Go Back   Home > Forums > Amplifiers > Chip Amps

Chip Amps Amplifiers based on integrated circuits

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 November 2012, 10:14 PM   #81
bigwill is offline bigwill  United Kingdom
diyAudio Member
 
Join Date: Dec 2004
Location: UK
It just occurred to me - under current swings the collector voltage on the driver transistors will be around 0.6v from the rail due to the base of the output transistor. If the VAS tries to pull the driver base higher than this, the driver obviously can't accommodate for this fact, can't turn on the output transistor any harder and becomes saturated.

The feedback diode seems to reduce this enough to stop the rail sticking, but doesn't cure the clipping spikes. An interesting problem though!
  Reply With Quote
Old 25th November 2012, 11:00 PM   #82
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
Try a schottkey diode if you haven't. At +V clipping there should not be a current spike through R5, this means Q5 is saturating. If Q5 saturates, negative feedback becomes positive and the amplifier latches until the input swings enough the other direction to reset the LTP.
  Reply With Quote
Old 26th November 2012, 09:26 AM   #83
bigwill is offline bigwill  United Kingdom
diyAudio Member
 
Join Date: Dec 2004
Location: UK
Quote:
Originally Posted by keantoken View Post
Try a schottkey diode if you haven't. At +V clipping there should not be a current spike through R5, this means Q5 is saturating. If Q5 saturates, negative feedback becomes positive and the amplifier latches until the input swings enough the other direction to reset the LTP.
Hi keantoken, I tried the diode first of all and it worked very well - I just happened to discover the other issues while probing around.

Regarding Q5 saturating - I'm not quite with you about this reversing the direction of the gain. Q11 can only pull 10s of milliamps through the base due to R5 - even at these high base currents, I don't see how the VAS voltage can fall. The current is going through the Vbe diode and Q5 is switched on as hard as it can. The output stage is already clipping at this point, so even if the gain was reversed somehow, we're already effectively in open loop conditions.

Last edited by bigwill; 26th November 2012 at 09:37 AM.
  Reply With Quote
Old 26th November 2012, 09:39 AM   #84
bigwill is offline bigwill  United Kingdom
diyAudio Member
 
Join Date: Dec 2004
Location: UK
I could see this happening with an emitter resistor in the VAS, but the only resistance is the inherent emitter resistance of the transistor, which is probably way too small to have this effect, and rise enough voltage to 'un-clip' the output stage to cause positive feedback
  Reply With Quote
Old 26th November 2012, 01:54 PM   #85
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
It does not matter how much gain there is as long as it is reversed. If it reverses, it's positive feedback period. It will cause the circuit to latch. Q11 may be limited by R5 but until that point readily saturates Q5.

What do the spikes look like? Those TIPs are super-slow and clunky, they are probably complaining very loudly when saturating. You might try probing R8 and R15.
  Reply With Quote
Old 26th November 2012, 02:52 PM   #86
bigwill is offline bigwill  United Kingdom
diyAudio Member
 
Join Date: Dec 2004
Location: UK
The issue actually appears to be the current mirror on the LTP - if I replaced it with the usual collector resistor, the sticking vanished, even with the VAS transistor still driven heavily into saturation.
  Reply With Quote
Old 26th November 2012, 02:59 PM   #87
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
That makes sense. Remember the VAS gain corner is probably in the 100s of Hz. This means the miller cap is effectively huge compared to the impedance at the VAS input. Therefore if this cap gets charged due to a glitch, it will take a long time before it can discharge and reset the circuit. Getting rid of the current mirror drastically lowers VAS input impedance.

This is one of the reasons I don't like large miller caps. One thing you can do if you think distortion is already good enough, is to degenerate the input stage in order to make the miller cap smaller. Furthermore, I believe you can also lower the miller cap if you trade the outputs for an MJE1503x pair which will be faster and cause less second-order loading of the frontend.

Last edited by keantoken; 26th November 2012 at 03:04 PM.
  Reply With Quote
Old 26th November 2012, 03:51 PM   #88
bigwill is offline bigwill  United Kingdom
diyAudio Member
 
Join Date: Dec 2004
Location: UK
The diode from VAS output to buffer input works great, but I'm going crazy trying to figure out WHY it works. Without it, the current mirror goes unbalanced during the sticking period, but with it, it feeds some current into the input side LTP. This brings the current mirror back into balance, somehow.
  Reply With Quote
Old 26th November 2012, 04:53 PM   #89
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
Interesting. Sometimes current mirrors are misapplied so that Q9 will saturate, but that is for the negative excursion and doesn't seem to apply here. Also, current mirrors in this position need transistors with good quasi-sat curves. BC5xx and BC3x7 are good. 2N5550 and 5400 are often used here but really shouldn't. The datasheets don't show Vcesat behavior for the 2N390x.

What is the Vbe mismatch between Q10 and Q9? It will suffice to probe between the emitters with a DMM while the amp is running with no input. Despite the 56R degen the CM will unbalance at very low currents, even if it is not saturating, as a result of any significant Vbe mismatch. But current should only get this low during clipping. How much does the current mirror unbalance?
  Reply With Quote
Old 26th November 2012, 05:18 PM   #90
bigwill is offline bigwill  United Kingdom
diyAudio Member
 
Join Date: Dec 2004
Location: UK
This has all been in simulation BTW.

Weirdly enough, it doesn't seem to be a drive issue. During the sticking period, the VAS driver abruptly stops feeding the base, until the voltage drops back within the rails, yet the VAS stays on for a period of about 4 microseconds. If I change the VAS transistor to a BC557, the rail sticking stops, even without a diode.

Last edited by bigwill; 26th November 2012 at 05:20 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
Class-AB meets Class-D: Yamaha's EEEngine Topology - where are this Diy Projects? tiefbassuebertr Solid State 29 14th April 2014 04:12 PM
Cannibalize a Class AB Stereo to Class A Mono block. owenhamburg Solid State 8 13th September 2010 08:55 PM
Configuring an AB class amp to work in pure A class jrme Tubes / Valves 5 23rd February 2010 06:31 PM
Can a Class AB PP amp be said to be operating in Class A at low signal levels? ray_moth Tubes / Valves 19 23rd January 2009 07:52 PM
Help! 50w Class A or 150w Class AB debate Toobstheway Solid State 12 17th September 2007 07:05 PM


New To Site? Need Help?

All times are GMT. The time now is 07:46 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