Gearing up to build the Amp

[JCM]

Just got my shipment of parts. Heavy duty bridge rectifiers, some polyester film caps, electrolytics, Zetex small signal mosfets, wire wound pots, and THE mosfets - IRFP044 - they are stout, indeed. (And I'm very glad to NOT require the very hard to find plastic grommets / sleeves, used to bolt TO-220 case devices to the heatsink !)

Now for the rest : I am currently looking at power transformers, power supply capacitors, and some miscellaneous items.

I will use a 24 V / 4 A transformer, maybe 24,000 uF total capacitance, for each channel. The current draw will be less than 1.5 A per channel. I figure on using one or two half-ohm power resistors for some RC power supply filtering. The amplifier is to be the ZEN 4 / Penultimate Zen.

I was planning on using a capacitance multiplier circuit. I figure V+ would be about 25 to 28 volts out of the capacitance multiplier. That's about enough for 10 Watts output. Then I started thinking. The original Zen didn't have any sort of regulation, and drew 2 amps with 20,000 uF capacitance. And for 10 Watts, I'm cutting it close. I figure I could have a good 30 Volts without the multiplier.

I will integrate the BoZ into the amp - and will attempt to pull a fast one, by creating a voltage doubler to feed it.

My thinking is becoming more conservative now that it's "down to the wire", so to speak. Simple is good. Some admonishments and dire statements discouraging the use of regulation for a power amp come to mind - though a multiplier isn't really a regulator, it's - sort of - a regulator. Maybe it's better and more natural to just let the reservoir capacitors do their job, remove the extra transistor, not making things any more complicated than need be.

I discovered I could buy Sprague computer-grade electrolytics from Mouser at these prices :

40 V / 7100 uF = $7.36 Each, 10 for $66.20
40 V / 8700 uF = $7.30 Each, 10 for $65.70 (Powerlytics)

and similar. Then there is a 50V / 10,000 uF unit for $8.08, ten for $72.70.

"Never use electrolytics caps. far above or far below their voltage rating". So 35 volt units would be acceptable, (however, not always available. Could just about get by with 30 V units)

I'm tempted to buy 10 for the reduced price, leaving some left over for my next amplifier project.

I could have sworn some of them were rated at 105 C - they aren't - all of them are rated at 85 C. I can't seem to find CG caps @ 105 C at any reasonable price, if at all. But the smaller, snap-lock types, and others, abound at that temp rating.

I searched for this topic, but haven't dug it out yet: What about using 85 C caps ? Are the 105 C units so much better ? I know they're preferred. But it just seems like a great big Sprague Powerlytic - 3 of them, per channel, actually - is better than some much smaller units.

I have a few parts left to track down yet - was hoping to find high quality DUAL pots - need two of them - maybe a Bourns unit - but those are elusive. I figured spending around $15 apiece = I just don't want to use those cheap two-dollar non-sealed carbon pots.

Comments are welcome !

I will have to put my Fostex / Jericho / J-Low &c. plans on hold, the amplifier must be built first, and will eat into my budget too much for now.

[JCM]

THE PREAMP At bottom of page

Modified BoZ. The mosfet is the same - IRF 610. Power feed through RC filter not shown. There is likely to be a tad higher noise, due to less filtering of the gate bias voltage. Yet the infinitesimal amount of noise NP quoted in his circuit was nowhere near what I was able to achieve. However, I used a simple voltage divider bias, simpler than shown here.

I imagine that the performance would not deviate significantly from the original BoZ circuit. Ever since I caught on to the Zen &c., I have sought to make it even simpler, one way is shown here. With this volume control arrangement, I can avoid a 15% reduction in output amplitude - a welcome development, since what follows is a simple EQ network. It also creates a lower output impedance, which is also good due to the EQ network loading.

I also have an oddball kind of - version of the Zen V4 - using the same - at first glance - principle as the Aleph current source. But I believe it isn't really the same thing. I plan to analyze it side by side with the Aleph, both versions very low powered. There is another version which is simpler yet, however the current source in that case is not a constant current source. I haven't much faith in it, and am looking to attach a few parts to it to make it act the way I want it to.

Looking at the schematic of a Zen+Aleph amp, one can immediately grasp some of the basic operation, and imagine alternate methods of implementation. Therefore, what I have cobbled together on the bench is not likely to be of any note or surprise.

One note is made here : It seems to me that the Aleph current source is deceptively clever. My initial thought was "regeneration - positive feedback" - and yet it isn't. Positive voltage going to the base of the small NPN turns the current source off. But there is more to the story. In a way I have not grasped, it creates this - ghost image, acting in oppostion (?) to the change in current the current source normally would counteract from the output swing of the audio driver mosfet.

[JCM]

Yet the infinitesimal amount of noise NP quoted in his circuit was nowhere near what I was able to achieve.

I meant, of course, that the noise in my circuit was far above the noise of NP's circuit. Ten or twelve micro-volts ?? I wondered if it was a misprint. The beam heighth of my scope is maybe 100 uV, hard to measure much below 500 uV. I checked my DMM, resolution is said to be 10 uV at most sensitive range (200mV) - this being a 4 1/2 digit unit. I have my reservations about it.

But in my experience, once I got the noise down around 1 mV, it was dead silent - the power amp noise is the harder one to tame.

Here it is again.

[Zen Mod]

go to http://www.apexjr.com
look at "new arrivals " or something like that

you'll find dirt cheap elkos

my bad that I have not surplus places like that in my country


edit:
his web sometimes is not functional-just be patient

[JCM]

Thanks for the link, I will see.

I got some Grab Bag specials from Alltronics a few years bag - a good deal, I thought. In the capacitor bag, (I ordered 2 - which is 10 pounds of stuff) there were a lot of computer grade caps, among other things. Of course, they seemed to be about 10 years old. I still have them, and they're still good, but brand new is better

I don't mind shelling out 7 to 10 dollars or more each for reservoir caps. But 40, 50, 60 $ for each - no.

This will be the first actual hi fi amp I put in a box, and not a running experiment.

I just need the reservoir caps to filter 1.5 Amps continuous - I figure maybe 3 of 'em, for each channel. Each channel will have it's own transformer, bridge, and reservoir caps.

Should I use 85 C computer grade caps or 105 C snap-lock or other caps ? Or maybe the difference would be slight. Just seems like it'd be nice to have some REAL filter caps in the supply - Sprague Powerlytics (85 C ).

***********************************************
Sprague is a darn good product, they've been around a while
(I still have over a dozen Sprague Atoms, from years back,40 and 30uF 450V units - never been used. The instructor in semiconductor class was going to throw them out - none of those other clowns wanted them - can you believe it ?? Sometimes college has it's benefits ).

Anyway, it just seems like 1.5 amps = actually less than that - isn't so much to ask a good capacitor. I have been the whole route, as we all have - accidentally shorted them out, (they do indeed MELT tips of screwdrivers, and create quite a spark), seen small caps blow (usually reversed polarity), seen them pop like confetti, all that, seen and heard stuff gush out of them. (Not that the Spragues would do that in my circuit ? -or, could I be mistaken ??)

I'm really nitpicking here, wondering if there's any good reason NOT to use those 85 C computer grades.

[jh6you]

If I give manufacturers claims a full trust, the life of ele cap very much depends on the rated operation temp vs. the actual temp. The decision of using 85C or 105C should go through that consideration. Anything else? I don t know well . . .