flywheel diode question

Thread Starter

Jassper

Joined Sep 24, 2008
84
I am using this MOSFET to control a Solenoid and I have it protected with this diode as a flyback. However once and a while I will still get a MOSFET to short from turning the coil on and off.
Oddly the Diode is still fine. Do i need a Larger Diode? the voltage level on the coil is +12v

This is the basic configuration, Imagine the Lamp as a coil.


I should also mention that the Diode is internal on the board with the MOSFET and the coil is external.

THANKS!
 
Last edited:

THE_RB

Joined Feb 11, 2008
5,438
Your diode link took me to a clunky slow Mouser page which then crashed. :(

What is the spec on your diode? If you are relying on it to handle PWM you need a high speed diode.

Please state the specs of the FET and diode, ie volts, amps, and details of the load ie coil amps and PSU voltage etc.

The more info you provide (without forcing people to go look it up!) the better the help you will get. :)
 

Thread Starter

Jassper

Joined Sep 24, 2008
84
Sorry, wasn't trying to "force" you to do anything, but it's kind of hard to post all the info from a Data sheet, so I attached them to this post.

The MOSFET is used as a switch, either ON or OFF, no PWM.

MOSFET
32 A, 60 V, RDS(on) = 35 mΩ(Max.) @ VGS = 10 V, ID = 16 A

Diode Schottky barrier rectifier
If 2A
Vr 60v
Vf 530mv Max
Max Surge Current 50 amp


The coil draws about 800ma when powered @12 volts.
The MOSFET is driven by a Micro-controller @ 5 volts on the Gate through a 680 ohm resistor.


@MaxHeadRoom
Thanks, but no, the coil is rated <1amp and the MOSFET should be able to handle at least 30 amps (with proper heat sink)



 

Attachments

Last edited:

crutschow

Joined Mar 14, 2008
34,201
The diode anode should be close to the MOSFET drain and the diode cathode should go directly to the plus side of the solenoid. The inductance of long leads from the diode to the transistor can generate spikes at the MOSFET drain.

Does the MOSFET get hot in operation? You should use a logic-level device as Max suggested to be sure it is fully turned on with 5V at the gate. The device you have needs 10V to fully turn on, as the spec you posted states.
 

Thread Starter

Jassper

Joined Sep 24, 2008
84
The diode anode should be close to the MOSFET drain and the diode cathode should go directly to the plus side of the solenoid. The inductance of long leads from the diode to the transistor can generate spikes at the MOSFET drain.

Does the MOSFET get hot in operation? You should use a logic-level device as Max suggested to be sure it is fully turned on with 5V at the gate. The device you have needs 10V to fully turn on, as the spec you posted states.

Thanks everyone - I will check those things out.
The Rds(on) of the MOSFET is 5v, sorry I posted the MAX value.
RDS(on) Static Drain-Source On-Resistance max/min
VGS = 10 V, ID = 16 A -- 0.027 - 0.035Ω
VGS = 5 V, ID =16 A -- 0.035 - 0.045Ω

I have the feeling that the issue is placement of the Diode as a couple of you mentioned. As I stated before, the Diode is on board with the FET, Anode to the Drain and Cathode to +12v however as it is installed the +12v for the coil may be coming from a long distance away - or possibly a different source, that helps confirm my suspicions.

Thanks again, I'll look at everyone's suggestions and report back later what I find.
 

THE_RB

Joined Feb 11, 2008
5,438
Thanks for stating all the device specs, it is appreciated. :)

32A 60v FET should be fine switching a 12v 800mA coil.

I think you might have some HV spikes on your 12v rail? Your schematic doesn't show any 12v decoupling caps, I would put a 2200uF 25v cap across the power wires at the PSB, and a 0.1uF poly cap in parallel with it.

Then a 30v TVS across the FET D-S like RonV suggested.

As this is switched at a very low frequency put a cap from the FET Gate to Source (gnd). That will make the FET turn on and off slowly, so you will get almost no inductive kickback and no HF noise on on/off transitions. Size the cap with the gate resistor (RC) to give about 50mS turn on or turn off ramp times. Then it wil lbe just about bulletproof.
:)
 

ErnieM

Joined Apr 24, 2011
8,377
The diode anode should be close to the MOSFET drain and the diode cathode should go directly to the plus side of the solenoid.
I would disagree with this. Instead I would put the diode as close to both ends of the solenoid as possible.

The solenoid itself is generating the back EMF, and I want to contain it to as small an area as I can, without any long leads in the build.

As "belt and suspenders" you might add a 2nd diode from drain to Vdd locally on the controller board.
 

crutschow

Joined Mar 14, 2008
34,201
I would disagree with this. Instead I would put the diode as close to both ends of the solenoid as possible.
..............
Well, I disagree with your disagreement. ;) You want to diode to protect against the wire inductance (which can generate a significant spike for a long wire length) as well as the solenoid inductance since the two inductances are in series. Look at the current path. It's though the solenoid, the interconnecting wire, and the switch. When the switch opens you want a bypass path for all the current to the switch to minimize the inductive effects of every element. Thus you want the diode at the switch.

In any case the back EMF is only about 0.7V with the diode so it won't radiate much energy. An unsupressed spike from the wire inductance would likely radiate more.
 

