Thermal-Paste Questions

[K4RT]

I will be replacing the two RF power FETs in my 100W solid state transceiver.  Can you solid state AM home-brewers or broadcast engineers recommend a robust type & brand of thermal paste/grease?  Anything to avoid?  Web searching turns up many options and opinions.  Any hints or suggestions for applying the paste?  Thanks.

[WBear2GCR]

The stuff sold for computer CPUs is very high performance. But I think some of it is conductive.
May or may not be an issue.

I use the old standby which is the white thermal compound and mica pads.
I've personally had zero luck with Kapton film, thermal conductive pads, or hard anodized pads.

You ought to be able to see what was under the devices when you pull them.

Application? Thin even layer, it gets squished when you tighten the devices down. Excess is best removed.

                            _-_-

[WD5JKO]

   Adding to what Bear said...

I service 3KW 13.56 Mhz solid state class E generators. I have tried several greases. Some that were once fine were not fine after RoHs compliance changes. The computer CPU stuff can be good, but can lead to breakdown if too much is used and it squishes out and puddles near the collector or drain lead.

The two surfaces need to be flat and polished. Many power semiconductors are NOT flat! I use a fresh stone to flatten the power FET's that we use, and the same on the heatsink. The act of drilling and tapping holes in the heatsink often results in a raised area around the hole; especially if the screw was over torqued.

If the surfaces are polished and flat, very little grease is needed. If it squishes out while torquing the screws, then you are using too much grease. Too much is as bad as not enough.

I use Dow Corning 340 grease. It seems to stay put and not migrate away from center like the more common Rohs stuff does.

http://www.amazon.com/Dow-Corning-Silicone-Compound-Lubricant/dp/B00CAVTGNE

I can put some in a baggie and send it to you.

GENUINE DOW CORNING 340 (DC340) HEAT SINK COMPOUND. A grease-like silicone material, heavily filled with heat conductive metal oxides. This combination promotes high thermal conductivity, low bleedand high-temperature stability. The compound resists changes in consistency at temperatures up to approximately 350ºF (177ºC). In doing so, it helps maintain a positive heat sink seal to improve heat transfer from the electrical/electronic device to the heat sink or chassis, thereby increasing overall efficiency of the device. White color. APPLICATIONS: Dow Corning® 340 Heat Sink Compound is applied to the base and mounting studs of transistors, diodes and silicon-controlled rectifiers. It can also serve as an effective thermal coupler for many heat sink devices where efficient cooling is required. It is also suitable for aircraft engines where heat sink properties are required. This product has also been used as a high-voltage corona suppressant, nonflammable coating in connections for flyback transformers and TV sets and similar applications. A proven and economical thermally conductive solution for low power applications. This grease can be used in a wide array of industrial applications including power components/power supply for Audio; Video (DVDs, TV Sets, etc.); broadcasting equipment; car-audio, frequency drivers (or frequency inverters) on automation equipment and some minor applications on temperature sensor for auto. Can be used as a thin interface or encapsulant. BEWARE OF LOWER PERFORMANCE PRODUCTS AT A LOWER PRICE! - 100% silicone grease does get in between all the voids like DOW 340 does but plain silicone grease does not have metal oxide additives which considerably improve thermal conductivity between surfaces. DOW 340 has more than 60% metal oxide content to promote high thermal conductivity. DOW 340 resists changes in consistency at temperatures up to approximately 350ºF (177ºC) and has a maximum operating temperature of 392ºF (200ºC).

Jim
Wd5JKO

[WA2SQQ]

I had a project using "Artic Silver" which is used on CPU coolers. I reached out to the REP and he suggested that after applying a "very thin layer" drag a new razor blade over it, like a squeegee to smooth and remove any excess. Been three years and so far no failures. "Clean" is the word!

[MikeKE0ZUinkcmo]

Like Randy, I also use the age old white aluminum oxide compound.

[WBear2GCR]

What Jim said...

Also Wakefield sells it, and Aavid - both heatsink mfrs.

