From:
ghost.crapper@gmail.com
On Monday, May 18, 2020 at 9:25:58 PM UTC-6, Xtrchessreal wrote:
Starting over: Concentrating on one stage at a time.
Under test: DIY 2204 V1 12AX7
2204 all component values are as stated on the following schematic: https://mhuss.com/MyJCM/JCM800_2204.pdf
Summary: Clipping on positive half of ac signal only.
Test performed:
I followed input signal from High Intensity input through V1b and then V1a.
100Hz, 250Hz, 440Hz, 1kHz, 10kHz signals were used. Preamp Volume up full.
Generated signal by MP3 file from computer of which the output signal gain
was not varied for any of the test signals. Initial 1kHz signal set @ 400mV at
the High Intensity input of the 2204 amp. Followed by 100Hz, 250Hz, 440Hz, and
10kHz. Though
the gain was not changed on the computer it was measured at varying levels for each of the test signals at the amplifier input.
100Hz @ 680mV
250Hz @ 680mV
440Hz @ 560mV
1kHz @ 400mV
10kHz @ 320mV
V1b plate showed good signal gain without any clipping for all test signals.
V1a plate output shows clipping on Positive half of signal only for all test
signals.
Values below were measured and recorded at High Intensity Input, V1b grid,
cathode, and plate then V1a grid, cathode, and plate. Last value of each line below is clipping point dialing Preamp Volume from full down. Preamp Volume was then returned to
full before next Freq signal was applied and measured.
100Hz 680mV@in, 540mV@V1b-grd, 400mV@V1b-cth, 20V@V1b-plt, 10V@V1a-grd,
8.5V@V1a-cth, 70V@V1a-plt =50V clp point
250Hz 680mV@in, 640mV@V1b-grd, 290mV@V1b-cth, 28V@V1b-plt, 16V@V1a-grd,
11.2V@V1a-cth, 100V@V1a-plt =48V clp point
440Hz 560mV@in, 550mV@V1b-grd, 180mV@V1b-cth, 29V@V1b-plt, 16V@V1a-grd,
12V@V1a-cth, 102V@V1a-plt =53V clp point
1kHz 400mV@in, 360mV@V1b-grd, 70mV@V1b-cth, 20.5V@V1b-plt, 13V@V1a-grd,
10V@V1a-cth, 78V@V1a-plt =50V clp point
10kHz 320mV@in, 200mV@V1b-grd, 25mV@V1b-cth, 10V@V1b-plt, 9V@V1a-grd,
8.8V@V1a-cth, 66V@V1a-plt =22V clp point
Is this information adequate to expose a defect in the V1 stage of the
amplifier circuit?
Why is only the positive half of the signal clipped? Tube is new, what
component and values could be replaced to correct this? What if any new values
should be added?
Thank you for your help
X
Well, I was going to fisk your post like I did last time.
but I think I'll just comment on a few things you wrote.
Using all those different frequencies won't really
give you anything you need, unless you're looking for
non-linearity - which you found, BTW. ;-) 10kHz is not
useful when testing an amp that will be connected to
a speaker which can generate no significant output above
around 7 kHz. Unless I'm looking for something that is
frequency specific, I do the vast majority of my testing
at 1kHz, with a small bit of 100 Hz (power checks on
bass amps, looking for rattles, etc.) and an occasional
foray down to the 4-8 Hz region, when doing shake testing
on speakers for voice coil rub. Also some low-frequency
sweeps for speaker and cabinet rattle testing.
So - you have what we call a half-wave problem. I don't
know why you started by looking at V1 - chasing this
particular problem usually starts at the other end of the
circuit. Here's why: half-waving almost always occurs
in the PI or output stage, because those stages are
push-pull and the phases operate fairly independently of
each other - in the PI, one triode for neg, the other
for pos, and then you have a power tube for neg and one
for pos (although most guitar amps run class AB, where
there is overlap between the phases) and then an output
transformer which - again - has half of the primary for
one phase and half for the other. A defect in any of
those areas will give you a half-wave on the output.
(It'll overlap slightly into the opposite phase, because
of the class AB thing.) The reason you always suspect
the PI/output stages for this problem is that all of the
preamp stages run in class A, and it's difficult to
make that kind of waveform there due to a fault, although
a radical shift in bias point might do it. So, I'm going
to forget all the testing you did around V1, and tell
you the procedure for troubleshooting half-wave output
on a tube amp.
1) Load the amp with a resistor of the correct value
for the amp's output, making sure that the resistor
is rated for at least four times the output power
of the amp you're testing.
2) If it has a master vol, turn it all the way up. If it
has a presence control, turn it all the way down. All
the tone controls should be turned all the way up. If
there's reverb, turn it all the way down. Ditto for
tremolo.
3) Input 1 kHz to the amp. The low-gain input is preferred
here. Turn the preamp volume control about halfway up.
Output level will be controlled by using the level
control on your signal generator as needed.
