Discrete, Class A Lateral MosFET Headphone amp (not another follower buffer!) - Page 3 - diyAudio
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 25th February 2012, 02:05 PM   #21
diyAudio Member
 
Join Date: Oct 2004
Location: Denmark
Quote:
Originally Posted by sgrossklass View Post

BTW, are you sure that symmetric drain resistors in the input stage give best performance? Normally I'd think that R4/5 < R15/18 would result in better symmetry.
Maybe they do, maybe not. To be honest I have not really looked into using different drain resistor values for R4/5 and R15/18.

We are already talking about a THD-20 that is less then 0.001% at 5.6V p-p in simulation.

Fiddling around with the resistors might a help a little but I personally feel that at these already insanely low THD-20 levels not much is to be gained.
  Reply With Quote
Old 27th February 2012, 09:08 PM   #22
diyAudio Member
 
Join Date: Jul 2011
Just came across this thread. The OP's design is very similar to what I am considering for a headphone amp. I tire of the op-amp based designs that seem to be the norm in headphone amps.
I am working on a very similar design, but I am including (at least the option for having) the driven input cascode on the LTPs. This is mainly because the gain is apt to be quite low. I haven't settled on the output drivers yet, other than they will certainly be biased in class A.
The big difference in what I am looking at is the power rails. I am thinking of about +/- 35 volts! I know many people are now calling me a complete whacko, but after a quick web survey, I see headphones range from about 4 Ohms (rarely) to as much as 600 Ohms (fairly often), with maximum power ratings of up to 1 Watt.
1W at 600 Ohms requires +/- 35 volt swing.
1W at 4 Ohms requires 710mA peak current, (355mA bias for class A).
So an amp that is capable of driving all phones to their maximum rating should be capable of +/- 35 V swings and have a bias of 355mA!!!!
I would be inclined to compromise on the low impedance side, since it looks like 4 and 8 Ohm headphones have gone the way of the dinosaur, but a higher bias may improve the sound even where the current isn't needed.
My present headphones are AKG701s, so this thing could be downright dangerous connected to them, so some sort of gain switch or maybbe even an "impedance switch" that could set the maximum voltage swing might be a requirement.
Does anyone know of a fairly comprehensive survey of headphone impedance and power requirements? The highest voltage requirement I have come across is the Beyerdynamic DT 100, which is available in 400 Ohms impedance with 1000mW power handling capability. This would require 57 Vp-p signal swing (+/-28V rails)! They also list the SPL at 94dB @ 1mW. So 1000mW would be 124 dB SPL, which would cause nearly instant hearing damage, so maybe no need to be quite that nuts.
Are there any 600 Ohm phones that are rated to take 1W?

Terry
  Reply With Quote
Old 27th February 2012, 09:31 PM   #23
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
there are 50 Ohm headphones that "need" 6 Wrms, ~= +/- 25 V, 500 mA

but that’s with pricey Orthodynamics and to reach 120 dB SPL peaks

its unreasonable to try to cover the 4 orders of magnitude that all categories of dynamic headphones sensitivity range over with one amp

since this is DIY you should be happy for the opportunity to build several amps customized to a small range of headphone sensitivity, Z
  Reply With Quote
Old 28th February 2012, 06:47 AM   #24
diyAudio Member
 
Join Date: May 2010
TerrySt,
I decided to drop the gain to about 5.5x and add the cascode setup to the input as well. Spice actually gave a slightly decreased THD20 with the cascode even though it apparently does not model the effect which causes the increased HF distortion at low gain.
The only downside to this that I can see is the increased noise from the larger feedback resistors; to decrease them you'd need to drop the input resistor as well to avoid upsetting DC balance and increase your input cap which isn't so great.
I'll probably settle on including the cascode and a gain of maybe 7-8. I'm going to use an opamp preamp so I could potentially decrease the amps feedback and input resistors (and therefor input impedance) to decrease noise, but I think I'll gauge the noise performance by ear when I'm building and hopefully I won't need to.
Attached Images
File Type: jpg Mosfet Headamp2.jpg (65.5 KB, 311 views)
  Reply With Quote
Old 28th February 2012, 03:33 PM   #25
diyAudio Member
 
