Sanguinololu, 27V and PCB Heat Bed

Hi,

I have bought a Sanguinololu and a PCB Heat Bed (MK2) and intended to run it on 27V, but have only just considered the ratings.

The Sanguinololu is rated for 35V but everyone who uses the PCB Heat Bed seems to use it at 12V.

At 12V (10A) it should dissipate 120W but at 27V (22.5A) it will dissipate 600W ! The hot-end should have the same problem too.

Changing the firmware to limit the PWM should be able to limit the average power so that it can be used like this but has anyone ever run these kind of setup on this voltage before ? (Sprinter has #define HEATER_CURRENT 255 in heater.cpp for the hot-end, but for the heat bed it doesn't seem to use PWM - How annoying.)

I wanted the higher voltage to get better performance from the motors.

Thanks,
-Andrew.

Edited 2 time(s). Last edit at 04/01/2012 08:29AM by yngndrw.

Even with PWM the peak currents would be too high. You need to use higher resistance heaters or return the positive leads to a 12V supply.

---

[www.hydraraptor.blogspot.com]

Thanks,

I'd forgotten that it just switches the ground and that I can just use a separate 12V power supply for the heaters, while still maintaining the 27V power supply for the motors.

It's a shame that the Sanguinololu board doesn't have separate heater and motor supplies.

The bed connector is severely underrated in my opinion so I intend to connect the bed directly to the tab of the MOSFET and the other wire to my 12V PSU anyway.

You might need a heatsink on the 5V regulator if you feed in 27V. Ideally that should be on the 12V feed also.

---

[www.hydraraptor.blogspot.com]

I might end up just pulling out the MOSFETs and replacing them with pin headers so I can have them external (Would allow me to break out both the signal and the ground in a single connector) but I will save that for later - For now I will just use it as-is, but with a 12V supply. I'm going to be pretty frustrated if the board doesn't just work with the PCB heat bed and 12V as that is what it was designed for.

The MOSFETs are also too close, which prevents you from mounting those nice little individual heat sinks.

I did look at a switching replacement for the LM7805, but the ones which could take 35V were very expensive. Might make one at a later date if it becomes a problem.

Having said that, I just searched again and found this:
[uk.farnell.com]

If you remove the MOSFET you can link drain to gate and then use the original connector to take the signal to the external MOSFET. An external one also saves the heavy ground current flowing through the board.

Quote

I'm going to be pretty frustrated if the board doesn't just work with the PCB heat bed and 12V as that is what it was designed for.

Yes I was pretty disappointed when I started looking at mine. Another issue is the MOSFET is not suitable for direct logic drive. Its only starts conducting at 4.5V worst case, and is nowhere near fully on with 5V. The same is true of GEN7. RAMPS uses a more suitable MOSFET.

---

[www.hydraraptor.blogspot.com]

The only problem with bringing the signal out of the main connector is that there is no ground connection to use as a signal reference.

Do you think that the PCB heat bed alone (I.e. with my own drive electronics) would be able to take the peak current from 27 / 35V PWM ? It would be a pain to have to use separate power supplies.

Thanks for pointing out the MOSFET issue, I'll have a look at the MOSFETs that RAMPS uses. I hate looking through pages of different MOSFETs (Or any component) trying to find the most ideal one - There are always so many similar ones and I seem to end up sinking hours into it just for one component.

Unless you use something like an opto coupler there is no separate signal reference on a MOSFET. The source pin is both the power ground and the input reference, so provided your power supply grounds are solidly connected, it should work. Always a good idea to put a series resistor close the gate pin to prevent oscillation.

The PCB will take high current surges provided the PWM is fast compared to the speed the tracks heat but you will get bigger losses in the wires and the MOSFET I think and you rely on the firmware not crashing or your bed may fry.

---

[www.hydraraptor.blogspot.com]

I'm a little concerned with how good the ground would be between the two power supplies, but I guess for this usage it doesn't really matter.

I'd probably have both the heaters controlled by a separate microcontroller so that if the main microcontroller does crash the heater will still maintain its temperature. (With a watchdog)

Hi nophead,

Having just had a problem with my connector frying I need to sort out an alternative. Have you modded your Sanguinololu yet? If so, will you be blogging about it (or sharing the details). As I said on my Blog, I'm no electronics guru so don't even know where to start.

Thanks,
Craig.

---

----------------------------------------------------------------------------------------------------
[blogger.kritzinger.net]

It looks like you have only connected one pin for each cable, halving the capacity of the connector.

You should connect pins 1 and 2 to the first cable and pins 3 and 4 to the second cable.

Thanks Andrew.

Nophead and Triffid_Hunter have also replied on my blog.

All advice is much appreciated!!!

I guess the same applies to the hot end!!

---

----------------------------------------------------------------------------------------------------
[blogger.kritzinger.net]

Most hot ends only need < 2A, so a single pin is sufficient theoretically but two wires won't hurt.

This is how I have my Sanguinololu wired at the moment: [reprap.org]

Edited 1 time(s). Last edit at 04/09/2012 04:43AM by nophead.

---

[www.hydraraptor.blogspot.com]

Nophead: I just noticed that you aren't using a heatsink on the heat bed MOSFET - How is that holding up for you or have you changed it since ?

I took a while away from the RepRap stuff so I have no idea what has changed since I was last here.

It holds up fine although it isn't on a machine that I use 24/7 and I only print PLA on it.

I am not sure why it works as calculations with the datasheet RDSon of 0.047R suggest it would dissipate ~5W, which would require a heatsink. The max you can dissipate with a T0220 without a heatsink is about 2W. It does get too hot to touch and I think the thick wire attached to the tab takes some of the heat away.

It may be that I am lucky and the actual MOSFET has an RDSon much better than the datasheet worst case.

When electronics is designed correctly it is guaranteed to work in all circumstances. When it is isn't it may work in some instances because semiconductors have a large spread of characteristics. It should have a heatsink or a MOSFET with a lower RDSon.

---

[www.hydraraptor.blogspot.com]

Ah okay, I ordered a 0.5C/W heatsink but after seeing yours I was wondering if it was worth leaving it with the aluminium double U channel which I've put on it for a heatsink. On one had I wouldn't have to desolder the MOSFETs but on the other hand with the large heatsink and a fan I should have no problems at all. I do intend on printing ABS in the future so I should probably just get it fitted.