PWM Driver Board and use with solenoids

I'm using the PWM driver board v1.1 from Makerbot in another project.

I understand there's a bug in the board, and that the solution has been identified.

Having the diode placement corrected, a picture of my board is shown here.

We did replace the TIP 120 with a MOSFET, as documented also in the last link, so that we would avoid having to use a heatsink. Everything else on the board is according to the RepRap directions.

The board should be rated for 5 amps per channel. We are using the board to power 3 amp solenoids. I have burned out 2 MOSFETS so far - I assume from back EMF from the solenoids. The boards worked well with a smaller set of solenoid valves, but now with this new set of solenoids referred to on the same page as the last link, I'm burning out the MOSFETS.

Does anyone have experience with using the PWM Driver v1.1 with solenoids as a load, or have suggestions on correcting the above problem? I was planning on using a larger MOSFET and a larger diode as the next step.

Hi Marcin,

Does the FET burn out because it gets hot, or does it just die for no apparent reason?

What MOSFET are you using? Naked MOSFETs are very easy to destroy. I would always recommend protected MOSFETS. The cost a bit more but they have built in ESD, over current, over voltage and over temperature protection. They are almost indestructible.

The diode in the picture looks like 1A. You should really use a bigger one for 3A loads but, unless the diode has gone open circuit, it should still be giving BEMF protection. Because the BEMF pulse is short you probably get away it.

ESD will destroy MOSFETs very easily, or any sparks due to loose connections. Also selecting one with a gate threshold too high for logic drive will cause it to get hot.

Another thing that can destroy FETS is that they can oscillate at very high frequency but, as long as you have a resistor in series with the gate, that is unlikely.

---

[www.hydraraptor.blogspot.com]

what fets are you using? how much current can they carry with Vgs=4.5v and no heatsink? I'm using the IRL3803 which can handle 16A under these conditions.

Also, 1N40* and 1N54* are next to useless for freewheel diodes- they're too slow! (but still better than nothing). you need a shottky type, like 1N5822

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

The slowness is only in switching off, so unless you are doing high frequency PWM (unlikely with a solenoid), Schottkys are unnecessary.

---

[www.hydraraptor.blogspot.com]

We are using these MOSFETS:

[search.digikey.com]

And these diodes:

[www.mouser.com]

They are attached on the circuit board as shown here:

[openfarmtech.org]

I found that as soon as the solenoid was turned on and off, the MOSFET fried. There is no visible signs of damage or heating on the MOSFET, nor on the diode. The MOSFET, measured right on the circuit, is conducting between all leads after it failed. The diode appears to be working still - measured on the circuit, it conducts one way and blocks the other way using a diode test on a multimeter. The measured current draw on the solenoid is 3.7 amps. There is significant sparking between the battery leads and the solenoid upon disconnecting, when the solenoid valve is tested by itself for an amp reading. The solenoid valves are turned on for steady state operation in real working conditinos, one at a time, for intervals of about 2 seconds.

Given all this, what specific diode and MOSFET should I use?

Thanks,
Marcin

I am using this MOSFET:
[search.digikey.com]
and this diode:
[www.mouser.com]
The control board looks like this after placing the diode in the correct location:
[openfarmtech.org]

The MOSFET has no signs of damage, it did not get hot, but fried right after the solenoid was turned on and off. The steady state amperage on the solenoid was measured at 3.7 amps. In real life,the solenoids are turned on and off for approximately 2 second intervals. When testing the solenoids independently with a 12v battery,they produce a considerable spark at their connection terminal when a power wire is detached.

With this said, what explicit MOSFET and diode would work?

Thanks,
Marcin

that mosfet isn't suitable because its threshold voltage is too low! at Vgs=5v the mosfet is only barely turned on and has a resistance significantly higher than 30mR. This condition will quickly fry it - The TO-220 package has 62C/w Tj-amb so you can only dissipate about 2 watts with no heatsink. Any more than this and the silicon junction will go over 175 celsius and start melting.

You need a mosfet that at least has Rds(on) @ Vgs=4.5v listed in the datasheet, and preferably a low enough Rds(on)@Vgs=4.5v that you don't need a heatsink.

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

Marcin,
I would recommend one of these: [gb.mouser.com] but they are out of stock, so one of these is overkill but should be bomb proof [gb.mouser.com]

I think the diode is OK because, although it is a 1A diode, the peak current rating is 30A and you are switching very slowly. With a protected FET it would actually survive without a diode, but the peak voltage would be enough to give somebody a shock, so best to have one.

---

[www.hydraraptor.blogspot.com]

Thanks a lot, Chris and everybody. I ordered the MOSFETs, and they are arriving Tuesday, and the problem should be fixed. I'll report on the results as soon as the circuits are tested with the solenoids.

The solenoids are controlled successfully with the Omnifet Fully Autoprotected Power MOSFET, VNP35N07 - suggested by Chris. Here is a video of the solenoids in operation - [openfarmtech.org] . Thanks, guys.