Is the UcD modulation scheme less than optimum? - Page 7 - diyAudio
Go Back   Home > Forums > Amplifiers > Class D

Class D Switching Power Amplifiers and Power D/A conversion

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 9th January 2006, 03:14 AM   #61
Account Disabled
 
Join Date: Feb 2004
Send a message via AIM to classd4sure Send a message via MSN to classd4sure
Quote:
Originally by analogspiceman:
I suspect this design approach is most often used in place of understanding rather than to enhance it.
I'll admit to that to a certain extent, I got real sick of trying to figure out all the little spice quircks and bugs, just to get a basic sim going rather than those of the circuitry itself. That's just a general spice comment, aimed at none in particular, but I did start with workbench, so welcome to that nightmare. The frustration was immense.

Having no scope or anything component level sims also served me well to get an actual amp working, which the idealized model would have never done. It took alot of research, trial and simulation in order to take an educated guess at what was really going on with it, reproduce it with spice, or sometimes the problem was already there and I just wasn't looking, and then trying to fix it. That, I think brought me to a new level of understanding.

Just to say, it has its place too, and can enhance understanding! I guess it depends on the user?

But you're right, and I regard your post as digital gold, it's what the books and websites don't really explain well.

Thanks,
Chris
  Reply With Quote
Old 9th January 2006, 04:32 AM   #62
diyAudio Member
 
Join Date: Sep 2004
Location: Yahoo, USA
Quote:
Originally posted by Chris (classd4sure)


I'll admit to that to a certain extent, I got real sick of trying to figure out all the little spice quirks and bugs, just to get a basic sim going rather than those of the circuitry itself. That's just a general spice comment, aimed at none in particular, but I did start with workbench, so welcome to that nightmare. The frustration was immense.

Having no scope or anything component level sims also served me well to get an actual amp working, which the idealized model would have never done. It took a lot of research, trial and simulation in order to take an educated guess at what was really going on with it, reproduce it with spice, or sometimes the problem was already there and I just wasn't looking, and then trying to fix it. That, I think brought me to a new level of understanding.
Never used it, but the "buzz" I've heard is that Workbench is worse than useless for serious work (supposed to be very friendly to those looking for an introduction to electronics and not familiar with SPICE syntax, though).

Spice, especially LTspice (which, because of its many enhancements, has never failed to run any circuit I've fed it) coupled with modern computing power, can just about replace the development lab, if one has some basic knowledge of electronics and is willing to wear a detective's hat now and then.

Attached is the output plot of the swept sine analysis posted earlier in this thread.  Note how the dc level modulates the frequency response.

Regards -- analogspiceman
Attached Images
File Type: gif sweptsine1plot.gif (38.2 KB, 569 views)
  Reply With Quote
Old 9th January 2006, 06:31 AM   #63
diyAudio Member
 
Join Date: Sep 2004
Location: Yahoo, USA
Here's an ac analysis for the component values in the feedback network of a UcD style design.  As mentioned previously, this analysis can be very helpful for ensuring that the lead network in the feedback path provides enough phase lead margin around the resonant point of the LC output filter operating into an open circuit.  Interestingly, most of the lead occurs past the sudden phase shift of resonance.

Regards -- analogspiceman

The attached image show an example of the above applied to an LTspice analysis of a UcD style amplifier.

Top Trace: the phase lead network response
Middle Trace: loop gain with nominal load (6 ohms)
Bottom Trace: open circuit loop gain
Attached Images
File Type: gif sweptsine2.gif (20.7 KB, 643 views)
  Reply With Quote
Old 9th January 2006, 08:18 AM   #64
diyAudio Member
 
Join Date: Sep 2004
Location: Yahoo, USA
Quote:
Originally posted by analogspiceman

Have you ever examined the consistency of the small signal, high frequency square wave response (say a couple of volts peak at 10kHz) of your UcD design when summed with a large signal, low frequency triangle wave (say nearly rail to rail at 50 hertz or so)?  This approximates the small treble riding on big bass of real music, but with a signal much more palatable to the "eye analyzer"
Okay, here is a simulation that does just that.  It shows the small signal (10kHz) square wave response of an idealized UcD180 style self-oscillating class-d amplifier in the presence of a superimposed large (rail-to-rail), slow triangle wave.  To give the amplifier a break and show only its audio band response, signals have been band limited to 32kHz (both the input waveform and the output switching distortion products are low pass filtered with a 2-pole, zero overshoot LCR section).

To make things really interesting, I have attempted to examine just the class-d switching effects with everything else subtracted out.  An otherwise identical non-switching amplifier has been constructed and the outputs of both are differenced against each other and examined with switching ripple blocked out.  Both amplifiers have the same gain and saturation characteristics, and both have nearly identical frequency shaping networks (the linear amp has no delay a very slight high frequency lead network to provide stability and keep it from switching).  If I've set up the simulation correctly, the only signals left in the difference plots are the switching-caused in-band audio artifacts.

Should be interesting, yes?

Regards -- analogspiceman

First the simulation schematic...
Attached Images
File Type: gif switchingdistortion1.gif (17.2 KB, 662 views)
  Reply With Quote
Old 9th January 2006, 08:36 AM   #65
diyAudio Member
 
Join Date: Sep 2004
Location: Yahoo, USA
Quote:
Originally posted by analogspiceman

[...] the only signals left in the difference plots are the switching-caused in-band audio artifacts.

