Thermally conductive silicone or adhesive that can fill 2mm gap (Solved)

[Agentfoot]

I have some Sony car amplifiers I am working on. Looks like they used a product to thermally bond the bias transistors to the heatsink. The product they used was gray and a rtv silicone like consistency. This fills a 2mm gap between the transistor and heatsink. Pictures attached for reference. Looking for recommendations on what to use here. All I seem to find is expensive products that only a few specially stores have in stock. 

[Zoli]

Thermal pad is what you are looking for; Digikey has from 0.037 to 20mm thickness n stock:
https://www.digikey.ca/en/products/filter/pads-sheets-bridges/218

[Agentfoot]

No that is not what I need. It is not for the outputs. It is for Q111 the small bias transistor. Pictured in IMG_4344.jpeg(or Q211 in the other pictures). Needs a product that can flow into the gap and cure.

[5U4GB]

Thermal pad is what you are looking for; Digikey has from 0.037 to 20mm thickness n stock:
https://www.digikey.ca/en/products/filter/pads-sheets-bridges/218

Are you going to get much thermal conductance over a 2mm gap? Why not use a 2mm copper shim coated with thermal grease and slide it into place?

[Agentfoot]

Thermal pad is what you are looking for; Digikey has from 0.037 to 20mm thickness n stock:
https://www.digikey.ca/en/products/filter/pads-sheets-bridges/218

Are you going to get much thermal conductance over a 2mm gap? Why not use a 2mm copper shim coated with thermal grease and slide it into place?

  I was just trying to do it the same way it came from the factory. I figured they did it that way for a reason. My other Sony amp has the bias transistors right next to the heat sink with no gap. Most other amps use a bigger package that can be mounted to the heatsink. This is the first time I have seen it done this way. I suppose the gray stuff could just be for screw retention. But I was under the impression that bias transistors needed to be thermally connected to the heatsink of the outputs.

[jc101]

I'd just use a thermal pad as mentioned earlier. Check the gap, and use a slightly thicker pad. It will squish down when the screw is done up. The pads have a recommended compression percentage, so it's just finding something that works.

[BillyO]

https://mgchemicals.com/products/adhesives/silicone-adhesive/gray-silicone/

Or

https://www.momentive.com/en-us/product-categories/formulated-products/thermal-management/silcooltia141gfthermallyconductivesiliconegapfiller

[Agentfoot]

I'd just use a thermal pad as mentioned earlier. Check the gap, and use a slightly thicker pad. It will squish down when the screw is done up. The pads have a recommended compression percentage, so it's just finding something that works.

This is the part everyone is missing. The bias transistor is the little FTR/FTL package in between the IRF9540 and the IRF540. It has no hold down screw or clamp. The clamp for the outputs(IRFs) does not touch the bias transistor. Just had some gray silicone like substance on it and the heatsink. Circled in red in these pictures attached.

[BillyO]

I'm not missing anything.  I gave you links to two products.  One is specifically a thermally conductive gap filler.  The other is probably what you took out, not quite as good, but good enough.

[Agentfoot]

https://mgchemicals.com/products/adhesives/silicone-adhesive/gray-silicone/

Or

https://www.momentive.com/en-us/product-categories/formulated-products/thermal-management/silcooltia141gfthermallyconductivesiliconegapfiller

First one does not mention being thermally conductive.

Second one looks to be like what I need. But it is not easily available and looks to be too expensive.

I found dowsill se 4420 and 4422. But they are also hard to get and very expensive.

[Agentfoot]

I'm not missing anything.  I gave you links to two products.  One is specifically a thermally conductive gap filler.  The other is probably what you took out, not quite as good, but good enough.

 

 You seem to be the only one that got it right. That post was to the others.

[BillyO]

The first one is not terribly thermally conductive (0.21 W/m*K), but may be what you had in there.  If you still have it around, use a soldering iron and your finger to check how quick it heats up across 2mm or so.

The other is far more conductive (4.1 W/m*K) but is likely intended for large scale industrial processes and only sold in large quantities to be used with specialized equipment.

[jc101]

It looked like a screw hole above the part, but the image with the gunk cleaned off looks like it isn't tall enough to reach the screw hole.

How hot does it become in operation? Without anything to apply pressure, such as a screw, it's difficult to see how adequate thermal transfer could take place. The grey stuff may just be to hold it in place, given a car can be a high vibration environment.

[Agentfoot]

It looked like a screw hole above the part, but the image with the gunk cleaned off looks like it isn't tall enough to reach the screw hole.

How hot does it become in operation? Without anything to apply pressure, such as a screw, it's difficult to see how adequate thermal transfer could take place. The grey stuff may just be to hold it in place, given a car can be a high vibration environment.

