Hi,
Still completely new to any DIY Audio stuff, so I'm sorry if I don't put things into words correctly.
I was wondering if it would be possible to use any of the currently available, class A or A/B, diy amplifier modules made by anyone and instead of just using a conventional power supply, use a class D amplifier in between to continuously varry the rail voltage just above the input audio signal. Obviously to keep the output devises as cool as possable and by reduceing the voltage across the output transisters you can use their full current capability or up to the full rated amperes of your output devices. Now before you reply, I think this is sort of how Bob Carver's "Tracking Downconverter" works, but my description is way over simplified and I'm not sure.
But back to my question, could this be done? As simple as it sounds to me and with my lack of knowlage in this field, I can't find anything where any person has tried this or even talked about it (in the DIY world).
Hopefully I can get some good (but please keep it simple) posts to this, that will help me to understand and feed me with some new ideas?!?!?!?!
Thanks for your replies,
Jay Meredith
P.S.
I have purchased 2 books by Randy Slone #1- High-Power Audio Amplifier Construction Manual / The Audiophile's Project Sourcebook. If any body has any other suggestions or recommendations to any other books or web sites that could help me learn more I would greatly appreciate it.
Thanks Again,
Jay M.
Still completely new to any DIY Audio stuff, so I'm sorry if I don't put things into words correctly.
I was wondering if it would be possible to use any of the currently available, class A or A/B, diy amplifier modules made by anyone and instead of just using a conventional power supply, use a class D amplifier in between to continuously varry the rail voltage just above the input audio signal. Obviously to keep the output devises as cool as possable and by reduceing the voltage across the output transisters you can use their full current capability or up to the full rated amperes of your output devices. Now before you reply, I think this is sort of how Bob Carver's "Tracking Downconverter" works, but my description is way over simplified and I'm not sure.
But back to my question, could this be done? As simple as it sounds to me and with my lack of knowlage in this field, I can't find anything where any person has tried this or even talked about it (in the DIY world).
Hopefully I can get some good (but please keep it simple) posts to this, that will help me to understand and feed me with some new ideas?!?!?!?!
Thanks for your replies,
Jay Meredith
P.S.
I have purchased 2 books by Randy Slone #1- High-Power Audio Amplifier Construction Manual / The Audiophile's Project Sourcebook. If any body has any other suggestions or recommendations to any other books or web sites that could help me learn more I would greatly appreciate it.
Thanks Again,
Jay M.
You could run the supply rails of a normal amplifier from a switch-mode converter. If the switching frequency of the converter is high enough it is possible to get the supply rails to track the voltage requirement of the amp. The problem is that even the fastest responding converter will still be far too slow to provide the increased voltage when it is needed.
You would need the incoming audio to be peak-detected, and feed this peak value signal to control the converter. You would then need a delay in the audio before it went to the amp so that the PSU had already 'anticipated' the increased power demand so that the power rails were at the right value for the audio that was about to appear at the input of the amp.
It's a lot easier to incorporate the converter and the amp in one, which is effectively what a class-D amp does.
By the way, the latest switching amp modules from D-2-audio have an impressive published spec. We'll be trying them here at work soon, so I'll let people know if they sound as good as the spec!
You would need the incoming audio to be peak-detected, and feed this peak value signal to control the converter. You would then need a delay in the audio before it went to the amp so that the PSU had already 'anticipated' the increased power demand so that the power rails were at the right value for the audio that was about to appear at the input of the amp.
It's a lot easier to incorporate the converter and the amp in one, which is effectively what a class-D amp does.
By the way, the latest switching amp modules from D-2-audio have an impressive published spec. We'll be trying them here at work soon, so I'll let people know if they sound as good as the spec!
Bob Carver's tracking downconverter is pretty straightforward in concept, a single switching supply tracking the audio envelope. A bridge amplifier is used so as to only need a single supply rail. A simple all-pass filter provides the delay.
