Issues with mosfet heat

More
5 months 3 days ago #12958 by drivemusicnow
I assembled the board and noticed that the mosfet was getting hot, when through everything and found a stray cut wire that was shorting two pins of the 4 pin connector. I fixed this, and the mosfet stayed relatively cool. In the mean time, I was waiting on some components to finish the wifi module assembly. Once I had everything together, I did not check the mosfet again, but it was functioning fully (with wifi). I let it run this way for less than an hour when I smelled smoke, the mosfet was very hot, and it had burned part of the board. I tried to remove the mosfet and of course also pulled up the trace between it and the 100mH inductor. I replaced the mosfet (IRF640N) (largely based on reading this thread: www.tubeclockdb.com/forum/12-arduino-nix...nixie-clock?start=12 )and was getting just a green LED and the power light to flash.

I figured the heat/power must have taken out the inductor as well, so I went ahead and replaced that with a slightly higher amperage and lower tolerance 100mH inductor. But I have the same problem. Only the first tube LED flashes green, and the power light flashes when I turn it on. No high voltage is running on the board. Not sure what else I could have damaged and how I should diagnose.

Some quick info: I hhave the modular arduino kit, I have a good multimeter, but I do not have a scope. I'm very much a "hobby" electronics person, with no real background in it. Thanks for any help you might have.

Please Log in or Create an account to join the conversation.

More
5 months 1 day ago #12961 by Ty_Eeberfest
Missed this post somehow. Sorry!

Are you saying that the red Power LED on the rear edge of the board near the power input jack is flashing?! That is not right and indicates a problem in the low voltage (Vcc) circuitry.

It could be your external power supply has gone bad, or it could be that the voltage regulator (5-pin TO220 device "IC1" next to the 3-pin MOSFET) got damaged when the MOSFET cooked. Note that if Vcc has gone unstable it will make the CPU "spazz out" which could be the reason the MOSFET cooked in the first place.

Check the voltage coming from your power supply and make sure it's what it's supposed to be (9v or 12v depending on what power supply you have). Check it under load, not just open circuit. Check Vcc to Gnd at the test points on the board. Should be a very stable 5 volts +/- 0.2 volts.

Which revision of Modular do you have: the one with the MOSFET & regulator standing up or the one with the MOSFET & regulator lying down flat??

Look into it later when the dust is clearing off the crater.

Please Log in or Create an account to join the conversation.

More
4 months 1 week ago - 4 months 6 days ago #12998 by vladca
Replied by vladca on topic Issues with mosfet heat
Hello.
I made 6-digit All-in-One clock with my own PCB design used Tesla ZM1082 nixies.

I have a problem with the temperature of the MOSFET (used IFR740, 100uH choke Coilcraft DR0810-104L, UF4007 diode, 2u2 / 450V Panasonic cap).
Calibration runs fine when finished automatically sets FET PWRon around 220-230. This seems to be too much, the transistor has about 100-120 degrees. If I manually reduce the PWRon value to 150, HV will drop to 145-150V and the temperature will drop to about 90 degrees. That's a lot, too.

Can it be caused by a high Rdson transistor (550mOhm) and a low coil impedance (220mOhm)? I'm going to make the following changes:
1. I increase the inductance of the coil to 330uH with an impedance of 1 Ohm
2. change the MOSFET to a smaller Rdson type ... about 200mOhm (SIHP25N40D or STP24N60DM2)

Do you think these changes could help? Which change should I make first?

I note that the clock works normally, the LEDs work, the RTC module also. Atmega works, the HVD signal has a nice course on the scope. But because of the high temperature, I don't want to keep them running for more than a few minutes.

Tahks for help ... Vladca
Last edit: 4 months 6 days ago by vladca.

Please Log in or Create an account to join the conversation.

More
4 months 6 days ago #13000 by Ty_Eeberfest
A MOSFET with lower Rds should help lower the temperature. In theory at least, the lower the Rds the cooler it will run. Just make sure the other major properties (VThreshold, Vds, Vgs, etc.) are similar to the IRF740 / IRF640 specs.

I'm not sure about the inductor. A moderate increase of inductance (220uH maybe) should do no harm and may be helpful - not sure, haven't tried that. And I have no idea how the DC resistance (what you called impedance) will affect things.

Also take a look at the voltage divider formed by the 390K (R1) and 4.7K (R2) resistors. If these resistors are out of tolerance then HV regulation will be inaccurate. If possible use quality 1% resistors here.

Look into it later when the dust is clearing off the crater.

Please Log in or Create an account to join the conversation.

More
4 months 6 days ago #13001 by vladca
Replied by vladca on topic Issues with mosfet heat
Hello.
In the local store, they only had the IRFP360 (Rdson 0.2 Ohm) in a TO247 package and a 330uH/0.5 Ohm coil. I bought them for tests.

I replaced the coil first, but the result was not quite good.
After replacing the MOSFET, great results came. HV will run at 180V much faster than before (a few seconds). It is held stably and the PWRontime calibration ended with a result of 178. And most importantly the temperature dropped to 32 degrees (measured using a contact PT100 probe). I can also use a 9V power supply without any problems.:)
Finally, I returned the original 100uH/0.2Ohm coil to see what it would do. It works. MOSFET temperature slightly higher, HV less stable and longer PWRontime.

I'll look for suitable replacements for the MOSFET and do other tests with 150-220uH coils, which will probably be better.

I'll write the results here, maybe it will help someone in the future.

Vladca

Please Log in or Create an account to join the conversation.

More
4 months 6 days ago #13002 by Ty_Eeberfest
Sounds good! Thank you for following up.

32 degrees seems very good to me. No matter what you do the MOSFET will always run a bit above ambient and that is okay. It makes sense to me that the higher inductance (within reason) would be more stable, but I must stress that I am not a boost converter expert!

Look into it later when the dust is clearing off the crater.

Please Log in or Create an account to join the conversation.

Moderators: AccutronTy_EeberfestIan
Time to create page: 0.100 seconds
Go to top
JSN Boot template designed by JoomlaShine.com