♫♪--My little posh Tringlophone--♪♫ - Page 3 - 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 16th October 2012, 04:16 PM   #21
diyAudio Member
 
Join Date: Jun 2004
Location: Beograd
No, I didn't.
But I thought you'd be more detailed in your description. You're making us want to build it.And it works.
  Reply With Quote
Old 17th October 2012, 05:01 PM   #22
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Let us have a closer look at the Tringlotron OP.

Compared to the basic topology described in the article, there are two main differences:
  • The polarity is reversed (PNP composites). This poses no particular difficulty of comprehension.
  • The current source is not fixed, it is modulated by the signal: this creates the adaptative SE class A.

The operation is illustrated in the first pic below: the green trace is the 2Â output current, which is easily supplied using only 1.6A average quiescent current: red trace, Q6's collector current. When a current excursion is needed, the peak reaches 2.6Â.
As a side effect, this also means that Q8 has a lighter job to do: the yellow trace shows an excursion only half that of the green trace.
However, it is important to understand that Q8 is the sole output device: it is "helped" by the current modulation, but does not request nor need it: if the modulation is turned off, the operation of the circuit remains completely identical, except of course that clipping takes place if the peak output current exceeds 1.6A.

This means that the Tringlotron operation is not affected by the modulation, and the modulating path doesn't need to be clean or linear: the noble path is under the full control of the three transistors in the tringle.
The modulation simply improves the efficiency.
Even with this trick, the actual efficiency remains pretty low: in theory, it could reach 50% at full power, but in fact it is closer to 25% due to the various overheads and margins required for both voltage and current.
Each transistor needs a minimal Vce to operate properly, and there has to be some current margin to account for the real-life impedance of the speaker: unlike class B, class A requires to plan in advance the use of the current.

To ensure proper operation and minimize the losses, the voltages along the tringle have to be set very carefully, in order to make an optimal use of the supplies.
This is shown in the next two pics, first the quiescent levels for each collector and emitter, then the same with a near-maximum signal applied.

In the tringle, the "error" (non-linearity) caused by Q12/Q13 is corrected by an anti-error appearing at Q10/Q11.
No it is not a typo: it looks counter-intuitive, but that is the way the circuit works: upside down.
When Q11's base is pulled lower, its collector current decreases: the opposite of a regular follower; this creates a negative input resistance and explains the need for R12/C6.
This means that ideally, Q10/Q11 and Q12/Q13 should be matched. In fact, the behavior of the NPNs is almost completely hidden by the PNP inputs, and as a consequence only Q11 and Q13 need to be matched (for maximum linearity; it will work perfectly without matching)
Attached Images
File Type: png TringloA1.png (104.0 KB, 363 views)
File Type: png TringloA2.png (89.2 KB, 356 views)
File Type: png TringloA3.png (98.0 KB, 338 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 17th October 2012, 10:33 PM   #23
diyAudio Member
 
Esperado's Avatar
 
Join Date: Jul 2011
Location: France
By the hell, Elvee,
why are-you using the awful 2N3055, the worse power device ever for audio ?
Experience have taught-me that quality of any amp is proportional to his speed. Why do not use the fastest devices you can find ?
Yes i know, Error correction is not so sensible to open loop bandwidth than global FB (shorter loop), but i'm sure that better BJTs will bring-you less TIM, and less high order harmonic, both with a better phase response.
__________________
Ultimate Protection and more.The Only Source of Knowledge is experience, everything else is just information” ©A. Einstein
  Reply With Quote
Old 17th October 2012, 11:24 PM   #24
diyAudio Member
 
Join Date: Nov 2007
Location: Dallas
Craptanium rather than unobtanium prove magic of the design, not of the parts.
Worst case devices are obviously not a requirement forced upon every builder.

But if you are exploring error correction, why not test your deviousness against
something with an obvious error to fix? 2N3055 becomes the optimal choice.

Last edited by kenpeter; 17th October 2012 at 11:33 PM.
  Reply With Quote
Old 18th October 2012, 12:00 AM   #25
diyAudio Member
 
Esperado's Avatar
 
Join Date: Jul 2011
Location: France
Hi, kenpeter. I have tried error correction long years ago, in a very different configuration. (Symetrical output stage, differential input/output comparator). Since that time, i have a great respect to the error correction principle. So i don't need to be convinced any further about that.

My remark was not ironic against Elvee, on the contrary. If his goal was to demonstrate, you're right.
But why not trying the best as possible from the start, trying to remove in the same time those sad condensers in the signal paths, and make a high end audio amp ?

I ask myself a question too. Considering the power stage with an error correction as a near perfect power device, what happens if we build a closed loop amp with it, adding, for example, global Current Feedback ?
How does the two feedbacks can react together and dispatch their action ?
__________________
Ultimate Protection and more.The Only Source of Knowledge is experience, everything else is just information” ©A. Einstein
  Reply With Quote
Old 18th October 2012, 12:36 AM   #26
diyAudio Member
 
Esperado's Avatar
 
Join Date: Jul 2011
Location: France
Quote:
Originally Posted by Esperado View Post
I have tried error correction long years ago
*Thanks to Bob Cordell :-)
__________________
Ultimate Protection and more.The Only Source of Knowledge is experience, everything else is just information” ©A. Einstein
  Reply With Quote
Old 18th October 2012, 06:52 AM   #27
diyAudio Member
 
Join Date: Apr 2010
Location: Coffs Harbour
Quote:
Originally Posted by Esperado View Post
....But why not trying the best as possible from the start, trying to remove in the same time those sad condensers in the signal paths, and make a high end audio amp ?.....
There are several good reasons not to, IMHO. We have one reason already in proving the design, not the parts or extraneous circuitry. Mostly, designs seen in a developed state here are just adaptations of old ones with new parts and the latest popular topology - nothing is learned except some exercise in reading datasheets, judging the symmetry of simulation traces and just talk about components.