First the simulation schematic...
The top trace is the output of the non-switching amplifier (note that it saturates to the rail very much like the UcD output below it).

The middle trace is the UcD amplifier.  It looks pretty clean except near the rails where the switching frequency becomes irregular.

The bottom trace is the audio artifact taken from the difference between the two amplifier outputs.  For signals not too near the rails, there is very little distortion, but things do seem to break up a little bit near the edges, eh?

Regards -- analogspiceman

Now, the simulation output...
Attached Images
File Type: gif switchingdistortion1plot.gif (17.0 KB, 570 views)
  Reply With Quote
Old 9th January 2006, 09:06 AM   #66
SSassen is offline SSassen  Netherlands
diyAudio Member
 
SSassen's Avatar
 
Join Date: Mar 2005
Location: Edam, The Netherlands
Analogspiceman,

Excellent posts, it looks like I'll be spending some time with LTspice again today.

Thanks,

Sander Sassen
http://www.hardwareanalysis.com
  Reply With Quote
Old 9th January 2006, 01:55 PM   #67
Account Disabled
 
Join Date: Feb 2004
Send a message via AIM to classd4sure Send a message via MSN to classd4sure
Quote:
Originally posted by analogspiceman


The top trace is the output of the non-switching amplifier (note that it saturates to the rail very much like the UcD output below it).

The middle trace is the UcD amplifier.  It looks pretty clean except near the rails where the switching frequency becomes irregular.

The bottom trace is the audio artifact taken from the difference between the two amplifier outputs.  For signals not too near the rails, there is very little distortion, but things do seem to break up a little bit near the edges, eh?

Regards -- analogspiceman

Now, the simulation output...

Hi,

I love when you get bored.

Seems like your results on this one confirm mine. From what I gathered with not much work (easily be wrong) it seemed the increased ripple amplitude was to blame. I was thinking of experimenting with limiters in the feedback network.

Regards,
Chris
  Reply With Quote
Old 10th January 2006, 07:16 AM   #68
diyAudio Member
 
Join Date: Sep 2004
Location: Yahoo, USA
Quote:
Originally posted by analogspiceman

If you mean this paper,

Hysteresis Controller with Constant Switching Frequency
Søren Poulsen and Michael A. E. Andersen, Member, IEEE
IEEE Transactions on Consumer Electronics, Vol. 51, No. 2, MAY 2005

it is still available from the IEEE. It is the one with the unreadable schematics that I mentioned earlier.
Quote:
Originally posted by soren

The thesis was sponsored by The Danish Energy Authority. ICEPower was one of two corporation partners, although the they didn't contribute much.
Quote:
Originally posted by analogspiceman

If I'm not mistaken you are the primary author of the papers I referenced in the beginning of this thread.  If so, I would very much appreciate it if you would email me the values used in the unreadable simulation schematics included in your papers that I referenced.  You must have noticed that Bruno more or less blew off the claim made in the paper with regard to the superiority of a linearized hysteresis self oscillating class d amplifier as compared to the UcD style class d circuit.  I would like to attempt to verify this claim using LTspice (hence the need for the values).
Never mind about emailing the values - Jaka Racman had access to a copy with better detail than the one I got from the IEEE, and he was kind enough to update me with the previously unreadable information. Thanks Jaka

I still had to guess at several pieces to the puzzle, though.  The inductor sense winding was not specified, the comparator and output stage voltage saturation voltages were missing as were their delay times.  Oh, and the load.

The following values yield a fair match to the waveforms of your paper: an 8 ohm load, a 20 to 1 current transformer turns ratio (inverting), +/- 7 and 40 volts for the saturation levels, and 10 and 300ns for the comparator and output stage delays, respectively.  (Without a realistic output stage delay the circuit doesn't seem to work at all.)

How'd I do, guess-wise? (Please feel free to jump right in and correct me.)

Regards -- analogspiceman

Here's the simulation schematic ...
Attached Images
File Type: gif sorencircuit.gif (12.7 KB, 649 views)
  Reply With Quote
Old 10th January 2006, 07:24 AM   #69
Account Disabled
 
Join Date: Feb 2004
Send a message via AIM to classd4sure Send a message via MSN to classd4sure
Hi spiceman,

I had no idea the circuit you were after was on that paper. The values in my copy are very readable as well!

Looking forward to your results!

Best Regards,
Chris
  Reply With Quote
Old 10th January 2006, 07:30 AM   #70
diyAudio Member
 
Join Date: Sep 2004
Location: Yahoo, USA
Quote:
Originally posted by analogspiceman

The bottom trace is the audio artifact taken from the difference between the two amplifier outputs.  For signals not too near the rails, there is very little distortion, but things do seem to break up a little bit near the edges, eh?

Regards -- analogspiceman

Now, the simulation output...
I inadvertently forgot to reduce the effective loop gain of the non switching amplifier reference standard in the above comparison.

The UcD design actually looks significantly better than indicated above.  When I get the time for it I will repost the correct results plus some more interesting data.

Regards and apologies to Bruno & Co. -- analogspiceman
  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
AD Vs BD Modulation in Class-D Workhorse Class D 87 27th October 2012 06:33 AM
Flux modulation 454Casull Multi-Way 71 2nd November 2011 11:21 PM
Glow modulation tubes engels Tubes / Valves 9 27th January 2007 07:22 PM
Class D Modulation Scheme pokka Class D 7 15th December 2004 07:33 PM


New To Site? Need Help?

All times are GMT. The time now is 08:09 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