Selenium rectifier question

[EricM]

All the "relay" does is to complete a low resistance path from the 600 V
voltage source, to the filter caps AFTER a time delay has (gently via 15k)
pre-charged the filter caps.

Disconnect the 600V lead from the PS to the Amp and see which way the
problem is. If the fuse still blows, trouble shoot the PS. You do have a
voltmeter?

Yes, the weird part about this is that before doing any work on the
power supplies (i.e. everything working) the output voltages with no
load were around 17 on the 12.6 legs, -38 on the bias, 6.7 - 7.2 on
the 6.3 volt legs, and nearly 930V on the "600" output. I used a
Variac to bring things up slowly after doing the modifications to
check the output and it was nearly the same on all outputs (again, no
load). Hooked the cables up to the amp and used a Variac again, and
the 17 was pulled down to around 9.7 once under load, the -38 remained
about the same, the filament voltages were all within 6.3 - 7.1V, and
the plate side of the output transformer read about 500V, but once the
voltage got nearly up to full input voltage and the relay closed
(110-120) the 5A fuse popped. If I disconnect the 600V connection at
the amp - the fuse doesn't blow. The amps and components test fine
and no changes have been made there since everything was working.

One thought though, and I'm not sure if it makes sense or not; I
measured the resistance of the selenium I removed in the bias circuit
- which shares the "common" with the 600V supply - and it was about
1.4 Meg. The resistance across the silicon diode I replaced it with
is only around 830K. Forward voltage drop aside - is it possible that
since the new component doesn't have the same resistance
characteristic as the new that adding a "dropping resistor" per se
would increase the resistance and balance out the circuit? I've put a
1/2 watt 330K resistor in series with one of the 10A 600V diodes that
I used, and the resistance then measures around 1.4 Meg, but I doubt a
1/2 watt unit would stand up to the circuit load. I'd probably have
to have at least a 10W unit, and 330K in 10W is hard to locate. Does
this make sense? I'm running out of ideas with this thing... :/

 

[Arfa Daily]

Here's a link to the rest of the device; the vacuum tube amplifier
http://www.mediafire.com/?zdmyygvyfxn It's a large image in order to
be able to scale it so you can see what's what so I needed to do
a .pdf.

Assuming that the power supply runs without blowing the fuse, when it is
disconnected from the power amp, then looking at the schematic, it seems to
me that tracking down the problem will be relatively simple. But before
going any further down that route, can we get a bit clearer on how exactly
you have gotten to the point that you are at now? First, did the psu / amp
combination ever work before you started work on it, or has it been blowing
fuses all along ? The LT selenium reccies that you replaced just for
reliability's sake can be discounted I think. What was the reasoning behind
replacing the HT rectifier packs ? As you imply that these are differently
wired from the originals, could this be anything to do with the current
problems ? I'm sure that you have described some of this reasoning elsewhere
in the thread in odd bits, but it would be helpful if you could just go
through the whole of how you have reached the point you are now at, all in
one go.

Arfa

 

[Arfa Daily]

Measured in or out of the circuit? Forward or reverse? At what voltage
was
the resistance measured? It's been a while since I worked on anything
using
a selenium stack, but....... They age. The forward resistance goes up.
I
was not ever impressed with the reverse resistance, and 1.4 Meg might be
ok.
When changing them out with silicon, I always added a series resistor. As
I
recall, I commonly used about 300 Ohms for a 75 to 100 mA selenium
change-out, but 50 to 100 Ohms might be more appropriate for this PS

The forward resistance of a typical silicon diode will be about 700 ohms to
1k ohm, measured on a typical analogue multimeter. Reverse resistance will
be, to all intents and purposes, infinity.

Reading again where you currently are, and how you got there, it seems that
before you started 'modifying' the rig, it worked. The only implication of
this is that the current problem must be down to something you have done
yourself. Looking at the sorts of values of resistor that you are trying to
work with, it is clear that your understanding of how this equipment should
work, is a little sketchy. Whilst tube based equipment is actually very
simple in design concept, and normally a breeze to troubleshoot, it is
actually quite 'specialist', and if you are not used to working on it,
regrettably, you may be getting yourself out of your depth on this one.

Unless you have a clear idea of exactly how to troubleshoot the problem, I
think that it might be time to abandon your efforts, and call in some
on-the-spot help. I'm not trying to deliberately put you off - or put you
down even - just offering some good honest advice with the experience of
many years at this.

