Go Back   Home > Forums > Amplifiers > Headphone Systems

Headphone Systems Everything to do with Headphones

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 5th March 2012, 05:12 PM   #31
diyAudio Member
 
Join Date: Nov 2004
Location: Germany
You guys may want to try a CCS instead of an emitter resistor on the first transistor in the Darlington VAS (R14). I found this entirely eliminated severe common-mode distortion introduced with the Darlington. (A differential VAS that I also tried suffered from the same problem, presumably because it only used a tail resistor.)

Yeah, Darlington + cascode may be pushing it a little.

Oh, and regarding LTP collector resistors: I found that once the LTP was properly balanced, resistor value on the unconnected leg pretty much does not matter.

My simulated discrete opamps so far are about on the level of Samuel Groner's SGA-SOA-2, with a Darlington VAS. Not bad and quite well-behaved when driving loads, but matching a good integrated opamp in transfer linearity seems difficult so far. Really a lesson in modesty.
  Reply With Quote
Old 5th March 2012, 07:46 PM   #32
diyAudio Member
 
Join Date: May 2010
Spent a few hours looking for the problem and found that by increasing the emitter resistors of the input pair to reduce their transconductance allows for the use of a more conventionally sized comp capacitor of about 150pf with stability into all the loads I tried. Increasing them to 220r worked for me. I was pretty sure it wasn't a vas problem, because when I cut out the output stage by connecting the output of the vas to the feedback node it was perfectly stable previously. Anyway I think the original value of 20r was just too close to the theoretical minimum, even though self suggests that its mathematically correct he still uses 100 or 220 in his designs
  Reply With Quote
Old 5th March 2012, 09:47 PM   #33
diyAudio Member
 
Join Date: May 2010
I like the idea of loading the vas darlington with a ccs, but since the vas is already wrapped solidly in global or local fb, do you think you'd get a measurable decrease in total THD? Also it would push the negative peak swing of the vas another .65v away from the rail voltage.
  Reply With Quote
Old 5th March 2012, 11:35 PM   #34
diyAudio Member
 
Join Date: May 2010
Heh...total THD... You know what I mean...
  Reply With Quote
Old 6th March 2012, 03:22 PM   #35
diyAudio Member
 
Join Date: May 2010
finally bothered to do the math, (which I really should have done earlier) to maintain a "safe" value of 30dB of NFB at 20kHz with the NFB network setting the closed loop gain at ~7.5 as it is and emitter resistors of 20Ohms giving a large gm for the input pair, Cdom would have needed to be just over 1000pF. Decreasing the input stage gm by increasing Re to 270Ohms allows a Cdom of 120pF to maintain this criteria. This was very close to the value for Cdom I found to be necessary for stability in practice, and seems a better solution to me to avoid potential slew limiting from a large Cdom.
  Reply With Quote
Old 6th March 2012, 07:52 PM   #36
diyAudio Member
 
Join Date: Nov 2004
Location: Germany
Quote:
Originally Posted by mnturner View Post
I like the idea of loading the vas darlington with a ccs, but since the vas is already wrapped solidly in global or local fb, do you think you'd get a measurable decrease in total THD? Also it would push the negative peak swing of the vas another .65v away from the rail voltage.
All I can say is that it did make a difference here when subjecting my opamp model to Samuel Groner's common-mode linearity and transfer linearity tests. With a 2k2 resistor, common-mode linearity was almost 20 dB worse (-43 dB vs. -62 dB, 10 kHz, ~+20 dBu) and in fact worse than with a simple VAS, which evened out when using an - admittedly ideal - current source. (It was a pretty standard pnp input LTP-VAS-EF topology otherwise.)

I don't think VAS voltage swing is affected by using a CCS load on the first transistor, as long as it is a simple CCS that is content with Vce,sat. It only ever has to go as low as Vbe of the 2nd transistor anyway.

One more funny effect I saw was that well-meaning filtering of supplies within the amp actually made distortion worse when it was faced with non-negligible supply impedance. Possibly VAS and output transistor are forming sort of a CFP.

I also simulated the first complementary (though bipolar) LTP input headphone amp of my own today, similar to the Rotel power amps of yore (I used a very richly biased EF2 output instead of the triple EF employed for driving loudspeakers, plus current mirrors on the LTP in spite of the doubts re: VAS current - a VAS emitter resistors seems to sort that out reasonably well). Getting the affair stable with a Darlington VAS proved a bit tricky, but even with a standard VAS (which was good enough for extremely low distortion at hundreds of watts), I must say performance is excellent as long as there's enough supply headroom - I went for +/- 22 V.
We're talking <0.001% at 10 kHz@20Vpp into 600 ohms and still very little distortion into lower-impedance loads, at a relatively modest 30 mA of output transistor bias. When unity-gain compensated and run at a gain of 2, 10 Vpp into 600 ohms is claimed to give 0.000034% THD, though I bet the unmodeled Early nonlinearity would mar this figure. Low enough, I say.

Seems to be a pretty efficient topology that's not too hard to get going, no wonder it's popular in higher-powered speaker amps. My folded cascode efforts so far, while finally working by now at least, aren't nearly as convincing at similar parts count (though I know opamps like the AD797 exist, so it must be possible to get first-rate performance). This complementary thing is the first concept I've tried that can at least hope to match first-rate opamps.

Last edited by sgrossklass; 6th March 2012 at 07:59 PM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
my Class-A MOSFET headphone amp davygrvy Headphone Systems 28 18th October 2011 06:42 AM
All Discrete Class-A Headphone Amp mnturner Headphone Systems 33 6th June 2011 01:39 PM
mos/bjt buffer in class A for headphone amp sunrise Headphone Systems 7 20th October 2008 06:34 PM
Lateral mosfet w Normal or Diamond buffer? fab Solid State 22 8th June 2007 04:23 PM
darlington/mosfet in emittor follower / headphone amp svokke Headphone Systems 11 8th June 2005 06:19 PM


New To Site? Need Help?

All times are GMT. The time now is 05:07 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2