Join Date: Jul 2011
I'm thinking of including the driven cascode in the layout, but of course if bench measurements or listening tests show it doesn't help or causes more harm than good, it can always just be removed. Simulations do show that there is a compromise between distortion and the transient response with respect to the value of the capacitor that is used to couple the drive signal to the bases of the cascodes.
I haven't started researching the noise yet. I suspect that noise could be a major issue with a headphone amp, so I'm not ignoring it, just haven't gotten that far yet.
I also have included a cascode for the VAS collector, although again, it would be easy to drop this if it doesn't really help.
Right now my design is basically from the Cordell book, but with the addition of the driven cascode on the input LTP. I haven't decided on the output stage yet, although it will certainly be class A. I may end up with something very much like your approach.
As mentioned a couple of times earlier, a gain switch is probably a good idea unless you want to target a certain headphone. Also, if I go with the plan to have high enough rails to drive 600 Ohm phones to high levels, some sort of "impedance" switch that limits the output voltage swing to a reasonable level based on the headphone impedance might be wise.

Terry
  Reply With Quote
Old 29th February 2012, 02:08 PM   #26
diyAudio Member
 
Join Date: May 2010
PCBs etched, just need to drill and populate them...
Attached Images
File Type: jpg headamp PCBs.jpg (919.1 KB, 253 views)
  Reply With Quote
Old 29th February 2012, 02:37 PM   #27
diyAudio Member
 
Join Date: Jul 2011
Nice. I'm very interested in how they sound. Especially the noise level and if you find any hum problems from power supply routing.
  Reply With Quote
Old 1st March 2012, 09:00 AM   #28
diyAudio Member
 
Join Date: May 2010
Got one populated, but i screwed up when I printed the transfer for the PCB's and didn't invert it correctly in one axis. Everything else was fine, just had to mount all the TO92 stuff 'backwards', but the mosfets pin out doesn't allow that. Might have to put the on the bottom of the board.
Yeah, it'll be interesting; the power supply routing is definitely sub-optimal, I pretty much went around the outside because I got impatient. Hopefully the current levels will be low enough that no hum gets into the signal...
Attached Images
File Type: jpg headamp.jpg (660.6 KB, 206 views)
  Reply With Quote
Old 5th March 2012, 03:11 AM   #29
diyAudio Member
 
Join Date: May 2010
Got it running, everything is as expected, DC offset is reliably 3mV without a DC servo. The issue that i came across was that oscillation was occurring unless I scrapped the twin pole comp and significantly increased the value of the comp capacitor. The oscillation was at about 1MHz and was only about 2vpk-pk with +/-15v supplies, so it was strangely controlled and reliable. I don't like the idea of having to roll off gain so early; i guess this could be due to the low net gain/ high feedback situation or poor pcb layout or a combination of both, any suggestions?
  Reply With Quote
Old 5th March 2012, 02:56 PM   #30
diyAudio Member
 
Join Date: Jul 2011
I'm still just in 'simulation mode', but I'm also fighting stability issues. Mainly because I added a cascode on the VAS collector also. The darlington VAS plus the cascode makes a fairly long local feedback loop. I have been able to stabilize it by adding a cap from the collector to the base of the VAS transistor (Q9 in your schematic above). I'm not sure I'm OK with that approach. I will continue to study it. You might try it and see if it lets you decrease the feedback around the VAS or maybe even get back to the Twin pole comp.
Also, I find that C15 value is critical for transient resonse (at least in simulations). Too large a value cause some bad overshoot on fast rise time transients. Too small reduces the distortion benefit.
Maybe some local bypassing near the amp would help?

You could try increasing the global feedback to see if that helps, but it sounds like it may be a local issue (around the VAS). Increasing the global gain will show if it is a global loop issue or a local one.

Terry
  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 03:30 PM.


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