A and T labs K6 SMPS for Class D amp use - 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 13th April 2005, 12:26 PM   #1
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Default A and T labs K6 SMPS for Class D amp use

Hello all,

I will continue reporting here on my progress with improving the A and T labs K6 SMPS for Class D use. I have started to report on this SMPS in the UcD400 thread. This SMPS was made for use with Class AM amps and is supposed to be able to deliver 1000W continous and 2000W peak. It should therefore be suitable to supply multiple UcD400 modules for use in an active speaker system. I'm working on it to make it suitable for use with Class D amps that have their unique issues with power supply pumping that regular Class AB amps don't have. Therefore, one of the topics is to increase the output LC filter caps with bigger ones, currently 2x470uF per rail. I want to increase that to 10.000uF per rail to reduce the effect of power supply pumping. However, to do that, the feedback loop has to be changed. I also want to improve the startup behavior since at startup, there is some overshoot of the supply voltage, this may trigger the protection circuit on the UcD and I like to avoid that.

Best regards

Gertjan
  Reply With Quote
Old 13th April 2005, 12:38 PM   #2
diyAudio Member
 
Join Date: Feb 2005
Location: Antwerp
Gertjan,

Couldn't you avoid the overshoot by turning on the UcD modules a few seconds after the SMPS starts?

You could switch the /on, or even use relays?
  Reply With Quote
Old 13th April 2005, 02:37 PM   #3
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by Yves Smolders
Gertjan,

Couldn't you avoid the overshoot by turning on the UcD modules a few seconds after the SMPS starts?

You could switch the /on, or even use relays?

The UcDs are not yet connected. The overshoot is caused by the fact that the feedback loop is working and tries to maintain a 63V output voltage while the softstart circuit in the SG3525 chip limits the startup speed. So both circuits are fighting against eachother resulting in the error amp giving its maximum 12V output voltage (clipping of the integrator).

The green line is the plus rail output voltage. The blue line is the output voltage of the error amp. It comes down when the output voltage exceeds about 63V, but it takes time since the integrator was clipped to 12V, therefore the overshoot. The error amp compares the reference voltage (5V) with the output voltage via a voltage divider and a compensation network. This 5V is applied to the error amp well before the supply really starts up, therefore, this effect of the clipping integrator (error amp output) takes places.

Gertjan
Attached Images
File Type: jpg error amp.jpg (86.4 KB, 2137 views)
  Reply With Quote
Old 13th April 2005, 02:44 PM   #4
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by Yves Smolders
Gertjan,

Couldn't you avoid the overshoot by turning on the UcD modules a few seconds after the SMPS starts?

You could switch the /on, or even use relays?

Well, the overshoot will always be there (as long as I don't fix it in some way). I could avoid that it reaches the UcDs by switching on the UcDs later, that is correct. However, the ELNA cerafines that I have are 63V, I think they will not have a problem with a 5V overshoot. Anyway, I need to lower the output voltage a bit, want to target about 60V, so the ELNA's would be fine, the overshoot would then be about 65V which should not be a problem for those caps, and also not a problem for the UcDs.

Best regards

Gertjan
  Reply With Quote
Old 13th April 2005, 10:12 PM   #5
Account Disabled
 
Join Date: Feb 2004
Send a message via AIM to classd4sure Send a message via MSN to classd4sure
Hi Gertjan,

Sounds like you need some derivative action in there, PID control. Think it'd be possible to mod it that way?

I take it you don't have anything like schematics to work with, or do you?

I'm speaking ignorantly here if course but I'd think the PWM action of the SMPS should be able to correct for any pumping effects, so a larger cap may not be a concern, but I guess with a bigger cap you'd get more energy stored as well.

Will that not drop your frequency, and so I guess that's what you'd be compensating for in the feedback loop?

Regarding the integrator clipping, what would happen if you delayed the reference voltage, by say, having it ramp up slowly. I see no need why it should need to turn on instantly, it should just be able to ramp up and settle, maybe that would solve your overshoot?

