Perhaps you've all missed my bolded average? 
I didn't pick 32degC at random, there is a 50% probability of a 32degC average daily temp occurring in any year. We've had at least 2 of them this year already, even though most of the summer has been reasonably mild. To get an average like that you're talking minimum ~25C min and ~39C max.
Random and useless information for most in the thread ...
I like to consider ambient conditions when selecting a heatsink.
On the topic of alu bars. Is anyone else concerned about how they are fixed in place? Its a low risk once its in place with thermal tape etc ... I'd really hate to have an alu bar work loose and end up floating around inside an enclosure though.
EDIT: Congrats to jazzm on getting to the bottom of the problems! I can't wait to hear my own, the rest of us will sit over on our respective sides of our respective ponds and gaze upon the Canadian contingent with jealousy in the mean time...
I didn't pick 32degC at random, there is a 50% probability of a 32degC average daily temp occurring in any year. We've had at least 2 of them this year already, even though most of the summer has been reasonably mild. To get an average like that you're talking minimum ~25C min and ~39C max.
Random and useless information for most in the thread ...
I like to consider ambient conditions when selecting a heatsink.
On the topic of alu bars. Is anyone else concerned about how they are fixed in place? Its a low risk once its in place with thermal tape etc ... I'd really hate to have an alu bar work loose and end up floating around inside an enclosure though.
EDIT: Congrats to jazzm on getting to the bottom of the problems! I can't wait to hear my own, the rest of us will sit over on our respective sides of our respective ponds and gaze upon the Canadian contingent with jealousy in the mean time...
Last edited:
hochopeper ah yes ok, average meaning you could wake in the night in a pool of sweat? yes i know what you mean lol.
yeah considering the alu bars are sitting right next to the tabs on the buffers which are all live, although all are at the same potential and it sounds like they would be isolated from the rest of the sink, its still an area of concern yes. If you consider what happens if you compress keratherm or other pads too much..... they can tear or at the least bleed, so using pressure alone to keep them there really isnt an option in my book.
You could use some sort of thermal adhesive like sekisui, but then when you start adding up all the little costs of extra fasteners, extra thermal interface, time etc it really starts to add up the cost
yes jazzm, you really should have the schematic out and have given it a good look while building. ALL the caps on the board are decoupling the rails, which means they connect to ground on one side, there are no caps in tyhe signal path of this amp.
the sound between this and a tube amp couldnt really be more polarized
yeah considering the alu bars are sitting right next to the tabs on the buffers which are all live, although all are at the same potential and it sounds like they would be isolated from the rest of the sink, its still an area of concern yes. If you consider what happens if you compress keratherm or other pads too much..... they can tear or at the least bleed, so using pressure alone to keep them there really isnt an option in my book.
You could use some sort of thermal adhesive like sekisui, but then when you start adding up all the little costs of extra fasteners, extra thermal interface, time etc it really starts to add up the cost
yes jazzm, you really should have the schematic out and have given it a good look while building. ALL the caps on the board are decoupling the rails, which means they connect to ground on one side, there are no caps in tyhe signal path of this amp.
the sound between this and a tube amp couldnt really be more polarized
Last edited:
Indeed, C87 is connected to GND. It's the set voltage bypass cap and it helps improve the ripple rejection of the reg.
If you're using a 25V cap in that location, then it was either a solder bridge, or you overheated the part very badly. Either way, be cautious while assembling the second board!
Glad it's up and running, and indeed, these things have been proving to be very robust! In this case the failure was in a pretty safe spot, and that failure mode perfectly explains the low level fuzzy audio. You'll get that when you only have one polarity!
Cheers,
Owen
If you're using a 25V cap in that location, then it was either a solder bridge, or you overheated the part very badly. Either way, be cautious while assembling the second board!
Glad it's up and running, and indeed, these things have been proving to be very robust! In this case the failure was in a pretty safe spot, and that failure mode perfectly explains the low level fuzzy audio. You'll get that when you only have one polarity!
Cheers,
Owen
I hadn't thought about fixing the bars. They could be drilled and tapped from the bottom so a screw inserted from the bottom of the amp would fix them in place. Not sure how the heatsink itself would be secured though...
Alternatively you could just drill clearance holes that screws from the heatsink would pass through. If you needed the bars to be isolated from the heatsink then some kind of shoulder washer could be used.
All in all though I think I am coming around to qusp's side that it might be more trouble than it's worth.
Alternatively you could just drill clearance holes that screws from the heatsink would pass through. If you needed the bars to be isolated from the heatsink then some kind of shoulder washer could be used.
All in all though I think I am coming around to qusp's side that it might be more trouble than it's worth.
We have a winner!
The question becomes: What thickness to make the pairs of aluminum bars?
My assumption about the machined heatsink offering was for the leg only to contact the board outside the devices / no direct contact with the devices. I see where my post has created the confusion.
