Welcome, Guest. Please login or register.

November 28, 2020, 03:30:30 PM

Login with username, password and session length

Recent Posts


Author Topic: Bias Question regarding Marshall 100watt 2203  (Read 2138 times)


  • Regular
  • **
  • Posts: 24
  • Chip Points: 1
    • View Profile
Bias Question regarding Marshall 100watt 2203
« on: February 13, 2017, 06:36:19 AM »

i bought a Marshall 2203 Clone.
100Watt, 4xEL34 3xECC83
It works flawless after thoroughly cleaning the dreaded Cliff-Type
Jacks (closing contacts).

The one thing that concerns me is that one of the EL34 looks way more
used than the other 3. After opening the Amp i discovered that the
loudspeaker-jack directly behind this particular tube-socket is slightly
melt, nothing drastic but it indicates that this tube has or maybe had some
sort of problem, it maybe draws more current was my first guess.
Or maybe the 1k5w screen grid resistor is not o.k.?
Note there is no sign of redplating or any other malfunction on this tube,
but all of the lettering is gone and i can see lots of little dents on the anode,
while on the other 3 the lettering is intact and there is absolutely no sign of
dents on the anodes.

These are original AEG (made by RFT, Eastern Germany) EL34 tubes, so i want
to keep those as long as they work.

I was checking the bias via the "Output Transformer Resistance Method".
My measurements as follows:

plate voltage: t1 - 440V (blue side); t2 - 440V (blue) ; t3 440V (red) ;
t4 - 440V (red) - !measurements were not stable, they varied between
439V to 441V over a period of 30 minutes, so i took the middle for all tubes!

output transformer centertap voltage: 440.5V ( varied between 440V to 441V)

voltage drop across half of the output x-former primary: 440.5V-440V= 0.5V

output x-former DC resistance measured with unpowered but warmed up amp:
red lead side=16.3ohms
blue lead side=17.6ohms

anode current blue side: 0.500v/16.3ohms=0.0307A or 30.7ma
anode current red side:  0.500v/17.6ohms=0.0284A or 28.4ma


If so then this amp is biased very cold (27% red side, 25% blue side)

WHY IS EL34 NR.1 (t1, next to phase inverter ECC83) MORE WORN TO THE

Any help welcome.


  • Legendary
  • ******
  • Posts: 2148
  • Chip Points: 262
    • View Profile
Re: Bias Question regarding Marshall 100watt 2203
« Reply #1 on: February 13, 2017, 07:38:22 AM »
Some things to check;
What do the screen voltages read?
Check the Screen resistors, are they within spec?
Also What does the bias voltage read? (at the valve pins)

Yes Divide by 2 for plate current,, and Yes it seems it's biased very cold.

BTW, Anode and screen voltages do drift so 439V moving around a bit to 441V is Normal.

Not sure I understand the dint's in the plate,, never heard of that one?? xP
Maybe others will have more info soon.


  • Legendary
  • ******
  • Posts: 1984
  • Chip Points: 216
    • View Profile
Re: Bias Question regarding Marshall 100watt 2203
« Reply #2 on: February 13, 2017, 09:50:04 AM »
Never forget, these are just guitar amps, not precision lab gear.  Your B+ was 440v.  Think about 441 or 439v, those are less that 1/4 of 1% of the total voltage.    Don't even think about reading something into that.  Also consider that your mains voltage hops around all the time.  Watch a big football game, like a playoff with your AC meter plugged into the wall.  Watch how the voltage drops when they run ads on TV.  EVeryone gets up, opens the refrigerator, nukes something in the microwave, water pumps run to flush toilet.   Mains voltage sags in response.  Now at 440v B+, for every volt change in mains, your B+ will change almost 4 volts.  So a 3v change in mains might alter the B+ by a dozen volts.  Fact of life, and nothing wrong there.

I tend to doubt the tube melted the speaker jack, but I'd easily believe someone laid their iron against the jack while soldering something.

One odd tube?   You can easily measure current now, but that doesn;t mean at some earlier point in time the amp didn;t have some issue causing the tube stress.  Maybe the socket that tube was in at the time lost its bias?


