|
Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets. |
|
Thread Tools |
11th May 2019, 2:33 pm | #1 |
Nonode
Join Date: May 2007
Location: Preston, Lancashire, UK.
Posts: 2,511
|
Power MOSFETs
A question for those familiar with modern solid-state components.
Can anyone advise whether there are any devices readily available which are electrically identical to the IRFP240 (N-channel) and IRFP-9240 (P-channel) power MOSFETS, but which are packaged in a way which gives superior heat-transfer from the device to a heatsink. These two devices are in TO-247AC packages, where the thermal path is only via a small metal annulus surrounded by plastic, including plastic around the single mounting screw. Interestingly the device data-sheets refer to them being mounted 'on a flat, greased surface', with no mention of an electrical insulating pad. Many thanks, Andy |
11th May 2019, 3:44 pm | #2 |
Dekatron
Join Date: Apr 2012
Location: Worksop, Nottinghamshire, UK.
Posts: 5,554
|
Re: Power MOSFETs
They may well have been superceded by an insulated tab part. No insulating washers will be required.
I usually replace an IRF part with an ST part due to there being a better range to choose from from UK suppliers. At the end of the day they are just a complimentary pair. Data sheet time. |
11th May 2019, 3:48 pm | #3 |
Dekatron
Join Date: Jun 2009
Location: Staffordshire Moorlands, UK.
Posts: 5,274
|
Re: Power MOSFETs
Not sure what you mean by annulus (ring?). IR will have got the heatsink area pretty right. Very common fet pairs in audio amps. Definitely need an insulator unless you're design is ok with live heatsinks; which many commercial designs already are. Heatsink live, then insulate its mountings.
__________________
Kevin |
11th May 2019, 4:21 pm | #4 |
Nonode
Join Date: May 2007
Location: Preston, Lancashire, UK.
Posts: 2,511
|
Re: Power MOSFETs
You're right, an annulus is strictly a 'round ring' - I meant a 'rectangular ring' as you've shown! The problem seems to be getting heat out of the devices into the heatsink via an insulating pad.
Andy |
11th May 2019, 4:34 pm | #5 |
Nonode
Join Date: May 2007
Location: Preston, Lancashire, UK.
Posts: 2,511
|
Re: Power MOSFETs
The problem revolves around using one of these modules. A pal of mine has been experimenting, and blown his up twice! I've bought one, but not put power on it yet as I'm still awaiting delivery of a heatsink.
He's on the forum, so hopefully will join in. Our application is a bit unusual, rather than audio use we want to drive it with a 1600Hz sinewave to produce power to operate vintage radar equipment. He reports very hot transistors, with the heatsink remaining cool. After blowing the first set, he discarded the aluminium angle (seen in the picture) and mounted the eight transistors directly on his big heatsink with insulating pads, but no improvement. Andy |
11th May 2019, 4:45 pm | #6 |
Dekatron
Join Date: Jun 2009
Location: Staffordshire Moorlands, UK.
Posts: 5,274
|
Re: Power MOSFETs
Interesting problem. They're used ever such alot in high power audio, bass guitar amps, PA amps, etc. I dont like those silicone based pads, prefer mica with heatsink compound, I believe it has much lower thermal resistance than the squidgy ones, and you can get them alot tighter without punching through.
Any chance something's going 'RF'? 1600Hz doesn't seem much compared to pro audio which are often specc'ed upto 25khz but some parasitic rf might couple to the heatsink somehow through the washers, I've heard about it before. I dont think there's any intrinsic problem with the IR devices or we'd have heard about it by now. Cloud Electronics for example use them, Ampeg, Mackie, just a few that spring to mind. An EBB amp (smaller version of that) spends every day powered up inside one of my PC monitor speakers, I did have an initial problem with hum and oscillation until I got the input tweaked/filtered....I posted the problem on here a few years ago,
__________________
Kevin |
11th May 2019, 4:51 pm | #7 |
Dekatron
Join Date: May 2008
Location: Derby, UK.
Posts: 7,735
|
Re: Power MOSFETs
Is it possible the resin encapsulation is standing ever so slightly proud of the metal plate, causing it not to make proper contact with the surface against which it is mounted? As the trapped air gets hotter, its pressure will increase, so worsening the contact.
Might be worth a few minutes going around the edges with a sharp blade. (Or a few seconds with a bench grinder, if you are feeling brave!)
__________________
If I have seen further than others, it is because I was standing on a pile of failed experiments. |
11th May 2019, 4:53 pm | #8 |
Guest
Posts: n/a
|
Re: Power MOSFETs
The "heatsink glue" stuff is very good, self shims to a very small gap and transfers heat very well, this sort of stuff https://cpc.farnell.com/loctite/outp...315/dp/SA00606
|
11th May 2019, 5:10 pm | #9 |
Nonode
Join Date: May 2007
Location: Preston, Lancashire, UK.