Arfa

 

[Heinz Schmitz]

The 1N1239 was a replacement for the
5R4 vacuum tube rectifier;

Oh, that explains it. Then the 5R4 ist a directly heated rectifier
tube and the 5 Volts AC are the heater power. This is, you may
disconnect 12 and 13 from the mains transformer when using
selenium or other semiconductor rectifiers.

I also located another online vendor that
sells 'C-Cap' rectifier replacements and he states that his 5R4
replacement should work. I'm debating whether or not to order some to
test though,

Investments into faulty equipment mostly turn out wrong, if a sound
reason for the purchase is not present .

And essentially what the relay does, is shunt resistor R5,

To give the amp tubes time to heat up before anode etc power is
applied. Very reasonable.

as pins 2
and 8 seem to be looped no matter what position the relay is in. This
one is really a pain - I've never encountered a PS like this one...

It looks quite ok once we understand everything .

Now, once the power supply seems to be ok, I would follow the 600 V
wiring into the amp with an Ohm-Meter and try to locate a low
resitance to ground.
Next I would check all anode-to-grid caps like eg C19 at V8b. If some
of those leak, the grid will not be negative enough and anode current
will go through the roof.
I'd disconnect T2 on the primary side and measure the resistance of
the windings and the resistance to ground, to prove that the
transformer is still ok.

Well, much everything like in a common radio repair .

Regards,
H.

 

[EricM]

Oh, that explains it. Then the 5R4 ist a directly heated rectifier
tube and the 5 Volts AC are the heater power. This is, you may
disconnect 12 and 13 from the mains transformer when usingseleniumor other semiconductor rectifiers.


Investments into faulty equipment mostly turn out wrong, if a sound
reason for the purchase is not present .


To give the amp tubes time to heat up before anode etc power is
applied. Very reasonable.


It looks quite ok once we understand everything .

Now, once the power supply seems to be ok, I would follow the 600 V
wiring into the amp with an Ohm-Meter and try to locate a low
resitance to ground.
Next I would check all anode-to-grid caps like eg C19 at V8b. If some
of those leak, the grid will not be negative enough and anode current
will go through the roof.
I'd disconnect T2 on the primary side and measure the resistance of
the windings and the resistance to ground, to prove that the
transformer is still ok.

Well, much everything like in a common radio repair .

Regards,
H.

Actually made some progress. Yes, a previous poster was correct that
my understanding is "sketchy" That's why I'm posting here...

I didn't replace the large Sarkes Tarzians, but did replace the
selenium stacks. I do intend to put 5R4's back in the PS for
rectifiers though, as due to their age they may be contributing to the
overvoltage in the 600V circuit (i.e. reading 900+ when all's supposed
to be present is around 600). Have since found out from another tech
that the silicon's I used elsewhere in the PS circuit are no problem
since they're not part of the HV circuit, and their output voltages
under load measure normal. Will add a small dropping resistor though
to prevent damage to the tubes. One problem was that there was a cap
in the 600V circuit on the amp side that was conducting - basically a
600V to ground. Not good, and was causing the fuse to blow. Replaced
it and got the amp functioning and not blowing fuses, but after
applying a test signal from a tone generator, something 'opened' and I
am getting some 60 hz buzz in the output, which I think might be one
or both of the two 80/450 caps in the power supply that were damaged
by the fuse blowing several times during troubleshooting. In
addition, I did find that two other caps in the amp were getting bad
(old can electrolytics) - made "new" ones and things got better, but
still 'buzzing'. I also suspect some of the old can caps (lower
voltages) in the power supply in addition to the large 450V units, as
at this point there isn't anything else in the amp that is old and
could be shorting or leaking. All of the signal caps have been
replaced with Auricaps. Been a lot of work, but I think I'm close to
completing the project.

 

[EricM]

Actually made some progress. Yes, a previous poster was correct that
my understanding is "sketchy" That's why I'm posting here...

I didn't replace the large Sarkes Tarzians, but did replace theseleniumstacks. I do intend to put 5R4's back in the PS for
rectifiers though, as due to their age they may be contributing to the
overvoltage in the 600V circuit (i.e. reading 900+ when all's supposed
to be present is around 600). Have since found out from another tech
that the silicon's I used elsewhere in the PS circuit are no problem
since they're not part of the HV circuit, and their output voltages
under load measure normal. Will add a small dropping resistor though
to prevent damage to the tubes. One problem was that there was a cap
in the 600V circuit on the amp side that was conducting - basically a
600V to ground. Not good, and was causing the fuse to blow. Replaced
it and got the amp functioning and not blowing fuses, but after
applying a test signal from a tone generator, something 'opened' and I
am getting some 60 hz buzz in the output, which I think might be one
or both of the two 80/450 caps in the power supply that were damaged
by the fuse blowing several times during troubleshooting. In
addition, I did find that two other caps in the amp were getting bad
(old can electrolytics) - made "new" ones and things got better, but
still 'buzzing'. I also suspect some of the old can caps (lower
voltages) in the power supply in addition to the large 450V units, as
at this point there isn't anything else in the amp that is old and
could be shorting or leaking. All of the signal caps have been
replaced with Auricaps. Been a lot of work, but I think I'm close to
completing the project.