  • Legendary
  • ******
  • Posts: 670
  • Chip Points: 80
    • View Profile
Re: Bias Question regarding Marshall 100watt 2203
« Reply #3 on: February 13, 2017, 01:55:18 PM »
Agree with Enzo about the speaker jack, if the tube overheats the heat will be outside the chassis, not inside at the socket where the jack is.
The way you are measuring at the OT will not give very much accuracy.  Instead of having one probe on ground, have one probe at the center-tap, and the other probe at the socket (use clip leads for safety).  This gives you the direct voltage reading without math, changing in real time.
Now, for test purposes you can run a pair of tubes at a time.  This way you can find out how well matched they are and whether the odd ball is idling at a higher current.
« Last Edit: February 13, 2017, 01:56:30 PM by g1 »


  • Regular
  • **
  • Posts: 24
  • Chip Points: 1
    • View Profile
Re: Bias Question regarding Marshall 100watt 2203
« Reply #4 on: February 13, 2017, 05:12:27 PM »
Greetings Enzo,G1 and Phatt!

Alright so i have to divide /2, thanks!

Some pictures.

I agree, it looks like somebody got careless with the iron.
Someone reflowed one half of the output-jack joints, these are
the only very shiney looking joints, plus one rectifier diode
has been changed sometime, probably earlier.

Probably just coincidence that this tube socket of the worn
"looking" Pentode was infront of that jack. There is no excessive
heat in that area, as of now. The socket looks fine like all others.

I understand the voltage drifting up and down alittle,
perfectly normal like you guys said.
The webpage i got my information from on "how to bias" said
i should take the voltage to 3 decimal places, which firstly my cheap
dvm doesnt do at high voltage and there is no point anyways if there is

I will get a bias probe socket and some good clipping probes
these days to compare.

Phatt you were right, there are no dents in the plates, looks like these are little
particles of the mica-material. Still,only on this one tube?

Here are the measurements:

Screen-Resistors: t1=985ohms/t2=992ohms/t3=1035ohms/t4=1010ohms
(checked without desoldering and with tubes in socket!) i guess they are fine?

Neg.Bias-Voltage: t1=-34.6V/t2=-34.6V/t3=-34.5V/t4=-34.5V (pin5 to grnd)

Screen-voltage: t1=435V/t2=435V/t3=436V/t4=436V (pin4 to grnd)

Shouldn't the screen-voltage be higher than the plate voltage?

Also, the L.C.R. Caps are made in 1982, have dimpled tops but still i can
not hear any unexpected hum...swap for new ones?

One thing puzzles me about this amp, it has 8 rectifing silicon diodes
were the Marshall has 4, everything else is the same. Can someone
try to explain to me why that is...easy beginner here  :).

Thanks for your insights!
« Last Edit: February 13, 2017, 05:17:28 PM by solidstate2199 »


  • Legendary
  • ******
  • Posts: 2148
  • Chip Points: 262
    • View Profile
Re: Bias Question regarding Marshall 100watt 2203
« Reply #5 on: February 14, 2017, 07:07:14 AM »
White powder is likely the remains of the getter (shinny chrome patch on the top or side of the Valve)
Cause is likely a Leakey envelope,, air gets inside and burns up the valve.
Because the bias is so cold it may work but likely will still fail soon. I'd guess the Valve is close to end of life. :'(

Look and see if the chrome patch is a different or darker shade of colour than the other Valves. That chrome patch will go white if all vacuum is lost.

Screen Voltage has to be lower than plate voltage otherwise you destroy the screen grid.  The Screen grid is the weakest part of power valves. **Especially with EL34** as it's just a wire fence. when they blow they light up like tiny fire crackers for a split second,, Wanna guess how I know this? 8|

re screen grids and over voltage;
Always remember that the Neg charged electrons emitted from the cathode will travel to the highest voltage potential,,, so if Screen was at a higher DC Voltage than the Plate then Most of the electrons would go through the screen and little would pass to the Plate. Result screen wires would melt.

When replacing EL34's check the Screen MAX Voltage some will have a Screen limit of ~420VDC even though the plate can take ~700VDC. I repaired 2 Marshall rigs last year and both were running over spec,, 480VDC Plate and 465 on Screen both amps had dead or dying power valves.  Why Marshall ?? Why? :loco :duh xP :grr :trouble.
From King tuts books you can read;
The 2 main failure modes of EL34,, Lost bias,, and over dissapation of screen grids.

hope it helps,, Phil.

Ed; A lot of good info on how they work can be found here,,with some history;

« Last Edit: February 14, 2017, 07:31:29 AM by phatt »