The idea is to remove heat from the output device. Spreading it over a wide area is the way to do it.
Because the output devices run so hot, it only makes sense that the heat flowing through a small area between the output device and the heatsink is more concentrated than anywhere else. Thus is it would seem that reducing the thermal resistance at that point would be most beneficial.
So let me get this straight in my head.
A heat spreader is designed to spread the heat from an output device over a wider area.
The only way that I can see a heatspreader doing that better than directly mounting onto a heatsink would be if the thermal efficiency of the heat spreader was much higher than the thermal efficiency of the heatsink.
e.g. If the heatspreader was made of say copper and the heatsink was say aluminium. In that case then I can understand why somebody would use this method.
That therefore begs the question; what material is the heatspreader made of, and; how efficient is the thermal transfer across the junction of the heatspreader and the heatsink.
The idea is to remove heat from the output device. Spreading it over a wide area is the way to do it.
Because the output devices run so hot, it only makes sense that the heat flowing through a small area between the output device and the heatsink is more concentrated than anywhere else. Thus is it would seem that reducing the thermal resistance at that point would be most beneficial.
and when the heatsink is not up for the job, let the spreader do it.
thats my point🙂 don't conentrate the heat on the sink. thats what you trying to avoide🙂
a big spreader does the job better then multiple smal ones. (and heatsinks with thin base)
the smal one's will not spread the heat very much.
So let me get this straight in my head.
A heat spreader is designed to spread the heat from an output device over a wider area.
The only way that I can see a heatspreader doing that better than directly mounting onto a heatsink would be if the thermal efficiency of the heat spreader was much higher than the thermal efficiency of the heatsink.
e.g. If the heatspreader was made of say copper and the heatsink was say aluminium. In that case then I can understand why somebody would use this method.
That therefore begs the question; what material is the heatspreader made of, and; how efficient is the thermal transfer across the junction of the heatspreader and the heatsink.
exactly 🙂
without any numbers or experiance to back it up. i think copper is the best way to go. you need a material that is softer then the heatsink.
Last edited:
I guess the problem with a heat spreader is achieving good thermal coupling with the main heatsink.
How do you guys manage that?
Graphite grease machined surfaces and about 10 hours of han lapping
(one surface against the other in place) first with carbide and finish off with baking powder till is all shiny and mirror like and the 2 pieces stick togheter as if magnetised (No I did not invent magnetic cooper it has someting to do vith the vacum formed between the 2 )
2 heat sink 2 heat spreders each is 40 hours of rubbing the pieces togheter by hand.
Then clamp the lot with loads of bolts for wich sink has to be drilled and elicoils fitted
and Spreader got to be drilled cuntersunk and oversiszed holes to alowd for expansion.
If one is in to self punisment and or self medication is easily done but the use of Pillar drill is strictley forbiden😀😀😀
I once thought of mounting the P type and N types devices on separate copper heat spreaders electrically isolated from the aluminum, so you can mount the outputs directly without any insulator reducing the case temperature even more.
I was thinking about this but same expert came up with the increased capacitance that the heath spreder wuld have in this case.
Rather than get in to the argument( I could not mathematicaly prove it wrong) I went for kerafoll onder the Mosfets.
hehe 😀 torture 😀 😀Graphite grease machined surfaces and about 10 hours of han lapping
(one surface against the other in place) first with carbide and finish off with baking powder till is all shiny and mirror like and the 2 pieces stick togheter as if magnetised (No I did not invent magnetic cooper it has someting to do vith the vacum formed between the 2 )
2 heat sink 2 heat spreders each is 40 hours of rubbing the pieces togheter by hand.
Then clamp the lot with loads of bolts for wich sink has to be drilled and elicoils fitted
and Spreader got to be drilled cuntersunk and oversiszed holes to alowd for expansion.
hehe😀 torture😀
If one is in to self punisment and or self medication is easily done but the use of Pillar drill is strictley forbiden😀😀😀
Okey dokey
Recap
Buy heat sink cost loads of dosh still base plate is only 10mm.