Forgot to mention that the 'buzzing' only starts after the relay
closes - so more evidence that it's either the 80/450's in the PS, or
possibly a problem somewhere around the voltage divider/0A2/6DR7
portion of the circuit that is fed from the 600V terminal in the amp.

 

[EricM]

Not possible that it is the relay itself that is mechanically buzzing
perhaps?

Peter

The relay itself isn't making any noise; it did 'buzz' a little
before the fuse blew prior to fixing that problem, but it's a 60hz
noise in the speaker now. Suspect the PS since most everything in the
amp that could leak 60hz has been replaced.

 

[Arfa Daily]

The relay itself isn't making any noise; it did 'buzz' a little
before the fuse blew prior to fixing that problem, but it's a 60hz
noise in the speaker now. Suspect the PS since most everything in the
amp that could leak 60hz has been replaced.

Is the buzz affected by the volume control ?

Arfa

 

[EricM]

Is the buzz affected by the volume control ?

Arfa

There is no volume control as this is a power amp only, however
increasing or decreasing the input signal doesn't affect the buzzing.
Just a 60Hz drone...

 

[Heinz Schmitz]

Well of course it does, because before that the amp isn't working at
all. A tube amp without anode voltage is just a refined heater.

Do not forget that this is a tube device. Some of the tube cathodes
are heated with 6.3 Volts alternating current of 60 Hz. If there is a
cathode-heater-leak (due to tube age) you get AC into the signal path.

Note that the preamp stages are operated with 400 Volts off the
stabilizer circuit around V9 - V10. So have a look at C20 / C2 - they
should neutralize AC ripple there.

I'm scratching my head about connectors 3 and 4 on TS2 named
regulator filament. Somehow TS2 in the ps-schematic doesn't seem
to be TS2 in the amp schematic.

Regards,
H.

 

[EricM]

Well of course it does, because before that the amp isn't working at
all. A tube amp without anode voltage is just a refined heater.

Do not forget that this is a tube device. Some of the tube cathodes
are heated with 6.3 Volts alternating current of 60 Hz. If there is a
cathode-heater-leak (due to tube age) you get AC into the signal path.

Note that the preamp stages are operated with 400 Volts off the
stabilizer circuit around V9 - V10. So have a look at C20 / C2 - they
should neutralize AC ripple there.

I'm scratching my head about connectors 3 and 4 on TS2 named
regulator filament. Somehow TS2 in the ps-schematic doesn't seem
to be TS2 in the amp schematic.

Regards,
H.

It's a separate 6.3VAC supply for just the 6DR7 tube which is in the
same circuit at the 0A2 voltage regulator, attached to 13 & 14 on the
TS in the amp. I've isolated the problem to the power amp circuit -
preamp section tests fine - and most likely something around the
voltage divider/6DR7/0A2 portion of the circuit. When the biasing pot
(25K unit in that circuit) is adjusted, the pitch of the buzz varies,
which leads me to believe that the problem is somewhere in that
circuit. I had replaced all the signal caps with Auricaps, following
the black lead in/red lead out scheme recommended by the manufacturer,
but C20 (I think) is a power supply decoupling cap and not just a
signal coupling cap. The manufacturer recommends that in this
application, the red should be the input and the black to ground. It
was reversed, I put it in the right direction, but it didn't stop
motorboating in the speaker. Auricaps aren't supposed to have
polarity, and originally the old caps that had been replaced by
another person were nonpolarized poly's and everything seemed to work
fine. Guess that's what you get when you try to "upgrade"...

