Maker Pro
Maker Pro

Just murdered a MOSFET. What did I do wrong?

Kelly Corcoran

Aug 6, 2017
6
Joined
Aug 6, 2017
Messages
6
Custom-designed pinball machine project here, trying to create a solenoid driver board. Decided to test it out with my first MOSFET driver before I solder all the rest of them, and I'm glad I did because apparently I did the first one wrong.

Transistor is a IRF510 N-channel MOSFET rated at 100V, 5A, trying to drive a 50V circuit. Didn't test the amperage of my solenoid circuit, but as it runs through a 2A fuse, it couldn't be much. Matched up the transistor's leads to the data sheets provided by supplier Jameco Electronics.

The gate is connected to an output pin of a Teensy 3.5, which fortunately was not harmed. It is also tied to ground through a 10K resistor, as I was advised to do to avoid floating gate voltage. Source connects directly to ground, and drain connects to my test device by way of an alligator clip.

When I tested the circuit using a 6.3-volt LED (driven by an appropriately rated wall wart), it worked exactly as intended. Unfortunately, when I tried to use the same setup to operate my flipper coil (with flyback diodes), nothing happened. The LED test no longer worked after that, until I touched another transistor to the first one to piggyback on it. So my HV circuit definitely killed the first transistor.

How can I solve this problem? Transistors in parallel? Darlington pair? Higher-rated MOSFETs? Different circuit arrangement?
 

Audioguru

Sep 24, 2016
3,656
Joined
Sep 24, 2016
Messages
3,656
Maybe the flyback diode was connected backwards?
The IRF510 Mosfet needs a gate to source voltage of 10V to fully turn on. What voltage did the "teensy" give it?
 

Kelly Corcoran

Aug 6, 2017
6
Joined
Aug 6, 2017
Messages
6
Definitely good thoughts.

The diodes do face the right direction. One of them didn't have such a good connection, but it seems to me like I'd have gotten at least one good flip before that problem would have caused an issue. The coils work fine when I do the "switching" by bridging them to ground with a jumper wire.

I should have given that data sheet a better look as I did not realize it expected 10V to switch it. The Teensy puts out 3.3 volts on the pins, although apparently this was enough to switch the LED on and off. Off hand, is there a better transistor to use?
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
You either need a logic level mosfet, a gate driver, or (probably "and") for you to read and understand the datasheet to some extent.

A small gate voltage, as long as it's above the threshold voltage for the MOSFET, will allow it to switch a small load. For heavier loads, higher gate voltages are required. Typically you would use a gate voltage significantly higher than the minimum technically required -- especially considering it can vary from device to device and with temperature.

Looking at the datasheet, the threshold voltage can vary from 2 to 4 volts, so your sample was probably around the middle or lower (considering it could switch a small bulb).

When asked to switch a couple of amps, several things may have happened.
  1. It may not have allowed significant current, yet still enough to cause an inductive spike when it switched off. An incorrect, failed, or absent protection could allow the device to be destroyed.
  2. The device may have turned on sufficiently to allow a significant current, but with a significant voltage across it. The load may or may not have operated. The current through the MOSFET may have lead to heating sufficient to destroy it.
  3. A combination of 2 and 1.
How did the MOSFET fail? Is it shorted (a typical failure mode)?
 

Kelly Corcoran

Aug 6, 2017
6
Joined
Aug 6, 2017
Messages
6
I suppose a bit of comprehension goes a long way. There have been enough posts from different sources mentioning how simple MOSFETs are to set up that I didn't realize they still do need a lot of research to get right. I'll check out the catalog and see if there is something more appropriate for this application.

Both of your failure guesses seem possible, especially (hindsight is 20-20) with the poorly attached diode. I've encountered shorted transistors before, but this one has the slightly unusual failure mode of being permanently open. I noticed it was a bit warm when I was trying to test it with the solenoid, if that matters.
 

Audioguru

Sep 24, 2016
3,656
Joined
Sep 24, 2016
Messages
3,656
An IRL510 is a "logic level" Mosfet and is fully turned on with its gate-source voltage at 5V (logic is usually 5V) and it does pretty well at 4V. Sorry, I do not know a power Mosfet that turns on well with only 3.3V but one might be available.
 

dorke

Jun 20, 2015
2,342
Joined
Jun 20, 2015
Messages
2,342
Well,
A good starting point is to provide a detailed schematic including all devices part numbers .

When using this kind of MOSFET to drive an inductive load like you do,
a fast clamp diode on the load is crucial.
Did you use one? which part?

That MOSFET is not suitable to be operated from a 3.3v logic at it's gate.
A min. of 4V is needed.

Looking at "dead" Mosfet:
1. Any signs of discoloration?
2. Is the device cracked?
 
Last edited:

Kelly Corcoran

Aug 6, 2017
6
Joined
Aug 6, 2017
Messages
6
Ask and ye shall receive. While it may be a crude pencil drawing, I made sure to include all the components I am currently using. Item numbers included where applicable. The full circuit is drawn in two sections - the main power supply/rectifier (top) and the coil itself with its logic switching circuit (bottom).

The transistor shows no outward signs of damage; the only symptom of its demise is the fact that it will no longer switch any current from drain to source.

The nature of my project gives me access to a 6.3VDC power source meant for the lighting circuits. How well would that work to switch my existing IRF510s? If I could expect reasonable results, I can bridge the gap with a 2N7000, which seems like it might be up to the task.