That conduct heat to fins that radiate heat to air.
Duble thikness of base plate it reduces the times the heat takes to get to the fins that radiate heat to air.
Heath travels 2 times as fast on cooper when compared to aluminium
Aluminium Radiate heat much better than cooper as cooper is like a mirror
Black radiate heat better than white
At least this is the symple explanation no need to complicate things.
Use Kerafoll 86/83 I do not know of better specs materiall
Kerafoll likes pressure duble the pressure reduces thermal impedence by half
There is a graph on data sheet and of the top of my head it show pressure up to 120 PSI for best performance but be carefull if you use proper screws with cuntersunk heads (IMO best advantagge with using heat spreaders) one may get to the point at whic the mosfect body crack with to much pressure.
Why I like to cuntersunk the screws heads on the heat spreaders?
Because the tread expecialy for M3 screws strips far to easy for my taste.
One has to use elicoil inserts to repair the tread or move the mosfets and drill new holes.
How many time did I change Mosfets on my boards / heat sinks?
Loads
how many new holes ?
Zero
Recap
Buy heat sink cost loads of dosh still base plate is only 10mm.
That conduct heat to fins that radiate heat to air.
Duble thikness of base plate it reduces the times the heat takes to get to the fins that radiate heat to air.
Heath travels 2 times as fast on cooper when compared to aluminium
Aluminium Radiate heat much better than cooper as cooper is like a mirror
Black radiate heat better than white
At least this is the symple explanation no need to complicate things.
Use Kerafoll 86/83 I do not know of better specs materiall
Kerafoll likes pressure duble the pressure reduces thermal impedence by half
There is a graph on data sheet and of the top of my head it show pressure up to 120 PSI for best performance but be carefull if you use proper screws with cuntersunk heads (IMO best advantagge with using heat spreaders) one may get to the point at whic the mosfect body crack with to much pressure.
Why I like to cuntersunk the screws heads on the heat spreaders?
Because the tread expecialy for M3 screws strips far to easy for my taste.
One has to use elicoil inserts to repair the tread or move the mosfets and drill new holes.
How many time did I change Mosfets on my boards / heat sinks?
Loads
how many new holes ?
Zero
demands 100% even surface on heatsink and heatspreader. witch is almost inpossible.
A surface grinder does 1 micron accuracy.
(me surface grinding press/extrusion die & mould parts at age 10)
A surface grinder does 1 micron accuracy.
(me surface grinding press/extrusion die & mould parts at age 10)
machin grinding🙂 you dont get that by hand grinding.
so the cost will for most be high. and cheeper to buy heatsinks with 15mm straight base🙂
Last edited:
Here's an inexpensive soft-start module on ebay. Probably cheapo relays but maybe can be replaced with some better ones.
Class A power delay soft-start temperature protection board 110V/220V | eBay
Class A power delay soft-start temperature protection board 110V/220V | eBay
Without direct access to a surface grinder, there's the option of stopping by at an engine head reconditioning shop (if desired), shouldn't cost an arm and a leg (or a child's foot for some).
Or go for BS' hand lapping routine. (mirror, mirror, on the wall)
Or go for BS' hand lapping routine. (mirror, mirror, on the wall)
Last edited:
Without direct access to a surface grinder, there's the option of stopping by at an engine head reconditioning shop.
yes it is. but it depends on price, and if they want to do it. and if there are any close by😀
around here they'r not on every corner😀 😀
Here's an inexpensive soft-start module on ebay. Probably cheapo relays but maybe can be replaced with some better ones.
Class A power delay soft-start temperature protection board 110V/220V | eBay
that one looks pretty good. if it had stanby/pushbutton and LED's i'll jump over it my self😀
Hi Jacco
Maybe you can explain properly why precision ground surfaces stick togheter.
Hand lapping does better
Use to do that at 12 helping out Franciscan Monk hand lapping telescope lenses.
And valve seats on marine diesels after that.
