Krell KSA 100mkII Clone

[seafire]

Good Josh I hope my post puts some spark back in you.Gees guys how long do we have to wait in suspense?

[Fix]

@Jozua:

hmmm then the preasure is on me now...
This is my first Amp project...

- - - -

UPDATE:

After some setbacks last night, I completed one channel.

Did some initial testing before today, and everythings looks good.
50VDC, 575mv over Emitter and 2.3Amp of current.
Heatsinks get about ~50* C without fan, after 15min

Will test with Fan tonight and run it for 2-3hours.


Will post pictures later today.

[Fix]

Testing....

Tested for about 1h... will let it be on for another hour tonight...



VCC: 50v - 2.54amp
Emitter: 625mv
Heatsink: ca: 36C with fan at full speed


In/Out signal 1Khz
In: 0.1v / square
Ut: 0.5v / square

Pictures are taken before I went up to 625mv

Question: is the emitter voltage supposed to drop 0.1-0.3v when getting hot ?

If I turn pot up again to 625mv then it after 5min drop 0.1-0.2v..

Guess something isn't stable in my powercube or signalgenerator.

** Just touched my powercube, it's super hot ... hahaha...
Hope it won't burn up... :-)


Question 2:

Is it normal that the emitter voltage is NOT the same on all transistors ?

625mv
618mv
630mv
622mv
.....

[Stuart Easson]

...if you didn't match the transistors and the emitter resistors you will see minor variations, and those are very minor.

I should add (in case it isn't obvious) matching the transistors without matching the resistors is completely pointless.

HTH

Stuart

[AndrewT]

Yes, the emitter resistors must be matched to better than 1% if you intend to use them as a measurement standard at a later stage.

I aim for better than 0.1% across all parallel pairs/sets of resistors.
Then I know I can rely on measurements thereafter. This really does show up transistor matching. Output sets must run with the base voltage equal for all. You can then measure emitter current and how it varies as bias is altered.

[Fix]

I'll keep it as it is, since the variation is minor.
Guess it would be hard to get 100% from both channels.

And I don't want to take everything apart once more....


- -

Anyway, I hooked up an old speaker 4ohm (just element) and used my computer as source..

It sounds crap...
When I use my signalgenerator it sound better, and not distorted.
But I can't get high output... ?!

When I raise the volym the Amp meter on my powercube lowers down to 0.5amp ( when speaker is connected )

1. I have serious problem with the amp.
2. My Powercube CAN'T drive my amp and that why I get those readings.

I guess the Powercube isn't build for driving a Class A amp...
But it's rated at 3Amp, guess that's why we need a 1000VA trafo rated 12amp @ 40v ?

Any ideas ?

[AndrewT]

Hi,
is the KSA100 Klone biased to 2.6A?

A push pull ClassA output stage driven to it's ClassA current limit, will draw ~5.2Apk from each rail alternately. That's 100W of ClassA into 8r0.

How the .... can you expect a 3A supply to even let the Klone run in ClassA?
What happens when you ask the amp to transit into ClassAB? Yes ask the Klone to deliver 55W into 4r0.
The rails currents become even bigger.

[BobEllis]

Another issue may be the way your power is connected. It looks like the power supply is connected to the rails without a ground. Floating the amp's ground may work for some instances, but using a single ended source, like the PC probably won't work out well.

If the two psu sides are truly independent (no terminals are connected to any other or ground when you select independent operation), you can create a proper split supply. Use the positive supply terminal to ground for the negative rail. Then tie the negative terminal of the other side to ground and you have the positive rail.

Of course, that doesn't alleviate the current limiting issue Andrew pointed out, but you should be able to listen at low levels.

[john65b]

For what its worth, the KSA-50 I built initially had the signal ground floating, and it sounded horrible too.

I think the bias and DC offset were not stable either without signal ground connected correctly.

Also, the amp ran quite hot when it was connected incorrectly too.

Once signal ground was correctly connected, all was perfect.

[AndrewT]

Bob & Fix,
that 5.2A ClassA current limit gets used up very quickly.
eg. 4ohm highly reactive speakers that draw (at low level output signal) enormous peak currents.
Let's assume that we have a peak current on fast transients of around ~ Volts peak/Load/0.35
For a 5Vac signal, that comes to 5.05Apk.
5Vac is only 6W into 4r0.
That is not much of a peak transient if you happen to be listening at an average level of 1W.

Yes, 4ohms speakers will run out of ClassA current at very low listening levels even when the Klone is biased to 2.6A through the output stage.

The 3A supply will similarly run out of current at very low listening levels. 3A has a margin of just 400mA above the bias current.
One rail could be supplying 3A and the other rail will have gone down to 2.2A. The difference between these rail current passes through the speaker and returns to audio ground.
800mApk is equivalent to 1.28W into 4r0. that is a transient peak of +1dB above 1W output.

All amplifiers need adequate current supplies to meet domestic music listening requirements. Don't skimp on the PSU because the ill informed have said it's not necessary.