A and T labs K6 SMPS for Class D amp use - Page 2 - 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 15th April 2005, 03:18 AM   #11
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by Pierre
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

The 130V pulses are referenced from the center tap. Between the two extremes the pulse is about double, so 260-280V. The 130V pulses go into an LC filter (350uH and 2x470uF) were you get about 63V out per rail (of course depending on the duty cycle of the pulses, now about 50%).

Gertjan
  Reply With Quote
Old 15th April 2005, 07:55 AM   #12
Pierre is offline Pierre  France
Banned
 
Join Date: Nov 2004
Location: Paris
I have a question: what if you had a power supply of that kind, with about 100uH + 1000uF filter (more or less), but WITHOUT regulation? Assuming that you have about 45% duty cycle, that is, after rectifying you have about 90% duty cycle, would the supply be estable?

I have done some simulations with a 50Vpp pulse with 90% duty cycle: My simulations show that the output would oscillate from 0 to about 100V at startup for several ms (the lighter the load, the longer it oscillates), until it is progressively reaching the final value, about 45V. For example, for a 1kohm load, it can be oscillating for about 1sec.

The only cure for this is to use a small capacitor value (about 1-10uF), so the filter is similar to the one in a Class-D amplifier. But this wouldn't allow any energy reservoir at the output of the supply.

So the final question is: can a SMPS of this type be unregulated with the only drawback of poor load and null line regulation?

Thanks!
  Reply With Quote
Old 15th April 2005, 12:13 PM   #13
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by Pierre
I have a question: what if you had a power supply of that kind, with about 100uH + 1000uF filter (more or less), but WITHOUT regulation? Assuming that you have about 45% duty cycle, that is, after rectifying you have about 90% duty cycle, would the supply be estable?

I have done some simulations with a 50Vpp pulse with 90% duty cycle: My simulations show that the output would oscillate from 0 to about 100V at startup for several ms (the lighter the load, the longer it oscillates), until it is progressively reaching the final value, about 45V. For example, for a 1kohm load, it can be oscillating for about 1sec.

The only cure for this is to use a small capacitor value (about 1-10uF), so the filter is similar to the one in a Class-D amplifier. But this wouldn't allow any energy reservoir at the output of the supply.

So the final question is: can a SMPS of this type be unregulated with the only drawback of poor load and null line regulation?

Thanks!

Hi Pierre,

I don't think I can answer all your questions. Anyway, I'll give you my thinking here.

1. I don't plan to use a smaller filter coil, the 350uH one that is used now looks very heavy duty so I leave it as it is
2. I plan to increase the output caps for a couple of reasons: a bigger cap gives your more power reserves for very quickly changing loads; it could also prevent/reduce power supply pumping; it reduces the Q factor of the LC filter which makes it easier to optimize the feedback loop since the phase response would be more gradual.

In fact, tonight I plan to measure the open-loop response of the SMPS by cutting out the feedback loop and apply a signal to the input of the SMPS input directly. This allows me to get accurate data about the gain which I now have estimated at about 130 and it also allows me to measure the response of the combination of the SMPS and LC output filter. I plan to measure the response with the standard caps and big caps and also with a few different loads (read lightbulbs). With that info I plan to do simulations to determinethe optimum feedback netwrok for my application.

I plan to keep using feedback since that way, the influence of mains voltage variation can be suppressed and also the mains ripple (100Hz) can be suppressed by the feedback. As a result, the increased mains ripple suppression will make it look like you have bigger output caps. The feedback loop will be active for lower frequencies, so for lower frequencies that need most of the power, the SMPS will actively regulate the output voltage to stay stable. For higher frequencies, the power is drawn directly from the output caps as the SMPS will no be able to regulate fast enough at higher frequencies.

With the SMPS I hope to create a very stable strong power supply that is able to deliver stable power and lots of current especially for the amps that drive the woofers. Since the current that is drawn by the amps for the woofers is relatively slowly changing, by proper design of the feedback loop the SMPS should be able to regulate that nicely.

