2 or 3 Channel, Single Ended Class "A" Design Thread - Page 6

The active ground with its relatively high output impedance can weaken the bass response.

Quote:

Originally Posted by jcx 

a little quick with the "thread crap" card there regal

 

if you had clicked on the link to my post you would have seen a SE Class A Pass design - seems like another Pass amp and the idea of scaling them for headphones was introduced by the op - is he dragging the thread off topic now because the later one Avro_Arrow posted was actually push-pull instead of single ended?

 

I think the class a Pass design was already scaled down for headphones ,  check the headwize archives,  still has the big output cap,

You suggested dropping the mofset entirely,  I am interested in this thread because the design is global NFB free more than anything as I think most are.
 

Quote:

Originally Posted by regal 

The active ground with its relatively high output impedance can weaken the bass response.

The output impedance of the active ground channel is exactly the same as the signal channel in most amps. 

The output impedance is doubled for that, but 2*~0=~0 The last schematic in this thread has global feedback so output impedance should be very low ~0ohms.

In the case of stuff without global feedback, like a simple szekeres which has a non-zero output impedance, is the additional output impedance from a 3or4-channel circuit worse than the output caps in a 2-channel version? A 2-channel Sezekeres with DC coupled outputs puts the power supply caps in the signal current loop between ground and V+/-. A 3or4-channel version with the outputs all at +6.000V breaks the signal current loops through the PS caps AND dumps the output caps. 

.

I built the VE and the mids were good,  but the noise floor relatively high and there was a bass rolloff when RMAA'd.

I assumed it was from the mofset ground channel.   I far as adding global NFB I would really lose interest,  just building an opamp at that point IMO.

 I knew I could fix the noise floor with a good shunt supply but the with the bass rolloff I lost interest.  Why do you think there was a bass rolloff?

Quote:

Originally Posted by nikongod 

The output impedance of the active ground channel is exactly the same as the signal channel in most amps. 

The output impedance is doubled for that, but 2*~0=~0 The last schematic in this thread has global feedback so output impedance should be very low ~0ohms.

 

In the case of stuff without global feedback, like a simple szekeres which has a non-zero output impedance, is the additional output impedance from a 3or4-channel circuit worse than the output caps in a 2-channel version? A 2-channel Sezekeres with DC coupled outputs puts the power supply caps in the signal current loop between ground and V+/-. A 3or4-channel version with the outputs all at +6.000V breaks the signal current loops through the PS caps AND dumps the output caps. 

http://phonoclone.com/diy-sze.html

Try a  szekeres VE with lm317 CCS, but also substitute the IRF510 with lm317 

Quote:

Originally Posted by Avro_Arrow 

 

Edit: Sorry, the feedback was messed up...fixed

If you want to use a feedback it will be better to use a bjt instead of Т1

Something like this:

Quote:

Originally Posted by regal 

I built the VE and the mids were good,  but the noise floor relatively high and there was a bass rolloff when RMAA'd.

I assumed it was from the mofset ground channel.   I far as adding global NFB I would really lose interest,  just building an opamp at that point IMO.

 

 I knew I could fix the noise floor with a good shunt supply but the with the bass rolloff I lost interest.  Why do you think there was a bass rolloff?

 

My good guess:

In other amps (from the design thread for the RJM twilight on head-fi, and experience with the MELOS) I have seen very non-ideal behavior from mosfets/jfets as followers. My instinct is that input impedance is related somehow to frequency, and drops as frequency drops. This combined with the output impedance of the POT and Rgate could account for drop if they were high enough relative to the new lower actual input impedance of the FET.