Aleph 30 PSU

[Scribble]

Hi,

I have recently re-commenced construction of my Aleph 30 amp which went on hold for a bit over a year when i moved sections at work .

I have used the Peter Daniel boards, got the first channel going today and it was great! but hot hot. Im pretty sure my rail voltages are too high. I got a custom made transformer at the time, a dual secondarys 21v 500VA toroidal. Why i went 21v at the time i cant remember .

Im running a simple capacitor bank containing 6x 10,000uF Nichicon caps per channel. One common bridge between the channels. Im getting around 28v DC @ 240v AC under load, one channel operating. Should i expect the voltage to drop slightly with the second channel attached?

After about 45 minutes of playing music the heatsink temp gets to about ~60c. This is in an aircon office @ about 26c. Im presuming lowering the rail voltages will reduce the temp a bit. What can i do, besides buying a new transformer, to lower the rail voltages without affecting stability/sound quality?




Thanks for your time!
Scribble


PS. The one channel sounds great!

[Scribble]

Ok today i added two bridge rectifiers in series per rail. This got my rails down to ~25v DC @ 240v AC.

I dont know if they have affected sonics as i have not listened to the channels yet with this mod. Is there any negative effects to what i have done?




Thanks for your time!

[lineup]

ambient temp 26 C (temp around amplifier)
heatsink temp 60 C

This is quite alright
A target value for how much difference should be
->>> is around 40 C

So if heatsink temp should be considerably above
26 + 40 = 66 C
then you should probably need a more effective(bigger) heatsink

Of course is no problem if the difference is lower than 40 C degrees.
In your case diff is 34 C

[Scribble]

Thanks for the reply!

I measured the temps on the surface of the IRFP240 packages and the hottest i could find was about ~75c. Running a small fan @6v (was inaudible) the surface temp of the devices got to around ~62c. This was with the extra bridges in series with the rails to get them down to about ~25v DC.

Would you know the expected surface temperature of the devices?

[Bill Fuss]

Hi Scribble,
I dont have direct experience with the A30 but have an A5 and F5. It sounds like you are not comfortable with the class A heat and I understand completely. Sounds like your build decisions have piled up on you somewhat, probably marginally inadequate heatsinks and too high a rail voltage. If the case temp of the 240s is 75* I would think that's pretty high and failure rate would be higher than a properly heatsinked unit, but that's just my opinion. If your heatsinks cant handle the disappation then you must either lower the bias current, use adequate fans, or get bigger heatsinks. As the sound of the Alephs depends on the high bias current, lowering it is not a good option but it can be done. Check out the WIKI for the info on changing the bias. Dropping the rail voltages will also help, but not as much as you would think.
The solution you decide on will undoubtably be worth the trouble and expense, they are truly great amps.

Best regards, Bill

[lineup]

To calculate actual
Tj, junction temperature (inside) of a transistor is easy when you have the:
Tc, CASE temperaure
Watt, Power that the transistor runs

The coefficient per watt: Rtjc
Rtjc for IRFP240, TO247 package is: 0.83C/W

Tj = Tc + (Watt x Rtjc)

For you:
Tj = 75 + (watt x 0.83)
-------------------------------------------------------------------------------

There are 2 ways we can say a temperture is too high (heatsinking not enough):

1. Temp close to or above The absolute Max.
For many power transistors the absolute MAX Tj is 150 C'
When getting to close this, we can have transistor damaged.

2. Tj is high for the transistor to perform good audio.
Here it is much a matter of each designers preferences.

Some, that does not want to use GIGANTIC heatsinks,
maybe can accept Tj ~ 100 C ....... 50 C below MAX
Here comes your wallet in, because big heatsinks are not cheap!

Others like Nelson, when money certainly is no object,
maybe run at Tj, junction temperature, like max 75-80 C .... safety margin 70 C below max !!!