1st Startup Rectifyer Flashover... HELP! plz :-)

[kheper]

Here is the manual for the stock st70. Scroll
down to page #9, 'Initial Adjustment'.

http://www.dynaco-doctor.com/st7_mnl.pdf

'While the tubes are warming up, set the bias
adjusting the potentiometers to the center of
their rotation. This is approximately the
correct setting and can serve as an
emergency operating adjustment if no test
instruments are available.'

I wonder how many blown rectifiers this
advice contributed to?

However, this is with the rectifier unplugged.
The manual, then, advises setting the
correct bias after the rectifier is plugged
in.

When Groovegrinder's rectifiers blew, was
enough current being drawn to make the
rectifier blow? He had the 'poor man's
variac' hooked up, did he not?


NOTE to Groovegrinder: There are other
initial procedures recommended like taking
an ohm reading on cap lug #2. If you do this
make sure, the cap is fully discharged
before trying!

[mesherm]

Well when Dynaco was in business they were shipping nice reliable Mullard EL34s with their amp kits so center position might have been pretty close to right. From what I have heard you could take any four Mullard EL34s out of a box of 100 and it would test like a matched quad.
We wont really know until he biases his tubes.

[kheper]

According to the 'dim bulb tester', ie. a
silly-ac.

http://www.hhscott.com/cc/1st_soft_start.htm

http://antiqueradio.org/dimbulb.htm

They advise 15, 25, 30, 40, 60, 75, 100,
and 150 watt bulbs in succession.

A 40watt bulb should not allow that much
voltage on primary. I wonder what
GrooveGrinder's primary and bias voltages
were when his rectifiers blew?

[Shannon Parks]

http://antiqueradio.org/dimbulb.htm

Thanks for the picture link!

[burnedfingers]

Disclaimer: I admit I've never paid attention to the light bulb load test descriptions that are prevalent and very, very useful.

Isn't a 40W bulb a 360 ohm resistor? A 360 ohm resistor at 400V will draw >1A. This would blow any 5AR4.

I'm confused here ...a 360 ohm resistor at 400volts?

The light bulb in on the input/AC side of the power supply transformer to limit current isn't it?

[GrooveGrinder]

First off, sorry for the delay in replying.
Came home and got the news I'd have to help
my G/F pack up and move out pronto, a month
before expected. (nasty landlord) Ugh.
Not finished yet either.
Probably won't be able to get back to the 70 till
the weekend, and apologies for the patience.
Having said that, I don't know with clarity
exactly what to do from here...

Should I:
1)First test the multicap for shorts, after discharging?
Or is that not needed since I've gotten this far?
2)Check the B+ to/from the 5AR4 socket pins?
3)Connect the CL-90?
(First, or after doing the above tests,
and after the bias check below?)

I understand that I should... and should have
check(ed) EL34 pin5 for maximum neg voltage.
With NO tubes?
Check at either EL34 pin5, and one on both
sides?

I'll keep the inputs shorted and spkrs connected.
I guess the 4 ohm tap is ok... since the junk spkrs
are pulls from someone's car.

Incidently, why is it better to do the preliminary
bias test at pin 5, rather than just full counter
clockwise rotation?

A 40watt bulb should not allow that much
voltage on primary. I wonder what
GrooveGrinder's primary and bias voltages
were when his rectifiers blew?

Isn't a 40W bulb a 360 ohm resistor? A 360 ohm resistor at 400V will draw >1A. This would blow any 5AR4.
I'm confused here ...a 360 ohm resistor at 400volts?
The light bulb in on the input/AC side of the power supply transformer to limit current isn't it?

This makes me think I should do the B+ and
multicap short test, rather than just setting bias
max-neg at pin 5.

BTW:
So far, the amp has always been plugged in to
the dumb-bulb-tester.
IE: never directly into wall AC.
I got as far as a 100 watt bulb (having no 150 on hand)
with only the 7199s and EL34s in....
The bulbs never lit.

Thanks as always!

[dcriner]

To my knowledge, CE is the only manufacturer of high voltage FP (twist tab mount) style multi section electrolytic caps currently in production. If you find something else, like a CD or Sprague TVL, it's NOS and should be treated as such.

RadioDaze has some new-manufacture twist-locks that aren't CEs: http://www.radiodaze.com/ (Drill down under COMPONENTS and CAPACITORS.)

But, their highest voltage is 450V. And their cans are a bright blue color, which wouldn't match a Dyanaco's style. (The voltage rating would be OK for antique radios, which is RadioDaze's main market.)

Close, but no cigar.

