Has anyone tried to add an isolated compensation voltage to the HT voltage to provide low ripple on the voltage used by the amp?
One thought would be to use the 5V winding (of an old amp) or a small PT as an isolated source, which is rectified and filtered to supply a 317 voltage regulator, which generates a ripple compensation voltage, which when added to the HT, provides a low ripple output for the amp.
The compensation voltage would add to the HT, and so would source the supply current for the amp. The raw HT ripple voltage could be capacitor coupled to maybe a 358 opamp to invert the signal and couple it to the 317 adjust pin in a standard voltage regulator configuration.
Seems like a simpler approach to active filtering than using a HV regulator per se, but then again it is only a thought !
Ciao, Tim
One thought would be to use the 5V winding (of an old amp) or a small PT as an isolated source, which is rectified and filtered to supply a 317 voltage regulator, which generates a ripple compensation voltage, which when added to the HT, provides a low ripple output for the amp.
The compensation voltage would add to the HT, and so would source the supply current for the amp. The raw HT ripple voltage could be capacitor coupled to maybe a 358 opamp to invert the signal and couple it to the 317 adjust pin in a standard voltage regulator configuration.
Seems like a simpler approach to active filtering than using a HV regulator per se, but then again it is only a thought !
Ciao, Tim
Schematic sort of like the link:
http://dalmura.com.au/projects/HT comp.pdf
The 317 would source current to supply the load current requirement, plus additional current to raise the compensation voltage during a raw HT ripple dip, and reduced current during a raw HT ripple peak.
http://dalmura.com.au/projects/HT comp.pdf
The 317 would source current to supply the load current requirement, plus additional current to raise the compensation voltage during a raw HT ripple dip, and reduced current during a raw HT ripple peak.
The only issue I could see is the response time of the capacitor charge Vs current drain after opamp response. Would it be easier to take an average set point and respond to change (sample and hold)? This is thinking about a rise time of the compensation voltage as the HT ripple falls.
Then again if the circuit responds to mV dip the time constant would be small.
When you going to show us the test circuit waveform?
Shame you can't just phase shift the ripple and add it to the supply!
Regards
M. Gregg
Last edited:
The ripple cancellation technique has been around for some time, re-patented repeatedly. Google search on "ripple steering" or "ripple cancellation". It's used for switching converters and class D amps nowadays for both input and output current smoothing.
An early readable reference for the output case is in Electronic Designer's Handbook by Landee, Davis, Albrecht page 15-21. The common design is to use a center tapped inductor with caps on both ends and the rectifier feed at the center point. It acts like an inverting autotransformer to cancel the two cap ripple voltages.
http://www.national.com/appinfo/power/files/f22.pdf
http://www.hamill.co.uk/pdfs/azrtatad.pdf
An early readable reference for the output case is in Electronic Designer's Handbook by Landee, Davis, Albrecht page 15-21. The common design is to use a center tapped inductor with caps on both ends and the rectifier feed at the center point. It acts like an inverting autotransformer to cancel the two cap ripple voltages.
http://www.national.com/appinfo/power/files/f22.pdf
http://www.hamill.co.uk/pdfs/azrtatad.pdf
Last edited:
Thats interesting!
Wonder why you don't see any chokes wound like this for HT "B+" for tube amps!
Regards
M. Gregg
"Wonder why you don't see any chokes wound like this for HT "B+" for tube amps!"
I wonder why we don't see any OTs made the right way either!
LMAO, well its because making them is such a pain in the ***...SOOO they can sell any Cr**!
Regards
M. Gregg
In my experience with removal of undesired effects (ripple, charge injection, voltage offset, and the like), it generally works better to not generate the offending signal in the first place. Nearly all cancellation circuits work really well on paper, but suck in practice. In many cases, they even make matters worse as the cancellation signal is very hard to get to match the offending signal and there usually ends up being a small phase difference between the offending signal and the cancellation signal that makes life miserable for the circuit designer.
That said, ripple is a relatively low-frequency signal. I say "relatively" as ripple voltage does contain higher frequencies than just 100/120 Hz. Due to the LF nature of the signal it would probably lend itself fairly well to some sort of cancellation.
But I can't help asking the question, "Why???!" A cancellation circuit will need to be able to supply the full load current and will be at least as complicated to design as a voltage regulator. So why not "simply" use a supply regulator. The floating regulator I have going in the lab currently, actually works quite well.
~Tom
Ripple cancellation was commonly used in valve domestic radio receivers, typically via a tapping on the OPT primary. It died out as a technique when large cheap electrolytics became available. As it is now not hard to reduce ripple to the required amount (set by PSRR), I can't really see what advantage cancellation has.
The topic is presently only a 'thought experiment' - so sorry no waveforms yet!
I agree that the main benefit would be for low-order mains harmonic cancellation, and not high-frequency noise alleviation, due to an initial aim of a simple circuit implementation. It wouldn't be a means for a wide bandwith and high slew rate regulator, nor would it be of any benefit for a low ripple voltage supply (eg. high capacitance filter supply).
My own focus would be aimed more at guitar type amps, where one design aim is to retain relatively small levels of HT filter capacitance (eg. 16uF), and valve diode rectification - as per clones or restoration of older classic amps.
The intent would be for a very simple added circuit. The current supplied by any such compensation voltage would be typically less than 1Apk, and the average power rating of the isolated supply would probably be less than 1W for a typical old guitar amp.
I think the isolated supply could even be generated off a 5V winding used in a directly heated dual diode cathode. This is where the germ of the idea came from.
I agree that the main benefit would be for low-order mains harmonic cancellation, and not high-frequency noise alleviation, due to an initial aim of a simple circuit implementation. It wouldn't be a means for a wide bandwith and high slew rate regulator, nor would it be of any benefit for a low ripple voltage supply (eg. high capacitance filter supply).
My own focus would be aimed more at guitar type amps, where one design aim is to retain relatively small levels of HT filter capacitance (eg. 16uF), and valve diode rectification - as per clones or restoration of older classic amps.
The intent would be for a very simple added circuit. The current supplied by any such compensation voltage would be typically less than 1Apk, and the average power rating of the isolated supply would probably be less than 1W for a typical old guitar amp.
I think the isolated supply could even be generated off a 5V winding used in a directly heated dual diode cathode. This is where the germ of the idea came from.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.