Cooling small chips Onkyo 5009 HDMI board

Hi peeps, I have an Onkyo 5009 AV Reciever with the very common problem of overheatin g HDMI board. You can see it just under the grill for cooling in the case. I have an infrared heat sensitive camera and discovered there are several s mall ic's that get very hot very quickly- i.e. 70-80c. The ICs are very sma ll, 10mmx10mm up to 10mmx16mm and thereabouts but there is space of about a n inch above so room for a heat sink maybe. I could Fit heat sinks to those chips(if there are any small enough) or fit fans above or both. But does a nyone have any experience cooling small chips like these?

Cheers

Steve

Reply to
Mr Sandman
Loading thread data ...

En el artículo , Mr Sandman escribió:

Yes. You do need some airflow though, not much.

Google "RAM sinks" or "RAMsinks"

formatting link

Reply to
Mike Tomlinson

ing HDMI board. You can see it just under the grill for cooling in the case . I have an infrared heat sensitive camera and discovered there are several small ic's that get very hot very quickly- i.e. 70-80c. The ICs are very s mall, 10mmx10mm up to 10mmx16mm and thereabouts but there is space of about an inch above so room for a heat sink maybe. I could Fit heat sinks to tho se chips(if there are any small enough) or fit fans above or both. But does anyone have any experience cooling small chips like these?

Yes. Either would help. You don't need a lot of tep reduction fro 80C to ga in longevity. Sitting a small low speed fan on it is presumably easiest.

NT

Reply to
tabbypurr

There almost certainly is small enough heatsinks, even if you use those intended for a Raspberry Pi. Attaching them with a decent thermal joint may be more of an issue. I've stuck small heatsinks onto the bridge rectifiers in my scope with "BERGQUIST BP100-0.005-00-1/1 - Thermal Adhesive Tape, Bond-Ply 100, 0.8 W/m.K,

0.005" Farnell order code: 8783721.

Fans really only work if there is a decent hot surface area to transfer the heat, they are noisey and wear out. I'd find the largest heatsinks that'll fit and sticke 'em on and then see what the temperatures are.

Reply to
Dave Liquorice

any flat bit of ali would do as long as it increases surface area significantly. Not much performance is required

superglue

true if you need a lot of thermal improvement, but this job doesn't

anything from totally silent to noisy

NT

Reply to
tabbypurr

Or the heatsinks used on the driver chips on my 3D printer (A4988). I glued mine on with a thermal two part epoxy and they have been fine ever since.

formatting link
formatting link

I found that because they were so small, using thermal tape (or the tape we used) allowed the heatsinks to move or even get knocked off. ;-(

Because the 'thickness' of the epoxy was mush less than that of the tape, I sandwiched a brass (reasonable thermal conductivity) washer the same OD as the size of the chip onto the top of the chip to raise the heatsink slightly to ensure it didn't touch the surrounding components.

+1 Ideally etc.

However, because I believe the 3D printer driver IC's actually 'sink' their heat downwards to the PCB, I fitted a silent fan into the end of the (3D printed) enclosure, with a grille at the other [1] ensuring air was also forced under the driver modules (that sit above the RAMPS board).

Cheers, T i m

[1] Part of an efficient cooling procedure can be fully managing airflow through an enclosure. Having fans that just blow the same air round and round isn't going to help. When I built my home server I used a std PC mini tower case and blocked up all the vent holes except those under the front panel and (of course) the PSU output. I also added ducting inside the case ensuring the cold air drawn in though the front went over the hard drives before being forced over the mobo (RAM / CPU / Bridge) before being extracted through the PSU and out the back. Therefore, natural convection would also help the process and the result is the whole thing runs very cool. I've since replaced the std ATX PSU for an external 'brick' and so the only thing in the PSU case now is a single 120mm slow speed (=silent) fan. ;-)
Reply to
T i m

Well, the thing is, what are their working temperatures supposed to be, and are they exceeding this, if they are then if they are still working maybe there is a power supply fault elsewhere, which will do for them in the end by pusing up their dissipation. Not all chips can be used with a heat sink unless internally the thermal conductivity is high enough to allow the heat to be dissipated in one. I assume these are surface mount chips, so the area available is going to be very small.

Certainly you could try just standing some on the chips with some hs grease in between but keeping them there is a whole other issue if there has been no provision for sinks in the first place. As for fanning do you really want the sound of a fan in your home cinema? Brian

Reply to
Brian Gaff

most silicon runs at up to 175 junction, and that can be easily 100C case on small chips.

Reply to
The Natural Philosopher

'Most' is not true. Some does - even if it's rated at 85'C - but you need to be aware of the limitations in both performance and longevity.

Cheers

Reply to
Clive Arthur

yes, running them that hot gives short life expectancy.

NT

Reply to
tabbypurr

not really.

Reply to
The Natural Philosopher

Yes, really. The rule of thumb is a halving of life expectancy for every 10'C increase above some reference point given by the manufacturer. Often 10,000 hours at 150'C is used as a starting guestimate if you don't have a reference point. That corresponds to

1250 hours at 180'C which is realistic.

Basically, diffusion and migration seem to follow the Arrhenius equation.

Cheers

Reply to
Clive Arthur

Yes, but there is a difference between 'short' and 'shorter'

10,000 hours at 150C is still a pretty long time at 80C

If we take a year at 150C that's 128 years at 80C

Experience suggests that unless you are running right on the limit, in which case the chips tend to unsolder themselves, your product lifetmime will be limited by other factors.

What tends to happen more is that chips go out of spec when hot.

However if the design calls for them to run at 80C then they will be in spec for that anyway.

Reply to
The Natural Philosopher

The rule of thumb equation doesn't apply at lower temperatures, only when diffusion gets going.

Lead free solder melts around 220'C, nobody uses ordinary tin/lead (~183'C) at high temperatures. (Although in extreme situations, ~95% Pb HMP solder is used which melts at around 300'C.)

I make a living designing electronics for use in 180'C ambient. It's fun, because of the constraints - you can't just throw a big processor at a problem, small geometries don't do well.

Cheers

Reply to
Clive Arthur

I agree, the heat sink is the first thing to try

In some equipment I've fitted a 60mm fan to the underside of the top casing blowing out through the existing ventilation slots and reduced internal component temperatures by 20C. The fans have been set to only run at a very reduced speed so noise isn't a problem [Pick a fan with a low operating speed and then reduce the supply voltage]. I've fixed the fan to the casing with cheap/thin/flexible cable ties around the solid bit between the casing ventilation slots.

My Onkyo has a recommendation for at least 4 inches of free space above the top of the casing which may not be achievable with many under TV shelves/cabinets.

Reply to
alan_m

HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.