JCM800 2203 Conversion from 6550 to EL34

  • Thread starter Thread starter Bxlxaxkxe
  • Start date Start date
B

Bxlxaxkxe

Amp ho
There is a lot of conflicting information when I google this. What needs to be done to get this conversion done properly? 1983 2203 specifically.
 
Rick Lee said:
I think it's just changing bias cap values, so it will bias within spec.
Click to expand...
From what I've read that's the bare minimum, but further changes required to get it to complete EL34 spec. But I'm a complete amateur so hoping the experts will chime in. Maybe I'm wrong
 
I just had it done a few months ago on my 4010, which is really a 50w JCM800 combo. I was already going to recap the thing with the ARS caps I had laying around. But it had some other issues I didn't think I'd be able to track down, so I had my tech do it all. I really think he just changed the bias caps. But it may have been more.
 
Sorry, you're right. It's resistors. However, this is the time for new bias caps too. If they blow, it's trouble. I like the 6550s a lot too, but they were getting to $70 each around the start of the Ukraine stuff and I have a lot of EL34s already stocked.
 
You need to change the negative feedback tap, as they differed between EL34 and 6550, you need to change series grid resistor values (220k vs 150k, I forget which is for 6550), you need to change bias resistor values (to get the 6550s biased correctly), and if it hasn’t been recapped, you need to change the bias caps as well ( two 10uF or 8uF, they used a pair of either one or the other).

Someone else can chime in on the grid resistor values to swap to, I can never remember it. You will need to change these as it will affect bias runaway conditions/stability.
 
I think there are five resistors to change and the UK Marshall's had the negative feedback on a different tap, like others mentioned.
 
glpg80 said:
You need to change the negative feedback tap, as they differed between EL34 and 6550, you need to change series grid resistor values (220k vs 150k, I forget which is for 6550), you need to change bias resistor values (to get the 6550s biased correctly), and if it hasn’t been recapped, you need to change the bias caps as well ( two 10uF or 8uF, they used a pair of either one or the other).

Someone else can chime in on the grid resistor values to swap to, I can never remember it. You will need to change these as it will affect bias runaway conditions/stability.
Click to expand...
EL34 Marshall's run 220K bias spillter(bias grid resistors). 6550 Marshall's from the factory used 150K bias splitter resistors(bias grid resistors) The fear of runaway bias never really became an issue. If you are a stickler for technicalities 6550 amps should run 82K bias splitters and .1uf PI coupling caps to pass the same bass frequencies as the 220K/.022uf configuration does. With all that being said I have been running my Splawn modded 79 Marshall with the 220K/.022uf bias splitters with 6550's with absolutely no runaway bias or any other issues for that matter since 2003. Use the 150K bias splitters if you want some piece of mind as it is what Marshall used in all the USA 6550 amps.

I also used the 82K/.1uf configuration in a 4 stage Marshall build and it worked and sounded fine...just like a Marshall. I would go with the 150K that way that resistor/cap combination is close enough you don't need to swap the .022uf coupling caps out. If you went lower to the 82K's then it would need at least .047uf or .1uf to mover the same amount of bass frequencies.

As for the bias supply circuit a stock factory EL34 Marshall is 27K to the diode then 15K then 47K before the bias pot.
A 6550 Marshall is 15K to the diode then a 15K then a 47K before the bias pot.

Marshall used the 8 ohm tap for EL34's and the 4 ohm tap for 6550's but really it's a personal preference. 8 ohm tap is slightly cleaner sounding and 4 ohm tap is a little rattier sounding.

If the 10uf bias supply capacitors are still original then I would change them out. If they have been replaced recently I would not worry about them unless you want to change them while you are lifting the board for the other work for peace of mind.
 
Last edited:
Why do I recall moving all the pins over one back in the day...
 
Exo-metal said:
Why do I recall moving all the pins over one back in the day...
Click to expand...
The tubes pinout the same if that's what you mean. There's probably another conversion you're thinking of that requires that.

