[Steven Everett]

I've finished building my first Szekeres MOSFET headphone driver and now need to finish the power supply. It has a low hum. I've decided to build Elliott's "Simple Capacitance Multiplier Power Supply For Class-A Amplifiers" I'm trying to find all the parts and I have no clue what a BD139 Driver is. I've looked through my electronic catalogs for that part number and can't find it. Can somebody give me a name that Mouser or Digi-Key would use? Another question; is there another name for the output transistors TIP3055? I've found this number but I would like to get all the parts from one company if I could. The Capacitance Multiplier project is located here:

http://sound.westhost.com/project15.htm

Thanks,
steve

 

[Apheared]

Hey, he means the driving transistor... calling it a driver when it's taken out of context; it's not a speaker, it's not a golf club, and it's not a guy with a touring cap.

Drivers. hehe. These and the output transistors they drive in thier config are just suggestions... I'm sure many many different ones would work fine.

What is your current supply like? I mean instead of building a multiplier why not stick a (traditional) hefty filter cap after the rectifier? a 6.8k-22k uF high ripple low ESR cap is about $3-10 usd. Place it after the bridge and see what it does for your hum. If you still want to build the cap multiplier circuit, you can still leave these caps in as filters.

For a power amp, sure. for a ~600mA headamp?
Your call.

 

[paulb]

Hey, Steve, I'm building the same thing you are including the capacitance multiplier.
A TIP3055 is a plastic version of the classic 2N3055 power transistor. Another version is the MJE3055.
You can use a single "Darlington" transistor instead of the BD139/TIP3055 combination; I'm using (something like) a TIP142. The BD139 could probably be replaced by almost anything NPN, even a 2N2222 for your headphone amp.
I'm using a cap multiplier mostly because I can use the same circuit to cut down my supply voltage from 25 to 15 volts so I can use my existing power supply. I'm driving the base of the transistor from an LM317 regulator, which I think should be better than using a zener diode and the passive 2-pole filter shown on the ESP circuit. I have a separate cap multiplier for each channel, but both are driven from the same regulator.

 

[Tomo]

Hey,

Wait a min. You can't just wontonly replace transistors. They have very different specs and they do not suit certain cases. It happens that you can use most any type of speedy power transistors for this circuit ONLY.

Tomo

 

[Steven Everett]

Thanks for the replies. I'm not sure how much current I'm using but my supply is 15 volt unregulated at 3 amps. I thought about adding a 12 volt regulator and a Darlington to boost the current but the plans for the Szekeres headphone amp called for 15 volts. I also thought that if I didn't have clean power my Darlington would boost the ripple. Elliott's articles on amps talks as if amps don't mind if the voltage drops a little when being pulled on as long as it doesn't clip. What amps do not like is a ripple on the voltage. I have 12,000uf of capacitance on my power supply right now and still have a hum. I bought a book on power supplies, from Radio Shack, and tried to figured out how much capacitance I needed to clean up my supply. I ran into trouble as soon as it ask the question on what percent of ripple I would allow. I didn't know. Since lambda power supplies are well spoken about I went to their spec. sheet. They have a .1% ripple. When I plug that figure into what I am using, 15v at 1amp (I'm not sure how many amps the Szekeres needs), and use a .1% ripple I ended up needing something like 160,000uf of capacitance. This is why I was looking at the capacitance multiplier. I see why many just buy their power supplies. In the end it is cheaper. But I am using these projects as my electronic schooling. I got to go through it to learn.
steve

 

[paulb]

Tomo's comments should be regarded; this circuit is not critical so substitutions are easy but this is not the case in general.
Steven, you'll drop some voltage across the transistor, especially if you use the Darlington arrangement Rod has. Darlingtons multiply the gain at the expense of more voltage drop, probably 1-2 volts.
If you're going to drop a volt anyway, you could just try a simple R-C filter; with say 3 ohms per channel (drops 1 volt at 330 mA, use at least a 1-watt resistor) you'd need maybe 4700 uF cap per channel. The cap multiplier would still be better, though.
It would be nice if you had a scope, you could measure the ripple under load. I wonder if your hum may be coming from somewhere else. Can you tell if it's 60 Hz or 120 Hz, if it's the latter it is probably ripple, if it's 60 it may be a ground loop or a wiring-related problem.
Or 50 / 100 Hz depending on your country.

