Inductors in Power Supplies

[u1022186]

Hi Everyone.

I started a thread in parts about capacitor choices for the power supplies I'm building for my new amplifier. I'm planning on using the Aussie Amplifier HPA-nxV500 R2 amplifier boards. I'm going to have separate dual mono supplies for the input & output, but all in the same large chasis. The required outut voltage is around 78 volts. I already have 2 - 1000VA 58-58 and 2 - 250VA 58-58 toroids that I bought from someone at a good price for both supplies. He bought them a couple of years ago for a project that never happened and they've never been used. That would make the voltage 81-82 on each rail which should be no problem. I've been looking over and reading all the posts on power supplies as well as many online articles. I see in most Class A amps, inductors are used between the caps to reduce noise. I was wondering if this makes an audible difference and why you don't see it in any of the class AB amps. I originally planned on using 8-Panasonic 12,000 uF 100v caps for each channel for the output and 6 of the same for each channel for the input. I was just wondering if adding inductors was something worth considering, and if the amount of capacitance I'm using would be enough to prevent any oscillations from the inductors? Also, do the inductors cause the output voltage to decrease? And how do you determine the proper value for the inductors. Thanks in advance for any help that can be offered.

Best Wishes,
Bob

[darkfenriz]

Inductors cause signal-dependant voltage sag in class AB

[AndrewT]

Yes, add an inductor between first stage smoothing and second stage smoothing. Do not make it too high in value.
Ensure the second stage smoothing can meet all the demand of the amplifier, i.e. size the second stage smoothing as if it alone were all that existed.

[CBS240]

A wire-wound resistor is as far as I will go for a class AB supply. Like Andrew says it should be between the capacitor stages. C-R(L)-C-C-Cfilm. In this circuit, the coils in the yellow rectangles are resistors. I needed ~100mOhms. 6ft of 22Gauge magnet wire is close enough, and cheaper than a power resistor since I had plenty of scrap magnet wire and stand-offs. It is wound around 6-32 size typical aluminum stand-off. As for inductance, it is negligent. Also it is important to seperate the return currents from the different capacitor stages. Each capacitor stage should have it's own return path to the star ground point so as not to mix power supply AC input currents with speaker return currents...(Bbbzbbzbzbbzbzzz)

[Loudthud]

An inductor will cause a sag in voltage in a Class AB amp, but the resultant DC voltage will be higher than the low part of the ripple would be without it. Also, the ripple will be close to a sinewave instead of a sawtooth.

[u1022186]

Thanks for the suggestions. I was wondering how much of a voltage drop the resistors would cause as opposed to inductors. Would something like 5 watt Mills wirewound resisters be OK. Also, how do I determine the proper ohms value to use. In a wiring diagram I saw in an older thread, it shows just as you said with an inductor after the first cap on both pos. & neg rails and lots of uF after the inductor and a film and bleeder resistor at the end. But it showed all the 0 Volt in center wired together along the enire string. So how would I separate the front and rear sections to ground. I'll actually end up with 4 separate supplies ( 2 mono lower current supplies for the input and 2 mono high current supplies for the output. I'd want to try to keep the voltage into the boards at least 78 volts for all 4. Wouldn't the resister approach cause a much greater voltage drop. I would need either 8 inductors or 8 resistors. Would it make more sense to do this just for the lower current input stage that might be more sensitive to noise, or would all 4 supplies be the best approach. I'm already nervous about how to ground all of this for no bzzzz or hum with 4 power supplies in the case. Would everything go through just one ground loop breaker circuit (bridge, resistor, cap) tied to earth ground? So many questions. While I can read a schematic and have a basic idea of what's going on, I don't have the knowledge to design my own circuits. But I love getting out the soldering iron and building a case and stuff like that. And when I started the rebuild of an old Hafler DH-101 preamp about 2 years ago, I started reading tons of old threads and books on audio design & theory. Some I understand, and some I'm still trying to figure out. The preamp came out great and I could hear a major improvement. I have an unmodded one and it's like night & day. The only change I might try in the future is either a Salas low voltage regulator based supply or one based around the super regulators. Of course it took me over 2 years to finish it. But I'm patient. I built a pair of mono Leach amplifiers around 35 years ago. That took me over 3 years. I remember how great they sounded. I gave them away about 15 years ago when we moved (they were really heavy) and now I'm looking forward to building something again of at least equal or better quality. The amp modules are not DIY as they come all finished, but all the rest is, and that's enough for me at this stage. Thanks Again for all the help. Very, Very Much Appreciated.
Best Wishes,
Bob

[AndrewT]

That long paragraph is difficult to read.

I gave up and rest my browser to narrow window, to get the wordy columns down to a readable width.

[AndrewT]

Use PSUdII to model your supply proposals.
Compare zero L and Zero R to a little L and/or a little R between the capacitance.

The results only show the mains' ripple attenuation. There is also the harmonics attenuation that is MUCH greater and should not be thrown away just because some say you will get rail sag.

Using 0r1 in the rail gives a bit of volts drop due to continuous current demand, but almost no volts drop due to transient current demand. A little uH will also cause a tiny Volts drop in the supply rail, but less than using R alone. The big advantage to using uH in the rail is the attenuation of the harmonics and especially the spikes coming from the mains and from the rectifiers switching off suddenly.

It is so easy and cheap to wind 50 to 100 turns of 0.6mm diameter enameled copper on a 10mm former between cheeks set 6mm apart, that I now adopt this for all my assemblies. The only variable is the wire diameter and I have a selection to suit the continuous current demand.

[u1022186]

Hi Andrew Thanks for explaining things clearly. I went to the site of PSD2 to try some simulations, but the instructions indicate it only works with older versions of Windows. All my computers have Windows 7 64 bit. Can I install it on my computers, or will it mess things up?. I don't have access to any cmputers with Windows 95, 98, or NT. I could borrow a neighbors laptop with Windows XP, but I wouldn't want to install the software if it could mess up his laptop, as XP is not mentioned as one of the accepted platforms. Thanks Again, Bob

[kroto]

most of 32bit apps can run on 64bit system.

if the program can't run in win7, try to change its compability, by:
~ righ click on the *.exe file,
~ Properties >> Compability >>
~ check on "Run this program in compability mode for:"
~ chose one
~ click apply/OK.

try it,,