Best regards

Gertjan
  Reply With Quote
Old 16th April 2005, 02:20 PM   #14
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by Pierre
I have a question: what if you had a power supply of that kind, with about 100uH + 1000uF filter (more or less), but WITHOUT regulation? Assuming that you have about 45% duty cycle, that is, after rectifying you have about 90% duty cycle, would the supply be estable?

I have done some simulations with a 50Vpp pulse with 90% duty cycle: My simulations show that the output would oscillate from 0 to about 100V at startup for several ms (the lighter the load, the longer it oscillates), until it is progressively reaching the final value, about 45V. For example, for a 1kohm load, it can be oscillating for about 1sec.

The only cure for this is to use a small capacitor value (about 1-10uF), so the filter is similar to the one in a Class-D amplifier. But this wouldn't allow any energy reservoir at the output of the supply.

So the final question is: can a SMPS of this type be unregulated with the only drawback of poor load and null line regulation?

Thanks!

Hi Pierre,

I did some measurements yesterday and today using the SMPS unregulated (cutting the feedback loop and apply a voltage to the input pin of the SG3525 chip.

I found that I need to apply about 1.7V in to get about 65V out, this is with a total duty cycle of about 50%. I checked the gain, with my relatively low mains voltage of about 100V, the gain around 60V out (checked with a square wave of 2Hz on the input with 100mV amplitude and offset 1.6V) is about 45. I had earlier estimated a higher value of about 130. However, the gain is not linear over the whole range, so I checked the gain around the working point where I plan to use the supply (around 60V) there it is about 45. Will be a bit higher when you use it with 115V in or 230V in as intended.

I varied the load current from about 0.4 to slightly over 4 A. I used dummy loads that I build with a collection of 82Ohm 20W resistors that I had bought the other day to make a dummy load for amp testing. Some of the resistor had to endure about 40W to be able to reach that 4 A output current, they survived, they were only stressed a short time, say 10-20 seconds at a time and all the resistors were glued to an aluminum plate.

The attached graph shows the unregulated output voltage of the SMPS versus the load current. You can see that the voltage drops quite significantly. With a load of about 4A (about 400W), the 100Hz ripple at the output has become about 2.8V peak/peak (1V RMS). With additional 8200uF caps at the output, the ripple reduces to about 140mV RMS.

So I would not want to use this supply unregulated.

You can also see that the Vout does not drop linear with Iout. This means that the output resistance is not constant and varies from about 20 Ohm or so for low currents while it goes down to about 1.4 Ohm at 4A. I estimated it will go further down to 0.6 Ohm or so at 10A. I estimated the resistance from the slope of the curve (dV/dI). So for modeling of the feedback loop, it is probably best to model the whole thing with an output resistance of 0.6 Ohm since that will give the highest Q factor of the LC filter and therefore the fastest phase change.

Best regards

Gertjan
Attached Images
File Type: jpg vout versus iout.jpg (52.0 KB, 1013 views)
  Reply With Quote
Old 16th April 2005, 04:05 PM   #15
Pierre is offline Pierre  France
Banned
 
Join Date: Nov 2004
Location: Paris
Thanks for sharing that exhaustive measurements.

There is a thing I don't understand: you have connected the supply unregulated, so the duty cycle has gone to about 50% with 1.7V input. I don't know if you refer to 50% each output (that is, one of the mosfets is conducting for every moment, except for a small dead-time), or you mean that each mosfet is conducting 50% of its max. time, so the supply can go to about +/-130V for max. input. Please clarify this conditions.

I built a car smps a long time ago, with 12V input and +/-38V output. It was unregulated but I remember that it was stiff, with very little variation with load (but null line regulation, of course). That's why I asked about unregulated off-line SMPS also.

