actually, in pro-audio there are now many amplifiers doing that, both with class AB and class D amplifiers. when i worked at Apt in the '80s, we were designing the Apt-2 amplifier, which would have been a class AB amp with a switching supply.... actually, since the power supply was going to follow the envelope of the audio, it would have been a class H design... a class AB with variable power supply rails.
the pro-audio amps that use this type of technology are known as "ultralight" amps.
the pro-audio amps that use this type of technology are known as "ultralight" amps.
By switching I assume you mean SMPS.
To be really effective, such an SMPS would have to be an off-line circuit. These are hazardrous and as such, outside the scope of many DIYers. For example, one of the many factors is radiated emission, something usually tested in a Faraday cage.
You'd also have to have a pretty high switching frequency to stop the noise getting onto the audio band.
To be really effective, such an SMPS would have to be an off-line circuit. These are hazardrous and as such, outside the scope of many DIYers. For example, one of the many factors is radiated emission, something usually tested in a Faraday cage.
You'd also have to have a pretty high switching frequency to stop the noise getting onto the audio band.
Probably best off buying a module such as this:
http://www.coldamp.com/opencms/opencms/coldamp/en/productos/SPS30
http://www.coldamp.com/opencms/opencms/coldamp/en/productos/SPS30
DIY SMPS
I agree with the previous thread that recommended buying an off the shelf supply. Switchmode power supplies are complex things that in most cases will involve designing and building your own transformers and/or inductors. Take a look at the SMPS in a typical PC and you will find over 50 components, and this is a design that has had every last possible component squeezed out of it.
Additionally, the voltages/currents on the rectifier and primary side are large enough to be lethal. Unless you have experience in designing and building high voltage electronics, I strongly suggest buying rather than building such a supply.
I agree with the previous thread that recommended buying an off the shelf supply. Switchmode power supplies are complex things that in most cases will involve designing and building your own transformers and/or inductors. Take a look at the SMPS in a typical PC and you will find over 50 components, and this is a design that has had every last possible component squeezed out of it.
Additionally, the voltages/currents on the rectifier and primary side are large enough to be lethal. Unless you have experience in designing and building high voltage electronics, I strongly suggest buying rather than building such a supply.
What I like to do is go on manufacturer websites such as ON Semi and look at evaluation boards/development kits designed around their ICs. That could be one way to start things off. I don't see much point re-inventing the wheel every time you want to get something done.
At the moment (in addition to all the other things I'm doing) I'm designing a power factor controller input stage, based on the "NCP165148VEVB" evaluation board. It would take me months or maybe even longer to do something like that from scratch.
Cheers,
Lech
At the moment (in addition to all the other things I'm doing) I'm designing a power factor controller input stage, based on the "NCP165148VEVB" evaluation board. It would take me months or maybe even longer to do something like that from scratch.
Cheers,
Lech
The limitations imposed by size and cost of large power supplies have also led me to look at SMPS for alternatives to linear PSU's. I was interested in this article http://www.stereophile.com/solidpoweramps/306chord/
about the Chord 14000 that apparently uses such supplies for what is far more amp than I can ever imagine needing. Here's a quote:
AC from the wall is filtered, then rectified to DC—some 300–350V worth—and stored in a bank of high-voltage capacitors. At this point, the current is still connected to the mains. It is then "chopped," using high-voltage MOSFETs running at 80kHz (footnote 2). From there, the resulting high-frequency waveform is sent to a custom-made ceramic-core transformer wound with multistrand Litz wire. As the size of a transformer decreases as operating frequencies increase, only a small transformer is required to re-rectify the 80kHz waveform. At the transformer's output a bank of very fast rectifiers, a small coil, and small capacitors convert the waveform back to DC for the final time, when the electricity is handed off to Chord's Dynamic Coupling system prior to being stored in a final bank of capacitors.
Even if I never attempt such a peoject I would like to know more about this and how it works. I guess that approach doesn't bring the cost down too much since these things still cost $75k/pair.
about the Chord 14000 that apparently uses such supplies for what is far more amp than I can ever imagine needing. Here's a quote:
AC from the wall is filtered, then rectified to DC—some 300–350V worth—and stored in a bank of high-voltage capacitors. At this point, the current is still connected to the mains. It is then "chopped," using high-voltage MOSFETs running at 80kHz (footnote 2). From there, the resulting high-frequency waveform is sent to a custom-made ceramic-core transformer wound with multistrand Litz wire. As the size of a transformer decreases as operating frequencies increase, only a small transformer is required to re-rectify the 80kHz waveform. At the transformer's output a bank of very fast rectifiers, a small coil, and small capacitors convert the waveform back to DC for the final time, when the electricity is handed off to Chord's Dynamic Coupling system prior to being stored in a final bank of capacitors.
