>
> similar idea. You need a topside copper pour to transfer heat from the
> dpak tab to the heat sink. That might be in the ballpark of 5 K/w, so
> adds to the overall theta some.
>
> What do those things cost? The hype level on the data sheet is high.
> There is no data below 200 fpm air flow, and the theta numbers look
> optimistic. All those tiny fins are useless at low air flow. The digs
> at Aavid are absurd.
>
> You can get pretty good dpak cooling by coupling through vias into pcb
> planes, and that's free.
>
>
>
Reply by Klaus Kragelund●April 20, 20202020-04-20
On Sunday, April 19, 2020 at 10:19:02 AM UTC+2, Piotr Wyderski wrote:
> Hi,
>
> does this make much sense?
>
> http://smartheatsinks.com/SHS_SMT.pdf
>
> The heat path between the transistor and the heatsink is awkward.
>
It looks interesting
I would perhaps be a little worried about if the heatsink reaches the correct temperature to enable correct soldering of both the heatsink and the SMD device
Cheers
Klaus
Reply by Jasen Betts●April 20, 20202020-04-20
On 2020-04-19, legg <legg@nospam.magma.ca> wrote:
> On Sun, 19 Apr 2020 10:18:56 +0200, Piotr Wyderski
><peter.pan@neverland.mil> wrote:
>
>>Hi,
>>
>>does this make much sense?
>>
>>http://smartheatsinks.com/SHS_SMT.pdf
>>
>>The heat path between the transistor and the heatsink is awkward.
>>
>> Best regards, Piotr
>
> Look at the competition - same issue.
>
> The AAVID parts are tinned copper and patented up
> the wazzoo.
>
> It would likely make more sense to turn the TO220 around, fix
> it to the extrusion, and then reflow the extrusion to the board.
>
> ..but then it wouldn't be a refow-only assembly.
form the leads for inverted surace mount
reflow the TO220 onto the heatsink.
reflow the assembly onto the board.
or place the TO220, place extra paste on the tab
place the heatsink on the tab.
> Fabs are allergic to mechanical fasteners.
not needed
the problem is getting DPAK with the leads bent away from the tab
This non-isolated buck module places the beastink at the other end
of a bunch of vias: https://www.aliexpress.com/item/32811528144.html
But it's only stuck on with glue (feels like some sort of silicone).
--
Jasen.
Reply by ●April 19, 20202020-04-19
On Sun, 19 Apr 2020 10:18:56 +0200, Piotr Wyderski
<peter.pan@neverland.mil> wrote:
>Hi,
>
>does this make much sense?
>
>http://smartheatsinks.com/SHS_SMT.pdf
>
>The heat path between the transistor and the heatsink is awkward.
>
> Best regards, Piotr
We use these,
https://www.dropbox.com/s/klpmvkfswq9tzmb/Surf_Mount_Heatsink.jpg?raw=1
similar idea. You need a topside copper pour to transfer heat from the
dpak tab to the heat sink. That might be in the ballpark of 5 K/w, so
adds to the overall theta some.
What do those things cost? The hype level on the data sheet is high.
There is no data below 200 fpm air flow, and the theta numbers look
optimistic. All those tiny fins are useless at low air flow. The digs
at Aavid are absurd.
You can get pretty good dpak cooling by coupling through vias into pcb
planes, and that's free.
--
John Larkin Highland Technology, Inc
Science teaches us to doubt.
Claude Bernard
Reply by Winfield Hill●April 19, 20202020-04-19
legg wrote...
>
> It would likely make more sense to turn the TO220 around,
> fix it to the extrusion, and then reflow the extrusion
> to the board.
>
>..but then it wouldn't be a refow-only assembly.
> Fabs are allergic to mechanical fasteners.
Yes, good idea. One idea, leave a bit of final
hand work, for after the automated assembly step.
--
Thanks,
- Win
Reply by legg●April 19, 20202020-04-19
On Sun, 19 Apr 2020 10:18:56 +0200, Piotr Wyderski
<peter.pan@neverland.mil> wrote:
>Hi,
>
>does this make much sense?
>
>http://smartheatsinks.com/SHS_SMT.pdf
>
>The heat path between the transistor and the heatsink is awkward.
>
> Best regards, Piotr
Look at the competition - same issue.
The AAVID parts are tinned copper and patented up
the wazzoo.
It would likely make more sense to turn the TO220 around, fix
it to the extrusion, and then reflow the extrusion to the board.
..but then it wouldn't be a refow-only assembly.
Fabs are allergic to mechanical fasteners.
RL
Reply by Phil Hobbs●April 19, 20202020-04-19
On 2020-04-19 04:18, Piotr Wyderski wrote:
> Hi,
>
> does this make much sense?
>
> http://smartheatsinks.com/SHS_SMT.pdf
>
> The heat path between the transistor and the heatsink is awkward.
>
> Best regards, Piotr
Reduces the board area required for thermal pours, and will work
slightly better. Their claims about better thermal performance look
like snake oil--theta will be dominated by the copper pour.
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510
http://electrooptical.nethttp://hobbs-eo.com
Reply by Piotr Wyderski●April 19, 20202020-04-19
Hi,
does this make much sense?
http://smartheatsinks.com/SHS_SMT.pdf
The heat path between the transistor and the heatsink is awkward.
Best regards, Piotr