First Pass project, need Aleph-M PSU advice please

[ssmith]

Hi all

Having dabbled in T-amp, cobbled together a gainclone, built frugel horns and a dac... the diy addiction has deepened. I strayed into the Pass Labs forum (I always used to avoid it because you're all so damn clever), where I stumbled across a Grey's highly entertaining Mini-A thread. Literally hundreds of hours of reading and searching later I'm putting together a Mini-A build.

My plan is to make two stereo amps to biamp two-way speakers. The Mini-A PCBs are on order, I have the transformers, huge heatsinks (just in case), and P-A's softstart/DC filter/rectifier PCBs.

I've almost got it all figured out (as best as a non-techie can hope to), except the PSU. The CRC/CLC business had me stumped for a while, but afte some searching I decided to go for CCRCCCCc (inductors are heavy to get posted to me and expensive).
My draft PSU schematic (WARNING: Microsoft Paint) is attached.

There will be 2 PSU's, each in their own chassis and each supplying two channels -- one high freqs (active xover point of 2,500hz) and one for lows. I assumed that going full dual mono would be a bit of a waste.

If anyone could comment on the schematic I'd be most grateful -- main question is the value and power rating of the resistor (I don;t want to have to heatsink it), and bypassing (I have some 2.2uf teflon speaker xover caps handy so just stuck those on). For the 10,000uf/35v caps I was planning on Nichicon KG.

thanks in advance!

 

[MEGA_amp]

I would parallel another 0.50R 20W to be supersafe, that should be fine, and if you want to increase bias you wont have to change your PS capability. With 15V secondaries you don’t need to drop much voltage. I'm currently running my Mini-A at +/-17.5Volts at 2.7amps with a trafo at 19volts. I have two paralleled 0.47R 25W per rail, and they get HOT, they're holding up fine and heatsinking isn’t necessary. The rest of the PS looks fine.



-john

 

[ssmith]

John,
Many thanks.
I found on Mouser some Vishay Dale 25w aluminium housed resistors so will go for those paralleled or one 50w.
As you said, the option to increase bias is nice to have. Hadn't thought of that. Would prefer for them not to get HOT, so can always add some extra little heatsinking to them anyway.
regards and thanks again,
stefan

 

[woody]

Why not just use an inductor instead of the power resistor. A
.5mh air core inductor from Madisound using 16 gage wire is
priced at $7 and has a dc resistance of .22 ohm. You would get
a little high frequency filtering thrown in for free.

 

[ssmith]

Hello Woody and thanks.
To be honest I was wary of using an air core inductor having read about saturation and it throwing off interference.

I also read this in the Aleph X wiki :

Aleph-X wiki
L vs. R... Some general notes: Large enough L values are difficult to achieve at typical Aleph-X power currents. Large and expensive coils are needed, and they have high series resistance. Due to the high R and low L values of such inductors, my simulations show little difference in performance between realistic CLC and CRC filters. Magnetic fields from inductors may also be a concern, especially for air-core inductors, so give due consideration to mounting and shielding. Many iron core coils used for speaker crossovers will not be suitable for power supply use, as they will saturate and lose their inductance due to the DC currents.

Although much higher currents are involved in the Aleph-X (4.5A) and this does not apply the the Mini Aleph (1.5A to 2.5A or so I guess), I do not have the knowledge to determine DC current handling on aircore coils and this is something you don;t see on published specs. But if you say a 16 or so gauge air core inductor would be fine, I guess that would be a preferable option.... as the resistors I mentioned are about the same price anyway....

 

[MEGA_amp]

sure no prob, I used the cement types, resistors can be run closer to their stated ratings without failure. I dont know why I said I ended up with +/-17.5 volt rails, its more like 21.5V, I shouldn't post as soon as I wake up.

-john

 

[juma]

This is how I did it:
(Iq=1.8A per channel, ripple=70mV peak-to-peak)

 

[GRollins]

main question is the value and power rating of the resistor (I don;t want to have to heatsink it)

Hundreds of hours reading the Mini-A thread? Goodness, that thread must have grown considerably since the last time I looked at it.
If I understood correctly, you're planning on running two channels of Mini-A (what's an Aleph-M?) off of each power supply. Assuming that the bias current is anywhere near the stock value, you'll see 1A per channel, 2A total per power supply. Unless I'm even shorter on sleep than I think I am, that's around 2W dissipation. Given a 20W resistor, I'd say you're pretty much good to go.

Grey

 

[ssmith]

Re: Re: First Pass project, need Aleph-M PSU advice please

Hundreds of hours reading the Mini-A thread? Goodness, that thread must have grown considerably since the last time I looked at it.

No. As I said, hundreds of hours reading and searching. Starting with that (your) Mini-A thread, the GB threads, the other (related) aleph threads, zen threads for good measure, then headed off at tangents -- crc, clc, biasing, mosfet matching, capacitance, heatsinking, Nelson's articles, electronics for dummies (I'm one of them)... and so on. Out of what I read, I probably vaguely grasp about a quarter of it at best, and truly understand far less than that. Still, I have a plan on paper and some parts.

what's an Aleph-M?

Blasphemy, apparently? Sorry.

Assuming that the bias current is anywhere near the stock value, you'll see 1A per channel, 2A total per power supply. Unless I'm even shorter on sleep than I think I am, that's around 2W dissipation. Given a 20W resistor, I'd say you're pretty much good to go.

thanks.

This is how I did it:
(Iq=1.8A per channel, ripple=70mV peak-to-peak)

Thanks.

I'm currently running my Mini-A... at 2.7amps

Is that per channel? And what kind of power output does that give you into 8ohms?

 

[GRollins]

Given the numbers quoted above--roughly 22W into 8 Ohms.

Grey

 

[MEGA_amp]

At 2.7 amps per channel, I'm also dissipating about 58 watts per output transistor. Just make sure you have enough heatsink if upping the bias is in the plans.

-john

 

[Ryssen]

I'm also dissipating about 58 watts per output transistor.

What transistors?Can they take it?

 

[Magura]

What transistors?Can they take it?

You can approach 100W per device (IRFP240) if adequate heatinking is added to the equation.

Magura

 

[MEGA_amp]

Each IRFP240 had its own 12"x9.75" heatsink.

 

[ssmith]

Hi again and sorry to bother.

I'm now stuck on what PSU caps to use --

1. a small number of 22,000uf/33,000uf screw terminal large can caps (like Cornell Dubilier)
2. a larger number of 10,000uf 35v snap-in caps (like Nichicon KG)
3. an even larger number of Panasonic FM 1800uf/35v caps (not too expensive from digikey).

The cost works out to be roughly the same, and I don't mind option#3 being a bit of a PITA to install and wire up, though I'm wondering if it would be worth the extra effort? From what I've gathered from searching here many people suggest using lots of smaller caps for low ESR and higher ripple, while others swear by the superiority of large cans.

Your thoughts would be appreciated.