How hot should a mosfet get?

[Trumpeter]

The one on my newly built kit gets pretty warm. Normal? I'm not an ee so go easy on me :)

[Trumpeter]

The HV generator is cranking out 171 volts, so it's working I just worry about component strain. This is the PV Kit with the 4xIN12 and 2 IN17.

[John Rehwinkel]

> The one on my newly built kit gets pretty warm. Normal? I'm not an ee so go easy on me :)

The MOSFET is used as a switch. When it's off, it shouldn't have any dissipation, when it's on, its dissipation depends on its on resistance. If your on resistance is low enough, the problem is probably the "in between" part while it's in the process of switching on or off. These high voltage, low on resistance MOSFETs necessarily have large gates, which have a fair amount of capacitance. Charging and discharging that capacitance in a hurry is they key to high efficiency, which translates to low wasted heat. Many of the circuits out there don't do a good job of gate driving, keeping the MOSFET fully on and fully off most of the time, minimizing the in-between times. These days, I tend to recommend using a dedicated gate driver chip, which is optimized for the task. Many of them can source and sink over two amperes, resulting in very fast switching times.

In short, check the on resistance of your MOSFET. There may be other parts out there with lower on resistance. The most likely problem is your gate drive circuitry: you need to SLAM that gate on and off as fast as possible, or the MOSFET will dissipate a lot of energy.

One last thing is that you need to have sufficient voltage to turn the MOSFET on all the way. If it has a high gate threshold, you may need 12V to fully turn it on. There are parts out there with low gate thresholds that are easier to turn on and off. However, when you want high voltage, low gate threshold, and low on resistance, you're getting into exotic expensive parts. Sometimes it's easier and cheaper to use the aforementioned dedicated gate driver chip to provide the high voltage and high current to get good performance out of a cheaper, more common MOSFET.

- John

[Trumpeter]

Thank you for this detailed explaination, even though some is over my head.

Should I check the resistance between the two joined pins and one of the two pins on the other side? Could the over heating mosfet damage other components or will it simply fail? I want to be sure there are not other factors contributing to the problem like me installing the wrong resistor but I am fairly meticulous during assembly. I assume if something was really out of whack the HV supply would not run at all. The clock is functioning but the mosfet is getting warm, and I notice it is mostly when all the tubes are running at the same time. Since I will be giving some of these as gifts I want to make sure everything is right.

[John Rehwinkel]

> Thank you for this detailed explaination, even though some is over my head.

That, unfortunately, is in the nature of switching power supplies. There be dragons.

> Should I check the resistance between the two joined pins and one of the two pins on the other side?

No, just look up the data sheet for the part you're using. It will give the on resistance, gate threshold voltage, gate charge capacitance, and so forth.

A low-end part will offer maybe 1 ohm of on resistance, but such parts will get pretty warm in operation. Less than 200 milliohms is preferable.

> Could the over heating mosfet damage other components or will it simply fail?

If it fails shorted, it can damage several other components in the process.

> I want to be sure there are not other factors contributing to the problem like me installing the wrong resistor but I am fairly meticulous during assembly. I assume if something was really out of whack the HV supply would not run at all.

Sometimes spikes from the high current switching can couple into nearby signals and confuse things in various subtle ways. However, the "hot MOSFET" issue, if you have reasonably clean HV, probably
isn't an interference issue.

> The clock is functioning but the mosfet is getting warm, and I notice it is mostly when all the tubes are running at the same time.

That makes sense, the supply is working the hardest then.

> Since I will be giving some of these as gifts I want to make sure everything is right.

Always a good idea!

- John

[NeonJohn]

On 09/28/2016 12:03 PM, Trumpeter wrote:
> The one on my newly built kit gets pretty warm. Normal? I'm not an ee
> so go easy on me :)
>

It depends. (always a good answer. :-) My rule of thumb is to keep the
maximum junction temperature somewhere between two thirds and three
quarters of the data sheet's absolute max. If you know the operating
condition then you can push closer to the max. But for our induction
heaters, for example, which might be operated in 100 deg ambient and
bright sun, I set the protective circuitry to around two thirds.

You must calculate backwards from the chip, through the substrate/chip
junction, through the substrate (package) to the heat sink and from the
heat sink to air. If you use an insulator between the device and heat
sink then you have two more junctions to account for.