Agree on the flatness, but for a radio with a factory supplied heatsink, likely not an issue.
The flatness of "plastic pak" transistors may or may not be an issue. some RF transistors are gold plated, so
I'd be likely to leave them alone, unless you put a precision flat on them and they are not flat for some reason. Then
I would suspect bogus asian parts though...

                       _-_-

[W3GMS]

What Jim said...

Also Wakefield sells it, and Aavid - both heatsink mfrs.

Agree on the flatness, but for a radio with a factory supplied heatsink, likely not an issue.
The flatness of "plastic pak" transistors may or may not be an issue. some RF transistors are gold plated, so
I'd be likely to leave them alone, unless you put a precision flat on them and they are not flat for some reason. Then
I would suspect bogus asian parts though...

                       _-_-

Like others, I always use Wakefield thermal grease with good results.

Joe

[WD5JKO]

   I used to use Thermalloy 250. Never had a problem with it. Then when that ran out I had to order 250G. Problems soon started and amplifiers were coming back. Removing the failed power devices, the grease had turned black, somewhat like carbon. I called Thermalloy, and asked them what the heck did they do? They said they made changes to be RoHs compliant. Perhaps they had lead particles in there? I don't know, but in a critical application where a lot of power is dumped in a small area, stay away from that stuff. They had another model meant to stay put, and it did....but the power transistor stayed HOT from the initial application.

  Temperature testing can be quite telling. On the devices we use, with the Thermalloy 250G, I measured 15 degree Celsius rise from the heatsink to the top of the devices when at full power. Between a 1-2 mill lap with the stone, and using Dow Corning 340, I reduced that rise to 2 degrees Celsius. The failures stopped. The devices have a nickel over copper coating, and the lapping just barely exposes some copper.

I attach a file from IXYS. See the last page concerning grease type, thickness, and how to tell if you are using too much.

Jim
Wd5JKO

[PA0NVD]

I produce 1300Watt 27.12 MHz RF generators for plastic welding and always use Indium foil of 0.05 mm. Considerably better heat conduction than grease and a good conductor, wich is extemely important for some modern power fets because the source is normally connected to the flange. The final amp uses a double D-Fet in balance.

[VE3ELQ]

I will be replacing the two RF power FETs in my 100W solid state transceiver.  Can you solid state AM home-brewers or broadcast engineers recommend a robust type & brand of thermal paste/grease?  Anything to avoid?  Web searching turns up many options and opinions.  Any hints or suggestions for applying the paste?  Thanks.

Depends on the application. 
In your case most RF FETs have a grounded source so electrical isolation is not required or even desirable.  A small amount of silver based CPU compound such as Arctic Silver works very well.

For electrical isolation such as in TO220 or TO247 devices I prefer mica pads and silicon grease, both excellent electrical insulators yet transfer heat fairly well.

For large area applications such a bolting a copper heat spreader to a heat sink I use a liberal amount of aluminum based anti seize compound from the auto parts store.  Its mostly very fine aluminum powder in grease and is a great heat conductor and its inexpensive. Ive used it on CPUs also with good results. It also comes in copper but I have found the copper stuff I tried to be gritty and does not give as good a contact as the aluminum stuff.

As others have said get it flat, burr free and clean clean clean before application.

73s  Nigel

[WBear2GCR]

Nigel,

What is the "binder" in the aluminum based anti-seize compounds?
I'm wondering.

Petroleum based might be a problem.

Wondering what the thickness of the final layer will be, and what the thermal conductivity is.
I would have expected that the semi industry would have tried all the possibilities, and used this
one if it were good enough?

I've found it concerning that they sell this stuff to put in steel to steel applications, since afaik
aluminum touching steel = electrolytic corrosion...

The never seize compounds from back in the day were lead and later nickel... so why did they go with
aluminum??

[K4RT]

Guys - Thanks for the advice, and Jim thanks for posting the link & IXYS guidance. I will review the IXYS information before ordering paste.