And another thing
I was going to make heat sink out of cooper but not enough scrap to make it possible in the UK
Cooper roofs are not that popular here but if there is a Cooper roof near you it means you can get the proper scrap
Silver braze he thin Cooper sheet to the thick base (afer you cuntersunk the bolt heads naturaly) and use a litle fan.
Easier than to weld Aluminium for me.
litle fan is must have for Cooper sinks (remember radiated heat is poor for mirrors)
Proper F5 turbo sinks material in Cooper is about £500 if one can not get scrap.
Maybe you can explain properly why precision ground surfaces stick togheter.
Hand lapping does better
Use to do that at 12 helping out Franciscan Monk hand lapping telescope lenses.
And valve seats on marine diesels after that.
And another thing
I was going to make heat sink out of cooper but not enough scrap to make it possible in the UK
Cooper roofs are not that popular here but if there is a Cooper roof near you it means you can get the proper scrap
Silver braze he thin Cooper sheet to the thick base (afer you cuntersunk the bolt heads naturaly) and use a litle fan.
Easier than to weld Aluminium for me.
litle fan is must have for Cooper sinks (remember radiated heat is poor for mirrors)
Proper F5 turbo sinks material in Cooper is about £500 if one can not get scrap.
The idea is to remove heat from the output device. Spreading it over a wide area is the way to do it.
Because the output devices run so hot, it only makes sense that the heat flowing through a small area between the output device and the heatsink is more concentrated than anywhere else. Thus is it would seem that reducing the thermal resistance at that point would be most beneficial.
I use mica insulators of 1-2 mils thickness. Here is a good reference about heatsinks, insulators, and transistor mounting: ESP - Heatsink design and transistor mounting
😀Why one would go to buy stuff from Ebay for then having to replace parts.
When even me with a few tips from intraweb can design its on thing.
And spend a few hours building things.
What pride one has at putting togheter IKEA like bits apart from telling its mates I have built this and did not even have to read the instruction manual😀
Maybe I shuld start selling stuff on Ebay insted of giving away info for free.
Only joking by the way so dont get oll titty lippy.
On a more serious note how would you like to buy my soft start I use 20 A rated relays?
And you know where to slag me off if things go wrong?
Ihquam that is most definitley a wery good link
You Lot certanly know that I am not that clever so you beter read ESP page I certanly did
When even me with a few tips from intraweb can design its on thing.
And spend a few hours building things.
What pride one has at putting togheter IKEA like bits apart from telling its mates I have built this and did not even have to read the instruction manual😀
Maybe I shuld start selling stuff on Ebay insted of giving away info for free.
Only joking by the way so dont get oll titty lippy.
On a more serious note how would you like to buy my soft start I use 20 A rated relays?
And you know where to slag me off if things go wrong?
Ihquam that is most definitley a wery good link
You Lot certanly know that I am not that clever so you beter read ESP page I certanly did
Last edited:
Hand lapping two bit of metal produces a finished surface (lots of care needed) that has very few 'peaks and valleys' - if you pull all the scratches out, your left with a surface which will expel air causing a vacume effect.........
and engine heads should not be ground this way as there will be an easy path for the fluids/gasses to flow through.......
realy realy basic eng. theory
Like the boards Toecutter, well done
realy realy basic eng. theory
Like the boards Toecutter, well done
that one looks pretty good. if it had stanby/pushbutton and LED's i'll jump over it my self😀
It can probably be modified to do what you need. The relays look like some made by Omron I've seen in appliances. I think they are universal standard sized, but not positive. Hard to beat for the money though.
Here's an inexpensive soft-start module on ebay. Probably cheapo relays but maybe can be replaced with some better ones.
Class A power delay soft-start temperature protection board 110V/220V | eBay
The price is scary ...... will it work .? 😛
Copper is pretty pricey, better to save the money , fets are cheaper to buy if you let the smoke out, take the cash and buy copper options ..🙂 Big sinks, 10 -15 mm heat spreader , if that's not enuff ... Tits up ....

- Status
- Not open for further replies.
- Home
- Amplifiers
- Pass Labs
- F5 Turbo Circuit Boards