BTW, some antique radio restorers have a way of opening old, original, cap cans, and restuffing them with modern, individual e-caps. - to maintain an original "look." More often, radio restorers disconnect the wiring from the original can's terminals and hang new, individual replacement caps under the chassis - leaving the original can there for ornamental purposes.

[kheper]

Should I:
1)First test the multicap for shorts, after
discharging?

Yes.

2)Check the B+ to/from the 5AR4 socket
pins?

Without any output tubes the voltage will be
very high and not representative.

I understand that I should... and should have
check(ed) EL34 pin5 for maximum neg
voltage. With NO tubes?
Check at either EL34 pin5, and one on both
sides?

You need to set the bias voltage to the most
negative possible, then put the output tubes
in and set the voltage so that both sides
are equally negative.

This should be done whenever the output
tubes are replaced.

Incidently, why is it better to do the
preliminary bias test at pin 5, rather than
just full counter clockwise rotation?

Do both. Make the voltage as negative
as possible without the tubes in, then put
the tubes in, and set both sides as equally
negative as possible.

If you have a set of known to be good
output tubes, use them. Bad tubes can blow
a rectifier as well.

[GrooveGrinder]

Well sort of anyway...

First off, I finally got back to it after a hectic few
weeks of no free time.
Only managed a few mins here and there in which
I did the tests a bit at a time, finally finding a
block of a couple of hours to get down to it last
night.
Apologies to all for dropping the thread/project
abruptly!
--------------

Here's how it went, and where its at now:

Ordered a new JJ 5AR4
Tested multicap = ok
Checked for shorts in B+ line = ok
Soldered in CL-90
Measured neg voltage at EL34 pin5 and
set for Max neg. (both sides=same)
(around -47V.. IIRC, may have been 43-ish)
This, with NO tubes in.

While at it, also measured at pins 2/7 = 6.6 v
incidently...
(no tubes in)
Do I need dropping resistors for long tube life,
or should I assume its in range with all tubes in?

Put in 7199s and EL34s and put power to 'er.
Checked at front panel bias points..
NO reading!

I figured thats likely due to no 5AR4 in but I
could be wrong, of course.
I decided to take a chance at that point, and try
putting in the last of the three 5AR4s I had
from before (before ordering the new JJ)
so as to not jepordize the JJ.
That older 5AR4 tested good when I had
2 of that group of 3 tested a few months back
(not used since).
I do note however, that one of those did flashover,
and that testers don't subject to voltage as high
as in use in the ST-70... or so I'm told.

I plugged in directly to wall AC, and held my
breath for about 2 minutes.

Awrighty then! No flashover!
Waited another minute or so then started to
ease the pots clockwise... a very small bit at
a time.
So far, so good.

"And for my next problem..."

At this point...
and I took so long to slowly ease the bias up
that it was well warmed up by the time I got
to center pot rotation....

I noticed... wait a minute... the decimal point
is before the 1... hmmmmm...

I forgot to mention... that I first measured a
fresh Duracell, with my new DMM.
I'm not familar with it and have mislayed the
instruction sheet... but...
The Duracell measured 1.61V
Decimal point correctly placed.
Hmm..

Also, before I noticed the decimal point thing..
I'd gotten it up to .140 (and more) volts,
and it varied a very great deal.
Bear in mind this point is very near center-pot
rotation.
Double hmmmm

I continued trying adjustments and even went
a bit over (I guess its) .156 volts...
It even hit around .170V or so.. (I backed it
way down right off)
Even tried backing down almost all the way,
and started again.

Given that the (apparently) .145 reading
is around center pot rotation, my first thought
is that I'm using the meter incorrectly.
The Duracell measures 1.61 with the DMM
on the same setting.

I pondered and pondered.

Well, there may be a fault somewhere, but its
stable.
Stable meaning no flashover or any other undesirable symptoms....
SAVE FOR the decimal point "in the wrong place"
AND the fact that the bias setting drifts considerably.

I suppose that if the decimal point is not in the
"wrong place" ...ie: less than a volt is the
actual value... that amount of drift isn't dramatic.
If instead, its just that I'm using/reading the meter
wrong... and the actual value is in the 1.xx
range... that much drift is a lot.
Enough swing that it goes into ranges warned
against in postings above!

For example...
I tried a conservative setting of .140(reading)
then waited a minute or 2... and 5 mins or so
too.
It would vary to .150 even .160...
OR downwards sometimes..
.130, .120, etc.

Well...
Stumped, I switched off.

Nothing bad was happening
(other than the seemingly benign stated above)
I decided to see if it would pass signal.
Just for curiousity. And since I thought it might
indicate the severity of the problem.