IIRC, the 6550 amps had both the bias supply resistor (the one before the diode) and the range resistor (the one before the pot) changed but most people just change the range resistor to convert to EL34s. I don't know if they did that because the cap's voltage rating would have been exceeded since 6550s need more bias voltage. The raw supply itself from a dedicated bias tap is typ ~95-100vdc.
 
harddriver said:
EL34 Marshall's run 220K bias spillter(bias grid resistors). 6550 Marshall's from the factory used 150K bias splitter resistors(bias grid resistors) The fear of runaway bias never really became an issue. If you are a stickler for technicalities 6550 amps should run 82K bias splitters and .1uf PI coupling caps to pass the same bass frequencies as the 220K/.022uf configuration does. With all that being said I have been running my Splawn modded 79 Marshall with the 220K/.022uf bias splitters with 6550's with absolutely no runaway bias or any other issues for that matter since 2003. Use the 150K bias splitters if you want some piece of mind as it is what Marshall used in all the USA 6550 amps.

I also used the 82K/.1uf configuration in a 4 stage Marshall build and it worked and sounded fine...just like a Marshall. I would go with the 150K that way that resistor/cap combination is close enough you don't need to swap the .022uf coupling caps out. If you went lower to the 82K's then it would need at least .047uf or .1uf to mover the same amount of bass frequencies.

As for the bias supply circuit a stock factory EL34 Marshall is 27K to the diode then 15K then 47K before the bias pot.
A 6550 Marshall is 15K to the diode then a 15K then a 47K before the bias pot.

Marshall used the 8 ohm tap for EL34's and the 4 ohm tap for 6550's but really it's a personal preference. 8 ohm tap is slightly cleaner sounding and 4 ohm tap is a little rattier sounding.

If the 10uf bias supply capacitors are still original then I would change them out. If they have been replaced recently I would not worry about them unless you want to change them while you are lifting the board for the other work for peace of mind.
Click to expand...
This is what I’ve been looking for. Appreciate this.

Now my question is, would there be an audible difference between leaving everything in stock 6550 form except for the piggyback resistor to bias up EL34s VS. making these other changes?

In other words, should I bother taking the extra steps beyond the piggyback resistor??
 
Last edited:
Bxlxaxkxe said:
This is what I’ve been looking for. Appreciate this.

Now my question is, would there be an audible difference between leaving everything in stock 6550 form except for the piggyback resistor to bias up EL34s VS. making these other changes?

In other words, should I bother taking the extra steps beyond the piggyback resistor??
Click to expand...
I am excerpting this conversion method out of the "TUBE AMP BOOK" 4th edition by Aspen Pittman page 763 Marshall w/6550 to Marshall w/EL34

For a 6550 amp to be converted to EL34's the bias needs to be decreased by about 10 volts roughly. The quick a dirty way alot of people do this is locating the 47K resistor which is directly above the and in line with the bias pot on a PCB board. Solder a 100K resistor in parallel with the 47K resistor. With the bias bias pot set to "MAX" and the output tubes OUT OF THE AMP, check to see that at least -38 DCV is present on the grid of the socket (PIN 5). If so, install the EL34 tubes and bias the amp. If there is more than -45 DCV, that's okay but much less may result in not being able to bias the amp properly. The value of the added resistor may be adjusted to set that voltage, with a lower value lowering the available bias.

Another important item to change when changing 6550's to EL34'2 or vice versa is teh feedback wire that connects to the speaker impedance selector from the output transformer. This feedback wire, which is purple on Marshall amps, should be connectes to the 4 ohm leg of the switch if you are using 6550's and the 8 ohm leg if using EL34's. 100 watt Marshall OT are as follows: 16 ohm green, 8 ohms yellow, 4 ohms black, common- brown. 50 watt Marshalls OT 16 0hm Gray, 8 ohms green, 4 ohms yellow, common-orange.

Excerpt complete:

I tried EL34's years ago and went right back to my 6550's, but alot of people love their EL34's. I hope this helps you out. Unless you are absolutley sure you want to make the conversion permament I would not worry about it. As long as the 100K is soldered properly you should not have any issues with it and it makes things alot easier not having to lift the circuit board. Most techs would probably do the conversion this way unless they are OCD. If the amp is a 6550 amp then it is running 150K bias splitters and that will not harm the EL34's at all if you don't want to change them and the differene is essentially negligible tonewise IMHO. In fact some people like the lower value Bias splitters with EL34 and I think it actually lessens the stress on the modern EL34's some. I think it will be fine to leave it as is and just do the piggy back 100K. Some may have another opinion on that so I will cede the floor to them.
 