Calculating the heat flow can be tedious so I've attached a spreadsheet
that automates all that tedium.

John



--
John DeArmond
Tellico Plains, Occupied TN
http://www.tnduction.com <-- THE source for induction heaters
http://www.neon-john.com <-- email from here
http://www.johndearmond.com <-- Best damned Blog on the net
PGP key: wwwkeys.pgp.net: BCB68D77

[NeonJohn]

On 09/28/2016 12:40 PM, Trumpeter wrote:
> Thank you for this detailed explaination, even though some is over my
> head.
>
> Should I check the resistance between the two joined pins and one of
> the two pins on the other side? Could the over heating mosfet damage
> other components or will it simply fail?

The problem is that power FETS *ALWAYS* fail shorted. So you need to
plan your circuit and any protection for that eventuality. I'd put a
PolyFuse in the circuit supplying the circuit. That'll protect your
transformer and power source when the FET lets its smoke out.

[Trumpeter]

It's an IRFD220 looks like it's resistance is .8 ohms.

[David Forbes]

The rule of thumb is that if it burns your thumb, then it's too hot.

The use of a properly designed circuit goes a long way towards preserving your
MOSFET. Also, if it's too hot and doesn't have a heat sink, then put a heat sink
on it. Even a small one will help noticeably.

Please reply with a photo or web link showing the transistor in question, so we
can see what the options are.

On 9/28/2016 9:03 AM, Trumpeter wrote:
> The one on my newly built kit gets pretty warm. Normal? I'm not an ee so go easy on me :)
>

--
David Forbes, Tucson, AZ

[NeonJohn]

You might want to look at some newer parts. I'm constantly getting
announcements from IR and IXYS about their 60 volt parts that go as low
as 10 milliohms.

John

On 09/28/2016 01:30 PM, Trumpeter wrote:
> It's an IRFD220 looks like it's resistance is .8 ohms.
>

[Trumpeter]

Looks like it's running around 110-113 deg F.

[gregebert]

Are you using a heat sink, and does the case have vents/louvers to allow air circulation (even without a fan, convection can provide sufficient cooling) ?

Heat sinks are not a "magic solution", but even a small heatsink can substantially lower the case temperature vs no heatsink at all. Just be careful that nothing (such as a wire, even if it's insulated) touches the heatsink when the case is closed. I always assume my heatsinks are electrically connected to the drain/collector of the device, which is usually high-voltage. 

If a wire rests against a heatsink, the insulation will eventually deform/melt causing a short. Even vibration can cause the edge of a heatsink to 'saw' thru the insulation of a wire.

[gregebert]

My bad...just checked the IRFD220 datasheet and it's not easily heat-sinked. I blindly assumed it was a TO-220 or similar case.

The theta j-c for this device is 120 degrees C per watt, so if it's dissipating 1W and you're measuring 113F at the case, your junction is probably overheating. The double-whammy is that this device mounts close to the PC board, so that will expose the PCB's fiberglass to excessive heat.

I'd change to a TO-220 case device with a small heatsink.

[David Forbes]

I consider the 'burn my thumb' temperature to be 120F, which is called 50C in
the data sheets. So it's probably OK as is.

On 9/28/2016 10:43 AM, Trumpeter wrote:
> Looks like it's running around 110-113 deg F.
>

[Dan Hollis]

I thought FETs always fail open, and it was SCRs which always fail shorted.

-Dan

> --
> You received this message because you are subscribed to the Google Groups "neonixie-l" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
> To post to this group, send an email to neoni...@googlegroups.com.
> To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/fbb5076d-89a9-df0a-ab3f-9646445982c4%40neon-john.com.
> For more options, visit https://groups.google.com/d/optout.
>

[NeonJohn]

I can't speak for low power FETs. The smallest thing I deal with is a
100 watt fan speed controller. I have a box under my lab bench that I
call my "L'il box of larnin' " that contains all the FETs and IGBTs that
have failed during development over the past 10 years or so. A thousand
at least. There's not a single device in that box that failed open.

John

[Dan Hollis]

Maybe the use case as well. The guys making spot welders always recommend
using FETs instead of SCRs due to FETs always failing open vs the SCRs
always failing closed.

So massive high current pulses vs contant duty cycle.

