Strange JCM800 2203 Biasing Problem

[2deaf]

You young folk always mess up the expressions.  It's douchebag, not just douche.  If you call me a douche, I envision an IV bag with a clear liquid in it and I am not offended.  I erased the example for a douchebag. 

I have experienced Marshalls with 500V's on the plates before, so 511V is not that shocking.  The 67V bias winding bugged me enough that I looked up the spec's on that PT and it is 98V @ 50mA.  The bias circuit draws less than 1mA.  I'm cleaning up my act, so I won't say anything about a bias winding rated for 50mA. 

The funny thing is;  98V with a 27K range resistor gives a range almost the same as 67V with a 10K range resistor.  So why is there only 67V on his PT?   

 

[Jaymz77]

A bit like dick and dickhead, I guess. It could be argued that it’s an abbreviation.
None the less, there are too many dicks/ dickheads/ douches/ douchbags around these days. I thought the idea of online forums was to help each other, not to be dick/ dickhead measuring contests.
Anyway, I do appreciate those who are trying to help.

[2deaf]

Did you read the serious portion of my post?

[Jaymz77]

Ok,
Pi caps changed, resistors changed, sockets tightened, bias circuit running all stock values with new caps... same problem.


I am getting -40v on the top of the cap after the rectifier in the bias circuit. I am getting -32.5v on pin 5 on stand by and -31.5 when running.


One pair of tubes is pulling 103mA per tube and the other pair 119mA per tube. So much for a matching quad.
There is about 458v on each of the plates.


I have not a clue...

[Jaymz77]

Yes I did read the serious potion 2deaf.
It’s clear to me the bias tap is running lower voltage.


Or did I miss something?

[pdf64]

Don't put tubes into that almost certainly damaging situation! The best case scenario is that dozens of hours of operational life have been wiped out in moments; the worst case is that they might never run right again.
Ensure bias issues are resolved before putting the tubes back in.

[sluckey]

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.

You're missing a fundamental. You are focused on 450V and 38mA as being the only correct bias condition. That's just not the case. A Proper bias condition is obtained when you have set the current and plate voltage to a safe plate dissipation for the tube. An EL34 has a max plate dissipation of 25W.This is static/idle conditions. No signal applied . Most people try to set the bias so the tube idles at 70% of max plate dissipation. For an EL34 that is 17.5W. THAT'S THE MAGIC NUMBER. This is a safe bias condition for the tube and it also sounds pretty good for a fixed bias amp.

Your 450V and 38mA gives a plate dissipation of 17.214 watts (68.9%). This is just shy of the magic 70% point. That's great but it's not the only voltage/current that will yield 70%. 511V and 38mA gives 19.4W (78%) which is a little hotter but is still acceptable. Many people like the feel and sound of an amp running a little hot and they are happy with replacing tubes every 9 months rather than every 12 months.

If you tweak the bias pot a bit more you should be able to set the current to 34mA.   .034 x 511 = 17.4W (69.5%). This is also a perfectly biased condition. I realize that plate voltage will increase a bit as you adjust the bias to lower the cathode current. It's a juggling act. Everytime you change the current you must recheck the plate voltage and do the math again. You may have to repeat this several times but you will eventually arrive at the magic 70% point.

That 450V and .38mA condition is probably obtainable with a particular set of tubes. But it ain't gonna happen with your current set of tubes. Quit focusing on that and focus on this...

     Plate Voltage times cathode current times 70% = good bias point. (should be 17.5W for EL34)

The above info has totally ignored screen current. Some people say the screen current is typically 10% of the cathode current (which is what you are measuring) IOW, if cathode current measures 40 mA, 4mA goes to the screen and only 36mA goes to the plate. So, if you ignore the screens there will be an error in your plate dissipation calculation. The actual plate dissipation will always be slightly less than the calculation. The err is on the safe side and that's OK with me and a lot of others.

I hope this clears up the bias adjustment for you. But by all means, quit running the current up to 100mA/tube. It will only take a few minutes to permanently damage the tubes by running at 100mA and 450V. That's 45W! From a tube that has a max rating of 25W.

[Jaymz77]

Thank you Sluckey for a detailed, helpful response.
Those high bias current numbers that I quoted are with the bias pot wound completely down. They only go up from there. Everything is so far out of balance.
Should I perhaps take the bias supply from the HT like on a 50w?
I understand what you are saying. But won’t 511v seriously eat up the tubes? 450v is supposed to be near the maximum recommend as it is.