Even if I never attempt such a peoject I would like to know more about this and how it works. I guess that approach doesn't bring the cost down too much since these things still cost $75k/pair.
gaetan8888 said:
This schematic is slightly off, I'd say. First of all, it shows something called a flyback converter, for these it is not possible to have full-wave rectification at the output, so you can't use a CT winding as shown. It also does not show the polarity, or if you will, winding direction for the windings, which is crucial. A similar schematic could be used to construct a 'forward' converter, with full wave rectification, but I don't see that in the schematic.
Also, it neglects to mention that both input and output filter caps need to be low ESR for SMPS applications, normal ones will literally cook themselves to a quick death.
As far as SMPS for power amps, it has been done for a long time - first I know of were some Sony amps ion the late 70's / early 80's, with fully discrete SMPS. I have one of those and it works just fine, but then, for EMI/RFI purposes, the supply itself is enclosed in an aluminium 'brick'.
Hi all,
I could be wrong but I think LAB gruppen used to have a flyback supply in their amps.
A teacup size pot-core that could have given the most rugged SMPS designer nightmares for life.
An SMPS for an audio amp is not necessarily cheaper nor smaller than a "linear" supply,
considering the overload capacity of a conventional X-former.
I was planning a 200W SMPS for an active speakers but I ended up with a 160W
standard toroidal instead.
However, going to higher power, then the SMPS probably start to pay off.
All control circuits, housekeeping supply etc starts to become less significant compared to the power stage, which is already small compared
to a "linear supply", as the power goes up.
Another reason to go SMPS is the formfactor. I can't see any problem
to fit a 1kW+ supply in a 1HE height case. Might need a fan though.
Regards / Mattias
I could be wrong but I think LAB gruppen used to have a flyback supply in their amps.
A teacup size pot-core that could have given the most rugged SMPS designer nightmares for life.
An SMPS for an audio amp is not necessarily cheaper nor smaller than a "linear" supply,
considering the overload capacity of a conventional X-former.
I was planning a 200W SMPS for an active speakers but I ended up with a 160W
standard toroidal instead.
However, going to higher power, then the SMPS probably start to pay off.
All control circuits, housekeeping supply etc starts to become less significant compared to the power stage, which is already small compared
to a "linear supply", as the power goes up.
Another reason to go SMPS is the formfactor. I can't see any problem
to fit a 1kW+ supply in a 1HE height case. Might need a fan though.
Regards / Mattias
I use vicor modules
I guess it's not the cheapest solution though. But they work very good for me..
It's important in any case to screen the smps even if it is vicor modules..
http://dsl.mine.nu/galleri/misc/amp2/DSC01027.JPG
http://dsl.mine.nu/galleri/misc/miniamp/mini5.jpg
(amp boards are my own construction of course if anyone wonders what's in there)
An externally hosted image should be here but it was not working when we last tested it.
I guess it's not the cheapest solution though. But they work very good for me..
It's important in any case to screen the smps even if it is vicor modules..
http://dsl.mine.nu/galleri/misc/amp2/DSC01027.JPG
http://dsl.mine.nu/galleri/misc/miniamp/mini5.jpg
(amp boards are my own construction of course if anyone wonders what's in there)
Use two.. or four.. I have four.. two for each channel. but not 95Volt's , what could I do with 95voltsm that would be a PA system not home amp..dhaen said:
and ilimzn, there must be a better schematic than that one?.. 80khz seems a little low for audiophile but i could be wrong of course, also I would like to se real opto feedback and secondary + and - windings.
In theory it should'nt be too hard putting something that works togeather but to get it small and good and safe could be a problem i guess, I dont want my amp go KABOOM in the middle of the night because I forgot to turn the amp off hehe..
I forgot which forum I was in.
The Vicors are neat but as said need some input output lytics filtering etc. Not cheap but nice and with approved isolation barrier.
However, stacking modules for higher voltages may need some precaution if not stated in the data sheet.
If you're going to current limit one module could reverse the output of the other thru the load.
Also secondary side Y-caps going to GND may experience a too high voltage then designed for.
Regards / Mattias
However, stacking modules for higher voltages may need some precaution if not stated in the data sheet.
If you're going to current limit one module could reverse the output of the other thru the load.
Also secondary side Y-caps going to GND may experience a too high voltage then designed for.
Regards / Mattias
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.