Plugged in a spare CD player and Preamp.
Turned on the source, then the preamp....

Hah! there's a CD in the drawer.
That player has been on a shelf for years.
THATS what happened to that disc!

Hit the switch on the amp..
Started up fine.....
Hit play... Slowly turned up VOL a little.

Music!!!!
Whoah!
Sounded darn good too!!!!!!!!!!
Turned it up a bit more...

Wound up listening to the whole CD!
(Santatna)
I know.... naughty... but, I couldn't resist.

Wierd, man...
I'm stumped.

Bear in mind I was using just some crummy
OEM car spkrs (4 ohm taps :-) so, not the ultimate
test of sound quality...
Sounded damn good anyway!
Only a tiny bit of distortion... randomly and
very infrequently, and only on peaks.
Drifting bias might be why?

I haven't done a thing since.
I still haven't tried the new JJ.
I'm heasitant to till I understand whats going on.

I'm wondering if another iffy 5AR4 that I've
used so far could account for this?
I don't think so, but I'm only begining to understand
this circuit. (mostly thanks to you all)

Also wondering if I should start all over again
with the bias setting at max-neg (pin5 no tubes)
before plugging in the new JJ 5AR4?

For that matter... I'm wondering what the hack
is going on overall!

Thanks in advance for any help that might be
forthcomming...
Would be nice to finally use the thing after all this.

Thanks also for all the help so far!!!!
I really can't thank you all enough!
I could never have gotten to this point without.

Greg

PS: I really can't account for the previous flashovers..
I'm assuming it was a combo of things....

[Shannon Parks]

Greg, did you have your DMM ground lead in a good spot? Maybe use the chassis itself or the speaker GND. I assume this is a 15.6 ohm stock bias (without re-reading the whole thread I forget). I really can see it working with those bias numbers, so I'm guessing it was a measurement flub.

[GrooveGrinder]

Hi Shannon...
Thanks much for your reply!

Greg, did you have your DMM ground lead in a good spot?
Maybe use the chassis itself or the speaker GND.

I put the negative probe in the vicinity of the
bias test point on the front...
(on the chassis)
so as to be able to hold both probes with one
hand, to have the other free for a screwdriver.
I hadn't thought about one spot being possibly
better than another.
However, now that you mention it, the reading
would change some when it slipped around,
even fractions of an inch, while holding it!
Hmmm...

Maybe I should use an aligator-clip test lead
clipped to the top of the main ground screw?

I assume this is a 15.6 ohm stock bias
(without re-reading the whole thread I forget).

It is indeed a 15.6 resistor.
This one from Triode:http://store.triodestore.com/15ohm2wcf.html
http://us.st11.yimg.com/us.st.yimg.com/I/triodeel_1990_1703526.gif
Though I don't suppose the type matters.
In the future, I will use the 10ohm thats been recommended.

I really can see it working with those bias numbers,
so I'm guessing it was a measurement flub.

(Assuming you meant can't)
The more I've thought about it, the more I think
its the only reasonable explaination.
It seemed so mysterious since it not only worked,
but sounded great.
So good, I can't wait to put it in my system, and
LISTEN!
Aarrgh! :o

Got to get it set properly first tho..

The only other thing that your reply makes me
think of, is grounding.
Maybe the main ground connection is less than
perfect??
Ie: connection between the ground lug/screw and
chassis?

One thing that really threw me, was that when
I'd move the test leads over to the D-cell battery,
it would give a rock-steady consistant 1.61 volts.

Gonna take another crack at it later this evening.

Thanks so much again!
G

PS: Can I just plug the new JJ rectifyer in and
let 'er rip, or do I need to worry about going back
to full negative voltage on the bias?

[Shannon Parks]

Greg, I've put the DMM ground lead in the wrong octal hole sometimes - either from soldering the bias line ground on the wrong leg when everything is upside down or just flubbing it outright (ie forgetting it). Put the ground lead on the chassis or speaker terminal, and then move the red probe around the octal socket until you get a reasonable voltage while reconnoitering, then readjust to free up a hand. Biasing is like that sometimes. :)

[GrooveGrinder]

Hi Shannon... and anyone else who might still
peak in on this thread.
Sorry to resurect it, but I've been feeling guilty
for spending rare free time lately, listening to
the amp (finally)rather than having popped on
to put some closure to it, and offer up thanks
to all once again.

Yes it works! Just fine too.
Sounds great.
Only minor problems left are minor.
A very slight hum... barely audible.. only from
near the spkrs and not at the listening chair.
I assume due to having a 3 conductor/pronged
line cord/plug.
The preamp has one too.
Haven't bothered cheating it as its to minor
to bother... not heard whatsoever with music
playing.