Best regards,
Pierre
  Reply With Quote
Old 17th April 2005, 12:27 PM   #16
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by Pierre
Thanks for sharing that exhaustive measurements.

There is a thing I don't understand: you have connected the supply unregulated, so the duty cycle has gone to about 50% with 1.7V input. I don't know if you refer to 50% each output (that is, one of the mosfets is conducting for every moment, except for a small dead-time), or you mean that each mosfet is conducting 50% of its max. time, so the supply can go to about +/-130V for max. input. Please clarify this conditions.

I built a car smps a long time ago, with 12V input and +/-38V output. It was unregulated but I remember that it was stiff, with very little variation with load (but null line regulation, of course). That's why I asked about unregulated off-line SMPS also.

Best regards,
Pierre
Hi Pierre,

It is a full bridge design, so with two mosfet pairs each mosfet pair conducts 25% of the time, however, since there are two mosfet pairs, effectively (together) they are conducting 50% of the time. So the effective duty cycle is 50%. In theory, when there is no dead time, they could effectively reach 100% duty cycle.

Best regards

Gertjan
  Reply With Quote
Old 18th April 2005, 02:01 AM   #17
diyAudio Member
 
Join Date: Jan 2002
Location: Wheaton, Il.
Default Unregulated operation

Hi,

Just to throw my 2 cents worth in (FYI, I am the owner of A and T Labs and the K6PS is one of my kits) - If tight regulation is not an issue, you can do a pretty good job of building an open loop switching supply running at 50% duty cylce. It eliminates any stability issues, and still gives you the benefit of a light-weight design. I published a Radio Electronics article a while ago for a 350 watt car amp that had a +/- 45 volt supply that was open loop. The key design requiremnt for doing so is that the transformer winding ratio has to be pretty much the ratio of your desired output voltage to the rectified input. You also then need minimal output inductance - mostly the filter can be just capacitance.

The K6PS transformer cannot be used in this mode, as it is fundamentally designed for pulse width modulation to achieve regulation, and expects an inductive output filter. that is to say that the pre-filter pulse amplitude is well above the regulated DC output, and operaton in 50% duty cycle open loop mode would result in substantially higher output voltage than the specified regulated voltage. In fact, the K6PS has a switch to set differnet output voltages, which are achieved by operation at differnet pulse widths (obviously also affected by load).

Reinhard Metz
__________________
R Metz
  Reply With Quote
Old 18th April 2005, 03:44 AM   #18
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Default Re: Unregulated operation

Quote:
Originally posted by rmetz
Hi,

Just to throw my 2 cents worth in (FYI, I am the owner of A and T Labs and the K6PS is one of my kits) - If tight regulation is not an issue, you can do a pretty good job of building an open loop switching supply running at 50% duty cylce. It eliminates any stability issues, and still gives you the benefit of a light-weight design. I published a Radio Electronics article a while ago for a 350 watt car amp that had a +/- 45 volt supply that was open loop. The key design requiremnt for doing so is that the transformer winding ratio has to be pretty much the ratio of your desired output voltage to the rectified input. You also then need minimal output inductance - mostly the filter can be just capacitance.

The K6PS transformer cannot be used in this mode, as it is fundamentally designed for pulse width modulation to achieve regulation, and expects an inductive output filter. that is to say that the pre-filter pulse amplitude is well above the regulated DC output, and operaton in 50% duty cycle open loop mode would result in substantially higher output voltage than the specified regulated voltage. In fact, the K6PS has a switch to set differnet output voltages, which are achieved by operation at differnet pulse widths (obviously also affected by load).

