Help fixing a cracked/ detached "ceramic heat-pad".

[Dan R. Hansen]

Hi.

Anyone here who know how to fix such a mess:

I was drilling some holes into a heatsink, and made a last second change to where the last hole would go. As it turned out, it ended up being right next to a mosfet, but still under this hard type of thermal interface. I do not know what this thermal pad material is called, but it's hard hard and brittle, like ceramic or something. Not only did it crack, but it also dethatched from the aluminum surface. In my infinite wisdom, I made a half-assed attempt to "glue" it back with some Loctite 243 I had on hand. I know after some thought, that this is stupid on the base alone of Loctite losing adhesion when getting hot and it's poor thermal properties.

I just don't know if there is actually anything I can do to make it stick again, or if I have to chip it of and use a thermal pad etc?


Cheers from Denmark.

[BrokenYugo]

Well, they probably didn't grow it on there. I think what you're after is a "thermal adhesive", thermally conductive epoxy or whatever, minding that electrical insulation must be maintained. Make sure that crack gets filled.

Loctite 243 is an anaerobic threadlocker, not a general purpose glue, clean that out.

[Wallace Gasiewicz]

Loctite 383 Output Thermally Conductive Adhesive

[Dan R. Hansen]

Well, they probably didn't grow it on there. I think what you're after is a "thermal adhesive", thermally conductive epoxy or whatever, minding that electrical insulation must be maintained. Make sure that crack gets filled.

Loctite 243 is an anaerobic threadlocker, not a general purpose glue, clean that out.

I understand that, I just made a late night brain fog mistake of desperation... Do you know how to remove it without damaging component and the PCB?

Loctite 383 Output Thermally Conductive Adhesive

Interesting. I hope it can bind to the surface the brittle pad left behind? Because the Loctite 243 didn't stick at all, and all it did was contaminate the surfaces...

[BrokenYugo]

I would think desolder the transistor+chunk of probably alumina, then clean up, glue with appropriate glue and clamp, resolder after glue cures. Not sure what solvent to use for cleaning uncured loctite, maybe acetone used sparingly on a q tip.

[Dan R. Hansen]

I would think desolder the transistor+chunk of probably alumina, then clean up, glue with appropriate glue and clamp, resolder after glue cures. Not sure what solvent to use for cleaning uncured loctite, maybe acetone used sparingly on a q tip.

I might be possible, but it will be a very tight fit. I think the PCB is coated too.

I haven't seen that sort of mounting material / application before though a couple of possibilities come to mind even
though I think it'd be uncommon for both from what I've seen:

https://en.wikipedia.org/wiki/Aluminium_oxide
https://en.wikipedia.org/wiki/Beryllium_oxide

The unit in question, is a Huawei R4850G2 power-supply/ battery charger with programmable voltage and current limiting output via a CAN interface:

[Wallace Gasiewicz]

The insulator (whatever it's chemical composition) is commonly used in mounting transistors in radios. They are advertised as Alumina. This is probably correct as Beryllium is dangerous.
Usually they are attached with a screw and insulating washer like using a Mica or Silicone insulator. They are commonly available.
As you know they are brittle.
Usually when I see this sort of installation, the chassis or heat sink has already been drilled and a ceramic insulator has the hole thru it. The assemblers just used glue instead of a screw. This is unfortunately not the case here.
The heat stable Loctite is expensive. There are cheaper thermal glues on ebay. I have not used them. Perhaps someone else has experience with them. I have seen the grey stuff used and it is not hard but does hold. The transistor is usually also held into place with a screw. Perhaps a goodly amount of it around the existing transistor gluing it to the heat sink would work. Should transfer the heat.
Perhaps you could remove the transistor and sand the surface of the insulator and the heat sink to remove old glue then use new glue.