So from what I understand is the bias transistors do not get hot. But they need to thermally connected to sense the heat of the outputs to prevent thermal run away. This is what I gathered after wondering if the gray silicone like stuff was there for a reason and doing a little research.

[Gyro]

It looks to me as if the temperature sensing transistor (yes, your understanding is correct, not just for thermal runaway but also optimal stable bias operating conditions) is meant to be in thermal contact with those nice clamping bracket holding the transistors down. The footprint on the PCB is wrongly placed in relation to the heatsink surface outherwise. By using the clamp, which also collects heat from the front surface of the output transistors, it will achieve faster thermal response.

Check visually with dry assembly but I think you are actually looking for is thermal compound that which bridges the tiny gap between the sensing transistors and clamps. Decent heatsink compound, applied more heavily than normal, may well do the job.

[Agentfoot]

It looks to me as if the temperature sensing transistor (yes, your understanding is correct, not just for thermal runaway but also optimal stable bias operating conditions) is meant to be in thermal contact with those nice clamping bracket holding the transistors down. The footprint on the PCB is wrongly placed in relation to the heatsink surface outherwise. By using the clamp, which also collects heat from the front surface of the output transistors, it will achieve faster thermal response.

Check visually with dry assembly but I think you are actually looking for is thermal compound that which bridges the tiny gap between the sensing transistors and clamps. Decent heatsink compound, applied more heavily than normal, may well do the job.

I cannot get in to measure but I would say the gap between the back of the clamp and bias transistor is around 1.5-2mm. So still too large for a normal thermal compound. I suppose a thick enough thermal pad could be placed in this location and would work.

Indeed the placement of the bias transistor is not optimal. I have another older Sony XM-2751 that uses the same bias transistors. But in that amp they are butted up against the heatsink.

[Gyro]

Ah ok, it looked as if it might be closer fit. Checking visually, you should be able to see whether the bias transistor is closer to the clamp or the heatsink. That's a clue to where it gets most of its heat from. As the clamp is removable, it ought to be possible to glue (maybe maintain location with a screw hole) an aluminium of copper shim to it to reduce the gap to virtually nothing so that you can use heatsink compound. The problem with a thermal pad is, as jc101 points out, that they need a certain amount of compression. With it retained only by PCB solder pads, the bias transistor is likely to bend away from the clamp (and stress the PCB pads).

Whatever you use, you need to tighten the clamps slowly to allow the thermal media to spread without stress.

[Agentfoot]

I did not think about stressing the pads or solder joints. But you are correct a thermal pad would put added stress to them. Which makes sense why the factory used what they did. Their product flowed until dry which would not stress the component and even helped secure the bias transistor in place. Now if only I can find an inexpensive product that does the same. The MG RTV167 BillyO recommended seems the best fit so far.

[Agentfoot]

The first one is not terribly thermally conductive (0.21 W/m*K), but may be what you had in there.  If you still have it around, use a soldering iron and your finger to check how quick it heats up across 2mm or so.

The other is far more conductive (4.1 W/m*K) but is likely intended for large scale industrial processes and only sold in large quantities to be used with specialized equipment.

You are probably right that they used a product like RTV167. These amps are far from top quality and from the early 2000s when Sony was getting cheap. Guess they figured its was good enough. Unfortunately I do not have any of the old product left to test the heat transfer.

[MrAl]

I have some Sony car amplifiers I am working on. Looks like they used a product to thermally bond the bias transistors to the heatsink. The product they used was gray and a rtv silicone like consistency. This fills a 2mm gap between the transistor and heatsink. Pictures attached for reference. Looking for recommendations on what to use here. All I seem to find is expensive products that only a few specially stores have in stock.

Hi,

Congratulations on the excellent photography.  We don't usually see pics that clear and explanatory.  I could not have taken better pics myself and I've done a lot of electronics photography in the past.

You see to have found a solution so I'll tell you what my first thought was before I looked at the pictures.
My first thought was to use one of those pads that are used to connect a M2 SSD to the case of an M2 style SSD external enclosure.  They must be at least 2mm thick.  Supposedly, then link the chips on the SSD to the case for better cooling.  I've never tested one though so I cannot speak for their effectiveness.

The other thing that grabbed my attention (a little sarcasm here)...

Gee, I can't think of a better way to keep my transistors cool than by covering the WHOLE heatsink with heatsink grease (ha ha).  Same goes for thermal epoxy.
Maybe the designers knew that the assemblers were very poor at getting the right transistor mounting placement in the correct spots (again, ha ha).
It looks really nasty, and dust can collect moisture which in turn can conduct in places we don't want any conduction.

