Symmetrical input stage - Page 2 - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

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 20th April 2004, 09:25 AM   #11
diyAudio Member
 
traderbam's Avatar
 
Join Date: Jan 2002
Location: Earth
Hjelm,
Ok, I see now. You don't mention the values of any of the resistors or the supply voltages.

Idea 1: Obviously the non-linearity comes from the transistors - so one way to make this more linear would be to make the resistors as large as possible. Can you use quite high supply voltages and large resistor values to minimize the contributions of Vbe non-linearities?
  Reply With Quote
Old 20th April 2004, 09:41 AM   #12
hjelm is offline hjelm  Sweden
diyAudio Member
 
hjelm's Avatar
 
Join Date: Nov 2003
Location: Västerås
Traderbam:
Afraid supply is limited to +-10v.
I am trying to replace an opamp in a given application so unless i have to i would like to avoid doing another PSU.

Seems i have to go back and do some homework on Vbe unlinearities and cascodes.
__________________
Hjelm
  Reply With Quote
Old 21st April 2004, 11:44 AM   #13
diyAudio Member
 
traderbam's Avatar
 
Join Date: Jan 2002
Location: Earth
Yes, I agree that you need to look at how to mitigate Vbe/Ic non-linearity, or in other words linearize the transconductance. Using cascodes is another kettle of fish and I'd recommend treating these as a last resort or you'll make the system too complex too early. I don't know what output V swing you require but cascodes will only make things worse if you just need a 1vp-p or so.

Idea 2: how do you make a transistor circuit have very linear transconductance? One of the most proven ways is a long-tailed pair arrangement, fed with a current source and with a current mirror on the tails. Rather than trying to cancel distortions using a symmetrical differentiator, why not use a LTP which will give considerably more linear results? The output can either be taken directly off the tail via a capacitor or you can add a second LTP/current mirror stage for extra gain.
  Reply With Quote
Old 21st April 2004, 11:57 AM   #14
hjelm is offline hjelm  Sweden
diyAudio Member
 
hjelm's Avatar
 
Join Date: Nov 2003
Location: Västerås
Really appreciated!

I do not know what a long tailed pair looks like but i will soon find out.

I have given up on the Vbe unlinearities they are Ic related and i want Ic to change so it was doomed from the start.

Thanks for the tip on the long pair.
The cascodes does not seem to be too popular in the input stages either so i that was probably a dead end anyway, tried it and got some results but they seemed to be more problem than the actual benefit you received.

If i am to be a bit lazy do you have any example of a long tailed pair on the net or any example schematic i would greatly appreciate it, if not i fully understand (its my homework).

Thanks
__________________
Hjelm
  Reply With Quote
Old 21st April 2004, 12:02 PM   #15
Electrons are yellow and more is better!
diyAudio Member
 
peranders's Avatar
 
Join Date: Apr 2002
Location: Göteborg, Sweden
Blog Entries: 4
Hjelm, you have approx three options as emitter loads.

1 Plain resistor (as you have chosen) = good

2 Current source, BJT, JFET = better

3 Current source with cascode = best

A very stiff current source will reduce distortion. This goes also for your current mirrors.

Cascodes are the salvation You should try it! The only disadvantage is loss of output voltage.
__________________
/Per-Anders (my first name) or P-A as my friends call me
Tube Buffered Gainclone in work |Thread || Diamond buffer |Thread for the group buy | Wiki
  Reply With Quote
Old 21st April 2004, 12:19 PM   #16
hjelm is offline hjelm  Sweden
diyAudio Member
 
hjelm's Avatar
 
Join Date: Nov 2003
Location: Västerås
Just trying to get the right thinking here.
If i use current sources on the input i get a fixed Vbe.

With the Vbe fixed and the current IC fixed i have to use a high ohmic next stage?

This would mean i could make decent buffer but how do i acheive gain?

As for cascodes i see the applicability to linearize the current gain but what if i want to do voltage to current gain, or am i just stupid?
__________________
Hjelm
  Reply With Quote
Old 21st April 2004, 12:35 PM   #17
Electrons are yellow and more is better!
diyAudio Member
 
peranders's Avatar
 
Join Date: Apr 2002
Location: Göteborg, Sweden
Blog Entries: 4
OK, I see what you mean but is the main goal here to have zero feedback? What about a current feedback amp with gain of 2? Then you can involve a couple of highly linear circuits and ther is more: low current consumption (if you want his), extremely high speed (10-30 MHz!), extremly low distortion.
__________________
/Per-Anders (my first name) or P-A as my friends call me
Tube Buffered Gainclone in work |Thread || Diamond buffer |Thread for the group buy | Wiki
  Reply With Quote
Old 21st April 2004, 01:08 PM   #18
hjelm is offline hjelm  Sweden
diyAudio Member
 
hjelm's Avatar
 
Join Date: Nov 2003
Location: Västerås
I would like to rank the wishlist this way

1: +-10v supply
2: Few components
3: low distorsion (below 0.01)
4: DC coupled
5: class A

I have even considered a single ended version provided the distorsion is small enough.
__________________
Hjelm
  Reply With Quote
Old 22nd April 2004, 10:15 AM   #19
diyAudio Member
 
traderbam's Avatar
 
Join Date: Jan 2002
Location: Earth
Quote:
I have even considered a single ended version provided the distorsion is small enough
You are making an assumption that a circuit that is symmetrical about the power supplies will be more linear than an asymmetrical circuit. You are not alone in this...it is a very common belief. But is it correct?

Eg: some people do deceive themselves by thinking that something that looks prettier on paper will sound better. As if nature favours symmetry. However, IMO electrons don't give a wet slap!
  Reply With Quote
Old 22nd April 2004, 12:37 PM   #20
hjelm is offline hjelm  Sweden
diyAudio Member
 
hjelm's Avatar
 
Join Date: Nov 2003
Location: Västerås
Well the reasoning was that if you could create a toplology that had an unlinearity one way a way to cancel it would be to invert the topology and double it so the errors cancel eachother.
I am aware of the fact that npn and pnp devices are very hard to match linearitywise and that simulations might deceive you in presenting false symmetri, where in reality the devices don't match.

To be honest i am looking into a single ended topology right now but i am not understanding what i am doing so it's slow work. I get no gain in the circuit.
__________________
Hjelm
  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
Counterpoint - A Fully Symmetrical I/V stage Russ White Digital Source 22 26th October 2008 01:36 PM
Adding a third input to a differential input stage? maudio Solid State 15 11th October 2006 02:16 AM
current bias j-fet input 2sk170 ,2sj74 at input stage YUTK Solid State 11 2nd June 2005 03:34 AM
jFET input ,full symmetrical class A circuit promitheus Solid State 28 20th November 2003 06:12 AM


New To Site? Need Help?

All times are GMT. The time now is 06:13 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