I'm going crazy here. I've got a bunch of these high thermal conductivity hard-anodized aluminum TO-3 transistor mounting insulators salvaged from old gear, and planning to use some. I figured I'd check the thermal conductivity specs on line, and be done. I quickly found they are likely Thermalloy brand type 4703A, 4720A, or 4726A (same deal but slightly different outline sizes), with drawings and such, but no thermal specs. I can't seem to find any specs for these anywhere, just lots of pictures and references and ones for sale, but no useful info. Thermalloy apparently disappeared into a conglomeration some time ago. There's stuff floating around, but somehow it escapes me - maybe a mental block or 24 hour insanity. I even found the alumina types and specs, but no anodized. Somebody must still use and make these things. Or maybe all old/obsolete package stuff has disappeared for some reason.You'd think at least the old data sheets would be out there. The only outfits I recall from the old days are Thermalloy and Wakefield, and can't seem to spot anything useful on the later either. I just downloaded a recent Wakefield-Vette (now called) catalog, and they seem to be almost entirely about heatsinks and greases, and no hardware except some Kapton insulators.

Anyway, does anyone have this sort of info, maybe in old catalogs, or know where to look? All I need is the thermal resistance in simplest form deg C/W in TO-3 configuration.

Ed

The only info with real specs I found on this topic is an old Motorola/ON App note, AN-1040D, it does discuss these anodized transistor insulators towards the end of the article. you can find it from ON semi here:

https://www.onsemi.com/pub/Collateral/AN1040-D.PDF

I would not be so keen to use these older anodized insulators in actual life, I see zero advantage compared to more
modern techniques, and I doubt the voltage breakdown is as good as even poor film, pads or mica.
all the best,
walter? (walter2 -at- sphere.bc.ca
sphere research corp.

Thanks Walter,

I did run across that same document, which briefly compares hard anodized with some others, but without specific part numbers. I was hoping to find actual specs for these parts, but it looks like info is indeed sparse. Your mentioning it made me look at the app note again, and more closely. By putting the "datasheet" and the app note info together, I think it will be close enough. The app note describes the hard anodized as 20 mils thick, and so does the datasheet, so that was probably the standard thickness for those things.

According to the app note charts on pages 5 and 6, the hard anodized is superior (but not by all that much) to mica, and quite a bit better than sil-pads of various types, in the TO-3 "greased" configuration. Only BeO and bare (non-insulated) are significantly better. I'll go with about 0.3 deg C/W nominal, for estimating.

I'll need to dissipate about 220 W worst case among three transistors. This is for my inductor test (in conjunction with the HP4276A) fixture project. I was originally going to use some modern big (TO-264 I think) MOSFETs that I have on hand, that are DC-SOA rated, but the shunt drain capacitance is huge, especially running with such low overhead - not good for this sort of thing. I found I have a bunch of nice old-school 2N6283 Darlingtons (TO-3), and decided to go with bipolar. The C-E operating limits and such puts it on par with the MOSFETs overhead-wise, but with about ten times less - and more predictable - capacitance. This is the nonlinear Miller C that I'll be trying to compensate for with some trick circuitry. There's also quite a lot of fixed C due to the insulators. I was quite surprised at how much it can be - I never considered or measured it until this application came up. With mica, I found about 200-300 pF from the TO-3 case to heatsink. Interestingly, the hard anodized landed in the same range.

I'm not too worried about the breakdown voltage - the maximum excursion of the collectors is limited to the -12 V supply, and possibly up to +40 V clamping level for any inductive kickback. Even if an insulator should fail, the collectors are the output of the current source, and would just short to earth ground. If the current setting is in the high (0-20 A) range, the fault tripper will kick it out of the range, to a safe level, to avoid melting any power cords or cables.

The thermal resistance is mostly important for quietness, actually. I have a nice big flow-through setup worked out, and would like to set the fan to a low, constant speed. It's capable of a helluva lot of cooling if I wear earmuffs. The better the transistors connect with the heatsink, the slower I can run it.? I can make it shift up if necessary, but more complicated.

Ed

Hi, anodized TO-3 oxide films on the insulators were easily punctured, resulting in shorts.

We used kapton and Bergquist TO-3 insulators.

For 220 W suggest, you bolt direct to your heatsink and float the entire HS.

I have designed tested and manufactured power magnetics since 1980s, never had to make such a fixture.

So very Interested in exactly what you are testing and the inductor design.

Bon Soirée

Jon

开云体育

Mike, ACK, I have seen the HS to ch?ssis plastic isolation standoff deteriorated.

The extruded HS vendors had slide on plastic clips

We used large dia teflon or porcelaine on occasion.

Curious about the OP need for 220 W dissipating to test an inductor.

Jon

There were some questions about what this is for. A quick reference is here:

/g/HP-Agilent-Keysight-equipment/message/115717?p=,,,20,0,0,0::Created,,HP4276A,20,2,0,82870949

If there's enough interest, I can open up a new thread about this specific project, with a lot more info on the plans and progress.

Ed

Re the hard anodized thermal interfaces.

We use dozens if not 100's of the things, for stud mounted 50 and 75W
Zenner diodes, with anything up to 400V across them, and they are run
right up to the envelope of their ratings, at up to 90'C!?? (The "heat
sink" is part of an oil cooled power amp, where the "coolant" can run at
that sort of temperature for hours on end..

The *Only* time there has been issues, is if when assembling it all,
there is dirt on something, that damages the anodising.?? Else, no
problems whatsoever.

Personally, when I assemble them, I use a 500V "Mega" to check the
resulting insulation.?? Anything that moves the needle, I dismantle,
clean and re-assemble with fresh compound.?? If it's still an issue, I
replace the interface washer.?? Often, when later examining them, there
is more often than not, a "light" patch somewhere, due to past
deformation and resulting wear of the anodised coating..

For the money, they do appear to have the best thermal vs cost
performance of all options (other than direct mounting.)

But, as in all things.? Care is needed in the installation and assembly
of such things.?? The only time I've seen insulation failures, has been
due to dirt, or undue movement during assembly (rotation of the stud
mount device, damaging the coating.)?? That would be "difficult" to do
with TO3 devices!

Yes, we use a smear of heatsink compound, and I mean a smear, no more,
on both sides of the anodised washer.

We have more trouble with the dry "SilPads" used for some large single
hole power FET's used in HV shunt regulators, they seem to fail at
random all the time.?? And yes, we religiously torque all such fittings,
as per manufacturers spec's.

All the best.

Dave G8KBV


--
Created on and sent from a Unix like PC running and using free and open source software:

OK. FWIW did I design drivers for old stepper motors about 10 years ago. They are unipolar and the mosfets need to absorb an dissipate the kickback. We used Keratherm RED and silicone grease and used high-strength M3 insex bolts that were tightened using a torque wrench to manufacturer's specs. The unit is still running...