Newbs will always ask if they can have PCB patterns or use craptanium parts because that's all they can buy, so it is great encouragement to beginners and experienced guys alike to get to work and do the polishing up, etch pretty PCBs and use hi-speccium parts for themselves, if they wish. That's sharing an opportunity for real DIY and learning, I think.
__________________
regards

Last edited by Ian Finch; 18th October 2012 at 06:55 AM.
  Reply With Quote
Old 18th October 2012, 12:38 PM   #28
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by Esperado View Post
Hi, kenpeter. I have tried error correction long years ago, in a very different configuration. (Symetrical output stage, differential input/output comparator). Since that time, i have a great respect to the error correction principle. So i don't need to be convinced any further about that.
There are very few true error-correcting schemes around.

Most so-called "error-correctors" are in fact some kind of NFB in disguise. This includes Cordell and Hawksford. This is not to say these schemes are worthless or ineffective, they do improve the linearity, but there is an intellectual dishonesty in calling them so: in principle, true error correction is capable of cancelling error completely, whereas feedback can only reduce it by the loop gain.
Yet the proponents of these schemes claim the total cancellation is possible, which it isn't.
In general, it boils down to a local feedback loop, or some bridge configuration, or a "comparator" as in your example: that's where the loop gain is hiding.
This means that these "EC" schemes are subject to the constraints of closed loop systems, in particular stability. No loop gain = no correction.
Obscuring the NFB aspect is dangerous, because it hides the issues associated with it, and it leads to needlessly complicated schemes, where accurate balance between passives is required for instance, when it is not actually needed if the same system is implemented in a more rational way, with NFB explicitly shown and accounted for.
That said, local loops and hierarchical nesting of NFB does have many advantages, but it is something else.


Quote:
I ask myself a question too. Considering the power stage with an error correction as a near perfect power device, what happens if we build a closed loop amp with it, adding, for example, global Current Feedback ?
How does the two feedbacks can react together and dispatch their action ?
That is perfectly possible: an error-corrected stage is a very clean and well-behaved building block, and if you add plenty of loop gain to a natively ppm level OP, you can end up with a ppb level amplifier (even using lousy 2N3055s as OP)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 18th October 2012, 01:40 PM   #29
diyAudio Member
 
Esperado's Avatar
 
Join Date: Jul 2011
Location: France
Quote:
Originally Posted by Elvee View Post
There are very few true error-correcting schemes around.
Most so-called "error-correctors" are in fact some kind of NFB in disguise. This includes Cordell and Hawksford.
Yes, my first experiments (long time ago) where in the spirit of Cordell. And, some time later, near this one:
PM-AB1 error correction amplifier
And, yes, thinking to that, it is "some kind of NFB in disguise", you're damn right.
To be honest the schematic you posted is pretty complicated (means not familiar to me) and i will need some time to analyze and understand-it.
May-i suggest-you to write a little 'theory of operation' ?
__________________
Ultimate Protection and more.The Only Source of Knowledge is experience, everything else is just information” ©A. Einstein
  Reply With Quote
Old 18th October 2012, 05:07 PM   #30
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by Esperado View Post
May-i suggest-you to write a little 'theory of operation' ?
The operation of the circuit can be demonstrated by simply writing down the Kirchhoff's laws, but it probably does little for the actual understanding.

Let us try to do it in a pedagogic and intuitive way. There is no need to mess up with the complete schematic for the moment, the principles remain the same as for the simplified circuit:
Click the image to open in full size.

Let us do it in two passes: first, the progression of the AC signal without consideration for the error signals, and then the analysis of the error voltages.

Starting from Vin, the signal is copied from Q1's B to E, is applied to Q2 through D1 and drives the top side of the load. Since Q3's B and E are at a zero potential in AC, the voltage is applied across the load and generates a current into Q3's E which is transferred to its collector and from there passes through Q1 to loop back at the top of the load: we have looped the loop encompassing Q1, the load and Q3.

Now, let us look at the error voltages generated along this path. The error voltages are caused by the BE non-linearity.
Two identical voltages are generated in Q1 and Q3 (neglecting the Hfe, the current is the same everywhere in the loop). The error voltage in Q3 is "normal", ie the product of the emitter current by the non-linear dynamic emitter resistance is subtracted from the voltage applied to the load.
The error voltage in Q1 is identical, but it is applied "upside down" to the top of the load: Q2's base is connected to the emitter of Q1.
This creates a subtraction of the error voltages as seen from the load.

There is still another possible error voltage in the loop, the BE of Q2. But since the AC current in Q2 is zero, so is its error voltage and the error cancellation between Q1 and Q3 is perfect. Neglecting higher order effects, like finite Hfe and Early effect, of course.
Compensations could be included for these effects, like in multi-transistor CCS's, but it is probably simpler to use devices like MOSFets having a negligible input current.

As a side effect, because the voltage appearing across the load is exactly Vin, independent of the output current, this implies that the output impedance is zero.
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪

Last edited by Elvee; 18th October 2012 at 05:11 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 little cheap Circlophone© ♫♪ Elvee Solid State 1223 23rd February 2014 04:36 PM
Posh SW8 drivers Original Burnedfingers Subwoofers 3 1st May 2011 09:02 PM
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 07:23 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