♫♪ My little cheap Circlophone© ♫♪ - Page 22 - 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 31st August 2011, 02:38 PM   #211
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by padamiecki View Post
Hi Elvee
was it a trick to fool the simulator
or will work in practice?
It is definitely not a trick played with the simulator.


Now, will it work in practice?
Yes, but your mileage may vary.

The crossquad is a pretty temperamental circuit, relying on a precisely controlled amount of positive feedback to function, and that feature plus the parasitics of real world circuits make an interesting mix...

If you build the circuit exactly as on the schematic, it will probably oscillate like mad.

If you try the usual recipes to tame the oscillations, like base stoppers and similar, not only will they not cure the problems, they will make them much worse: at the bases of the input transistors, a negative resistance of Re-e/Hfe is created.
If that negative resistance is parallelled with a positive one (the source impedance) larger than itself, the resulting equivalent resistance is negative too, with results you can imagine.

By adding base stoppers, you increase the source resistance, making the problem worse.
And because the Hfe drops at high frequencies, input resistances much smaller than expected can cause problems.
One solution is to bypass the base with a small value cap, to reduce the apparent impedance at high frequencies.
That's just an example. There are a number of other mechanisms leading to oscillations.

The compensations of the amplifier itself need to be adjusted, to take into account the effects of the Xquad in the overall loop.

It is not a simple task, but it is doable. I have made functionning designs based on a crossquad (not in the Circlophone though).

Some other pecularities have to be remembered: the phase is reversed compared to a standard LTP, but when it saturates, it degenerates and reverts to the normal phase.
And it only takes several hundreds mV to saturate, however large is the tail current and/or the degeneration resistor.

This leads to "interesting" behaviors...
Which is why there are back to back schottkys at the input.

If you surmount all those problems, the actual performance you can obtain will be strongly dependent on the matching of the transistors, but with low power devices, a good match is not difficult to achieve.
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 31st August 2011, 04:55 PM   #212
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by Elvee View Post
If you try the usual recipes to tame the oscillations, like base stoppers and similar, not only will they not cure the problems, they will make them much worse: at the bases of the input transistors, a negative resistance of Re-e/Hfe is created.
Small typo there, I should have written "a negative resistance of Hfe*Re-e", Re-e being the gain-setting inter-emitter resistor.

There is a variant of the cross-quad, created by Eva (I think so, I have never seen it elsewhere, and it is typical of the circuits she is capable of creating); it is better behaved than the 4 transistor version, but it doesn't deliver the raw performances of the naked Xquad.
I named it the EvaQuad, but EvaHex might be more appropriate:
Reinventing the N-channel wheel?

A working example of a crossquad can be found in the voltage gain stage of the Tringlinator:
The Tringlinator: a MOS-based Tringlotron amplifier
But there, it is used with a well-defined gain of 7 rather than as a pure high gain block, making it rather less finicky.

Note that the Tringlotron, the power stage of the Tringlinator is also a kind of vertically-stretched crossquad.
All this explains the remarkable performances of the complete Tringlinator, despite the absence of GNFB.
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 1st September 2011, 12:33 AM   #213
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
So, you saying we just need DIACs to stopper the bases, right?
  Reply With Quote
Old 1st September 2011, 01:39 AM   #214
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
I been playin w. evquad (plus active tail) on sim.

This sems to me no different than a quad with a
current mirror output instead of (or in parallel with)
the standard outputs... Why is this one said to be
any more stable? Sims OK, but all the same worries.
  Reply With Quote
Old 1st September 2011, 08:17 AM   #215
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by kenpeter View Post
This sems to me no different than a quad with a
current mirror output instead of (or in parallel with)
the standard outputs... Why is this one said to be
any more stable? Sims OK, but all the same worries.
Rule N°1: never misunderestimate Eva, she is smarter than you, me, and a good part of the forum's members put together.

Look at the first pic: with an input of 700mV^, a saturation occurs, but the EvaQuad shows no phase reversal, unlike the standard circuit.

In the standard quad, Q1's collector waveform is in phase with the base voltage: this means that any feedback caused by a combination of input resistance and Ccb capacitance will be positive, increasing instabilities.
In the Evaquad, Q5's collector has 0V AC potential, and this couldn't happen.

Moreover, should Q6 be used as an input, voltages at out2 and tp1 will be antiphase, meaning that Q9's capacitance will play its normal role of Miller and contribute to stabilize the thing.