In the multichannel home theater amplifer the supply tracks the channel with the highest audio level.
Do we even need this switching supply?
Not really.
The Carver M1.5T put out 1200W/8R in bridge mode, no heatsink, no fan, and from a 3.5" chassis. 16lbs, a couple of those from the 3/16" thick front panel.
You could build a 6 X 200W piece for home theater using the same ideas.
You could even use a heatsink for cosmetic purposes if you wanted, or make the chassis larger.
Switching amplifiers have been around for about 50 years now, about every ten years or so a bunch of noise gets made about them.
Maybe this time around there will be some that: sound good, don't blow up in your face, and don't cost an arm and a leg.
You are welcome to hold your breath.
PS:
Battery powered equipment doesn't count in my book, I don't own any, YMMV.
In the multichannel home theater amplifer the supply tracks the channel with the highest audio level.
Do we even need this switching supply?
Not really.
The Carver M1.5T put out 1200W/8R in bridge mode, no heatsink, no fan, and from a 3.5" chassis. 16lbs, a couple of those from the 3/16" thick front panel.
You could build a 6 X 200W piece for home theater using the same ideas.
You could even use a heatsink for cosmetic purposes if you wanted, or make the chassis larger.
Switching amplifiers have been around for about 50 years now, about every ten years or so a bunch of noise gets made about them.
Maybe this time around there will be some that: sound good, don't blow up in your face, and don't cost an arm and a leg.
You are welcome to hold your breath.
PS:
Battery powered equipment doesn't count in my book, I don't own any, YMMV.
millwood said:
The ideal would be an isolated mains-input SMPSU whose output can drive the speaker directly. (with the audio signal being summed in with the PSU voltage feedback signal).
Perhaps feasible for bass drive, but the SMPSU would need a very high switching frequency to give good performance over the full audio range.
Perhaps someone has tried out the concept?
It can work, but for the DIYer, better quality sound is easier by means of either a straight class D or linear. I simulated one without the delay, and for most music, the supply can keep up with audio demands. That was using a regular response switching supply. The switching supply design could be modified to respond faster, but would work with normal response for powering a sub amp.
Hi phase_accurate, how is you switching amp project coming along?
Hi phase_accurate, how is you switching amp project coming along?
millwood said:
But as a PSU it wouldn't need to track the individual waveform of the signal. It simply provides power to track the envelope of the signal. The output caps on the suppply remove the need for it to track the voltage rail needs on a cycle-to-cycle basis. The delay on the audio signal allows for the PSU to charge up the caps to provide enough headroom for the signal burst that's about to appear at the input to the amp.
I have been working on the idea of using just a switching power supply also but am not happy that if active components are not used to switch the voltage on the secondary side, the positive and negative modulated outputs short together. However, I am considering using one switching power supply and half of a class D amp to produce two halves of an audio waveform. I think it is possible to make a simple decent quality class A style class D amp (Does that make sense? ) if the power supply is set to produce constant current. Then only one fast N-channel mosfet is needed on the lower power supply rail.
) if the power supply is set to produce constant current. Then only one fast N-channel mosfet is needed on the lower power supply rail.
Ouroboros said:
it is the tracking of that envelope that's probamatic: at any given point, you cannot predict where the next "wave" would be. the "envelope" is more like averaging of historical "peaks". As such, it is delayed.
for example, if the last 5 seconds the highest peak is 20v. you add some margin and set your rail at 20+10=30v. what if the next peak turns out to be 40v? you get clipping.
If you increase the margin, then you sacrify efficiency. if you decrease the margin, you sacrifice sound quality.
Might as well go with a full switching mode amp.
One has just to guarantee that the signal envelope doesn't change faster than the tracking supply is able to follow. This is achieved by a lowpass at the input of the linear amp and an envelope detector with faster rise- than fall- rate (which is of course controlled by the input signal before the lowpass).
But I by myself do still prefer a pure class-d solution.
