Any Single Low Impedance Fast PS PCB?

[kokoriantz]

I am looking for a fast low impedance single supply 45v regulator PCB. I like the dadod's but it is dual. Any proposition?

[kokoriantz]

I suppose there is not any available, so I designed one by myself.
LM1875 SUPPLY2.JPG
It has an impedance of 0.012 ohms resistive up to 20khz. Requires a shameful 800uF in compared to dadod's 220uF.
This is my first ever design of such PS, any comments is hepful.
Hayk

[geraldfryjr]

Hi there kokoriantz !!

Here is an interesting thread that may be of use to you, this link is of a design I did a while back and is still working perfectly to this day.

Lm3886 problem

It is very fast and very low noise as I had measured it, but I don't recall the exact switching recovery time but it is fast for sure when switching from a 1 amp load to a 4 amp load.

As built it is only used up tp 22Vdc output right now and I am actually in the process of building another one to be used at higher voltages and higher currents as specified in these variations of the same design.

Lm3886 problem

My next build will use a 35Vrms transformer for an input voltage of 49Vdc and a current rating some where between 5-10 amps or so.

Do read through the whole thread as I think that you will find some very useful information on such current boosted IC Regulator types and I explain why/how some of the circuits that are all over the net, but just don't work good, for several reason as discussed in the thread.

Mainly because I have built those other versions just to find out how poor the performance of those circuits were and why I blew out many good transistors because of such a poor design !!

When used properly the LM317HV is an awesome device.

I never got as far as current sensing/limiting circuits for it but they will be employed in my next build for sure.

I don't know exactly what the regulator circuits output impedance is yet, but it is low enough to drag down my current 18v transformer pretty hard and still give me about 3.5-4 amps out of it at about 14Vdc or so, that data may be already posted in the thread.

Cheers !!!

jer

[russc]

If you like the one in your opening post why not just build half?

[kokoriantz]

This PS is intended for eventual LM1875 based PBTL, to propose. Dadod sells his PCBs with very little margin to help him add some butter in his soup. I will feel guilty to copy his circuit and see others make profit of it.

[kokoriantz]

@geraldfryjr, I will try out on sim. your circuit to see the speed. How much uF capacitor you have on the output?

Hayk

[kokoriantz]

I simulated the circuit. First LM317 needs more than 3V drop out to function, so I increased to 26Vin. The impedance is extremely low in uohm range, and totalt inductive.original.jpg
I modified the circuit, first the 47uF on the opamp output goes away. A 0.01 ohm resistor necessary to limit a resistive component, which can be the wire linking the supply to amp, 100uF capacitor at arrival. The reference is TL431 2.4V, The result supplyfry.jpg .
Thank you, I like the idea to use the LM317 as driver.

[geraldfryjr]

Nice data!!

The schematic I posted is exactly how it is built with the .22uf cap on the output.
I am sure that it could be improved but as is it works very nicely.

As I had mentioned I will be working on a larger version shortly and I am sure that a few values will be changed.

I only used as much capacitance as was needed on the output to keep it stable nothing more.

If you don't use the NFB as in the other versions found on the web, under high current conditions the LM317 will try to apply to much Base voltage to the output transistor and it will blow it every time without it because it doesn't know what it is trying to track.

A hard lesson learned the hard way !!

The Dropout Voltage I stated in the thread was as measured in the real world as well.
LTspice was very very close to what I measured and was under 3v.

The 100uf I used on the Vref was as per the data sheet info of that device too lower the residual noise even more.

jer

[kokoriantz]

I am very thankful for your posting. As playing around with your circuit, I got new ideas and applied them. Your very low impedance source gives the means to supply two channels, as the wiring resistance is several folds higher. I decided to supply each channel with independent regulator to be very near to the channel and also limit the current of a single. I will apply as dadod's speaker protection to shutdown by SCR until reset.

[geraldfryjr]

Here is a screen shot of the noise floor out of a Class A amp to I am working on using that power supply circuit set at about 1.5-2 amps of current through the output stage of the amp.

The rising of the noise floor with frequency has to do with the amp itself I am designing and not the power supply.

I have observed the noise floor before under static tests up to 3.5 amps and were as stated in the originating post.

FWIW

jer