Hi,
.....All this in the age of carbon nano tube conductors (and semi-conductors)? 😛
Cheers, 😉
.....All this in the age of carbon nano tube conductors (and semi-conductors)? 😛
Cheers, 😉
Scott, as you can see, I completed that AD815 power amp.
Even changed the screws as per your instructions 🙂
Awaiting delivery of my dScope III for measurements.
But it sounds better than anything I have ever heard!
Jan
My goodness, I'll pass this around with the boys that remember that package! Going to be published? Possibly a good companion to Doug's mass of 5534's 😀
Last edited:
That explains why I couldn't find them. Ah well, maybe there's something else suitable?
Fine work as usual.
Fine work as usual.
The existing package is fine, but would need some other way to extract the heat from the PCB.
I've been thinking about a suitable way, but that would need some mechanical engineering, the downfall of most audio diy projects 😉
Jan
I've been thinking about a suitable way, but that would need some mechanical engineering, the downfall of most audio diy projects 😉
Jan
One can use multiple AD815s in parallel, as detailed in the datasheet.
On top of that, 6 of the leads are connected thermally to the die.
The do provide a path to get heat out from the die directly.
Maybe Scott can tell us whether the SOIC package has the same die size as the power package.
And what the Rthjc is from the die to the case, but also to the 8 thermal heat tabs.
Patrick
On top of that, 6 of the leads are connected thermally to the die.
The do provide a path to get heat out from the die directly.
Maybe Scott can tell us whether the SOIC package has the same die size as the power package.
And what the Rthjc is from the die to the case, but also to the 8 thermal heat tabs.
Patrick
Last edited:
To get the heat out from the thermal heat tabs, you just need to solder some copper litze (e.g. solder wick) onto each of those 8 SOIC legs closest to the case.
Then solder the other end of the litze (say 20mm long) to a cable lug.
Then screw the lugs (with as large a bolting surface as possible) directly onto a piece of copper.
The copper can be say 20x30x2mm.
This can then be bolted onto the heat sink through proper electrical insulation.
(I believe the thermal pads are connected to +Vs.)
Patrick
Then solder the other end of the litze (say 20mm long) to a cable lug.
Then screw the lugs (with as large a bolting surface as possible) directly onto a piece of copper.
The copper can be say 20x30x2mm.
This can then be bolted onto the heat sink through proper electrical insulation.
(I believe the thermal pads are connected to +Vs.)
Patrick
Last edited:
https://www.rocelec.com/parts/results/all/?s=ad815
(Still expensive .......)
And link to pdf (better for Scott to post pdf here .....)
http://pdf1.alldatasheet.net/datasheet-pdf/view/48416/AD/AD815AVR.html
Patrick
(Still expensive .......)
And link to pdf (better for Scott to post pdf here .....)
http://pdf1.alldatasheet.net/datasheet-pdf/view/48416/AD/AD815AVR.html
Patrick
Last edited:
I probably still have a few tubes somewhere, and yes Rochester is our official discontinued part distributor so they are real.
I think it was Bonsai that found someone selling parts removed from service, which considering the package are very hard to fake.
There is only one die so the DIP is the same part with the heat spreading pads directly connected to the paddle. It is hard to get at the real thermal numbers because JEDEC standards don't allow you to add any cleverness when you characterize the thermal performance.
If you would tell me the die size in mm2, I can estimate the Rth to the legs.
Small legs though, since package is unfortunately SOIC and not DIP.
Rthjc can be rough guessed using numbers from other SOIC24 packages .....
Patrick
Small legs though, since package is unfortunately SOIC and not DIP.
Rthjc can be rough guessed using numbers from other SOIC24 packages .....
Patrick
This looks OK even though the fonts are wrong, I actually can't find one here anymore except maybe a hardcopy buried in some drawer. I did hand some out at BA a few years ago with a few tubes of samples.
I'll have to get that from the databank at work. I think I mentioned this before our cost on the big package was $0.95 way out of line with current costs.
EDIT - This part is junction isolated, in this case the back of the die is/must be connected to Vcc.
Last edited:
You can probably find some ground cables with a lug and a flatcable the same width as the 4 pin spread.
Jan
See page three ,"heat tabs are tied to positive supply". I admit the OCR on this DS is hard to read.
Last edited:
The SOIC24 costs already 6 USD at 1000 pcs.
Might as well buy from Rochester and save all the trouble.
You can use Jan's PCB as a bonus.
🙂
Patrick
Might as well buy from Rochester and save all the trouble.
You can use Jan's PCB as a bonus.
🙂
Patrick
- Status
- Not open for further replies.