[pdf64]

Maybe the 1 ohm resistors aren’t actually 1 ohm any more?
Hence messing up your current measurements.
Consider replacing them.

What info source are you getting a 450V plate recommendation /limit from? see p4 of https://bms.isjtr.ro/sheets/129/e/EL34.pdf or http://frank.yueksel.org/sheets/163/e/EL34.pdf
Maybe thinking of the screen grids?
Whatever, it's the electrode voltages at heavy / full load that really matter; exceeding voltage limits at idle or low loads isn't an issue, providing that the idle dissipation is reasonable.

[sluckey]

Should I perhaps take the bias supply from the HT like on a 50w?

no

I understand what you are saying. But won’t 511v seriously eat up the tubes? 450v is supposed to be near the maximum recommend as it is.

Voltage don't eat tubes. Current eats tubes. Where did you get that "450V" number? Link please.

Voltage ain't a magic number. Current ain't a magic number. THE PRODUCT OF PLATE VOLTAGE AND CATHODE CURRENT IS THE MAGIC NUMBER!

Look at this. Look closely at VaMax…

     http://tdsl.duncanamps.com/show.php?des=EL34

[pdf64]

This chart comes from the JMP era I think but even so http://www.kbapps.com/audio/schematics/tubeamps/marshall/ampchart.php

[Jaymz77]

I don’t know... To be honest I dont know where I got the 450v vamax from.
I just remember reading somewhere that anything too much above that is not so good for tubes.
So in other words any balance of 70% is fine as long as plate voltage is below 800v?

[sluckey]

BINGO!

[shooter]

Maybe the 1 ohm resistors aren’t actually 1 ohm any more?

my last thought b4  falling asleep!

[sluckey]

Maybe the 1 ohm resistors aren’t actually 1 ohm any more?

Why? Even with 100mA flowing through them they are only dissipating 10mW. I bet they are fine.

[shooter]

only dissipating

I wasn't thinking power, was thinking Vdrop.
I didn't go back n look, but seems Jay indicated they might  not be original
I only had 10ohmers the last amp I did, knew it was coming, but my heart still skipped a beat seeing 200mV til my brain did the /10 

[sluckey]

Jay indicated they might  not be original

That's a safe statement. The original JCM-800 2203 had NO 1Ω resistors.

[pdf64]

If we assume that the resistor values are accurate, the recent indicated dissipations of 10mW doesn’t inform us of their dissipation history. Maybe this amp has been running its power tubes very hot for a long time; that can promote catastrophic failures, and such events can result in a lot of cathode current and hence the potential for significant over dissipation of the cathode resistors.
Best not to assume anything is  good in a wonky old amp.
As dedicated low ohm meters are really necessary to accurately measure 1 ohm resistors, it may be best just to replace them.

[PRR]

> where I got the 450v vamax

The 800V number is mighty bold. However many, even most, EL34 amps run near 480V. I had an Ampeg ran them at 590V, and I have seen other amps in that zone. 511V is not a tube-eater.

EL34 at 511V plate and screen targeting 50mA needs nominal -44V G1 bias for 40mA. This *will* vary +/-20% between different tubes. So you need to be able to dial-down to -53V from the bias supply. With typical 50VAC bias winding (I forget if you measured that) you should have over 65V raw DC available for filtering, no problem hitting -53V at one end of your bias trim.

You mention that voltage at pin 5 reads less than from the bias supply. This is meter loading. It would be good to know your total grid resistance and the make/model of you meter which will lead to some meter specs. If you have 225k of grid resistor and a 1Meg meter, it will read 82% of the true value when the meter is not poking the grid. This is exactly your "-40v on the top of the cap ... -32.5v on pin 5" reading. Many tube-guys favor 10Meg meters which will show less drop (~~2% in this case).

[2deaf]

With typical 50VAC bias winding (I forget if you measured that) you should have over 65V raw DC available for filtering, no problem hitting -53V at one end of your bias trim.

He has 67VAC on his bias winding with a stock Marshall 27K/15K/47K/22K pot bias circuit.  He gets -40V at the first capacitor and -32.5V at the second as expected.

The spec. on that transformer is 98VAC @ 50mA.  The 67VAC reading is the only thing that is out of the ordinary on this amp.

[sluckey]

The spec. on that transformer is 98VAC @ 50mA.  The 67VAC reading is the only thing that is out of the ordinary on this amp.

That could very well be due to a cheap meter.

[2deaf]

That could very well be due to a cheap meter.

