A dive into the past - 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 2nd March 2014, 03:25 PM   #1
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Default A dive into the past

I have begun sorting my "hardware archives": there was no room left in the dedicated cupboard, and this meant it was time to part with my less valuable prototypes (or the most useless ones, depending on the point of view).

Basically, each time I have some silly idea, I build a quick and dirty prototype, made of junk material I happen to have handy.
Sometimes it works satisfactorily, and I dump it into the "archives", and sometimes it doesn't, and it follows the same way, pending further investigations.
The result is identical however: they all end up undisturbed, layer after layer.
The first pic shows some the "good" (functional or partly functional) prototypes I managed to sort so far.

I have completely forgotten about most of these circuit (although I also have a corresponding heap of paper documentation too), but I picked up a nice little gem I remember of: it is essentially an opamp (TLO82) driving a pair of darlington transistors in the crudest way, without bias or anything.

Yet, it manages an acceptable quality, as the oscillogram shows (7.5V@1KHz pp into 5Ω).

What trick does it use?
It uses HF bias, just like magnetic recordings: the other half of the opamp is configured as a ~150KHz oscillator, and it superimposes a few volts to the PA.
An inductor and a series LC circuit block most of the HF in the output, but this is sufficient to overcome the ~2.4V of the unbiased darlingtons, eliminating practically all of the horrific Xover distortion.
A small residue at the dithering frequency is still visible.
It is a kind of linear class D amplifier......
Attached Images
File Type: jpg Dump1.jpg (577.4 KB, 455 views)
File Type: jpg Dump2.jpg (276.9 KB, 448 views)
File Type: jpg Dump3.jpg (425.3 KB, 435 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 2nd March 2014, 04:57 PM   #2
diyAudio Member
 
Join Date: May 2011
Location: Silicon Valley
I guess you could also build a bias spreader to set the NPN Darlington's Vbase = +1.9VBE and to set the PNP Darlington's Vbase = -1.9VBE, then AC couple the bias oscillator(s) waveform into these two bases. You would still have zero bias current in the output stage.

You could arrange for the AC waveforms at the two bases to be perfectly in-phase, or perfectly out-of-phase (180 degree shift), or perhaps somewhere in between (90 degrees?).
  Reply With Quote
Old 2nd March 2014, 10:11 PM   #3
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
I believe Self has some comments on the HF bias in his new edition. One of many clever tricks that do seem to work, but never caught on as other issues arise.
  Reply With Quote
Old 3rd March 2014, 12:54 PM   #4
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by Mark Johnson View Post
I guess you could also build a bias spreader to set the NPN Darlington's Vbase = +1.9VBE and to set the PNP Darlington's Vbase = -1.9VBE, then AC couple the bias oscillator(s) waveform into these two bases. You would still have zero bias current in the output stage.
Yes, if the initial dead zone is reduced it will improve matters undoubtedly, but the main attraction of such a scheme is its simplicity and straightforwardness combined with acceptable performances: straight connection between bases, between emitters, and direct drive by the opamp output.
If spreaders are added, they will require bias too and the initial simplicity will be mostly gone


Quote:
You could arrange for the AC waveforms at the two bases to be perfectly in-phase, or perfectly out-of-phase (180 degree shift), or perhaps somewhere in between (90 degrees?).
I think the in-phase option is the best, because it maximizes the level of the carrier available for the intermodulation to take place.
If that condition is not met, some of the signal will be lost in the fight between the two signals, and ultimately there is also a risk of cross-conduction.

Refinements are certainly possible, but it will always be at the expense of simplicity.

As it is, the circuit has the advantages of a class D in low power mode, whilst retaining the linearity and bandwidth of a linear amplifier for large signals. Not that bad for such a simple and problem-free circuit
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 3rd March 2014, 05:40 PM   #5
effebi is offline effebi  Italy
diyAudio Member
 
effebi's Avatar
 
Join Date: Nov 2010
Location: Milan
I am not sure I understood how it works. Would you please put up a schematic, even a sketch on paper ?
Is it like the sound signal (LF) is AM modulating a RF signal? But even so I do not see how it overcomes the Xover zone.
  Reply With Quote
Old 4th March 2014, 11:11 AM   #6
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
I managed to find an approximate schematic, but it isn't very helpful for comprehension, I'll try to explain it in a simplified way, first in the time domain because it is more intuitive.

In this example, the Xover distortion is simulated using just diodes.
The first pic shows the 1KHz input signal (green) and the distorted output at point a (red).
The second pic shows the result of linearly adding a 2Vpp, 10KHz squarewave: point b, blue.
The choice of squarewave and 10KHz is for clarity of the graphics; in reality, it would be a poor choice.

