Re: The KISS AMP: a progress report

Peter:

It is quite clear to me that you are totally out of your depth here and
in your other posts in this thread. Here are just some samples:

1. You appear not to understand the rationale behind the various power
supply topologies and in particular how a choke input works or why it
is traditionally built with two sections, meaning two chokes.

2. As we saw earlier, you appear to think that a resistor will give off
heat exactly matching its power rating, which is totally untrue; the
heat dissipated is related to the resistance value, the current flowing
in the resistor, the mounting and the air flow over the resistor.

3. You appear not to understand, or to have overlooked despite being
told (if you do it again I will assume you are doing it maliciously and
react accordingly), the fact that the T68 was made from a 300B proto
merely to prove the driver stage of the KISS amp, and therefore uses
the 300B amp's power supply regardless of its precise suitability. It
is not presented as the ne plus ultra of anything. If you seriously
want to discuss power supply design, discuss the T39 circuit, also
published at the same time.
http://members.lycos.co.uk/fiultra/T39-KISS-300B-Ultrafi-crct.jpg
Giving you the benefit of the doubt, I shall assume you're still
working your way down to. Work faster.

4. Your questions reveal that you do not to have the faintest idea of
the conventions guiding SE amp design towards the desired result.

In the light of these misunderstandings, and because I don't want to
repeat myself if and when we come to the T39
http://members.lycos.co.uk/fiultra/T39-KISS-300B-Ultrafi-crct.jpg
my answers to the rest will be shorter than usual. (Hallelujah chorus
starts up!)

pfjw@xxxxxxx wrote:
It would be useful to know the secondary voltage at that transformer
before the resistors are added, so as to get the 180V rectified voltage
you require. Yes, you have 500V ~500ma at the bottom, I assume that is
across the entire two coils (5-to-7) and not to the (virtual)
center-tap, but that too would be helpful to know.

The convention is that a 250-0-250 transformer is described as 500Vct,
usually abbreviated on the schematic to 500V. The centre tap is not
virtual but real.

Assuming that you are getting 250V/Coil at your stated input, the way I
read your output from the rectifier tube, that would come to some 415V
or so out of the 5AR4/GZ34, excepting the dropping resistors. So, how
did the value of 660 ohms of total AC resistance get chosen, especially
in light of the 4700 ohm bleeder resistor further down the line (at 25
watts) to get to the required 180V?

It'll be in a chapter of the KISS files on my Fiultra site when I get
around to writing the chapter. Basically you decide the current
required by the tubes, allocate a bleed (which in the absence of a
swinging choke must be higher than a certain amount in a choke input
design -- not essential in my design with the 11 second delay on the
GZ34 filament but I always put in a hefty bleed for stability, safety
and just in case someone else plugs in an instantly-on rectifier), add
the two together, work back over the copper drop of the chokes,
transform the voltage and current via the rectifier tube backwards from
DC to AC, and then apply to Colonel Ohm for the value of the ballasts.

And 'adding a small capacitor' to
adjust voltage... this is an iffy (albeit nearly heatless) expedient,
'cause if that cap should fail, *poof*.

You do talk a lot of manure, Wiecky. The cap will have the same rating
as any of the other caps, just lower value. It is in the same position
as the cap in a proper cap input filter. There is no reason for it to
fail or precedence for it failing. (The idea of the voltage-tuning
small cap in front of a choke input filter was given to me by Gordon
Rankin, whose caps would *dare* fail...)

So, it would be useful to know
the required/expected voltages right down the line so that the correct
choices might be made.

They are written on my circuit, every single one of them. What are you
talking about? (We might note that I am one of the few designers who
write all the necessy information on the circuit.)

Would it not be much simpler to drop the primary side to ~220V or so,

No, no, no, never! Your winder will come cut you, pal.

thereby almost eliminating the need for filament-voltage dropping
resistors excepting getting 5V on the rectifier?

Peter, you have these resistors on the mind. The desire to eliminate
them arises from your ignorance about how much heat big resistors
radiate when only a little current flows, and cheap production
paradigms. In good audiophile tube design, and especially DIY work,
such resistors are treasured for the advantages they bring, which
Sander has tried to describe to you.

SWAG, assuming 230
single-phase hot/neutral at the primary, and about a 2A (allowing for
losses) total load shouldn't take much of a single dropping resistor to
get that 220V, or even slightly less. This way, individuals in other
countries with different standards could work to the required
_secondary_ voltages based on their wallplate voltage. If one is fond
of iron, either a small autotransformer or a small bucking transformer
could be used ILO a resistor for less heat.

No.

This might also greatly simplify the requirements for your B+ side?

Again, this is a temporary chassis for a proto for a part-amp. If you
want to discuss power supplies in terms of simplicity, look at the T39
design, which is the whole amp.
http://members.lycos.co.uk/fiultra/T39-KISS-300B-Ultrafi-crct.jpg

Two chokes? Why?

Because that is how one builds a choke input supply. It cannot be
wound/specified/built more economically (Patrick?) or efficiently in
terms of smoothing than as two (equal) sections.

The 5AR4 is a pretty good, quite rugged (if expensive)
rectifier.

So what? I'm specifying a couple of grand's worth of tubes already.
Anyway, it sounds better than semiconductor diodes and lasts longer too
unless you heatsink diode bridges properly.