And..
The bias still drifts a fair amount.
I'll explain but I did want to mention how it all
turned out.

Soon after my last posting I took another crack
at it.
All I did was:

Having first grabbed my trusty old DMM,
which was at another location, while there...
One I used for years and was much more familar
with than the one used till this point.
I thought to test the battery of the newer DMM
with the old trusty DMM, and it read 8 volts..
instead of 9.
Ah Ha! I thought..
(even though the newer DMM had consistantly
read a D cell at 1.6x V)

Then:
I re-set the bias pots to max neg voltage,
then put in the brand new JJ rectifyer tube.
Thats it.

Powered up and very slowly... still "walking on
eggshells", turned up the bias.
Passed the 1.0 volt mark sucessfully..
Everything still A-ok.
Ever upward slowly, a bit at at a time, letting
it stabilize a while each incriment.
Eventually decided to let it stay, for the time
being, at around 1.45V.
(Since it was all over the map and I couldn't
be certain what the real value was, I stayed
conservative)

Just for the heck of it, I decided to try the Newer
DMM (that I'd used up till this attempt).
It also read 1.45!
Hmmmmmmmmmmmm...
I dunno..

It still varied quite a bit, as in my last attempt.
(and STILL does)
Like.. when set to 1.45V it can vary upwards
as far as 1.65, and down by roughly the same
proportion.
Lots of smaller incriments in between.
This over the space of a minute or 2.
It never holds still for more than a second or 2.

I eventually learned that if you wait and watch
the meter long enough, you can sort of find a
center point (ie: 1.56) and set it for that with
a fair degree of reliability.

Anywayz...
I decided that it was ok enough to try it in my
system (rather than on the bench as till this point).

I had also remembered the day before that I have
that system plugged into a pwr conditioner.
A Tripplite LCR-2400... that has sat there for
years silently doing its job..
Save for the ocasional relay click as it protects,
and its LEDs shifting up or down to indicate
under/over voltage conditions.

I remembered that it puts out 115V rather
than the 120 + from the wall.

Its other function is to stabilize voltage... to some
degree, anyway.

Good to 2400 watts draw on a 20 amp line..
1800 on a 15 amp line.
Did the math, and realize it'd be ok.

I'd never thought of plugging a power amp into
it before!

Well.. all that is fine, but the bias voltage remains
all over the place.
Just as described above.
I just eventually realized, this is the way its
gonna be... and it sounds great, so...
I sat there several times staring at the meter
for an hour or 2 at a time.. to verify that I'd
found that center point............
And I've been listening happily ever since.

Dunno how long the amp or tubes will last...
nor why exactly its that way..
Though I do remember the amp in its original
state wasn't like that.
Granted, I never used it at this location, and it
was in pretty sad shape.

I can only conclude that the pwr in this condo
is a disaster... what with ACs and fridiges, computers,
microwaves, and heaven knows what else switching
on and off at random intervals.

It does seem to make the case for PS regulation,
I guess

As far as the DMM issue... I'm stumped.
If it WAS a DMM issue.
The only other change was plugging in the new
JJ 5AR4.
I wouldn't think that would have anything to do
with the funky sub 1V bias readings I mentioned
in the above posts..........
But plz feel free, anyone who might still read
this..
to explain how/why it might!

In any case, I've got about 200 hrs of very happy
listening on it now...
and for that, I want yet again, to thank all of
those who helped make it possible!!
And much more for the incredible education it
gave me!!!!!

Sorry I can't identify what the source of the
various probs was...
I'd be glad to hear thoughts on it though.

Greg

[kheper]

Only minor problems left are minor.
A very slight hum... barely audible.. only from
near the spkrs and not at the listening chair.
I assume due to having a 3 conductor/pronged
line cord/plug.
The preamp has one too.
Haven't bothered cheating it as its to minor
to bother... not heard whatsoever with music
playing.

Remove the preamp. Use 2 rca shorting plugs. If it still hums, the problem is inside the amp.

And..
The bias still drifts a fair amount.

Did you completely rebuild the bias supply? It could be an old cap.

Did you replace the old selenium diode?

[GrooveGrinder]

Did you completely rebuild the
bias supply? It could be an old cap.

Did you replace the old selenium diode?

Yes and yes.
I really think its just the unstable power in this
building.

Remove the preamp.
Use 2 rca shorting plugs.
If it still hums, the problem is inside the amp.

I'll do just that.

Thanks kheper...
For all your help, I'm greatful!