4) Connect your scope across the test load. Set the sweep
and amplitude as required,
5) Advance the level control on the sig generator until
problem occurs - plus a bit more. (You don't want the
bottom half to clip in this case, but you need to have
enough output to work the amp a bit.)
6) OK, now that you've verified you do indeed have half-
waving going on, start looking for it in the PI stage.
Sub in a different PI tube. No change? Check to make
sure that both plates have B+. (An open plate load
resistor is one of the most common reasons for half-
wave operation.)
7) Run the amp back up to the level you had it at in
step 5; scope the two coupling caps at the output
of the PI stage. Scope each cap on both ends - the
+DC ends and the -DC bias ends - a bad cap here
will cause half-waving. OLD PRO VALUABLE INFO:
You will (or you should have) set the scope for AC
coupling when point-to-point shooting in a tube
amp; this can cause some problems, because the AC
coupling function is accomplished by inserting a
capacitor in series with the scope probe. Because
you're looking at the AC signals, using AC coupling
will leave the previous *DC* voltage from the
stage you just tested on the blocking capacitor.
If the next stage is at the same or nearly the same
voltage it's probably not going to matter, but if
there is a large difference in voltage there can
be a problem - as there will be in a PI stage where
one side of the output cap(s) will be at the PI tube's
plate voltage and the other will be at the negative
bias voltage...say, +100V on the plates and -50V on
the bias side. So if you just took a look at the
plate of one of the PI triodes, 100VDC will be left
on the scope's blocking cap; you move next to the
bias side of the coupling cap, WHICH IS CONNECTED
DIRECTLY TO ONE OF THE POWER TUBE CONTROL GRIDS,
and when you touch the probe to that spot, KERBLAM!
You just hit the power tube with a whopping transient,
which may or may not blow the shit outta something...
especially a speaker, if you're dumb enough to use
one instead of a testload. SO, here's the tip:
when doing stage-to-stage testing where the DC
voltages can be all over the place and you have
your scope set to AC coupling, drop the probe tip
onto the chassis after every reading to discharge
the blocking capacitor. You my see a spark now and
then; be brave, be a man and ignore (or enjoy) it.
8) This one is the one I do first, but I wanted to show
you the proper sequence for shooting this problem:
Swap the power tubes into each others' sockets. If
the problem still looks the same on the scope, the
fault could be in the PI stage or the output xfmr.
You eliminated much of the PI stage in steps 6 and 7,
so now you'll have to take a closer look at things:
are both grids receiving signal? If not, find out
why. Pretty simple circuit, check for bad caps and
open resistors.
9) Couldn't find anything wrong with the PI stage? OK,
back to step 8, where you swapped the tubes. Did
your problem turn upside down? (Neg half is now
clipped.) This means you have a bad power tube.
While we're still at this point (if the waveform
didn't flip over) check for anything that would
cause one of the power tubes not to function: open
screen resistors, open grid swamp resistors
(although in this case, one tube would be red-
plating), a wire broken off at pin 3, no screen
voltage for some reason, which would almost have
to be a broken lead, etc. Could even be a bad
filament connection on one tube. Check it all.
10) This leaves us with only the Expensive Shit to
look at: the output transformer. For this part
to cause half-waving, one of the primary phases
has to be open, or gone high-resistance, or have
a bunch of shorted turns. You'll find zero or
even a bit of negative voltage on pin 3 of one
of the power tubes, sometimes a greatly lowered
voltage instead of the full amount, which you'll
see on the other phase.
I'd like to see what you come up with when you use the
above instructions; I know you're seeing half-waving at
V1 but I'm not sure why, so do it my way and let's see.
I haven't had a chance to pint your order sheet out yet;
I've got a '68 Deluxe Reverb on my bench that's been
giving me fits...been worked on by a total amateur who
thought you had to use "vintage" HV capacitors, so the
amp has a variety of old Fender caps removed from equipment
in it, different date codes, "close" values, etc. This
idiot also replaced SEVEN of the 9 pots the amp uses;
never in my entire 50+ years at the bench have I seen
an amp that needed that many pots replaced unless it was
in a fire or a flood. He decided to use "modern" Fender
pots, those cheesy-ass little Chinese things with the
four-legged mounting frames, like they use in the Hot
Rod series, Blues and Pro Juniors, etc. The bushings
are the wrong size, too (metric crappola), so the knobs
form a ragged line on the front of the amp because metric
bushings won't center in the good old American 3/8" SAE
holes. And, of course, his wiring is a total fucking
mess. My bench is really the only place I have to set
up my printer at this point, so if I want to print something
I have to do it between jobs. Pain in the fucking ass, see?
But I expect to finish it up this afternoon, so I'll print
your picksheet and price it out for ya then. Patience,
patience. ;-)
Lord Valve, ThD
Expert (fuck you)
--- SoupGate-Win32 v1.05
* Origin: www.darkrealms.ca (1:229/2)