You might take the transistor out, break off the insulator and sand the surfaces, both the transistor and its heat sink to remove the old adhesive,
Then drill and tap a hole and use a mica insulator and insulating washer and thermal paste. You can trim the mica to fit around the remaining ceramic insulator on the other transistors and perhaps move the transistor a few mm to clear the remaining ceramic insulator on the other transistors. Or drill thru the chassis heat sink and use a screw and nut.
Unfortunately you cannot drill thru this ceramic stuff, you would need a diamond drill.

[Dan R. Hansen]

I haven't seen that sort of mounting material / application before though a couple of possibilities come to mind even
though I think it'd be uncommon for both from what I've seen:

https://en.wikipedia.org/wiki/Aluminium_oxide
https://en.wikipedia.org/wiki/Beryllium_oxide

The plot thickens.

I came to a realization of sorts, yesterday.

While I was asking for suggestions here on the forum, I had already ordered a new unit, waiting for et to arrive at my doorstep.
This however didn't happen, because there was an issue with the package at the airport shipment handling. I asked the Chinese seller, what the issue was this time, since it wasn't the first problem with this particular purchase.
In broken translated English, the seller said this:

"亲 带电带磁 不可以空运 严格的来说 部分物流不能承运
It is not allowed to be transported by air with magnetic wire. Strictly speaking, part of the logistics cannot be carried"

I was puzzled by this, since I had already received a unit like this before, though shipped with DHL and not AliExpress Premium. I simply couldn't figure out what part inside this unit, could possible be a problem- Could it really be something magnetic? Maybe?
Then it hit me. What if it wasn't a question of something magnetic, but rather something radioactive and/or hazardous?

Could the mysteries thermal pad, actually be consisting of Beryllium Oxide?... Oh boy!

[thm_w]

Can you link the item page?

Seller is just making up an excuse for why they don't want to ship to your country. There may be various legitimate reasons, shipping cost too high, current delay too long, etc.
Definitely not radioactive. Unless the product specifically is advertised as such.

[BrokenYugo]

You mostly only see beryllium oxide insulators, as I understand, in RF transistors, ancient magnetrons, etc. And with warnings at that in anything remotely modern. I wouldn't expect it in this case, but I am about as far from an expert as you can get.

[Dan R. Hansen]

Can you link the item page?

Seller is just making up an excuse for why they don't want to ship to your country. There may be various legitimate reasons, shipping cost too high, current delay too long, etc.
Definitely not radioactive. Unless the product specifically is advertised as such.

I do actually believe the seller, because I could track the package, and it stopped in the airport, where it got delayed and rejected. It had been through many checkpoints before that, so I think it is safe to say, that they did in fact ship the item.
How, there were some "fuckery" going on before that. It was sold as free Fedex shipment included. But it ended of beign shipped with AliExpress "Premium" shipping.

I was not happy about it, at all. But there is practically nothing I could do about it. In the end though, the seller actually refunded all of the money the received from the transaction. AliExpress however, did fuck me over by keeping the mandatory VAT, that they were required to charge at the checkout. Fucking scumbags...
Last time I bought a unit like this, the seller didn't include the proper paperwork and documentation, resulting in me having to pay VAT twice. AliExpress dispute "judges" didn't understand the problem, and kept referring me to the rule, saying that seller is not responsible for extra fees beign applied by other governments...

You mostly only see beryllium oxide insulators, as I understand, in RF transistors, ancient magnetrons, etc. And with warnings at that in anything remotely modern. I wouldn't expect it in this case, but I am about as far from an expert as you can get.

Well, it is cell tower equipment and the unit are made for outputting 56.1A, at 53.5 volts, so I assume the mosfets do need some insanely good heat transfer, which Beryllium pads are, if I understand correctly?
And it seems like Beryllium pads are a good choice since it never dries out, because it is a hardened ceramic, which I believe is a plus in always on equipment that must not fail often, being blasted with air from al kinds of environments and conditions.

https://great-ceramic.com/beryllium-oxide-beo-ceramics/

[RoGeorge]

It happened to me a couple of times to get fake tracking numbers, so following that number might be irrelevant.