Reinhard Metz

Hello Reinhard,

Very good to see you are here on this board. I like your supply since it seems to be very robust and it has survived so far all my experiments (and more important, I also survived taking utmost care to keep the experiments safe). I plan to use it as a regulated supply of course, just cut open the feedback loop to get a better idea of the open loop performance and gain to improve the simplified simulation model that I use in Switchcad. Because I plan to use it with Class D amps, I want to use bigger output caps (to avoid/reduce effects of power supply pumping) and I need to make sure that the feedback loop stays stable with those bigger caps. I`m thinking of redimensioning some of the parts around the error amp (IC2) possibly adding a 2nd zero by putting an RC in parallel with the current R at the input of the error amp. This could improve the phase margin and may make it even possible to use the amp with a wide range of output caps. I`m also thinking of moving the poles that define the highest cut-off frequency of the error amp down since when I add a zero, I get more gain at higher frequencies. The gain goes down after the 2 poles are becoming active but at that point, the gain will now be higher than it was before (because of the added zero) and I don`t really want to increase the gain at the high frequencies say in the area of 50-100kHz too much because of possible negative effects of high frequency noise in the feedback loop.

Another, more simple solution that probably also works (albeit with less phase margin) would be to lower the current zero and pole to a lower frequency to reflect the lower LC resonance frequency that you get with bigger caps. However, then the loop gain at 100Hz will be lower than it is now, so the 100Hz mains induced ripple will be less suppresssed by the feedback network, although it will of course be suppressed by the bigger caps. That is why I prefer to add a zero, so keep the gain at 100Hz to get the optimum in 100Hz ripple reduction.

By the way, maybe you have read this in an earlier mail. I added an RC (15k, 22uF) delay for the reference voltage that goes to the error amp to avoid the overshoot in the output voltage when starting up.

Thanks and best regards

Gertjan
  Reply With Quote
Old 18th April 2005, 07:45 AM   #19
Pierre is offline Pierre  France
Banned
 
Join Date: Nov 2004
Location: Paris
What Reinhard says makes a lot of sense.
Ghemink tested the supply in open loop, but forcing the PWM to be less than 50% by inserting a given voltage in the SG3525 input. So the transformer+filter wasn't working in an optimal way, because it was not regulating neither it was not working at 50%, that's why regulation was so poor.
I am almost sure that, if you tie input to GND, so the PWM controller goes to its maximum duty-cycle, you will get a higher output voltage (can you confirm what voltage exactly, Ghemink?), but you will have much better load regulation, right?

I have also some experience with DC/DC for cars (12V to +/-40V), unregulated, and it was so stiff that regulation was not needed, at least from the load point of view (of course line regulation is null). But it is the same with a conventional offline linear power supply, isn't it?
  Reply With Quote
Old 18th April 2005, 12:00 PM   #20
ghemink is offline ghemink  Netherlands
diyAudio Member
 
Join Date: Jun 2004
Location: Japan
Quote:
Originally posted by Pierre
What Reinhard says makes a lot of sense.
Ghemink tested the supply in open loop, but forcing the PWM to be less than 50% by inserting a given voltage in the SG3525 input. So the transformer+filter wasn't working in an optimal way, because it was not regulating neither it was not working at 50%, that's why regulation was so poor.
I am almost sure that, if you tie input to GND, so the PWM controller goes to its maximum duty-cycle, you will get a higher output voltage (can you confirm what voltage exactly, Ghemink?), but you will have much better load regulation, right?

I have also some experience with DC/DC for cars (12V to +/-40V), unregulated, and it was so stiff that regulation was not needed, at least from the load point of view (of course line regulation is null). But it is the same with a conventional offline linear power supply, isn't it?

With 0V at the input of the Sg3525, nothing will happen, it will have 0 output. The sg3525 starts to work from about 0.9V at its input, I think it will give a 50% duty cycle at around 2V in or so. Effectively, the duty cycle will be almost 100%, so output voltage would be about 130V, this will blow my caps and I don't want to try that.

I will soon try to modify the feedback loop in the simplest possible way. Not adding an zero, just keep it as it is, just increase the size of the cap in the feedback loop to compensate for the lower resonance frequency of the LC filter with the bigger output cap. This should then still work with the original caps as well.

Best regards

Gertjan
  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 10:22 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