Hi,
a while ago I reported about my Aleph 4 DC offset details. Even with perfect matched input stage FETs an offset of some ten mV occurs. It is not a real problem, but I want to optimize anyhow… For Details see http://www.diyaudio.com/forums/showthread.php?threadid=10718 .
Now I found the time to analyze the cause - and the extremely easy solution. Something had to be wrong with the currents in the diff pair FET stage. I use 0,1% tolerance resistors – so by measuring the voltage over them the precise current values are given. The sourcing current shall be around 20mA, but it was 26mA. The reason: 9,45V at ZD9,1V and just 3,7V at FET S-G unction. So 221R got 5,75V equals 26mA.
The positive input FET current is forced by 392R, seeing at my Aleph just 4,3V (from IRF240 G-S junction plus B-E voltage of npn transi). That causes 11mA. So the other, negative input FET got the rest, 15mA. This different currents of 11 vs 16mA cause the offset increasing with unequal warming of the two FETs. My Aleph started with 15mV and drove up to 45mV – in spite of 3mV close matched FETs.
I prefer fighting causes not their reactions. So the easiest thing was to bring the source current down to 22mA. This equals 4,86V over 221R plus ca. 3,65V over the FET = 8,5V at the ZD. So I replaced the 9,1V ZD by a 8,2V ZD – and perfect! 22mA, two times 11mA and a DC offset stabile below 3mV. So If you feel to have trouble with DC offset maybe check a 8,2V Zener…
Tweaky regards
Klaus
a while ago I reported about my Aleph 4 DC offset details. Even with perfect matched input stage FETs an offset of some ten mV occurs. It is not a real problem, but I want to optimize anyhow… For Details see http://www.diyaudio.com/forums/showthread.php?threadid=10718 .
Now I found the time to analyze the cause - and the extremely easy solution. Something had to be wrong with the currents in the diff pair FET stage. I use 0,1% tolerance resistors – so by measuring the voltage over them the precise current values are given. The sourcing current shall be around 20mA, but it was 26mA. The reason: 9,45V at ZD9,1V and just 3,7V at FET S-G unction. So 221R got 5,75V equals 26mA.
The positive input FET current is forced by 392R, seeing at my Aleph just 4,3V (from IRF240 G-S junction plus B-E voltage of npn transi). That causes 11mA. So the other, negative input FET got the rest, 15mA. This different currents of 11 vs 16mA cause the offset increasing with unequal warming of the two FETs. My Aleph started with 15mV and drove up to 45mV – in spite of 3mV close matched FETs.
I prefer fighting causes not their reactions. So the easiest thing was to bring the source current down to 22mA. This equals 4,86V over 221R plus ca. 3,65V over the FET = 8,5V at the ZD. So I replaced the 9,1V ZD by a 8,2V ZD – and perfect! 22mA, two times 11mA and a DC offset stabile below 3mV. So If you feel to have trouble with DC offset maybe check a 8,2V Zener…
Tweaky regards
Klaus
Long time ago I solved such problem with applying “the right resistor in the right place” in the differential current divider resistors, also by non-polaring feedback capacitor. I used to add VR to have a flexible constant current but never thought it could be used to solve such problem (Thank’s Uli)
) in, so no way to trimm it. So I solve the ohter way around - both leads to the same perfect result.
Nickolas K. said:
Hi Nick,
you ment R11 or R14 - not R12.
In principal you are right. If e.g. the current through R14 would be to high, its value could be increased. Like I had to reduce the current through R11. Many ways.
But my interest was to adjust the currents w/o changing any R. So the Z-voltage had to be set right.
Since I have outed myself as a "resistor sound difference listener" I can also tell you that - checked in an A-B-direct comparison - I found that the mono block with FETs working really identical at 2 x 11mA and nearly no DC offset sounds the last % better. It sounds like "less distortion". Sweeter, the more neutral. Nice for one new cheap Z-diode.
Klaus
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.