Since there is a piece of chipped material, and if you are still curious about the ceramic material, you may try to identify the material, for example by its density.  I have no experience doing that, found it by searching for "identify beryllium oxide":  https://www.eevblog.com/forum/chat/beryllium-oxide-vs-alumina-is-there-a-way-to-identify/

[Dan R. Hansen]

It happened to me a couple of times to get fake tracking numbers, so following that number might be irrelevant.

Since there is a piece of chipped material, and if you are still curious about the ceramic material, you may try to identify the material, for example by its density.  I have no experience doing that, found it by searching for "identify beryllium oxide":  https://www.eevblog.com/forum/chat/beryllium-oxide-vs-alumina-is-there-a-way-to-identify/

I know they give you false tracking number, but this one doesn't fit the usual pattern. It does seem legit. Times and dates match, and a package is actually being sent to my destination. Seller also offered to send the package again with DHL, free of charge, after granting me a full refund.
It doesn't fit the usual shady pattern.

[Gyro]

...
Well, it is cell tower equipment and the unit are made for outputting 56.1A, at 53.5 volts, so I assume the mosfets do need some insanely good heat transfer, which Beryllium pads are, if I understand correctly?
And it seems like Beryllium pads are a good choice since it never dries out, because it is a hardened ceramic, which I believe is a plus in always on equipment that must not fail often, being blasted with air from al kinds of environments and conditions.

https://great-ceramic.com/beryllium-oxide-beo-ceramics/

It's not Beryllium. If it was, there would be big bold warning labels on the heatsink / enclosure. Beryllium was only ever used in small quantities in the packages of specific RF semiconductor. There is some debate as to whether it was ever used in industrial magnetrons or other high power tubes (certainly not consumer).

Alumina and Beryllium still require thermally conductive grease (or maybe adhesive in this case) to fill the microscopic voids caused by surface irregularities on the semiconductor packages and also the surface machining of the heatsink. They are not a 'dry' solution. They are no different to Mica in this respect - the grease can still dry out. They just have higher voltage withstand than Mica or silicone pads [EDIT: and better RF properties).

[richard.cs]

It's not Beryllium. If it was, there would be big bold warning labels on the heatsink / enclosure. Beryllium was only ever used in small quantities in the packages of specific RF semiconductor. There is some debate as to whether it was ever used in industrial magnetrons or other high power tubes (certainly not consumer).

I agree beryllia is very unlikely here as it seems to be a pretty modern device, but it didn't used to be that rare. It was certainly once possible to buy beryllia semiconductor mounting washers similar to (but thermally better than) the alumina ones available today (https://www.mascera-tec.com/product/to247-to220-al2o3-alumina-ceramic-wafer). I have an RF load with a BeO resistor in, specifically this resistor: https://www.richardsonrfpd.com/docs/rfpd/RFP-150-50TCGF.pdf. The thermal path is BeO but the top cover is alumina, it's obsolete and replaced by an all-alumina part, but only in the early 2000s (from memory).

[Dan R. Hansen]

It's not Beryllium. If it was, there would be big bold warning labels on the heatsink / enclosure. Beryllium was only ever used in small quantities in the packages of specific RF semiconductor. There is some debate as to whether it was ever used in industrial magnetrons or other high power tubes (certainly not consumer).

Alumina and Beryllium still require thermally conductive grease (or maybe adhesive in this case) to fill the microscopic voids caused by surface irregularities on the semiconductor packages and also the surface machining of the heatsink. They are not a 'dry' solution. They are no different to Mica in this respect - the grease can still dry out. They just have higher voltage withstand than Mica or silicone pads [EDIT: and better RF properties).

It is Huawei after all. China operates in mysterious ways.  It's differently not an "end user consumer product", so there is always that little chance of oddities.

[Dan R. Hansen]

I found this documentation of some Huawei equipment. Beryllium Oxide is mentioned at section 2.1

The age of this equipment seems to be from 2007 or so. I am not sure. So Beryllium Oxide is not out of the question...