It's very unlikely because everything else is right in there for a 67V winding.

I think what needs to be addressed is why it only has 67V.  It was wound that way?  It has an unseen load on it?  It has partially shorted?  Is this a progressive phenomenon that is going to cause more problems in the future?

[shooter]

It has an unseen load on it

I got a small bet placed here

[Jaymz77]

Thank you all for the input while I was sleeping.
My meter is a fluke 117. While not a top of the range fluke, still good.
67v on the bias tap is definitely on the low side. But if 511v on the plates is no issue then I’ve clearly freaked out over nothing.
When I read that 511v on the plates I had a 10k resistor as the series bias resistor and I was reading 38mA with the bias pot at minimum. In order to get the correct 70% i am going to have to lower it even further, maybe to 5,6k.
I will change the 1 ohm resistors while I’m at it and see where my balance point lies then.

[Jaymz77]

Shooter, you think there IS or ISNT unseen load? If there is where would it come from?

[shooter]

If it's me, I NEED to see the bias range from 30 to 60-ish with NO tubes
once I got that, I want a plate V*I that I'm comfortable with.  IF, I can't get that, I go lookin down all the rabbit holes

[Ritchie200]

What I cant get past is that this used to play well. The only life changing issue is a couple of red plating tubes? Now all of a sudden the bias is so far off we need to drastically mod that circuit? I be thinking the same that there is a draw coming from somewhere. Or those coupling caps!

Jim

[Jaymz77]

Well, I changed the bias series resistor to 5k6 to give me the right bias range, powered up, listened to to HT fuse pop, replaced the fuse with a 2A since I didn’t have any 1A, powered up, measured the bias tap and got basically zero volts.
Clearly the bias tap is dodgy.
Can I up the bias series resistor and take a bias supply from the HT on a bridge rectified amp like with a 50w amp? Or is there more to it?

[sluckey]

While you are messing with bias modifications pull the output tubes and don't put them back in until you have the proper negative voltage range.

Well, I changed the bias series resistor to 5k6 to give me the right bias range, powered up, listened to to HT fuse pop, replaced the fuse with a 2A since I didn’t have any 1A, powered up, measured the bias tap and got basically zero volts.

When you say bias tap, are you referring to the wire from the transformer? If so, remember that is AC volts. Disconnect the transformer bias wire and measure the AC voltage on the dangling end of that wire. This will tell us once and for all if that winding is dodgy. And it will provide a clue about which direction the troubleshooting will take. What AC voltage do you measure?

Can I up the bias series resistor and take a bias supply from the HT on a bridge rectified amp like with a 50w amp? Or is there more to it?

It's a little more complicated than that. It requires a Class X capacitor and a resistor. Look at the simple circuit on the bottom left of page 5 of this pdf...

     http://sluckeyamps.com/misc/Amp_Scrapbook.pdf

I don't advise doing this. The bias circuit is probably the most critical circuit in that amp. A lot of expensive parts are relying on the bias supply to be very reliable. That's why the 2203 uses a special separate bias winding (not a tap) to provide a simple very reliable bias circuit.

Report the voltage measurement I asked for and we'll go from there.

[pdf64]

While you are messing with bias modifications pull the output tubes and don't put them back in until you have the proper negative voltage range.
...

It's rather frustrating that Jaymz77 keeps getting advised about this and keeps ignoring it. May as well try to fix the amp by dumping it off a 10th floor balcony.

[2deaf]

It's rather frustrating that Jaymz77 keeps getting advised about this and keeps ignoring it.

What is even more frustrating is the constant onslaught of suggested causes that have already been precluded by the data given.

OP:  23K to 200K resistors in parallel with the 47K resistor results in cathode currents of 120mA or more.
#8:  Bias cap's have already been changed to Sprague Atoms.
#11:  Cathode current was measured with a bias probe and by using 1 Ohm cathode resistors.
#20:  -32.5V on all four 5.6K resistors.
#23:  Measured 67V_ac on the bias winding.
#27:  No change with the bias filter cap's replaced.
         No change with the 220K resistors changed.
#33:  Measured 511V_dc on the plates.
#34:  Changed the 27K resistor to 10K and measured 38mA @ 511V.
         Measured greater than 99mA @450V with the 27K resistor in place.
#41:  Measured -34V maximum with the 27K resistor in place.
#49:  Still hasn't replaced the coupling cap's.
#53:  No change with coupling cap's changed.
         No change with resistors changed.
         No change with the sockets re-tensioned.
        -40V on the first filter cap after the rectifier with the stock bias circuit.
        -32.5V on pin 5 on standby.
        -31.5V on pin 5 with the amp running.
         One pair of tubes idling at 103mA @ 458V.
         Other pair of tubes idling at 119mA @ 458V.
#73:  A Fluke 117 was used for the reported measurements.
#77:  1A HT fuse blew when the 27K resistor was changed to 5K6.
         Bias tap was measured at basically 0V with the 1A HT fuse replaced with a 2A fuse.
#79:  It was suggested that his approach is equivalent to dumping the amp off a 10th floor balcony even though it has already been abundantly demonstrated that he does not respond well to derogatory statements borne out of exasperation. 