Or in the same line of thought, some sort of derivative feedback onto that 5V reference.

My last ignorant idea for the day,

Should you not have the schematics perhaps you could contact A and T and ask them for it in order to post them here so they're supply could be modded by the ol community for use with the common amp modules used. Or maybe they'd rather just mod it themselves... some food for thought.

Good luck with it anyway, I'll now be following this thread passively

Cheers,
Chris
  Reply With Quote
Old 14th April 2005, 03:31 AM   #6
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by classd4sure
Hi Gertjan,

Sounds like you need some derivative action in there, PID control. Think it'd be possible to mod it that way?

I take it you don't have anything like schematics to work with, or do you?

I'm speaking ignorantly here if course but I'd think the PWM action of the SMPS should be able to correct for any pumping effects, so a larger cap may not be a concern, but I guess with a bigger cap you'd get more energy stored as well.

Will that not drop your frequency, and so I guess that's what you'd be compensating for in the feedback loop?

Regarding the integrator clipping, what would happen if you delayed the reference voltage, by say, having it ramp up slowly. I see no need why it should need to turn on instantly, it should just be able to ramp up and settle, maybe that would solve your overshoot?

Or in the same line of thought, some sort of derivative feedback onto that 5V reference.

My last ignorant idea for the day,

Should you not have the schematics perhaps you could contact A and T and ask them for it in order to post them here so they're supply could be modded by the ol community for use with the common amp modules used. Or maybe they'd rather just mod it themselves... some food for thought.

Good luck with it anyway, I'll now be following this thread passively

Cheers,
Chris

Hi Chris,

Thanks for your tips. A few answers.

Yes, I have the schematics and you can download them from the internet. The error amp does indeed not have a derivitave action. I plan to add that to increase the phase margin when I hook up bigger caps. So I expect this to improve the startup behavior as well.

I plan to check effects of pumping later on (weekend???) by hooking up an UcD400, use a dummy load on it and supply 20Hz at various amplitudes.

Adding bigger caps will not drop the switching frequency of the SMPS. The SMPS oscillation frequency is determined by an R and C that are connected to a sawtooth generator in the SG3525 chip. The feedback loop only works at relatively low frequencies.

The oscillation frequency was stated in the documentation as 75kHz, however, since it is full bridge design, I get 150kHz pulses (with about 130V amplitude) before the LC output filter.

Indeed as you say, I wanted to delay the reference voltage with an RC delay. Therefore I started some measurements to check that out. There is a big delay between the time that the Vref comes up and when the SMPS starts oscillating, that would need an RC delay for Vref of several seconds, a bit long. I need to do a bit more measurements to find out why is that. I have some suspicion that it is because I made the softstart cap bigger. I will check that later by monitoring a few more pins on the the SG3525 (I have the datasheet). I assume that I should make the softstart cap smaller so that the SG3525 starts to oscillate faster after startup, then I can use an RC delayed Vref to the error amp to get a very nicely controlled ramp-up of the output voltage.

Best regards

Gertjan
  Reply With Quote
Old 14th April 2005, 01:39 PM   #7
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by classd4sure
Hi Gertjan,

Sounds like you need some derivative action in there, PID control. Think it'd be possible to mod it that way?

I take it you don't have anything like schematics to work with, or do you?

I'm speaking ignorantly here if course but I'd think the PWM action of the SMPS should be able to correct for any pumping effects, so a larger cap may not be a concern, but I guess with a bigger cap you'd get more energy stored as well.

Will that not drop your frequency, and so I guess that's what you'd be compensating for in the feedback loop?

Regarding the integrator clipping, what would happen if you delayed the reference voltage, by say, having it ramp up slowly. I see no need why it should need to turn on instantly, it should just be able to ramp up and settle, maybe that would solve your overshoot?

Or in the same line of thought, some sort of derivative feedback onto that 5V reference.

My last ignorant idea for the day,

Should you not have the schematics perhaps you could contact A and T and ask them for it in order to post them here so they're supply could be modded by the ol community for use with the common amp modules used. Or maybe they'd rather just mod it themselves... some food for thought.