> To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/8f61d21d-adf9-c134-f559-0bc0170238db%40neon-john.com.

[Charles MacDonald]

On 16-09-28 02:00 PM, gregebert wrote:
> My bad...just checked the IRFD220 datasheet and it's not easily
> heat-sinked. I blindly assumed it was a TO-220 or similar case.

looks like they expected the designer to use the side with the double
pins to be heat sinked by the PC board. Might be able to attach a sink
on the top of the chip... But yes a TO-220 is much easier to play with.

--
Charles MacDonald Stittsville Ontario
cm...@zeusprune.ca Just Beyond the Fringe
No Microsoft Products were used in sending this e-mail.

[Trumpeter]

Should I install a new mosfet and see if a different one runs cooler? Worth a try?

How about these for heat dissipation?

http://www.ebay.com/itm/Enzotech-MOS-C1-Copper-MOSFET-Cooler-10PK-/252414079757?hash=item3ac50d330d:m:mb_3I4NWrjGw-R3DDEL2YEA

[gregebert]

I'd choose a device designed for higher power dissipation and that can accommodate a heatsink, such as a TO-220. Be sure your replacement device has a similar or slightly lower Vgs (on), and a lower Rds. I don know what the DC-DC converter design looks like, so make sure Vds is greater than the output voltage, probably 200V min.

The device you have now will only have marginal improvement with a heatsink because it's package was not optimized to use a heatsink. There isn't enough info in the datasheet to calculate the thermals with a heatsink.

[Nicholas Stock]

Folks, you can find the schematic for the clock on Pete's website under the kit instructions for Frank 3 clock (http://www.pvelectronics.co.uk/index.php?main_page=page_2&zenid=20158eaf309af0e37312e9962a16f589). The mosfet does get warm on this but I've never had an issue with any of these clocks for several years. I did have one IRFD220 on a larger tube clock fry on me but that's because I was driving the HV too hard at 185V. You can replace it with an IRF640 or similar I believe if you want to heatsink it (I'd check with Pete on this one beforehand...he's very helpful!) but I don't think that's necessary...

Nick

On Thu, Sep 29, 2016 at 7:00 AM, gregebert <greg...@hotmail.com> wrote:

I'd choose a device designed for higher power dissipation and that can accommodate a heatsink, such as a TO-220. Be sure your replacement device has a similar or slightly lower Vgs (on), and a lower Rds. I don know what the DC-DC converter design looks like, so make sure Vds is greater than the output voltage, probably 200V min.

The device you have now will only have marginal improvement with a heatsink because it's package was not optimized to use a heatsink. There isn't enough info in the datasheet to calculate the thermals with a heatsink.

--

You received this message because you are subscribed to the Google Groups "neonixie-l" group.

To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+unsubscribe@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/87f146d0-07dc-4dea-88aa-620eed0527b1%40googlegroups.com.

[Roddy Scott]

I have had a couple of these IRFD220s heat up on me enough to cause concern and they were replaced by a different batch that solved the problem. 2 out of 7 clocks have had this issue but Pete pretested the replacements and since then no issues. He is always helpful with any problems

[Trumpeter]

Hello all,

First off, thank you all for providing input. Truly a great collection of knowledgable folks here.

I ended up removing a mosfet from another kit and installing it in this clock. It now runs at a normal temperature, around 90 degrees F or so. I guess it must have been a bad part. Perhaps I fried it when I was soldering...though I am careful to limit heat soak.

[petehand]

It's an unfortunate characteristic of MOSFETs that the Rds(on) increases with temperature. As they get hotter, they dissipate more power and get hotter still. I've had MOSFETs get so hot they actually unsoldered themselves, though remarkably, they still worked once they cooled down. But it's always a good practice to run them as cool as possible.

If you have a HV converter running hot, it's nearly always a consequence of the driver not turning it off quickly enough. That results in the FET absorbing a lot of power from the inductor, which goes off in heat instead of high voltage. You need a driver that can bring the gate down hard and fast with power to spare, and then things will run a lot more efficiently. The turn on time is far less important, as the inductor limits the current rise time anyway. The popular MC34063 is a particularly bad choice for a HV supply, because it has a powerful high side driver to turn the FET on quickly,but no active pull-down to turn it off.