 

[Steven Everett]

I don't know if it's a 60hz or 120hz hum but I think it is 120hz hum. What leads me to this conclusion is when I tore apart the power supply, looking for a ground loop, this is what I found. There is a transformer with two lead on the primary side that plugs into the wall. No ground. There are 4 leads on the secondary side. Two of these leads have diodes on them and two leads twist together and go through a fuse. There is a resistor and a capacitor across the output. There is a soft hum always being played regardless of the input. Whether the CD player is on or off, loud or soft, it is always there. When I remove the power supply and take it out to my car, I hook it up to my car battery the amp is dead silent. What do you guys think? Just add a whole bunch of capacitance to the output? Should I accept the voltage drop and regulate? Should I accept the voltage drop and add a cap multiplier?
steve

 

[Apheared]

There is a transformer with two lead on the primary side that plugs into the wall. No ground. There are 4 leads on the secondary side. Two of these leads have diodes on them and two leads twist together and go through a fuse.

Ah. Could this be one of those 13.8V car battery eliminators? What it is, is a dual secondary run series and center-tapped full-wave rectified... and the fuse is probably a thermal fuse style... (does it look more like a light bulb inside than a fuse?)

If it's a standard laminated transformer it could be radiated noise; no matter how you clean up the power it'll hum. Also is this in a steel case? Is the frame of the transformer grounded to the case?

You can test the radiated noise easy enough... make some 6' power extension cables and see if it lessens.

 

[Tomo]

Hello,

If I were to use 4700uF capacitance after the rectifier bridge, I should get about 0.5Vrms ripples. This is not bad. You shoud NOT use massive capacitance. This will become problematic since when the capacitor charge and discharge, the surge of electricity will damage your components.

Regulators will make it 10 ~ 100uV. That is very good. Most power supplies used in audio are about this quality even for high level hifi equipments. Many people try to go much below this point but their PSU end up noisy, costly, or wimpy.

Capacitance multipliers are not a good choice. Its benefit is low power dissipation but here we are taking less than 1A. The power dissipation is NOT our problem. It is very important, however, to have the voltage very highly regulated for our purposes. Thus you should stick with linear regulator design.

Note if you use opamps, you do not need high regulation since it has its own regulator stage within nowadays.

Tomo

P.S. 100dB ripple rejection Walt Jung regulators are the favorite among HeadWizers at the moment. (Except for me who already built and stocked up LM317/337 PSUs. I found it more than satisfactory for opamps. I will need better one if I were to build discrete design without PSRR.)

 

[Steven Everett]

Apheared were you looking over my shoulder?

Ah. Could this be one of those 13.8V car battery eliminators?

That's it exactly, right down to the little light bulb looking fuse and the steel case. I don't know if the transformer is grounded to the case but the transformer is bolted directly to the case. Okay, I won't use this power supply and I'll build another one. I have a couple of 25v 2amp transformers I salvaged from some old equipment. How can I tell if they are standard laminated transformers and therefore noisy?
Tomo, thanks for information. I didn't have a reference voltage to work with before your reply. Now I know what to aim for (10 ~ 100uV) to make a good clean PS. Where can I find Walt Jung regulator articles? What do you guys think of the switching power supply written by Stuart Rubin in this months AudioXpress magazine? Would you try it?
Thanks,
steve

 

[Apheared]

Not watching over your shoulder, but I've been down that (bumpy, dirt) road... And Tomo's criticism of heavy capacitance aside, you can't filter out everything if the transformer is radiating it.

I built one from a standard enamelled surplus transformer with 480,000uF - no multiplier, but 4 120kuF paralleled... Tomo's grumbling about the mega-amps of inrush this made on that poor diode bridge, but it didn't matter... it was still hummybuzzy at loud volumes. You couldn't hear it @ 100mW or so, and most people don't listen anywhere near this loud anyway... but it does make the sound grittier/harsher even though you can't directly hear it until ear-splitting volume.

Use a split-bobbin or toroidal transformer. Nothing else. Not for most things audio, and certainly not for a Szekeres.

re: switcher... We all think switching power supplies suck for audio, especially things for a Class A amp!? But times they are a-switching, and eventually someone will make a switcher that everyone loves for audio. Could this be it?
(I don't have a subscription to aXp, haven't seen it)