You notice that if you mentally average the squarewave, you find the original sine, even in the Xover region, because the square amplitude is larger than the Vbe of the diodes.
The 10KHz can also be removed by a filter: third pic.
For a fair comparison, the uncorrected output is also filtered in the same way.

On the blue trace, you notice that the step in the 0V region has been completely polished off. The 1KHz amplitude is also somewhat larger.
The sinewave doesn't look clean: that is because of the filtering, and the 1 to 10 ratio of signal to carrier frequency which is much too low and generates additional intermod products.
With a proper choice of carrier frequency, as in the real circuit, the waveform is much cleaner.

The fourth pic shows the whole thing in more details, with a larger input amplitude and a different scale.

In the frequency domain things are not as intuitive, but it also provides useful informations.
The non-linear junctions can be seen as a modulator: with an input spectrum containing Fs and Fc, it will create Fc+Fs and Fc-Fs. But things won't stop there: it will also generate harmonics of both the carrier and the signal (they are the ones that create the Xover distortion), and it will go even further: all of the above will mix and combine, and generate tens of supplementary intermod products and so on.
Most of these products will be located around or above Fc, (and they will be filtered out) but the low frequency ones are interesting: they will reinforce the Fs fundamental, and cancel its harmonics.
Some higher order products will be located between Fs and Fc, just like in class D, which is why it is important to chose a high enough frequency.

We have seen the voltage aspect so far, but the current side is as important: the current at Fc must be sufficient to drive the output load, which is why the post filter must not present a too high impedance at the carrier frequency.
The fifth pic is an approximate schematic of the circuit presented; L2 is not too large, and a series resonant circuit is added at the output, to maximize the current at the carrier frequency and accessorily help remove unwanted residues.
It is one possible implementation, but there are many alternatives: the carrier frequency can also be injected separately, see last pic for example
Attached Images
File Type: png HF1.png (76.0 KB, 245 views)
File Type: png HF2.png (87.0 KB, 228 views)
File Type: png HF3.png (82.2 KB, 53 views)
File Type: png HF4.png (67.9 KB, 49 views)
File Type: png ClassW.png (70.1 KB, 85 views)
File Type: png ClassW4.png (71.8 KB, 101 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 4th March 2014, 12:46 PM   #7
diyAudio Member
 
Join Date: May 2011
Location: Silicon Valley
Is schematic #5 (copied below) correct? Do you modulate using the RC bi-exponential sawtooth wave at red node "B", as shown? Or perhaps do you connect R2 to red node "A" instead, thus modulating with a fast-edge-rate square wave?
Attached Images
File Type: png RC_sawtooth.png (14.8 KB, 95 views)
  Reply With Quote
Old 4th March 2014, 12:56 PM   #8
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by Mark Johnson View Post
Is schematic #5 (copied below) correct?
Yes, here are the asc's of the circuits in question:
Attached Files
File Type: asc HFbias.asc (3.5 KB, 8 views)
File Type: asc classW2amp.asc (3.7 KB, 6 views)
File Type: asc classW4amp.asc (3.2 KB, 9 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 5th March 2014, 06:56 AM   #9
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
One or two things I forgot to mention about these circuits: they are not intended as constructible amplifiers, but as objects for study of the concept, ie. to show as much as possible its deficiencies: for example, in W2, the feedback R4 is taken from the output of the opamp, while the sensible thing to do is to take it from the emitters, to include the transistors in the feedback loop.
But doing so would obfuscate most of the transistor's bad behavior, and that's precisely what we want to examine.

Also, the 2P transistors are just bogey types, ideal complementaries of the 2N with just a sex-change operation in the spice definition.
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪
  Reply With Quote
Old 27th April 2014, 11:31 PM   #10
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
This is a nice idea. It may work better with modern transistors that are faster.

However, would it pass FCC? It would need to be in an unrestricted radio band emitting low enough power without harmonics that would interfere with higher bands.
  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
Dumpster Dive Maggies Standalone Planars & Exotics 15 22nd February 2010 02:43 PM
Dumpster Dive widman Multi-Way 9 8th March 2009 12:43 PM
Complete Sub Newb looking to dive in alexmoose Subwoofers 9 4th June 2008 11:17 PM
Happy dumpster dive! HIPCHECK Parts 6 19th September 2006 07:51 PM


New To Site? Need Help?

All times are GMT. The time now is 12:46 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