parallel banks in series, which should give me the desired 4 ohm (for 24v) but it measured 6ohm.... I thought it must just be a bad connection where I have modified it so I soldered it back to double check the original circuit...It reads 2ohm... Am I miscalculating somewhere???

Edit: Yes I am... I calculate that each track is 8ohm so when in parallel you get 2ohm. I can't seem to find a permutation which allows me to double the resistance to 4ohm while keeping all 4 tracks (for equal heating). Can anyone help me here?

Edited 3 time(s). Last edit at 04/18/2014 10:48AM by metas.

If you really want to do it with this board, you could cut one of the four traces into thirds (8/3), and add a third of that trace to each of the full traces(8+8/3), then with three parallel circuits you should get 1/(1/R+1/R+1/R) = 1/(3/(8+8/3))=3.5556 ohms, so V/R=I gives 24/3.5556=6.75A and V^2/R=P gives 24*24/3.5556=162W. Each branch would have the same resistance, amps, and heating.