Hi Bob,
The Aleph 30 bias is quite a bit lower and that makes all the difference. Look at the older designs and you will see numbers more in line with my thinking. For instance, the Aleph 5 uses 221 and 75. I believe NP has stated that as low as .4V at the base will cause distortion, and I can attest to it. Even the F5 uses 1K and 150, it will go into AB though but the concept is identical.
The Aleph 30 bias is quite a bit lower and that makes all the difference. Look at the older designs and you will see numbers more in line with my thinking. For instance, the Aleph 5 uses 221 and 75. I believe NP has stated that as low as .4V at the base will cause distortion, and I can attest to it. Even the F5 uses 1K and 150, it will go into AB though but the concept is identical.
Last edited:
Are we talking about original Aleph J schematic?
With 1.3A bias per FET at idle, the drop on R18 (0.47 Ohm) will be 0.61V. R13 and R14 are a voltage divider which will give 0.4V drop on R14. Am I wrong?
With 1.3A bias per FET at idle, the drop on R18 (0.47 Ohm) will be 0.61V. R13 and R14 are a voltage divider which will give 0.4V drop on R14. Am I wrong?
Right, Frags, but when the output swings negative, more current flows in the outputs, increasing the source resistor voltage and turning on the ZTX450, clipping the drive signal.
Yes, Bob, but R24/C3 feedback reduces the current flowing from the current source (Q5/6) at negative swing.
Q3 is not only current protection but soft clipping as well.
Q3 is not only current protection but soft clipping as well.
Bob, you made the point before that only 1ma will flow at .6V, but that is already 25% of the available current from the front end shunted to the minus rail. Even at .4V I can see there would be enough current shunted to cause a fair bit of distortion, and it was easily heard. I couldn't put my finger on the problem when I first started using the new amps, but I knew something was wrong.
Bill, let consider it this way:
The maximum output current for Alpeh J is 2.5A according to the service manual, so the maximum current via Q7 and Q8 will be 2.5A plus half of Aleph current source = 3.75A or
1.875A per FET. It will give 0.56V drop on emitter-base - this is the point where the clipping starts. May be something wrong with some Q3s?
The maximum output current for Alpeh J is 2.5A according to the service manual, so the maximum current via Q7 and Q8 will be 2.5A plus half of Aleph current source = 3.75A or
1.875A per FET. It will give 0.56V drop on emitter-base - this is the point where the clipping starts. May be something wrong with some Q3s?
Maybe I'll learn something here. How can feedback from the output to the current source affect the voltage Q3 sees at it's base? The output fets don't care where the current is going. Or coming from, depending on your outlook.
Last edited:
I have no argument with your math, but your statement that Q3 does not conduct until the base sees .6V suggests that it is like a light switch. .6V or .63V or whatever is the saturation voltage, unless I'm seriously in error. It will conduct at lower voltages.
Frags, I don't see it. The Aleph current source adds current on the positive side of the output signal.
When the output signal goes negative, The Aleph source sends it's static value and the output devices see that current plus the current in the speaker.
When the output signal goes negative, The Aleph source sends it's static value and the output devices see that current plus the current in the speaker.
The feedback affects it indirectly. The current through the load is the sum of current via Q7/8 and the current source (Q5/6). W/o Aleph feedback (R24/C3) the portion from the current source would be always constant. But the feedback increase current source current at positive swing and decrease it at negative one decreasing the total current via Q7/8 as well. This leads us to a reduction of voltage drop at R18 and emitter base of Q3.
No, Aleph source reduces its current at negative swing because of R24/C3 feedback. Otherwise the power consumption will increase when signal level increases.
Frags is right about that, the feedback modulates the signal at Q4, raising and lowering the current, but it still has no direct effect on Q3. The output fets will pull current from the current source or from ground through the speaker or a combination of both, they don't care where it's coming from.
Yes, they don't. But w/o Aleph feedback the voltage at Q3 would be higher for the same output level.
Please excuse my bluntness, but that is irrelevant. I was merely reporting the obvious distortion I was getting because the current limiter voltage divider was incorrect.
This IS a lively in depth discussion of the inner workings of an Aleph amp though.
This IS a lively in depth discussion of the inner workings of an Aleph amp though.
Yes, but it would no longer be an Aleph, just SE class A, and it would clip when the fets shut off.
OK. I've just measured Vbe and Ic parameters of one ZTX450 from my box:
0.40V - 1uA 0.45V - 5uA
0.475V - 13uA 0.5V - 34uA
0.525V - 87uA 0.55V - 230uA
0.575V - 600uA 0.6V - 1.5mA
So, It may have some noticeable effect when Vbe reaches 0.55V. This is possible when the current via both Q7 and Q8 reaches about 3.7A. This will give us about 2.45A output current. So I may say that Q3 has no effect up to the maximum specified current 2.5A for Aleph J.
0.40V - 1uA 0.45V - 5uA
0.475V - 13uA 0.5V - 34uA
0.525V - 87uA 0.55V - 230uA
0.575V - 600uA 0.6V - 1.5mA
So, It may have some noticeable effect when Vbe reaches 0.55V. This is possible when the current via both Q7 and Q8 reaches about 3.7A. This will give us about 2.45A output current. So I may say that Q3 has no effect up to the maximum specified current 2.5A for Aleph J.
Last edited:
I didn't bring up the fact that I use MPSA18s, did I. Regardless, why take the chance, use bigger resistors.
It was a subjective observation, substantiated by the obvious improvement when I replaced the resistors. Night and day. I don't have a distortion analyzer, but I sure wish I did.
It was a subjective observation, substantiated by the obvious improvement when I replaced the resistors. Night and day. I don't have a distortion analyzer, but I sure wish I did.
Here we go. The saturation Vbe for ZTX450 is 1.1V but 0.7V for MPSA18.
I'm not surprised you got some noticeable distortion with that one.
There are other factors too but the simple fact is people use many different transistors in these amps and there is no reason to set the current limiter that close.
I did find NPs statement about ZTXs causing distortion even at .4 Vbe, it is in the F5 article.
I did find NPs statement about ZTXs causing distortion even at .4 Vbe, it is in the F5 article.
Just reading back.
I don't agree that the current through the signal fets decreases with the decrease in Aleph current, they just pull more current through the speakers.