Thread Starter

Jassper

Joined Sep 24, 2008
84
I would disagree with this. Instead I would put the diode as close to both ends of the solenoid as possible.

The solenoid itself is generating the back EMF, and I want to contain it to as small an area as I can, without any long leads in the build.

As "belt and suspenders" you might add a 2nd diode from drain to Vdd locally on the controller board.
Thanks, that was my thoughts as well. I have a diode reverse biased on a test coil now - so far so good. I was hoping to get away with not needing one - but of course this is only a problem when the installer doesn't want to follow directions. The output was actually intended to switch a relay coil that has a suppressor diode across the coil such as a Bosh automotive cube relay. I'll still make necessary changes to my board.

Thanks again everyone for the suggestions, much appreciated :D
 

DC_Kid

Joined Feb 25, 2008
1,072
i suspect if you scope it you'll see some ringing, so a small rc snubber will smooth it out. this is of course you wish to smooth it out.
 

DC_Kid

Joined Feb 25, 2008
1,072
that fet probably needs to be driven harder



is your fet dying? if so the voltage is probably climbing too high, thus add a zener.
to control flyback i use 33V 3A zener with small resistor to gnd, and a snubber to kill the ringing

before and after snubber (the wide part is zener in action, etc). the inductor + cap in fet make it ring.
this is a fuel injector being turned off fast at 12vdc, the green marker is the 0v (fet drain when on), and then you see it climb.

 
Last edited:

Thread Starter

Jassper

Joined Sep 24, 2008
84
Thanks guys, I'm pretty sure the issue is Flywheel. I've protected against ringing - but I will check for that. If I disconnect the coil I can actually see and hear the arc as I pull the wire off the coil - so that's a pretty strong EMF. Getting the FET to fail is pretty easy to repeat as well. With a diode across the coil - no more arc and FET stays happy.

There are 4 other FETs on the board switching resistive loads with no issue, one of them spikes > 1 amp at start up. This FET is the only one that switches an inductive load.

It will be a while before I can get the board redesigned however, so an external diode across the coil might have to suffice until then.

This is basic representation but the pdf shows what I believe I need to correct.
 

Attachments

Last edited:

DC_Kid

Joined Feb 25, 2008
1,072
i would clamp the Drain of fet to GND with a 33v 3A zener in series with ~20ohm.
have a read http://www.eao.com/global/en/Downloads/technical_articles/EAO_TA_Inductive-Loads.pdf

you are probably frying the fet from over-voltage transient spikes.

from another doc
"...and can blow out the driving MOSFET, as they are very sensitive to voltages greater than Vds(max)."

that inductor is likely spiking well above 60v when it turns off hard.
 
Last edited:

ErnieM

Joined Apr 24, 2011
8,377
Well, I disagree with your disagreement.
Your point is well taken.

When you have a significant length of wire between the switch and the load a local diode would also be required.

I'd use two diodes for that situation. Yep, the one at the FET side is the most important one so that is what I would use when I'm only using one.

However, the back EMF may be well above that 0.7V you say. I once worked for a company that made the suppression networks for relay manufacturers. Got to design one myself too. They require a 1,000 volt diode plus a 36V zener (and sometimes even a fuse). We made devices for several relay manufactures this way so it is a common technique.

So the relays with internal suppressors (right at the relay where I want mine too) will spike up to 37-40 volts above your power supply rail. That's 52V on just a 12V system, and if you thought you'd be safe with a switch device rated at 30 volts....
 

Thread Starter

Jassper

Joined Sep 24, 2008
84
i would clamp the Drain of fet to GND with a 33v 3A zener in series with ~20ohm.
have a read http://www.eao.com/global/en/Downloads/technical_articles/EAO_TA_Inductive-Loads.pdf

you are probably frying the fet from over-voltage transient spikes.

from another doc
"...and can blow out the driving MOSFET, as they are very sensitive to voltages greater than Vds(max)."

that inductor is likely spiking well above 60v when it turns off hard.
I will probably do that too - I don't have the proper equipment to measure the spike (time to invest) But if you go by the "rule of thumb" it can be 10 times greater than Vcc so it is probably seeing a 120v spike, at least.
 

DC_Kid

Joined Feb 25, 2008
1,072
I will probably do that too - I don't have the proper equipment to measure the spike (time to invest) But if you go by the "rule of thumb" it can be 10 times greater than Vcc so it is probably seeing a 120v spike, at least.
which will toast the fet, which is why i suggest you add a zener. the voltage can still climb with just a diode across the coil, etc. clamp it down to something below Vds(max), etc. 50% is a good #, ~33v zener will do. the diode across the coil allows the coil to stay "active" for a short period. you can leave it and add zener if the circulating coil current is not a issue for its function, etc.

let easy math find your voltage spike (is the coil a std part # ?, need to know its inductance value if you dont have a Z meter)
http://www.learnabout-electronics.org/ac_theory/inductors02.php
the fet specs can be used to estimate on/off times.... btw, how is the nfet gate being turned on and off?? is there a pull-down somewhere?? the current is 12v across the dc ohms of the coil.
 
Last edited:
Top