Rule N°2: do not misunderestimate me too much either.
The second pic compares the linearity of the two versions: the raw Xquad is superior, because the cancellation effects are more accurate.
Attached Images
File Type: gif XquadVsEva1.GIF (56.2 KB, 456 views)
File Type: gif XquadVsEva2.GIF (57.8 KB, 457 views)
Attached Files
File Type: txt XquadVsEva.asc.txt (3.8 KB, 24 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 1st September 2011, 09:09 AM   #216
diyAudio Member
 
Join Date: Mar 2004
Location: Krakow
Quote:
Originally Posted by Elvee View Post
Rule N°2: do not misunderestimate me too much either....

do you have the idea hot to prevent this circuit from the saturation?
I attached a linearizion idea circuit which is a bit similar to Xquad,
o found this somewhere on the net...
Attached Images
File Type: jpg Obraz7.jpg (49.6 KB, 438 views)
__________________
regards, Pawel
  Reply With Quote
Old 1st September 2011, 11:35 AM   #217
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
I never saw any saturation with the plain quad.
Then again, I have that active tail thing goin on...
Keeps the quesscent roughly equal all times.
.
Raw xquad is working better in this instance than
Evaquad or Evaquad wth driver collectors parrallel.

I'm also driving an emitter drop to the rail, which
isn't a huge voltage swing on the collector.

Last edited by kenpeter; 1st September 2011 at 11:46 AM.
  Reply With Quote
Old 1st September 2011, 12:13 PM   #218
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
Where is the problem, cause I'm just not seeing it.
No transistor of this quad loses collector voltage,
even in moderate clipping, its all good...

The two waves in the center (blue maybe little hard
to see) are actually all four voltages of the xquad.
The left and right closely overlap, both sides look
as-if transistors bias like diodes. Differential currents
vary, but differential voltages are held nearly equal.
Its definitely doing something though, distortion is
unmistakably reduced by presence of the xquad.

No collectors here ever dip below the base voltage.
I have no idea what is this "phase reversal" you say
when the collector saturates. Do you simply mean the
quad would stop reversing things working as a quad?
Saturated behavior would revert to normal LTP ???

I don't think this tail allows the latched condition to
exist. Even if you hit it with a transient huge enough
to make one transistor saturate, and phase revert to
uncrossed LTP. It would quickly be put back to right.
Attached Images
File Type: gif nosat.gif (31.2 KB, 425 views)

Last edited by kenpeter; 1st September 2011 at 12:33 PM.
  Reply With Quote
Old 1st September 2011, 12:34 PM   #219
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by padamiecki View Post

do you have the idea hot to prevent this circuit from the saturation?
If it is included in an amplifier having global NF, this will be implicitly realized.
At least under normal working conditions.
But to prevent a latch-up during some freak event, there has to be some form of safety net, it would be suicidal to leave the circuit do whatever it wants, that's the reason for the diode clipper at the input.

But when a larger dynamic is required, as in the Tringlinator, a solution is to increase the Vbe of the lower transistors, by means of resistors or zeners f.e.
If you look attentively at the schematic, you will see that that's what is done there.
Incidently, it leaves more Vce space to breathe for the transistors.

Quote:
I attached a linearizion idea circuit which is a bit similar to Xquad,
o found this somewhere on the net...
It is the poor relation of the Xquad then.

It simply mimics the effect of a LTP, but with a diode to replace the other transistor.
The compensating effect is already crude in a LTP, it is equivalent to assume that exp(-k*V)=-exp(k*V), but with two heavily dissimilar components, it will be worse.

The only advantage that I can see in this circuit is the fact that the bias current is recycled, thus halving the current consumption compared to a LTP.
But it costs two big e-lytics, and a similar trick could be used with transistors.
Not of a great practical value.
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 1st September 2011, 12:50 PM   #220
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by kenpeter View Post
Where is the problem, cause I'm just not seeing it.
No transistor of this quad loses collector voltage,
even in moderate clipping, its all good...
No that's what I explained in my previous post, the NFB keeps the input voltage <0.6V.

That's in sim. With a real circuit, you have 99% chances the circuit locks up as soon as it is powered.
I know, I've been there.
In the sim try to put an .ic V(base of Q2)=2v, I think it will latch up.

Quote:
No collectors here ever dip below the base voltage.
I have no idea what is this "phase reversal" you say
when the collector saturates. Do you simply mean the
quad would stop reversing things working as a quad?
Saturated behavior would revert to normal LTP ???
Yes, exactly.
Quote:
I don't think this tail allows the latched condition to
exist. Even if you hit it with a transient huge enough
to make one transistor saturate, and phase revert to
uncrossed LTP. It would quickly be put back to right.
Dunno, try the .ic experiment to find out.
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  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
Need help with OB design - cheap, cheap, cheap djn Multi-Way 38 30th May 2010 06:26 AM
Scosche SPL meter: Cheap find with potential, or cheap junk? theAnonymous1 Everything Else 5 11th October 2006 03:40 AM
Alpine cda-9831 vs. 9833 vs. 9835 cbecker33 Car Audio 3 10th December 2004 12:47 PM


New To Site? Need Help?

All times are GMT. The time now is 03:40 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