A trafo based linear supply has no control loop, it relies on the bulky smoothing caps for regulation.
So, on similar lines, cant SMPS perform the same way? No control loop, just an active switcher at a high frequency, say 60Khz or whatever with fixed 50% duty cycle.
This question is with audio amp use in mind.
Thanks and Regards,
WonderfulAudio
So, on similar lines, cant SMPS perform the same way? No control loop, just an active switcher at a high frequency, say 60Khz or whatever with fixed 50% duty cycle.
This question is with audio amp use in mind.
Thanks and Regards,
WonderfulAudio
It probably could be done but since there is no control it could run away and have too high voltages and nothing to prevent that.
The transformer will not likely have much fluctuating primary voltage. Even if it has the transformation factor will make the impact less. There is no device switching where timing has a direct influence on output voltage, it is dependent on the stability of the grid.
The transformer will not likely have much fluctuating primary voltage. Even if it has the transformation factor will make the impact less. There is no device switching where timing has a direct influence on output voltage, it is dependent on the stability of the grid.
Last edited:
have seen a switcher with no regulation. This was in a tube guitar amp to make plate voltage from 12V DC. A very simple circuit. An oscillator to make a square wave, a Mos, an inductor, rectification. No regulation. Not my design, it was working fine.
The LLC-converter with fixed, i.e. unregulated frequency, is the pendant to the iron xformer simple PSU. Works nice and is less bulky😉
There are cheap chinese made smps for audio amplifiers ,they have no feedback loop ,output voltage has some 100hz ripple and drops a little with load ,like traditional supplies with big transformer. Personally I'm using such one to power my amplifier .
That's not Regulation. (Just Filtering.) If my wall voltage drops from 124VAC to 62VAC, the DC will also drop to half. (It DOES sag from 125V to 105V on laundry days.)
In switch-mode we must add many parts to control side-effects of switching. There may be large kick-up at low current. And there is little penalty in making more voltage than we need and reducing it. As long as we can can do all that, we may as well set the voltage =constant=.
Yes, it can be done and has been done, in the 70's and 80's, mostly: telecom or minicomputer supplies, instruments from HP for example
It has also been done, and on this forum you can find examples posted by members.
You may need some searching to locate them, but I am sure I saw a few
I only see now that OP makes a mistake by stating linear power supplies have filter caps for regulation. Filter caps don't regulate, filter caps filter and regulators regulate. If that assumption makes OP think SMPS also regulate without control loop....
The SMPS without control loop will also not regulate if I am not mistaking. One would think that the lower ripple and the regulation with control loop as in "standard" SMPS would both be benefits of using an SMPS. Just sayin'
The SMPS without control loop will also not regulate if I am not mistaking. One would think that the lower ripple and the regulation with control loop as in "standard" SMPS would both be benefits of using an SMPS. Just sayin'
Last edited:
The old vibrator and transformer high-voltage supplies for valve car radios were basically uncontrolled switched-mode supplies used for an audio application.
A SMPS can definitely work the same way.
Personally I think it just has to do with the fact that it's very desirable to have a control loop for most other non-audio circuits.
These things will always give some issues with (big) dynamic loads (like audio)
I think a good approach would be an active PFC circuit, followed by just a simple IR2153 or similar.
Personally I think it just has to do with the fact that it's very desirable to have a control loop for most other non-audio circuits.
These things will always give some issues with (big) dynamic loads (like audio)
I think a good approach would be an active PFC circuit, followed by just a simple IR2153 or similar.
I realise I should not have used the words 'trafo based linear supply' instead I should have used 'trafo based unregulated supply', the older kinds with just trafo-rectifier-DC buffering caps.
If one makes an uncontrolled SMPS (just constant 50% duty cycle switcher), will it not perform the same way as the older unregulated supplies?
It would perform a little different . For higher switching frequencies no longer needed big capacitors ,because they may just overload power supply output when discharged . In example, for 50hz are used 4700uf and more per output , depends how low ripple expected ,for 67khz is more than enough 1000uf to suppress that switching frequencies ripple . But if you make a high current load , transient voltage drop may be higher in smps , because capacitor will discharge faster (lower capacity) . Low freqency power supplies mostly have only power switch and fuse in high side . SMPS have inrush current limiting resistor ,filter resistance ,diode bridge drop depends on load ,and with load DC voltage drops on bulk capacitor (300v),so switcher will output less voltage to output . Like in big trafo supplies ,everywhere are internal resistances , which cause output voltage drop depending on load . Regulation is good thing ,but feedback loop need to respond fast ,but not oscillate if load changed quickly .The higher is switching frequency ,better can be reguliation in response to dynamic load . I think good thing should be properly made PFC in front of unregulated smps , it will compensate dc drop with load on high side capacitor ,and output will no longer depend on mains voltage fluctuation ,which is a some regulation too ,also 100hz ripple should decrease alot . Correct me if i'm wrong .
The topology of a SMPS also makes a difference as a full bridge will be semi regulated but a fly-backs voltage is almost totally dependent on the load.
Last edited:
Take a look at Hypex PSUs suplied to matcheir Class-D amplifiers.
SMPS, but they don't regulate the output voltage.
SMPS, but they don't regulate the output voltage.
Well, yes and no.
The higher switching frequency only helps against the ripple and noise.
Which is actually A LOT easier to get rid of with higher frequency.
On the other side you need enough "buffer" to provide enough energy for the dynamic loads, aka for the audio signals themselves.
That just always has to come from big enough capacitors and/or a properly made control loop.
But there is only so much you can do with a control loop.
So yes, a PFC would work very well, also just to create an universal mains input.
Hypex SMPS aren't very special, just another TL494, only with a lot more buffer storage (capacitors) vs standard SMPS