Krell KSA 50 PCB

attachment

The emitter resistor values for those amps are 1 Ohm, not 0.5 Ohm.
325mA quiescent current per output device in the KSA-80 sums up to 3.9A peak class A output current (2 banks of 6 output devices).
3.9A is ~120W peak class A or some 60W continuous Class A power.

Neither value really fits the bill for a nominal 80W Class A power amp.
Reality is that the bias level is determined by the power handling capability of the Krell design heatsink in the KSA/KMA series.
Still today a very efficient heatsink for it's cubic volume and mounting footprint imo, but bottomline remains that it can handle no more than 80W at an acceptable temperature limit.

REAL Class A power amps are biased to the nominal level for both 8 and 4 Ohm operation.
1.8A bias for the KSA50 means that it not only does 50W Class A in 8 Ohm, but also in 4 Ohm loads (as in : 3.6 x 3.6 x 4 = 50W without entering the cross-over area)
 
Last edited:
325mV on .5 ohm emitter resistors is 650mA per device

KAmps said Max class A is limited by the bias current of 690mA per device...

Aren't you both saying same thing? How is he over driving his devices (40mA?)

I have mine at 610mV on .649 emitters, getting around 940mA per device...are you saying I am just wasting heat and there is no Sonic benefit??

Your attachment was not for KSA-50, but KSA-80/100/160/200. I remember the KSA-50 design was two devices driven to over 900mA per device.

Nelson Pass' position of raising bias as much as your sinks can take to better sound doesn't apply here (I know different design - single ended and he uses MOSFETS)??

Not trying to cause trouble...just interested in finding out. I wouldn't mind turning down the bias on my Krells...if there was a benefit...




Got my copy...thanks.


Got mine too 😎

Guys the Class-A calculations have been vetted by a host of experts on the forum including the one and only himself. PC walley... as far as sonics go, thats a debate that will not be settled anytime soon. I can tell you this, it is not merely about only eliminating x-over distortion... (though that is one of the benefits) heavy Bias has many other effects on the system.

One last thing, when I joined the forum a few years ago, to my own detriment I did not sometimes give everyone the respect they deserved. The amount of collective knowledge that the forum has; is simply astounding, consider yourself blessed.
 
REAL Class A power amps are biased to the nominal level for both 8 and 4 Ohm operation.
1.8A bias for the KSA50 means that it not only does 50W Class A in 8 Ohm, but also in 4 Ohm loads (as in : 3.6 x 3.6 x 4 = 50W without entering the cross-over area)
oops.

The maximum ClassA current for a 1.9A bias is ~3.7Apk.
This will result in a maximum ClassA power into 8r0 of 3.7 * 3.7 * 8 / 2 = 54W.
The maximum ClassA power into 4r0 is 3.7 *3.7 * 4 / 2 = 27W

The KSA50 is biased for maximum ClassA power into 8r0 not 4r0, nor 6r0.
 
Hi,
I finished may KSA50 but I'm not really happy. The sound isn't I expected. I feel something like fog. Trebles are not very good and i think that there are to few details in music. Do you think it'll change?

Bias 1A/transistor
2xMJ15003 & 2xMJ15004 per channel
Power supply 2x47000uF Rifa & 2x45V per channel
Everything is like in an original schematic of ksa50.
 
Hi,
I finished may KSA50 but I'm not really happy. The sound isn't I expected. I feel something like fog. Trebles are not very good and i think that there are to few details in music. Do you think it'll change?

Bias 1A/transistor
2xMJ15003 & 2xMJ15004 per channel
Power supply 2x47000uF Rifa & 2x45V per channel
Everything is like in an original schematic of ksa50.

KSA-50 runs warmer (soundwise) than other amps thats why people like them. So if you are used to more HF then an adjustment needs to happen... (and not necessarily to the KSA-50 😉 ).

What drivers do you have? The MJ1503x are warmer sounding than Japanese trannies in my experience e.g. d669 b649s sound very clear but at high bias levels they may be inadequate...
 
The RCA input ground to board input ground (assuming your using the Pimkmouse boards?) and no where else. The signal input ground should not terminate into chassis.

Also a jumper on board from neg signal input to ground since you indicated RCA input (Single ended) and not balanced (XLR).

What I can't remember if your supposed to have continuity (zero resistance) from RCA barrel ground to power supply ground ....I think so...

Also, as Neychi says, check your bias. Too much or little will affect sound - although can't see too much being bad if your sinks can take teh heat.
 
Yes, but obviously there is a reason for that. KSA 50 Mark II had two editions, with and without mosfets. Mark I had also two editions, with 36,5 V rails (1,8 A bias) and later with 48 V rails (1 A bias). That, later edition of Mark I, had same layout as earlier Mark II but with cooling fan on the botom.
 
I wrote mistake - I have (1A / 1 transistor) so I have bias = 2A because I've got 2 pairs of transistors. Yesterday i set bias 2,2A and sound is better. Today I try set 1,2A and I'll see. How much power in class A will I have with 48V and 2,4A bias?

Where should I wire ground with chassis?
I've got another PCb than Pikmouse. My PCB is from forum audiostereo.pl from Poland.
Of course ground from RCA is connected with ground on PCB. I also connected RCA ground with chassis. And you say it's wrong...How should I do this?