Jim what is the abrasive & grit # of the stone you use?  Do you grind by hand?  Use a lubricant?

Brad

[WD5JKO]

Guys - Thanks for the advice, and Jim thanks for posting the link & IXYS guidance. I will review the IXYS information before ordering paste.

Jim what is the abrasive & grit # of the stone you use?  Do you grind by hand?  Use a lubricant?

Brad

Brad, I use a high grit diamond stone that was my fathers from 50 years ago. It stays precision, and has no wear pattern that I can detect. Looking around here is a link that might have suitable stones:

http://www.eze-lap.com/product/diamondstone.htm

The grit size might vary depending on need, but I would finish with 1200 grit.

Nigel,

   There are several You Tube videos that solder those Freescale /NXP LDMOS transistors to the copper heat spreader in a trench. It is amazing that those devices can take that kind of heat during this hot plate method of soldering. Also looking at anti sieze compounds, Loctite has a silver based product:

https://www.grainger.com/product/LOCTITE-Anti-Seize-Compound-2VFF3?functionCode=P2IDP2PCP

That silver product might have possibilities..

Jim
Wd5JKO

[VE3ELQ]

Nigel,

   There are several You Tube videos that solder those Freescale /NXP LDMOS transistors to the copper heat spreader in a trench. It is amazing that those devices can take that kind of heat during this hot plate method of soldering. Also looking at anti sieze compounds, Loctite has a silver based product:

https://www.grainger.com/product/LOCTITE-Anti-Seize-Compound-2VFF3?functionCode=P2IDP2PCP

That silver product might have possibilities..

Jim
Wd5JKO

Jim the BLF188 amp I built a while ago was flow soldered to the copper.  I really had to force myself to do that to a $300. device but it survived and works FB.  I have a new one nearly finished using 2 BLF188s only this time they are bolted down with 4-40 cap screws and Arctic Silver thermal compound, not powered up yet so the jury is still out.
Thanks for the link that silver stuff looks great, have not seen it for sale up here but may order some.

Bear,  Dont have the answers to your post above. The carrier seems to be a heavy oil probably petroleum based. Been using it for years with no issues, but I'm just a simple HAM builder and never consider MIL specs and the like. It just works.

73 Yall  Nigel

[PA0NVD]

For the BLF188 you better use Indium foil, no grease whatsoever. I make amps of 1300 Watts and ALWAYS use Indium foil, which also helps for a perfect grounding of the source and gives better heat conduction.

[WD5JKO]

For the BLF188 you better use Indium foil, no grease whatsoever. I make amps of 1300 Watts and ALWAYS use Indium foil, which also helps for a perfect grounding of the source and gives better heat conduction.

   I attach a data sheet from the Indium Corporation. Looks like good stuff, but that maximum temperature of 140 Celsius worries me a bit. That heatsink better not be undersized, because if allowed to get that hot, the Indium might liquefy. Then again if that happened the transistor Die temperature might take out the device, depending on the Tj temperature.

Jim
Wd5JKO

[PA0NVD]

The max temp is indeed low, and several people use it to solder. So, when it melts, it mostly improves slightly, not degrades. But, when the heatsink and flange get to 140 degrees, you have quite a problem, than the dye temp may be too high!!
My amp reaches, when properly loaded, approx 60 degrees C at continuous 1300 Watt output, 91% efficiency at 27.12 MHz
Indium Corp has nice stuff, even a little better than pure Indium when properly tightened, but quite expensive. Suitable for the industry but too expensive for many hams.
Indium can be found at Aliexpress. When you do a search you can purchase Indium foil of 0.05 mm for very acceptable prices. For 50 bucks you can mount approx. 50 mosfets. It is a little tricky to handle because it is extremely ductile. You double a corner before you know. And it works when it is flat.
If you like to try, I can sent you a sheet of 2 x 2 inches Jim, sufficient for 4 p. BLF188.

[K4RT]

Jim, thanks for the information & links.

Thank you all for responding.

Brad