This is what's happened so far - the replacement of the seleniums (how
this all got started in the first place) really wasn't a problem; the
filament voltages which is all they affect are fine. When the PS/AMP
were originall fired after replacing all signal caps with Auricaps and
the large electrolytics with new units either separately or ganaged
per a multi-section unit being replaced, the fuse in the PS blew after
starting the amp, waiting for the relay/cap charge delay, and the b+
circuit engaging per the relay. Found that C21 was conducting,
replaced it, and everything worked fine with no signal. Applied a
1khz signal to 4-5-6 on the feedback input terminal strip of the amp
(goes directly to the power amp), something crackled briefly and the
amp started "motorboating". When started now, the output is clear
other than a bit of presence once the heaters heat up, but when the b+
kicks in the noise starts. I'm suspecting it might be the 80/450 caps
and possibly the 100K resistors in the power supply that might have
been damaged by some sort of issue with the b+ during the fuse blowing
fiasco. I'm about to measure the frequency, but I think it's 120 hz
and not 60 which as stated might indicate the power supply.

 

[Arfa Daily]

It's a separate 6.3VAC supply for just the 6DR7 tube which is in the
same circuit at the 0A2 voltage regulator, attached to 13 & 14 on the
TS in the amp. I've isolated the problem to the power amp circuit -
preamp section tests fine - and most likely something around the
voltage divider/6DR7/0A2 portion of the circuit. When the biasing pot
(25K unit in that circuit) is adjusted, the pitch of the buzz varies,
which leads me to believe that the problem is somewhere in that
circuit.

<snip>

Altering the bias changes the standing current in the output tubes. If this
changes the *level* of the buzz, then that indicates a power supply problem,
most probably - but not necessarily, given the on-going work that's been
done around the rectifier circuitry - to do with the filter caps. If
altering the bias pot actually does affect *pitch* - ie frequency - of the
buzz, then you've got a problem, as this would indicate that something is
oscillating at low frequency ...

Arfa

 

[EricM]

<snip>

Altering the bias changes the standing current in the output tubes. If this
changes the *level* of the buzz, then that indicates a power supply problem,
most probably - but not necessarily, given the on-going work that's been
done around the rectifier circuitry - to do with the filter caps. If
altering the bias pot actually does affect *pitch* - ie frequency - of the
buzz, then you've got a problem, as this would indicate that something is
oscillating at low frequency ...

Arfa

It changes the frequency not the amplitude. Changed out *all* caps in
the power supply, replaced the sarkes tarzian units with 5R4 tubes,
double checked the replacement silicon diodes - they're fine, all
power outputs are spec at the amp under load. It's gotta be something
in or around the 6DR7/0A2 part of the circuit. Used a 2.0uf Auricap
for C20 but C21 is an electrolytic. Maybe replacing C20 with a
2.0@450 would do the trick... Checked all resistors in and around
the power amp circuit and all test good and within tolerance.

 

[Sjouke Burry]

It changes the frequency not the amplitude. Changed out *all* caps in
the power supply, replaced the sarkes tarzian units with 5R4 tubes,
double checked the replacement silicon diodes - they're fine, all
power outputs are spec at the amp under load. It's gotta be something
in or around the 6DR7/0A2 part of the circuit. Used a 2.0uf Auricap
for C20 but C21 is an electrolytic. Maybe replacing C20 with a
2.0@450 would do the trick... Checked all resistors in and around
the power amp circuit and all test good and within tolerance.

Low frequency oscillation, once called motorboating,
might be caused by feedback along the supply path.
Try a fat capacitor added to the supply to see whether
that occures.
Also check other de-coupling caps, shunt them with a bigger one
to test.

 

[EricM]

Low frequency oscillation, once called motorboating,
might be caused by feedback along the supply path.
Try a fat capacitor added to the supply to see whether
that occures.
Also check other de-coupling caps, shunt them with a bigger one
to test.

Been troubleshooting this with an oscilloscope and I see AC on both
12.6V lines - as well as on the 600V line. But - if I place a 300-600
ohm resistor across the monitor amp output (15/16 on the amp schematic
on the term. strip) the buzz goes away and both power amp and monitor
amp work perfectly. The -38 from the PS checks clean and the 6.3 AC
sources are, well, AC... but I can't see why I'm getting AC taint in
the 12.6 lines and the 600 plate supply. It just doesn't make sense.
The 80/450 caps are NEW and have been lifted, checked, tested, etc.
and they're good - I also replaced all the other caps in the circuit
and STILL get the motorboating noise when the HV relay closes. If I
ground certain points in the circuit - particularly the .047 uf caps
in the monitor amp circuit the buzz also goes away (using a 10 uf 600V
to the chassis), so I'm thinking the real problem is in the amp
somewhere in the monitor circuit where it connects to the EQ card -
when the card is out, everything works fine but when it's plugged in
the buzz comes back. This one's got everybody stumped I guess...