Posts: 2,511
|
Re: Power MOSFETs
I've no experience of that 'heatsink glue' - does it guarantee to provide electrical insulation, what stops it all squidging out?
|
11th May 2019, 5:11 pm | #10 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 14,007
|
Re: Power MOSFETs
Julie_m has a good point there: I've seen this issue in the past with some plastic-pack 2N3055s, and solved it by gently 'lapping' the backs of each transistor using a piece of plate-glass and fine valve-grinding compound until the whole thing was smooth and the metal contact-area was a nice, uniform matt-grey.
|
11th May 2019, 6:56 pm | #11 | |
Guest
Posts: n/a
|
Re: Power MOSFETs
Quote:
|
|
11th May 2019, 6:58 pm | #12 |
Guest
Posts: n/a
|
Re: Power MOSFETs
Technique, apply some of the gloop to the transistor, brush on the activator to the heatsink, press home and hold for a few 10's of seconds, done.
|
11th May 2019, 7:36 pm | #13 |
Dekatron
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,996
|
Re: Power MOSFETs
The problem with thermal pads is that they require an absolutely flat surface, and surface finish at least as good as that of the transistor. Extrusions, like the angle, are not good enough, and finned heatsinks in particularly are banana shaped.
I flatten mine laboriously on a piece of plate glass with belt sander abrasive strip stuck to it. Then I use mica insulators and heat sink compound. Anyway you get two bad interfaces with angle - thermal pad to a wonky surface on the angle, and then angle to a banana shaped heatsink. Put a straight edge (AKA decent steel rule - Rabone, Moore & Wright or similar) on a finned heatsink and look at it up to the light. So either do the abrasive trick and spend an hour of fun and games, or sweet talk someone on the list with a milling machine, or a local friendly machine shop to machine the back of the heatsink flat with a large diameter tool and slow feed. Craig |
11th May 2019, 7:37 pm | #14 | |
Nonode
Join Date: Mar 2010
Location: Sunderland, Tyne and Wear, UK.
Posts: 2,476
|
Re: Power MOSFETs
Quote:
Never seen this stuff either, sounds good to me.
__________________
I don't suffer from Insanity. I enjoy every minute of it. |
|
11th May 2019, 7:45 pm | #15 |
Dekatron
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,996
|
Re: Power MOSFETs
There are a number of helpful apps notes about this general topic which cover surface finish etc. Attached
Craig |
11th May 2019, 7:47 pm | #16 |
Dekatron
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,996
|
Re: Power MOSFETs
Another one
|
11th May 2019, 8:19 pm | #17 |
Nonode
Join Date: May 2007
Location: Preston, Lancashire, UK.
Posts: 2,511
|
Re: Power MOSFETs
Some interesting articles, a bit of reading to do!
Many thanks. Andy |
11th May 2019, 10:16 pm | #18 |
Nonode
Join Date: Dec 2012
Location: Stockport, Greater Manchester, UK.
Posts: 2,117
|
Re: Power MOSFETs
2SK413 or 414 and complement 2SJ118 or 119 are available in the "Plastic TO3" (TO3P) case with more metal than plastic on the heatsink side: 100W at 25°C, intended for audio applications
__________________
- Julian It's good here |
12th May 2019, 6:31 am | #19 |
Heptode
Join Date: Oct 2009
Location: Melbourne Australia
Posts: 901
|
Re: Power MOSFETs
Andy, have you and your friend been able to confirm how much power is being dissipated by each device, and what variation you are getting in power dissipation between each device, and whether that aligns with case temperature (from a small thermocouple or IR) and the heatsink general temp, and the anticipated Rc-hs, and how that correlates back to Tj, and whether Tj is sufficiently derated below the rated max Tj=150C ?
That all also needs to be compared against expected DC power supply consumption, just in case you have another unexpected power loss going on (like hf oscillation that doesn't show up in an rms or dc meter reading at the device). It looks like you are operating the FETs in the linear region. Is the source ballast resistance sufficient to force acceptable current sharing as device parameters change with temperature? Is your bias able to adjust with changing device temp? Is your gate stopper not too high in resistance so as to lose control of gate voltage due to transient or fast changing drain voltage (causing Cdg to influence Vgs)? |
12th May 2019, 7:26 am | #20 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,902
|
Re: Power MOSFETs
Taking a step back, perhaps a 1600Hz AC power supply might be a case where the efficiency of a class-D amplifier could save a lot of heat? The problem would then switch to finding one and filtering its output.
I wondered about a car/truck alternator minus its rectifier. The load would have to be split up, if possible, and distributed amongst its three phases. But a rough estimate suggested it would have to be driven at over 12,000 rpm depending on the number of rotor poles. This is likely too noisy and dangerous for the intended application. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |