Strange JCM800 2203 Biasing Problem

[Jaymz77]

Hi all, long time no post...


I have a JCM800 2203 that im looking at for a friend. When it came to me 2 of the output tubes were red plating badly. Turns out it was a bad pair of tubes. Anyway, now i have a set of new Svetlanas in it and i can't seem to get the bias in the right window. I have changed the value of the resistor in series with the bias pot and i can only seem to raise the bias current, not lower it. I have been soldering various resistors in parallel with the 47k to see what sort of value it requires, i have tried values between 23k up to about 200k and nothing seems to lower the bias current below about 120ma.
Any suggestions on what it could be? Clearly whatever this issue is is the cause of the red plating tubes that were in it.


Cheers

[pdf64]

Reducing that resistor value is reducing the bias voltage, whereas you need to increase it.
Rather R29 or R30 value need to be increased.

[sluckey]

Reducing that resistor value is reducing the bias voltage, whereas you need to increase it.
Rather R29 or R30 value need to be increased.

R29 is not in the bias circuit. R30 and/or R27 values need to be DECREASED to increase the negative voltage which will decrease the bias current (make tubes run cooler)

[Jaymz77]

Thanks for the replies guys.
Do either of you have links to a more modern schematic? I only have the old ones. In fact the one i have doesn't even have an R30!
Cheers

[sluckey]

https://el34world.com/charts/Schematics/files/Marshall/Marshall_jcm800_lead_mstvol_100w_2203.pdf

[pdf64]

Doh, many apologies for my typo, of course sluckey is spot on.

[Jaymz77]

Thanks gents, ill have a look at those 2.
Why would i need to tho? If they are the standard values why would they need to be changed to give the correct voltage/ current?


Cheers

[sluckey]

Because the resistors in the amp now don't give you the bias range you desire.

Or maybe the bias caps are dragging the bias voltage down. Not a bad idea to replace them while you're dancing around them.

Or maybe the PT is not putting out enough voltage for the bias circuit.

Or maybe the new tubes are more finicky than old production tubes.

A resistor change will get you where you want to be. But consider changing the bias caps too. There's a lot of expensive stuff depending on that bias supply to be right.

[Jaymz77]

Thanks Sluckey,
it looks as though the bias caps have already been changed to Sprauge Atoms. They should be ok.
I did notice however that the bias supply seems to be coming from a dedicated transformer tap. Should i try taking a new supply from one side of the PT secondary?


Cheers

[Ritchie200]

As Sluckey said, replace those bias caps. Don't let the new look sway you!  I'd do that first before you change anything else.

Jim

[2deaf]

I have a JCM800 2203 that im looking at for a friend. When it came to me 2 of the output tubes were red plating badly. Turns out it was a bad pair of tubes. Anyway, now i have a set of new Svetlanas in it and i can't seem to get the bias in the right window. I have changed the value of the resistor in series with the bias pot and i can only seem to raise the bias current, not lower it. I have been soldering various resistors in parallel with the 47k to see what sort of value it requires, i have tried values between 23k up to about 200k and nothing seems to lower the bias current below about 120ma.
Any suggestions on what it could be? Clearly whatever this issue is is the cause of the red plating tubes that were in it.

Is this a real 2203 or a reissue 2203?
What kind of tubes do you have in it?
What is "bias current" and how did you measure 120mA?
What is the negative voltage going to the grid leaks?

A bias issue at the supply would usually red-plate all four tubes, not just two.

[Jaymz77]

2Deaf,
Yes its an original.
What do you mean "what is bias current?" The current the tubes are pulling on the bias circuit. I measured with a bias probe as well as across the 1ohm resistors that i don't believe are original.
I haven't measure the screens yet.
Yes that's the strange part, only 2 were red plating.


Cheers


Ritchie200
Ill replace them if nothing else fixes the issue. I would have to order some.


Cheers

[2deaf]

I measured with a bias probe as well as across the 1ohm resistors that i don't believe are original.

So there are 120mA going through each cathode of the still unknown power tubes?  If that's the case, I would take those tubes out until the bias issue is resolved.

[Jaymz77]

2Deaf


I've only let that happen for a short period of time while i check the changes I've made to the resistor values.
Ive never seen a 2204/ 2204 style of amp with its own bias tap on the PT before, have you?


Cheers

[2deaf]

All of the 2203's, even the reissue, have a separate bias winding because they use a bridge rectifier that doesn't lend itself to taking the bias supply from one side of the HT winding.

[Jaymz77]

Ok.
To be fair I'm far more experienced with 50watters, than 100watters.


I have just ordered some bias caps. If they dont fix the issue ill be at a loss as to why the current/ voltage is out of spec when all the component values are as standard.


Cheers

[pdf64]

Yes, take all the power tubes out and check their terminal 5 control grid Vdc.

If the red plating tubes both on the same push-pull 'side' then the coupling cap to them may be leaking dc, or their grid leak resistor dodsgy somehow.

[Jaymz77]

PDF64,


Yes they were both on the same side. Ill try the bias caps, check the grid leak resistors (they are the 1k ones on the back of the tube socket, right?) and if it isn't any of those, ill try the coupling cap.


Cheers

[pdf64]

The 1k 5W R35-38 are screen grid current limiters.
The grid leaks are R24,25 220k.
The coupling caps in question are C14,15 22nF.

[2deaf]

There is an orange wire and a green wire coming out from underneath the board.  Each wire connects to two 5.6K resistors.  Each of these four resistors connects to a pin 5 on the power tubes.  With the power on and the amp in standby mode,  what is the DC voltage on those 5.6K resistors?  Speaking of the power tubes, what kind are they?
 

[Jaymz77]

2deaf,


Sorry for the late reply. Ive had some family dramas going on.


The new tubes are Svetlanas.
Voltage on all 4 5.6k resistors is -32.5VDC

[Jaymz77]

This is very strange indeed.
My research says that voltage on pin 5 is about right.

[Jaymz77]

maybe a touch low

[Jaymz77]

Ok,
i tried paralleling another resistor over the 27k in the bias circuit to lower its value and it did bring the bias current down closer to where it should be.
Then i thought to measure the voltage on the bias tap from the PT. It only measured 67vAC or so. this seems quite low to me.
In order to get the right bias current range I'm going to need to swap the 27k with a 10k, or lower.
Is the PT on its way out? or at least that tap?


Cheers

[sluckey]

67vac on the bias winding is fine. Have the bias caps been changed? How much negative voltage do you measure on the bias cap that connects to the diode?

[2deaf]

My research says that voltage on pin 5 is about right.

Ah.  Then you must have EL34's.

Then i thought to measure the voltage on the bias tap from the PT. It only measured 67vAC or so. this seems quite low to me.
In order to get the right bias current range I'm going to need to swap the 27k with a 10k, or lower.

-32.5V with the bias pot maxed is consistent with a 67V_rms winding.  Replacing the 27K resistor with a 20K resistor is likely to yield a bias range of -31V to -38V.  Paralleling a 77K resistor with the 27K resistor will make a 20K resistor.  Don't have a 77K resistor?  Neither do I.  Try 68K or 82K if you have those.  Replacing the 27K resistor with a 10K resistor will likely give you a range of -40V to -48V which is kind of cold for EL34's. 

[tubeswell]

Besides getting your bias voltage sorted, its best not to replace 2 tubes at a time in these amps. better to replace all 4. The chances are that you'll end up with a pair of tubes that run hotter than the other pair, and that will smoke the PT when the amp is cranked with a square wave signal

[Jaymz77]

last night i tried putting in a 10k and while it does help the bias voltage it also drives up the plate voltage, so its not a real solution.
Why is this happening? If the bias tap is fine and all the components are stock values, where is the problem?
i have replaced the caps. its not them.
i have replaced the 220k's on the PI output. its not them.
I have fiddled with the 47k and 27k resistors. the 47k doesn't help and the 27k drives up the plate voltage.


I just want this thing fixed!!

[shooter]

it also drives up the plate voltage

yup, OHM says so, IF you make bias cooler - more NEG, then you have less current, hence more B+

[Jaymz77]

I realise that it is a natural reaction but it’s not how the amp should be operating. Plate voltage should be around 450vdc with the correct bias voltage. Why can’t I get that?

[sluckey]

Quit questioning everyone and just do what they suggest. THEY ARE RIGHT!

[shooter]

correct bias voltage

I'm a self biased guys, so...
the voltage value, is a "ballpark thing" that effectss current, a real thing is tube current, Ip
hopefully each tube as it's own 1ohm R
set the bias to MAX NEG, measure the 1ohms(value = current), and plate volts
calculate plate dissipation and compare to tube datasheet for "typical" fixed bias Ip

low, bump bias volts a smig, recheck.  where the amp lands is it's "normal" operating point.
I typically set customers amps between 60 -70% of max plate dissipation idling, NO signal.
new tubes will change daily as they find their way in the world

[2deaf]

last night i tried putting in a 10k and while it does help the bias voltage it also drives up the plate voltage, so its not a real solution.
Why is this happening? If the bias tap is fine and all the components are stock values, where is the problem?

There is a wide range for the characteristics of those tubes.  The same bias voltage, screen voltage and plate voltage will cause different EL34's to idle at different currents.  There are sets of EL34's that will idle fine with your stock setup, you just don't happen to have one of those sets.

More current draw from your power transformer means more voltage sag and less current draw means less sag.  When you adjust the bias for less tube current, there is less sag and the plate voltage rises.  Then you have to recalculate and adjust some more and then again and then again until you are willing to call it close enough.   

[Jaymz77]

Sluckey, what is your problem? Who is questioning anyone? I have done what was said and the problem still exists.
100+ mA bias current is not normal. 511vdc plate voltage is not normal.
With a normal plate voltage of 450vdc the bias circuit is pulling 100+ mA. When I lower the bias current by raising the bias voltage the plate voltage goes up to 511vdc. Neither of these situations is normal. I don’t understand what problem is causing me to not be able to get a normal plate voltage of 450vdc at the same time as a normal bias current of about 38mA. There is clearly some other condition that is causing this. So instead of being narky offer some suggestions. Geez.

[Jaymz77]

A plate voltage of 511vdc requires a bias current of about 34mA. I can’t get any lower than 38mA. This is with the 10k resistor in place.
With the 27k resistor in place I have a plate voltage of about 450vdc but I can’t get the bias current any lower than about 100mA.
Neither of these situations allows for correct biasing.
So my choice is either a red hot biased tube or a tube with plate voltage that is way too high and will eat tubes constantly. Which of these is the lesser of 2 evils?
Unless there is something else causing this which can be fixed.

[shooter]

I have a plate voltage of about 450vdc but I can’t get the bias current any lower than about 100mA.

It would help me if you list each tubes current, I'm assuming 100/4 = 25mA per tube * 450vdc = plate dissipation of 11.25W/tube plenty good for a start.

Now if EACH tube is at 100mA and you have new tubes in, I'd say you have a partial short in your OT

[Jaymz77]

Shooter, thanks for that.
I should have been more clear. I’m measuring 100+mA across each 1ohm resistor that has been installed by someone else.

[Jaymz77]

So you reckon a possible OT problem?

[2deaf]

Now if EACH tube is at 100mA and you have new tubes in, I'd say you have a partial short in your OT

Ha! I get it.  You're making fun of me because of my critique of that D-Lab guy.

[shooter]

there's a lot possible, If it came to me I'd make sure with no tubes my bias can adjust between -30 to - 60
I'd make sure my PS IS rock solid
I'd surf till I found dcr values for my OT
i'd make sure I had a known good quad of tubes
i'd make sure I had a known good speaker at the proper load for the amp
i'd spend lots of time ohming everything in the PI and PA, re-flowing solder in the same area
i'd verify there is no DC creep coming in from the PI
i'd verify the tube sockets are solid
IF ALL that checks out
tube it check it call customer with my suspicions and concerns, then spent their money 

[shooter]

Ha! I get it.  You're making fun of me because of my critique of that D-Lab guy.

 
I'm not that witty, and I smoked too much to be funny

[Jaymz77]

2deaf,
I can tell you that before I played with the bias series resistor I could get a maximum of -34vdc. Nowhere near -60vdc. That’s why I thought the bias tap might be faulty somehow.
The only bits you mentioned that I haven’t tried is the pi coupling caps, retensioning the sockets and reflowing all solder joints.

[sluckey]

Clearly you need a new OT. And a new PT. And a thicker skin! 

You should be able to find those for under $500. When you get it installed come back and re-read this thread. Maybe listen to some of the stuff you've been told!

[Jaymz77]

Sluckey, were you born a douche or did you have to work at it?
What part am I not listening to? You have not offered anything of assistance and I have done everything others have said, so why are you being such a dick?

[shooter]

while Jay's workin on the amp, and I AM hoping you solve it
I'd like to 

2deaf, don't let you're head get to big for SG's helmet, that Answer sheet is laid out as well as any test I've takin, and I've takin ALOT, you and Steve could make good $$'s,  your works are the standard!
back to regular programing;


get it fixed yet?

[sluckey]

Everything I've posted has been right on target.

[Jaymz77]

No everything you have posted is just a narky arrogant smart ass load of shit. Not one helpful suggestion has come from you.
Everything that has been suggested by others I have tried.
Maybe it’s you who should re read the thread. It wasn’t me who suggested the ot was gone.

[Ritchie200]

Jaymz,


I am the resident Dousch and it is a hat I wear proudly. (this is where SG posts another picture of Ritchie Blackmore wearing a witch hat)


PDF64 suggested the following:


"The grid leaks are R24,25 220k. The coupling caps in question are C14,15 22nF"


You mentioned that you have not replaced the coupling caps yet.  Change those and the resistors (if you haven't).  Definitely do the cheap stuff first, eliminate any question marks before you start looking at the PT. Shooter's SOP should be everyone's SOP.


Jim

[2deaf]

2deaf, don't let you're head get to big for SG's helmet, ...

Not that witty?  I was lucky to catch the first one.  This one just blew right over my head.

[Jaymz77]

Thank you Richie2000, that is what I call a helpful response.
Yes to correct myself, I am yet to try the coupling caps and retension the sockets. I will do that tonight when I get home from work.
I just don’t get why Sluckey had to start carrying on like a flog. There is just no need for it. He started out being helpful and then turned like a cut snake.
It was someone else who suggested the PT, I merely asked if the 67v I measured was within spec.
I have some Mallory’s at home so I will replace those crappy LEGO blocks tighten the sockets and see where I’m at then.
I think there is probably miscommunication involved here from all concerned, including me.
I do understand ohms law quite well as I am an electrician of 20 + years, so I get that lower current equals higher voltage and vise  versa. What I don’t get is why I have a condition that means I can’t bias correctly. I should have about 450v on the plates with a bias current of about 38mA. Whatever the condition is is causing the bias current to be over 100mA ON EACH TUBE at 450v plate voltage. Efforts to fix the high current by lowering the bias voltage have tipped the scales in the other direction and ended up with a bias current of 38mA but with a plate voltage of 511v.
I’ll try the caps and the sockets and report back.