PMik said:Luke
Just use 0.2 Ohm value if you prefere to do that
I prefere 0.1 or 0.15 Ohm emitter resistors,
AuroraB
True class A for 50watt/8Ohm demands 3.5 Amps idle current
A more "human" approach is to go for 2.5 amps idle
Crunched some numbers:
Inputs
Number of output devices: 2 Pairs
Voltage rails (per rail): 45 volts
Emitter resistance (per device): 0.5 ohms
Bias voltage per Emmiter resistor: 450 mv
Idle bias per device 900.0 mA
Speaker ohms 8 ohms
Results
Idle bias per device: 0.900 Amps
Total Amplifier bias (per rail)** 1.80 Amps
Total Amplifier bias (both rails) 3.60 Amps
Total Dissipation (per channel) 162.0 Watts
Dissipation per device pair at idle 40.5 watts
Class-A output: Peak 103.7 Watts peak
Class-A output: RMS 51.8 Watts RMS
Efficiency 32.00 %
Max Class-AB RMS (1 kVA Toroid p/CH 5% reg) 115.16 Watts RMS
Or if you want to use 0.22 ohm ER:
Number of output devices: 2 Pairs
Voltage rails (per rail): 45 volts
Emitter resistance (per device): 0.22 ohms
Bias voltage per Emmiter resistor: 195 mv
Idle bias per device 886.4 mA
Speaker ohms 8 ohms
Results
Idle bias per device: 0.886 Amps
Total Amplifier bias (per rail)** 1.77 Amps
Total Amplifier bias (both rails) 3.55 Amps
Total Dissipation (per channel) 159.5 Watts
Dissipation per device pair at idle 39.9 watts
Class-A output: Peak 100.6 Watts peak
Class-A output: RMS 50.3 Watts RMS
Efficiency 31.52 %
Max Class-AB RMS (1 kVA Toroid p/CH 5% reg) 115.16 Watts RMS
Max Class-AB factor (100%=ideal load, Krell/565's = 91%, 555=70%) 91 %
K-amps said:
Paralelling 0.1 is 0.05 ohms not 0.4 ;-)
Remember, many people still think the KSA-50 is one of Krell's best sounding and one of Stereophile's top 10 power amps of all time. ;-)
I would stick with the stock ER's unless you want to play around.
Your arithmetic is correct. Read my post again (and the one I replied to) and you'll see that mine is too.
What sort of resistors are you getting, Luke?
Slap me if this is off-topic, but would it be sensible to reduce the supply voltage and increase the bias to suit lower impedance speakers, or just lower the voltage to reduce dissipation/cost?
Would the driver circuitry be happy with, for example, 30V rails? Or would it make sense arranging separate supplies for driver and output (45V and [whatever])?
Cloning this well-regarded and hugely expensive amp is worth doing, for sure, and I take K-amp's point that it's too lovely to change, but how down-rateable or tweakable do y'all think it could be?
Would the driver circuitry be happy with, for example, 30V rails? Or would it make sense arranging separate supplies for driver and output (45V and [whatever])?
Cloning this well-regarded and hugely expensive amp is worth doing, for sure, and I take K-amp's point that it's too lovely to change, but how down-rateable or tweakable do y'all think it could be?
bremen nacht said:
I would tweak it somewhat. Like I would use newer devices with more stable gain. (the Mj2119X are good choices) But I would not assume that my mods will make an improvement, if it does, Good if not I go backcken.
Also I would probably not reduce the rails too much, after all headroom is another thing Krells are known for. But to your point, reducing rail might make the sound a little less harsh maybe.
with 30v rails (I am assuming, fully loaded rails the are 30v not 30v unloaded) you will get a max of 50 watts with nada headroom. You will also need a huge Toroid. Just a few things to consider.
In the end it may depend on what parts you have scavenged.
30V supply is too low, - 50W into 8 ohms is 28,2V peak. 35 V is minimum in my opinion, but then of course idle dissipation increases as supply goes up.
K-amps... your calcs are OK, but the combined 1.8-2 A idle will flow through both halves... theoretical minimum idle will be 2,5A for 50W into 8 ohms in true class A, which makes my toroids just about good for a 50W amp, as the trafo is spec'ed at full load. Obviously, a larger trafo would make for a stiffer supply voltage, but at a cost,- price....
Maybe the frontand should be ragulated ?? any thoughts ??
ACD, - which design voltage was used for the driver section ??? Are the component values true to the original, except for the transistors ??
K-amps... your calcs are OK, but the combined 1.8-2 A idle will flow through both halves... theoretical minimum idle will be 2,5A for 50W into 8 ohms in true class A, which makes my toroids just about good for a 50W amp, as the trafo is spec'ed at full load. Obviously, a larger trafo would make for a stiffer supply voltage, but at a cost,- price....
Maybe the frontand should be ragulated ?? any thoughts ??
ACD, - which design voltage was used for the driver section ??? Are the component values true to the original, except for the transistors ??
AuroraB said:
I wasn't hoping to get 50W out of a reduced supply voltage, rather to get similar performance but at a lower power.
By the way, there are a couple of inconsistencies in the schematic/layout:
-on the layout Q108 is labelled Q106
-notes on schematic state that Q107 and 108 must be fitted with suitable heatsink. This should read 'Q109 and Q110'. What thermal rating is required?
AuroraB said:
Agreed, Also a class-A amp loaded rails at 30v are much stiffer than a class-AB PSU with the same 30v voltage. (This ofcourse assumes the PSU is designed according to dissipation requirements).
Also to get 50w from 30vdc, you will need a huge trafo... 800VA to 1kva otherwise you will not get 50 watts.
It's possible, but I'd rather have rails of 42-45v and have some headroom... don't like clipping too much here...
K-amps said:
Agreed, Also a class-A amp loaded rails at 30v are much stiffer than a class-AB PSU with the same 30v voltage. (This ofcourse assumes the PSU is designed according to dissipation requirements).
Also to get 50w from 30vdc, you will need a huge trafo... 800VA to 1kva otherwise you will not get 50 watts.
It's possible, but I'd rather have rails of 42-45v and have some headroom... don't like clipping too much here...
As I said, I'm not expecting 50W from 30V, just something smaller but similar sound. Call it a KSA-25 if you like. I also mentioned that I realised that the driver might have to stay at 45V, that I was prepared to try just reducing the volts to the output stage.
AuroraB implied that even in the original KSA-50 the drivers and output might not be the same rails - can anyone confirm what the driver stages require?
I don't understand 'Also to get 50w from 30vdc, you will need a huge trafo... 800VA to 1kva otherwise you will not get 50 watts.', not that it matters because I'm not trying to get 50W.
Apologies if it looks like I'm hi-jacking the thread (cloning the KSA-50); should I move it to a new topic called 'baby KSA' ?
btw...I am right in thinking that this is a class-A amplifier...?
.