[Agentfoot]

I have some Sony car amplifiers I am working on. Looks like they used a product to thermally bond the bias transistors to the heatsink. The product they used was gray and a rtv silicone like consistency. This fills a 2mm gap between the transistor and heatsink. Pictures attached for reference. Looking for recommendations on what to use here. All I seem to find is expensive products that only a few specially stores have in stock.

Hi,

Congratulations on the excellent photography.  We don't usually see pics that clear and explanatory.  I could not have taken better pics myself and I've done a lot of electronics photography in the past.

You see to have found a solution so I'll tell you what my first thought was before I looked at the pictures.
My first thought was to use one of those pads that are used to connect a M2 SSD to the case of an M2 style SSD external enclosure.  They must be at least 2mm thick.  Supposedly, then link the chips on the SSD to the case for better cooling.  I've never tested one though so I cannot speak for their effectiveness.

The other thing that grabbed my attention (a little sarcasm here)...

Gee, I can't think of a better way to keep my transistors cool than by covering the WHOLE heatsink with heatsink grease (ha ha).  Same goes for thermal epoxy.
Maybe the designers knew that the assemblers were very poor at getting the right transistor mounting placement in the correct spots (again, ha ha).
It looks really nasty, and dust can collect moisture which in turn can conduct in places we don't want any conduction.

It was determined a thermal pad will just add stress to the bias transistor and solder joints. It is a small transistor. Just about every amp I repair has had tons of thermal paste everywhere. You would think they could save some by just placing where needed but probably quicker to slather it on all over. My standard treatment is to clean inside and out, replace bad components, and new thermal paste.

[Agentfoot]

Any reason not to use the GE advanced silicone 2? It is neutral curing, same working temperature, a bit softer at 15 shore A hardness, and should have the same standard RTV silicone thermal conductance of .2-.3 W/m*K. Its also almost a quarter of the price of the RTV167.

[Doctorandus_P]

Simplest is to desolder, then screw it to the heatsink with a normal thermal pad and solder the pins again.

You can also add a small piece or aluminum to fill the gap if you wish. There are also ceramic thermal pads, but those are not always easy to source.

[Agentfoot]

Simplest is to desolder, then screw it to the heatsink with a normal thermal pad and solder the pins again.

You can also add a small piece or aluminum to fill the gap if you wish. There are also ceramic thermal pads, but those are not always easy to source.

That is a solution, but far from the simplest. That would require new bias transistors with longer leads. Then you would need a different package transistor that has a screw hole to screw down or fabricate a clamp for the original ones.

The simplest way is to keep it as is. Just need to know for sure if the gray stuff is needed and is it standard gray electronics silicone or some type of thermal product.

[MrAl]

I have some Sony car amplifiers I am working on. Looks like they used a product to thermally bond the bias transistors to the heatsink. The product they used was gray and a rtv silicone like consistency. This fills a 2mm gap between the transistor and heatsink. Pictures attached for reference. Looking for recommendations on what to use here. All I seem to find is expensive products that only a few specially stores have in stock.

Hi,

Congratulations on the excellent photography.  We don't usually see pics that clear and explanatory.  I could not have taken better pics myself and I've done a lot of electronics photography in the past.

You see to have found a solution so I'll tell you what my first thought was before I looked at the pictures.
My first thought was to use one of those pads that are used to connect a M2 SSD to the case of an M2 style SSD external enclosure.  They must be at least 2mm thick.  Supposedly, then link the chips on the SSD to the case for better cooling.  I've never tested one though so I cannot speak for their effectiveness.

The other thing that grabbed my attention (a little sarcasm here)...

Gee, I can't think of a better way to keep my transistors cool than by covering the WHOLE heatsink with heatsink grease (ha ha).  Same goes for thermal epoxy.
Maybe the designers knew that the assemblers were very poor at getting the right transistor mounting placement in the correct spots (again, ha ha).
It looks really nasty, and dust can collect moisture which in turn can conduct in places we don't want any conduction.

It was determined a thermal pad will just add stress to the bias transistor and solder joints. It is a small transistor. Just about every amp I repair has had tons of thermal paste everywhere. You would think they could save some by just placing where needed but probably quicker to slather it on all over. My standard treatment is to clean inside and out, replace bad components, and new thermal paste.

Hi,

Oh, after I saw the pics I ruled out the thermal pad entirely.  Whatever is used has to be able to take on the right shape as well as conduct heat.  This is a little rare, but not without precedent as we see this in power strips too where the body of the transient suppressors have to make good thermal contact with the thermal fuse.  I don't think we could have a thermal pad solution for that unless somebody came up with some sort of standardization.  Same with a lot of the AC fans with the little thermal fuses.