Choke regulated power supply for S/S amp

[Matt Rowland]

Can anyone tell me how to design a choke regulated power supply for a solid state class AB power amp of around 120watts?
Musical Fidelity seem able to do this with good effect, yet I can't find any DIY examples, but I'm sure someone must have done it!

Cheers

Matt

[MRupp]

I built a choke input (LC) powersupply for a SS amp, it has single polarity (50V) and is for a Class A amp at 2A DC bias, but the same principles apply and I am planning on building a bipolar PS next. So basically this can be done but it costs some money as chokes can become really big.

Now, 120W output is not a meaningful piece of information, you design an LC power supply for a given voltage and minimum fixed current (your bias current). The basic formula to calculate the minimum (aka critical) inductance is:

Lc=Vout/Iout at 60 Hz or Lc=1.2*Vout/Iout at 50 Hz, Iout is in mA

First of all, Vout is 0.9 times Vrms whereas it is SQRT 2 * Vrms in a capacitor input filter, and if your current draw goes below the critical value your voltage will rise quickly, so use capacitors with a high enough voltage rating to avoid disaster. If well (over) dimensioned the time constant of the power supply should be lower than the lowest audio frequency and you should not have an issue with AC current fluctuation, but more bias current is better (or Class A), and choose higher inductance values if you can.

Lastly: Get PSU designer from Duncan amps, and also search the forum as there is much more information to be found. Sorry, don't have more time now.

[MRupp]

P.S. I am assuming that you really mean a choke input filter. There are off course simpler filters that have an additional choke after the first input cap (LCL), but I would not call them choke regulated.

[Matt Rowland]

Vielen Dank MRupp.
I do mean choke input (LC) style. I will check out the Duncan amps psu designer. It does seem that there are some other issues to look out for other than the basic calculations though.

Cheers

Matt

[analog_sa]

An obvious issue is what happens to the core once the amp draws 5A. Do MF use proper choke input indeed?

[MRupp]

I think the low idle current draw (compared to class A) may be the major issue. You need a high inductance and still have to pass high currents at full load (I would concur that you then come close to 50 % of the max current draw). I am using (overdimensioned) 300mH chokes rated at 4A rms and with 0.7 Ohms DC resistance and they are big, 150b core and weighing more than 10 kilo a piece.

A potential solution might be to use swinging chokes with high inductance at low currents but I do not have any practical experience.

Oh, and you can off course combine two chokes into one for a bipolar PS by using two windings, which might save some space. I saw such commercial chokes but doubt that they have the inductance that you need, so a custom wound choke may be in order.

[analog_sa]

Martin, your chokes seem perfect for choke input. Having basic knowledge of MF products i suspect they are indeed using swinging chokes.

[woody]

You might try using a transformer as a swinging choke.
True I have head a transformer can't be used as a choke
because it doesn't have an air gap to prevent saturation.
But I found that at least one defination of a swinging choke
said it doesn't have an air gap. I tried a couple of standard
transformers in a choke input power supply and they worked
fine.

[AndrewT]

one can take apart an EI transformer core and reassemble it with all the plates facing the same direction. It's then very simple to add a paper or plastic spacer into the E legs to create the required flux gap.

[av-trouvaille]

Hello,

In a standard C-L-C supply filter the L will see much DC-current (depending on the value of the first C). Then the DC-properties of a choke will matter; in case of a mains transformer the reassembling of it with all the plates facing the same direction as indicated by Andrew T is needed.
With a choke input filter (L-C) the L will see almost no DC current, while the smoothing occurs after the C. So the properties of the L might not be so critical compared to a standard C-L-C filter. Eric Juaneda even uses a toroid in his choke input filter (the toroid type is known for hardly being able to handle DC current). Furthermore, in his view the output of the L should preferably be a sine-form, from an audible point of view. See his site at:
http://tech.juaneda.com/en/articles/lm317.html

A comment can probably be raised on the type of rectification. A full bridge would deliver 100 Hz and the choke should be able to cope with that (mains transformers normally are limited to a small frequency range around 50 or 60 Hz).

Anybody care to comment?

Best, Arjen.