I think it is expecting too much for all things to come together in such a way that the tops of the devices and the bars are in the same plane; better to rely upon aluminum contact only... & Isolation would occur at board level, using the keratherm.
The question becomes: What thickness to make the pairs of aluminum bars?
Last edited:
opc mentioned the above quite early in the thread (12th/13th Jan). There is some discussion on it following the post quoted (see post 96 onwards)
Probably applies to your suggestion too but with alu bars you no longer need REALLY thick thermal pad, just a thick-ish one to sort out the installation/manufacturing tolerance of the buffers and then you should be able to bond to both the buffers and the pcb.
Bar thickness probably needs to be as close to the 0.18" height of the buffer as possible.
Last edited:
I remembered this post when i was talking with them last night and here it be, i've already asked for a quote on it along with the kerafoil. its called softtherm and can be up to 5mm thick, is pliable, but the ability to transfer heat is directly disproportionate to the pliability. Its available with adhesive backing available from 0.5-5mm attached is 3 datasheets ranging in thermal ability, probably somewhere in the middle would be best which means about the 86/320 or 86/450 with Rth of 0.5 and 0.3 K/W respectively and W/mK of 2.5 and 4.2. its available up to 7 W/mK but this is probably not only rather pricey, but not so pliable.
not sure if the attachments worked as they are above the limit for size, but if not here is the keratherm datasheets page
and in order of thermal rating
softtherm 86/250 and 86/255
softtherm 86/300 and 86/320 and 86/450 (this is the most suitable range i think)
softtherm 86/500 and 86/525 and 86/600
also HYPERTUNE: how would the CNC go cutting out a stencil if i got the TIM films as a large sheet? or would i be best to just print out a stencil myself and cut with a sharp blade? there isnt that many of us, so wouldnt be too much work
Attachments
Last edited:
In Germany, many Keratherm products are available at an electronics merchant named "Conrad" in small quantities. They would ship worldwide.
Conrad Electronic SE - Your Online Shop for Multimedia, Computers, Cars, Components, Modelling, Batteries & Tools
Conrad Electronic SE - Your Online Shop for Multimedia, Computers, Cars, Components, Modelling, Batteries & Tools
thanks, that might be good for the future, but we dont want individual cutouts and they only have smallish sheets that are a bit on the pricey side then (and i mean no disrespect) i expect to get erm..charged properly... for international shipping from a commercial shop in Germany. they also dont have everything i want from kerafoil. the US and Asian distros got back to me pretty quick, just waiting to check with Mark and opc as to what might be the best way forward and i'll work out a price. buying direct from the level Conrad are buying from is better. i'll probably throw in the extra to get a largish amount and keep the rest to use on other projects.
have you tried the kerafoil KP12 silicone free thermal paste? any good?
have you tried the kerafoil KP12 silicone free thermal paste? any good?
Conrad has a reputation of being an expensive source for components among German DIYers. This might be due to the fact that they don't sell exclusively via the internet. Conrad maintains also a department store chain with shops in several major German cities - just like Radio Shack in the US, but the store net is thinner compared to Radio Shack.
But they carry some products not easily found elsewhere, such as electrolytic caps from Fischer & Tausche and some products made by Keratherm. When comparing prices, please consider that the prices shown include 19% VAT. This tax will not be charged if you would buy from outside the European Union.
No, I haven't yet used the paste you mentioned.
But they carry some products not easily found elsewhere, such as electrolytic caps from Fischer & Tausche and some products made by Keratherm. When comparing prices, please consider that the prices shown include 19% VAT. This tax will not be charged if you would buy from outside the European Union.
No, I haven't yet used the paste you mentioned.
thanks mate, we'll see what we come up with for the best shape and best use of materials to match your thermal heat 'bridge' and if its reasonable to do by hand. youre probably right, a spindle would cut it to shreds without producing a shape with an edge, as it has a cross-linked fiber sort of structure.
I would like to liaise with you, hochopeper and opc over the coming days so i can get back to the distributor over the weekend with a final order. i wont need much of your time, but its best that we are on the same page so i know if i'm taking on too much. the problem is with strips there is no way they can overlap in any way or it will compromise the pressure or compress 2 layers in the overlap too much, or risk a small exposed area.
I would like to liaise with you, hochopeper and opc over the coming days so i can get back to the distributor over the weekend with a final order. i wont need much of your time, but its best that we are on the same page so i know if i'm taking on too much. the problem is with strips there is no way they can overlap in any way or it will compromise the pressure or compress 2 layers in the overlap too much, or risk a small exposed area.
@qusp, I'm working on a drawing that should present what I think are perhaps 2 alternative sizes of template that will be required for the alternative heatsink arrangements discussed here.
I think we need pictures to make sure we're all talking about the same thing, hopefully won't take too long.
I think we need pictures to make sure we're all talking about the same thing, hopefully won't take too long.