[Ritchie200]

Oops! Sorry I missed the “PI caps changed”.

Jim

[Jaymz77]

pdf200, you are right, i do keep missing that.


I shouldn't need to explain myself but i will anyway.
My Mrs left me 2 weeks ago and i have been like a ship without a rudder ever since. I have not been in a good way. Its only the last few days where i have stopped bursting into tears at the drop of a hat.
So you are all correct, i am all over the place at the moment, but i have good reason, so i would appreciate a bit of understanding.
On top of my home issues i have committed to fix this am for a friend of mine. The amp belonged to my friends cousin who recently died, he just wants the amp to be in working condition so he can sell it.


So, unloaded bias tap voltage is 69.5vAC. Yes, i was measuring DC, rookie error, but as I've explained above, I'm not exactly thinking clearly at the moment.
Perhaps the problem is obvious to some of you, and maybe it would be to me if i was thinking clearly, but its not, and i would appreciate all the help i can get to put this issue behind me.


Thank you.

[sluckey]

So, unloaded bias tap voltage is 69.5vAC.

There should be 100VAC on that bias winding. The proper fix is to replace the power transformer. Then you'll be able to put the original resistors back in the bias circuit and have the proper bias adjustment range.

I'm sure the four output tubes have been damaged due to the 100mA they've had to endure, so a new set of tubes would be proper also.

[Jaymz77]

Thank you Sluckey,


It seems i may just have a PT sitting here.
About 2.5 years ago i purchased a 1976 JMP50 1987 Marshall that had its original PT replaced. Anyway long story short, the amp had an issue at one point which i can't remember and during the course of repairs i became aware that someone had used a JCM800 style 100w PT in it and converted the 50w amp to run with a bridge rectifier.
Anyway i pulled that transformer out and replaced in with a Marstran PT running the year correct voltages. The amp kills now.


Anyway i still have that PT sitting here. Ill make sure its suitable first and then maybe try installing it.

[pdf64]

Yikes, sorry about your situation; working on amps can be good therapy but if your mind drifts from the task in hand (as may be understandable) it would be best to take a break.
Building yourself a light bulb limitier would be a really good idea before the next power up https://robrobinette.com/5e3_Modifications.htm#Light_Bulb_Current_Limiter

[Ritchie200]

Wow, yikes is an understatement. One foot in front of the other in life and it seems, this amp as well. Sounds like you have acceptable iron for replacement. Shame to put all this work in and have it go down the road - although at this point you might be ready to help it along!  Hang in there.


Jim

[Jaymz77]

Thank you all.
I can report that by sheer fluke, the PT I had sitting here was the correct one. I have installed it and the amp now works fine. Plate voltage about 447v, biased at 39mA. I’m going to hang onto it for a couple more days just to test it but it seems it’s problem solved.
Thank you all for your help, sorry for my tantrum. Now to sort my life out.
Thanks again.

[shooter]

Now to sort my life out.

been there, must here probably have
slow n steady, learn n work through the 5 stages
when you get there, take up rock climbing and fast cars 

[mresistor]

YAY !       

[Jaymz77]

Shooter,
I’ve always wondered what a mid life crisis would look like for me since I already have 4 guitars, 4 amps and 2 motorcycles!
Maybe I get all sensible in my old age?


Thanks your your help and advice.

[shooter]

Maybe I get all sensible in my old age?

I was sensible the 1st half, didn't work out so well, but once I found Rockclimbing, Jeeps, rallycars, and down in their luck street alcoholics to help, I was in the zone 

[2deaf]

Maybe I get all sensible in my old age?

Nope.

[shooter]

Nope.

spoken like a true jaded 'ol man? 

Jay, need more PM me

dave

[Jaymz77]

Thank you very much