Re: The KISS AMP: a progress report


Sander deWaal wrote:
"pfjw@xxxxxxx" <pfjw@xxxxxxx> said:

[silly bit of transformerless table radio gimmickry about silicon
diodes on one winding snipped]
If a suitable transformer were not available, I would
[another silly idea snipped]

Even a blind squirrel must find a nut occasionally. I looked up the
chicken scrawl on the clipboard from my workbench. I used a
250-0-250Vac secondary transformer with 330 ohms of ballast in each
leg. According to my notes the alternatives were:

a) special-order a proper power tranny 240-0-240. Gee, just 10 volt
difference?

b) use the tranny already on the chassis and drop c110mA in the
bleeder; this lets you use the GZ34 with a standard spec sheet 100 ohms
of ballast in each leg.

c) use the tranny already on the chassis and boost the ballast
resistors to finetune the voltage, as the makers of the GZ34 clearly
intended you to do because they limited the capacitance you may use.

Of course I went for option c! Those ballast resistor are supposed to
be there, regardless of how they bother Peter Wieck.

Sorry, Peter, the other blind squirrel got the nut.

While the resulting DC voltage will be lower in that case, the ripple
is certainly going to be far greater than in case of the dual-diode
power supply design.

Getting rid of ripple means big caps, pushing the average DC voltage
up again.

More ripple also means higher charge/discharge currents, which in turn
can cause hum induction in the actual amplifier circuit.

Dropping resistors don't load the windings, bleeders do.
A vacuum rectifier must see some DCR in the anode circuits, to prevent
cathode stripping.
A beefy tranny has a lower DCR, so series resistors *must* be added
to avoid large cathode peak currents caused by the electrolytics.

In case of a class A amplifier, burning some volts and watts away in
series resistors isn't a big sin IMHO, unless one is designing to
utmost efficiency.

A fellow of Scottish extraction owned one of my favourite small tube
manufactories. He delivers a set of transformers to me, of which he
says he has plenty on his shelves. I am to design an amplifier using
these transformers. He also gives me one unit of the amp currently
built with them and says I must use the stainless steel topplate from
that. So I reverse engineer this existing amp and discover that the
other designer (a much-respected Englishman) ran the poor power supply
tranny to within 1mA of total rated capacity. This is clearly a
monstrous miscalculation but I can't find out where I went wrong, so I
call the tranny designer and say to him, "According to my calculations,
this tranny is stressed out to within a milliamp of its life." And the
tranny designer says, "Don't tell Mr X! He will make you design for the
last milliamp as well! And don't put the amount of current you want to
bleed on circuit or he'll try to bargain you down!"

Peter must have learned that sort of "production economy" somewhere and
had it impressed on him for life in order for him at this stage of his
life, in ultra-fi amps at that, to be so worried about 10 volts and
possibly three or four buck wasted in over-rated resistors -- on a
disposable part-prototype to prove a sub-assembly!

Looking at the schematic again, I noticed that André actually uses the
Lundahl LL1638 choke as a swing choke.
As this choke wasn't designed as such, it won't regulate as good as a
correctly designed swing choke (without air gap).
It will provide some additional voltage drop and series resistance,
though.

I've answered this in my first reply to this letter but there is an
additional fine point about the chokes that is easily overlooked unless
you already know about the trick. It is mounting a split choke with two
coils on common iron to straddle both voltage rails; this is a super
means of achieving that extra edge of common mode rejection that in the
long run will make the amp so much more satisfying that it is worth any
amount of money for chokes with this capability.

--
- Never argue with idiots, they drag you down their level and beat you with experience. -

I dunno. I would never have discovered half this stuff if I weren't an
idiot who would ask any question and not care if the clowns sneered and
jeered as long as someone gave me the answer.

Andre Jute
Visit Jute on Amps at http://members.lycos.co.uk/fiultra/
"wonderfully well written and reasoned information
for the tube audio constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site
containing vital gems of wisdom"
Stuart Perry Hi-Fi News & Record Review

.

Relevant Pages

  • Re: The Future of US Kids Making Stuff...
    ... To your own design -- or using a layout from a magazine or other ... Almost certainly your own design of circuit, ... taken as verbatim as you can manage, then "assembling". ... paper to make resistors, and using tinfoil and wax paper for ...
    (rec.crafts.metalworking)
  • Re: Prefered resistor range
    ... >> But after designing the Nth precision active filter, ... >> for a short list of resistors. ... >ranges - smt design optimisation usually involves minimising the number ... And the 5% caps we were using were similarly close to ...
    (sci.electronics.design)
  • Re: SMPS design ...
    ... In the drain of the FET, is a single primary winding up to the raw rail from the input bridge. ... On every one of the examples sent to me, the two resistors are chalky and very discoloured to the point where you can't read the bands. ... As far as I can make out, doing some on-line reading about SMPS design, this network across the transformer primary, is a simple snubber. ...
    (sci.electronics.repair)
  • Re: Discussing audio amplifier design -- BJT, discrete
    ... and biases the output transistors to idle current using a couple of ... Parallel the emitter resistors with diodes. ... Discrete-transistor audio design, being such an ancient practice, ... Hey, Mr Audio, show us a power amp that you designed. ...
    (sci.electronics.basics)
  • Re: Tempel Magnetics
    ... the stack height, tongue width, window length or window height is. ... change, but in a real tranny design, going from 2A/sq.mm to 3A/sq.mm would cause an enormous design change to size and turns. ...
    (rec.audio.tubes)