The aforementioned peavey amp (plus corresponding patent) is in fact an SMPS that outputs audio.
Regards
Charles
But I by myself do still prefer a pure class-d solution.
The aforementioned peavey amp (plus corresponding patent) is in fact an SMPS that outputs audio.
Regards
Charles
Sometime during the 70's, Sony produced a working prototype using Class D. It used a switching supply for the plus and minus DC rails. For the SMPS Class D audio section they switched at 300 KHz and had no feedback, if my memory is working today. I never remember them producing it for the public. The problem I see is bandwidth. The higher the switching frequency the higher the bandwidth. Using modern ferrite cores for the magnetics, core loss becomes an issue at the higher frequencies. Eddy currents and related proximity effects causes power losses in the copper. Using planer magnetics design reduces the copper power losses. A rule of thumb for usuable bandwidth in SMPS is 1/10 the switching frequency. This probably hold close to true for the Class D audio amp version of the SMPS. This makes me prefer to use the Class D for the base and lower midrange frequencies only. This is where most of the power is needed and where the Class D's efficiency is best. For the higher frequencies that require less power, a linear (non-switching) amp seems best. Of course, this means either bi-amping or tri-amping. Another potential problem with a Class D amplifier, especially at higher power levels is their switching noise. This can be conducted around the circuits, coupled electrostatically, or electromagnetically. It is not trivial to deal with unless you have a lot of experience in dealing with it.
I am new to this but I can clearly see that.... If the switching supply becomes too slow to keep up with voltage/frequency then it's probably a good idea to have a back up system where an almost instantaneous voltage (short duration only) is injected in the power supply output side to pull up the voltage fast enough till the supply caches up. Frequency dependant and only for a few milliseconds to hold the voltage at the level needed. I had this idea a long time ago. I believe Yamaha has a patent that sounds just like what I am describing however I have no idea how their system works (electronically)
can this be a logical solution?
can this be a logical solution?
Hi,
Interesting topic. I'm working on a PSU directly controlled by the audio signal (prototype already exists) measures are good, but is geared to sub for high ratings.
I would like to speak my mind, about the rapid response of a SMPS, after many tests, even with CPU usage, the best solution to speed response and sound very clean, sound identical to a toroidal with large capacitors (bass hot) is it is shown that the choice of the linear regulator voltage on transformer. This helped not to vary (as PWM) cycles on the transformer. This fact is important, normal pwm on transformer, the inductance is constantly changing, this causes many problems in the case of an amplifier with audiophile performance.
Under these conditions, you may work the best curve in the transformer, optimizing the losses on a fixed frequency. Certainly not as simple as they say, the linear regulator is very fast (150ns) and we had to think of a trigger to increase efficiency. Result is the same as a 500w transformer with 2x22000uF outputs (at + /-70V) and excellent voltage regulation line in just 500 grams weight.
PWM is good on continuos load PSU, it can offer very good line regulation/load also.
Regards
Roberto P.
Interesting topic. I'm working on a PSU directly controlled by the audio signal (prototype already exists) measures are good, but is geared to sub for high ratings.
I would like to speak my mind, about the rapid response of a SMPS, after many tests, even with CPU usage, the best solution to speed response and sound very clean, sound identical to a toroidal with large capacitors (bass hot) is it is shown that the choice of the linear regulator voltage on transformer. This helped not to vary (as PWM) cycles on the transformer. This fact is important, normal pwm on transformer, the inductance is constantly changing, this causes many problems in the case of an amplifier with audiophile performance.
Under these conditions, you may work the best curve in the transformer, optimizing the losses on a fixed frequency. Certainly not as simple as they say, the linear regulator is very fast (150ns) and we had to think of a trigger to increase efficiency. Result is the same as a 500w transformer with 2x22000uF outputs (at + /-70V) and excellent voltage regulation line in just 500 grams weight.
PWM is good on continuos load PSU, it can offer very good line regulation/load also.
Regards
Roberto P.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.