Remaining:
- Flyback diodes are 1N4003. They seemed to work before, but then I'm not sure I've used those coils in a logic driver circuit before.
- I recall hearing a suggestion to put another flyback diode from source to drain on the MOSFET. Would that help protect it?
 

Attachments

  • IMG_20170807_181554209.jpg
    IMG_20170807_181554209.jpg
    81.6 KB · Views: 109

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Have a look in the resources section. There is a circuit in there which will allow you to drive your mosfet gate from a higher voltage (12v would be ideal, but you'll probably get away with 6v)
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
I'm not sure what happened to my last post, but look here for a gate driver:

http://www.piclist.com/techref/mimedecode.asp?url=piclist\2004\06\21\220147a.txt&part=2

I've described this corrupt a number of times here. 12V would be best, but 6v should be sufficient.

It's a bit of overkill, but it will do the job from your low voltage input.

Beware that it is inverting!

However I'm not convinced the problem is solely the low gate drive. For the fault you're seeing, it seems more like a failure of the gate oxide near the source. This is caused by a high voltage spike being presumed at the gate. It is worth measuring the resistance from gate to source and from gate to drain and see if you get anything other than an open circuit
 

Audioguru

Sep 24, 2016
3,656
Joined
Sep 24, 2016
Messages
3,656
You can buy hundreds of iRF510 Mosfets, test them all and maybe find one that turns on good enough with a gate-source voltage that is less than 10V. Maybe none are available.

A 1N4003 is designed for 50Hz to 120Hz. It switches too slowly to protect the Mosfet from the very fast high voltage spike produced by the coil when it is turned off.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
The issue with diode speed is how fast they turn off, not on. (Assuming your not using PWM)
 

Bluejets

Oct 5, 2014
6,901
Joined
Oct 5, 2014
Messages
6,901
May not be relative to the fault, but your 50 vac is not 50v dc as shown, more like 70v dc.
 

dorke

Jun 20, 2015
2,342
Joined
Jun 20, 2015
Messages
2,342
There are a few issues with your design:

1. Secondary voltage of the transformer is too high to generate a 50VDC for the solenoid.
it will create a DC voltage of around 65V(with peak that can be at around 74V!) .
This will create a larger "ON" current which may damage both MOSFET and the solenoid!

2. The primary fuse should be lower in value than the secondary one!

3. You should use a single fly-back schottkey diode ,
connecting 2 diodes in parallel is worthless anyways.

4. Is the solenoid built like you have shown? 2 coils?

5. One solution to the insufficient driving voltage is to use a power Darlington .
Which needs lower voltage to be on and may be driven directly.
As a minimum, It should be an NPN,80V,3A type.

6. An advanced issue,
What is the rise time/fall time of the control board driving signal?
 

Kelly Corcoran

Aug 6, 2017
6
Joined
Aug 6, 2017
Messages
6
Wow, I did not get notified of my new messages. Had other parts of the project to work on anyway.

I will have a look at the alternative components you guys recommended and see what works.

Addressing the notes:

I will test the actual DC voltage to find out what it is, but some games do actually use around 70VDC for the high power coils so this may be okay.

Did not know about the fuse values. What rating should I give the AC fuse?

The two parallel diodes are as they come factory-installed on the solenoid, which is indeed two separate strength coils wrapped up into a single package. Normally they ship with 1N4004 flyback diodes pre-installed, but they must have either broken off or been (inexcusably) removed at some point. My arrangement of the circuit is designed like other flipper circuits on production games.

I definitely like the driver suggestion. That might be my solution.

I'm not sure about the signal time or how to find this information, but I believe it can switch in under 10 microseconds based on tests I've done with "dimming" devices on the pins.
 

dorke

Jun 20, 2015
2,342
Joined
Jun 20, 2015
Messages
2,342
Assuming the 2A fuse at the transformer output is the correct value needed(it's location is not "traditional" though) ,
the rating of the AC-line fuse should approx. be 2A *50VAC/115VAC=0.87A.

It is important that you measure the solenoid "on current" !
or better yet, post it's datasheet here or part number and make.

In your application,are you using more than one "solenoid channel ",if yes ,how many?
 

Kelly Corcoran

Aug 6, 2017
6
Joined
Aug 6, 2017
Messages
6
Thank you for the fuse tip! I will keep that one in mind.

I'll measure the solenoid current when I get a few spare minutes to hook up the test circuit again. I looked up the part number (FL-11753) and did not find any data sheets. This seems surprising as it is a very common part in this industry. I did, however, find resistance values for the two inside coils. Pasted:

Primary winding (power) ~9.8 ohms (no diode)
Secondary winding (hold) ~165 ohms (no diode)

By solenoid channels, are you referring to separately-fused voltage rails? I wasn't planning to do this, but I probably could if it seems like a good decision.
 

dorke

Jun 20, 2015
2,342
Joined
Jun 20, 2015
Messages
2,342
The FL-11753 solenoid is 50V /5A.
You shouldn't operate it with 70V! make sure that your P.S supplies 50VDC.
The MOSFET you are using should be rated higher current than 5A (8A?)

That "Flipper assay" drawing is something you made yourself from an existing device?
or did you build it from scratch?
It doesn't look correct.
Please post a good photo of it ,showing parts and connections.
Especially the EOS connection.
 
Top