In the T39 the GZ34 will be replaced with a GZ37, which will really
hurt if you worry about costs!

And it would seem simpler to add additional capacitance on
the line-side of the first choke, thereby eliminating the need for the
second?

See above about efficiency and cost (cost at least in days of old --
maybe today a humongous choke could be cheaper).

Furthermore, the power supply is part of the transfer function. The
signal literally runs through it. In your mind's eye draw a line from
the bottom of the last filter cap on the T39
http://members.lycos.co.uk/fiultra/T39-KISS-300B-Ultrafi-crct.jpg
to the ground of the RCA socket, and you will get the idea. In an SE
amp you want control over all the time constants, and therefore in the
power supply too. You can't just throw around extra capacitance without
considering the specification of available chokes, the time constants
you want further on, and so on, even onto your speakers (you can't
output too much below 32Hz in Lowthers, for instance, so you that is a
time constant in the 300B circuit which influences time constants all
the way back to the power supply). Again, this is an example of why
Patrick despairs of repair hacks who won't hit the RDH hard ever
learning anything really useful about tube amps. I mentioned Rankin
above for a purpose; he is a very thorough engineer and he says that
*after he had designed an amp* there were an additional 100 or so
checks he makes to be sure the thing will work right.

With all your prior losses, you are running well below the
saturation level of a single unit (400/575ma). 500V caps are cheap
these days, even 50uF or better, certainly cheaper than iron. And if
all you want to filter is 50/60Hz chop, one will do fine.

No, no, no! See above about time constants in the filters, efficiency
in the choke sections, etc. Then, as a practical matter, check a
catalogue for the physical size of the polyprop caps I have specified,
and you will be absolutely horrified at the floor space already on the
amp.

As a small
aside, the 5U4 rectifier is much cheaper, albeit not quite as rugged
and without the slow-start characteristics that distinguish the 5AR4.

I have a box full of 5U4 sitting three feet to my left as I write this.
If I wanted a 5U4, I would have specified and fitted it. Peter, put any
consideration of cost out of your mind unless and until *I* or another
*established* audiophile come to mention it; cost considerations
simply do not apply to DIY ultra-fi amps. The GZ34 is there because it
sounds right and, important this, to protect the 300B in DIY-built amps
with its slow-start filament. It is being replaced with GZ37, cost no
object, because that sounds better still.

But in any case, a single choke with sufficient capacitance would drop
any chop way below other noise sources.

No. RTRDH.

Are you familiar with Mad Man Muntz? He would go into his skunk-works
whenever his engineers decided they had a new product to develop, and
start cutting out parts (he was an excellent EE, so this was not done
at random) until whatever-it-was stopped working. *That* was the magic
point from which to design.

I question from the perspective of one who is usually presented with
smoking ruins and given the mandate to 'make it work'.

I have never had an amp fail to work on switch-on.

Many times
without benefit of schematics, or if they are available, facing so many
whimsical modifications to make the schematics of dubious value. So,
"expected" and "actual" values are important things to know.

The voltages and currents are on my circuit, together with the value
and rating of every other component, together with the maker's name of
every component, and in may cases the name of a supplier if the
component is rare. Again, you'd do better to study and discuss the T39
circuit.

At times,
I thank my lucky stars that this is a hobby.... so I do not have to
make a living at it. Otherwise, not 1 unit in 10 that I see would be
worth the time and trouble to fix.

I find it cheaper at the hourly rates my time is worth to design and
build a new amp rather than fix a broken one if much more than the fuse
is gone. It seems to me that amateur troubleshooting of high-voltage
devices is an ulcer-making pastime. Anyone who doesn't find it
ulcer-making lacks the imagination to be afraid of high voltage and
should find another hobby.

Peter Wieck
Wyncote, PA

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: My quad EL84 build, schematics included.
    ... "Converting a quad-EL84 power amp into a guitar amplifier". ... friend tried it as a hifi amp, really did sound weak into 8 ohm ... otherwise with a more typical 470K resistor there. ... I didn't have another high-voltage cap to put in. ...
    (alt.guitar.amps)
  • Re: My quad EL84 build, schematics included.
    ... "Converting a quad-EL84 power amp into a guitar amplifier". ... friend tried it as a hifi amp, really did sound weak into 8 ohm ... otherwise with a more typical 470K resistor there. ... I didn't have another high-voltage cap to put in. ...
    (alt.guitar.amps)
  • My quad EL84 build, schematics included.
    ... "Converting a quad-EL84 power amp into a guitar amplifier". ... friend tried it as a hifi amp, really did sound weak into 8 ohm ... otherwise with a more typical 470K resistor there. ... I didn't have another high-voltage cap to put in. ...
    (alt.guitar.amps)
  • Re: My quad EL84 build, schematics included.
    ... "Converting a quad-EL84 power amp into a guitar amplifier". ... otherwise with a more typical 470K resistor there. ... I didn't have another high-voltage cap to put in. ...
    (alt.guitar.amps)
  • Re: OLD style JTM 45 power supply
    ... cap and feeds the O/P trans center tap. ... the choke as the first node feeding the B+ to the O/P trans center tap. ... I have confirmed this design with an area collector and his ... note that the donor amp was an EL34 amp. ...
    (alt.guitar.amps)