Good luck with it anyway, I'll now be following this thread passively

Cheers,
Chris

I set the softstart cap back to its original value (10uF) and I cut a trace on the PCB to be able to implement an RC delay for the Vref that goes to the error amp. I took an R of 15k and a C of 22uF giving an RC time of about 330ms. The result is attached. Now the softstart is faster than the reference voltage and there is no overshoot. Well there still is, but it is at a much lower level as you can see. The softstart results in the output voltage ramping from 0 to 46V in 100ms, after that it drops a bit and then gradually goes up to 63V. Basically following Vref (blue trace). The noise on Vref is due to the scope probe and GND wire of the probe being seperated about 15cm or so and thus picking up switching noise.

Since the output first rises with a rate of 46V/100ms, this would mean that the supply needs to supply a current of 4.6A when I hook up 10.000uF caps to each rail. This should be not much of a problem.

By the way, after some measurements (not shown here), I concluded that the input of the SG3525 switches from 0% duty to 50% duty (with is 100% duty in case of a full bridge that this SMPS has) for an input voltage of about 0.9V (0%) to 2V (100%). So this means in my case that the gain from input of SG3525 is about a factor 130-140 as my rectified mains reach about 260-280V and the output transformer is a 28:28 center tapped. This is usefull info to improe my simulations with switchcad.

Best regards

Gertjan
Attached Images
File Type: jpg reduced overshoot.jpg (78.5 KB, 1902 views)
  Reply With Quote
Old 14th April 2005, 02:01 PM   #8
Pierre is offline Pierre  France
Banned
 
Join Date: Nov 2004
Location: Paris
Do you have any info on the winding of that transformer?
By the moment we know that it is a ETD49 with 28+28 turns in the secondary. Do you know how many turns does the primary have? Is it gapped? (difficult to know without disassemblying, that's true)
BTW: rectified mains reach 260-280V only? What is your nominal AC voltage, then? 230VAC should give about 315Vdc, while 125VAC should give about 170Vdc, right?

Best regards,
  Reply With Quote
Old 14th April 2005, 02:12 PM   #9
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by Pierre
Do you have any info on the winding of that transformer?
By the moment we know that it is a ETD49 with 28+28 turns in the secondary. Do you know how many turns does the primary have? Is it gapped? (difficult to know without disassemblying, that's true)
BTW: rectified mains reach 260-280V only? What is your nominal AC voltage, then? 230VAC should give about 315Vdc, while 125VAC should give about 170Vdc, right?

Best regards,

It is indeed an ETD49 according to the schematic. Although the type number on the transformer itself says 486-060394 and 486-0420.

Primary has 28 and the secondary is 28 center tapped, so 14 per winding.

I have 100V here in Japan, the SMPS uses a voltage doubling circuit, so that gives me about 260-280V rectified depending on the amount of lightbulbs I use as a load :-)

It must be 14 per winding on the secondary as I get about 130-140V pulses out at the secondary.

best regards

Gertjan
  Reply With Quote
Old 14th April 2005, 02:38 PM   #10
Pierre is offline Pierre  France
Banned
 
Join Date: Nov 2004
Location: Paris
Yes, that matches quite well.
You get 130-140V pulses in the secondary? I assume you mean between the two extremes, not between each extreme and the center tap, right?

thanks for the info, I am gathering info for my next project.
Best regards,
Pierre
  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
SMPS for class D module NewTesla Class D 40 16th May 2010 07:38 PM
SMPS vs. Class-D. Dag Class D 17 2nd March 2007 05:23 PM
JLH class a with a SMPS?? Possible? JZatopa Solid State 2 27th March 2005 04:47 PM
help with class D amp and SMPS Evan Shultz Class D 9 23rd February 2005 12:36 AM
SMPS and class D parts gearheaddruid Swap Meet 0 31st January 2005 05:08 AM


New To Site? Need Help?

All times are GMT. The time now is 12:54 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