[thm_w]

I do actually believe the seller, because I could track the package, and it stopped in the airport, where it got delayed and rejected. It had been through many checkpoints before that, so I think it is safe to say, that they did in fact ship the item.
How, there were some "fuckery" going on before that. It was sold as free Fedex shipment included. But it ended of beign shipped with AliExpress "Premium" shipping.

I was not happy about it, at all. But there is practically nothing I could do about it. In the end though, the seller actually refunded all of the money the received from the transaction. AliExpress however, did fuck me over by keeping the mandatory VAT, that they were required to charge at the checkout. Fucking scumbags...
Last time I bought a unit like this, the seller didn't include the proper paperwork and documentation, resulting in me having to pay VAT twice. AliExpress dispute "judges" didn't understand the problem, and kept referring me to the rule, saying that seller is not responsible for extra fees beign applied by other governments...

Asking again, can you link the item page? Sounds interesting.
It isn't Huawei gear is it, that is banned from a number of countries for import.


The VAT thing I've heard reported a lot, no experience with it myself.
https://www.reddit.com/r/Aliexpress/comments/oxoncq/why_is_aliexpress_collecting_vat_and_not_paying_it/
https://www.reddit.com/r/Aliexpress/comments/plh1ml/double_charge_vat/

[Dan R. Hansen]

Asking again, can you link the item page? Sounds interesting.
It isn't Huawei gear is it, that is banned from a number of countries for import.


The VAT thing I've heard reported a lot, no experience with it myself.
https://www.reddit.com/r/Aliexpress/comments/oxoncq/why_is_aliexpress_collecting_vat_and_not_paying_it/
https://www.reddit.com/r/Aliexpress/comments/plh1ml/double_charge_vat/

The link I used is down, but you can just search "R4850G2" on Taobao or Aliexpress. They are sold everywhere. But please know, if you all start searching for it, the prices will go up permanently.

https://youtu.be/_IKTpq7bXoI

I looked at those reddit links before. It seems a lot of people are frustrated with the amount of shit Aliexpress and some Chinese sellers come up with.

[thm_w]

The link I used is down, but you can just search "R4850G2" on Taobao or Aliexpress. They are sold everywhere. But please know, if you all start searching for it, the prices will go up permanently.

https://youtu.be/_IKTpq7bXoI

I looked at those reddit links before. It seems a lot of people are frustrated with the amount of shit Aliexpress and some Chinese sellers come up with.

R4850G2 is Huawei gear, but it doesn't look like Denmark has banned it.
https://www.beyondlogic.org/review-huawei-r4850g2-power-supply-53-5vdc-3kw/
https://www.youtube.com/watch?v=U9LE8FThbPU

""亲 带电带磁 不可以空运 严格的来说 部分物流不能承运"

This translation seems to mean, charged batteries and magnets are not allowed. This is neither, but, maybe it was incorrectly labeled for customs (eg as a battery charger, but charger part cut off, etc.).
You can always call up Fedex and ask for specifics.

What are you using it for?

[Wallace Gasiewicz]

Dilbert thinks it is made from:

"Elbonium"

[Dan R. Hansen]

R4850G2 is Huawei gear

I know, I mentioned it many times.

You can always call up Fedex and ask for specifics.

As I said, it was supposed to be shipped with Fedex, but ended up being shipped with "AliExpress Preminium Shipping".
So no, I can't contact Fedex .   

[thm_w]

As I said, it was supposed to be shipped with Fedex, but ended up being shipped with "AliExpress Preminium Shipping".
So no, I can't contact Fedex .   

Ah. Understand now.
If you plug the tracking number into https://trackitonline.org/ does it tell you the carrier at all?

Some people mentioned it was DHL but it could involve multiple handoffs.

[Dan R. Hansen]

The tracking number has been removed after the dispute ended. Right now, the purchase status is "The order is frozen due to security reasons. Please wait patiently.".
This has to be the most oddly eventful purchase I have made in recent years.