Last edited:
harddriver said:
I am excerpting this conversion method out of the "TUBE AMP BOOK" 4th edition by Aspen Pittman page 763 Marshall w/6550 to Marshall w/EL34

For a 6550 amp to be converted to EL34's the bias needs to be decreased by about 10 volts roughly. The quick a dirty way alot of people do this is locating the 47K resistor which is directly above the and in line with the bias pot on a PCB board. Solder a 100K resistor in parallel with the 47K resistor. With the bias bias pot set to "MAX" and the output tubes OUT OF THE AMP, check to see that at least -38 DCV is present on the grid of the socket (PIN 5). If so, install the EL34 tubes and bias the amp. If there is more than -45 DCV, that's okay but much less may result in not being able to bias the amp properly. The value of the added resistor may be adjusted to set that voltage, with a lower value lowering the available bias.

Another important item to change when changing 6550's to EL34'2 or vice versa is teh feedback wire that connects to the speaker impedance selector from the output transformer. This feedback wire, which is purple on Marshall amps, should be connectes to the 4 ohm leg of the switch if you are using 6550's and the 8 ohm leg if using EL34's. 100 watt Marshall OT are as follows: 16 ohm green, 8 ohms yellow, 4 ohms black, common- brown. 50 watt Marshalls OT 16 0hm Gray, 8 ohms green, 4 ohms yellow, common-orange.

Excerpt complete:

I tried EL34's years ago and went right back to my 6550's, but alot of people love their EL34's. I hope this helps you out. Unless you are absolutley sure you want to make the conversion permament I would not worry about it. As long as the 100K is soldered properly you should not have any issues with it and it makes things alot easier not having to lift the circuit board. Most techs would probably do the conversion this way unless they are OCD. If the amp is a 6550 amp then it is running 150K bias splitters and that will not harm the EL34's at all if you don't want to change them and the differene is essentially negligible tonewise IMHO. In fact some people like the lower value Bias splitters with EL34 and I think it actually lessens the stress on the modern EL34's some. I think it will be fine to leave it as is and just do the piggy back 100K. Some may have another opinion on that so I will cede the floor to them.
Click to expand...
Perfect info- thanks man.
 
Bias ranges for Marshall amps are as follows:

Marshall 100 watt w/EL34's -36 to -45
Marshall 50 watt w/ EL34's -34 to -40

Marshall 100 watt w/6550's -44 to -55
Marshall 50 watt w/6550's -38 to -48

This the negative bias read on PIN 5 of the power tube sockets.
 
I only ever do the bias supply to 220k and the bias resistor. Never change the NFB, etc.
 
harddriver said:
Bias ranges for Marshall amps are as follows:

Marshall 100 watt w/EL34's -36 to -45
Marshall 50 watt w/ EL34's -34 to -40

Marshall 100 watt w/6550's -44 to -55
Marshall 50 watt w/6550's -38 to -48

This the negative bias read on PIN 5 of the power tube sockets.
Click to expand...
Now, I understand ideal bias range for 6550s are higher like you posted BUT I've also read it's safer to bias the 6550s like EL34s; since the transformers are designed for them and not 6550s. Less stress etc on the transformers.
 
Racerxrated said:
Now, I understand ideal bias range for 6550s are higher like you posted BUT I've also read it's safer to bias the 6550s like EL34s; since the transformers are designed for them and not 6550s. Less stress etc on the transformers.
Click to expand...
I have not heard this myself I do believe 6550 must be run as any other vaccum tube within their bias operating range specific to that tube referring to the - bias voltage on pin 5 in the amplifier circuit in relation to the output transformer impedance too operate properly.

Now Marshall transformers are actually spec'd out impedancewise for EL-34 that's why Marshall was worried the 6550's might have runaway bias/instability issues that's why they used the 150K bias splitters as a happy medium to ward off any issues. From what I gather those runaway bias issues never really became an issue for whatever reason. I personally have never had an issue running 6550's in my Splawn modded 79 Marshall an Scott installed 220K bias splitters back in 2003 when he installed one of his boards when he rebuilt the amp he knew it was running 6550's. That amp is running a Drake Power transformers and a Mercury Magnetics Output transformer.

NOS GE6550's are rated for 42 Watts dissipation. For newer production 6550's/KT-88 I would use 35 watts dissipation to be on the safe side. Now I guess you could use the EL34 watts dissipation of 25 watts (if that is what you are referring to) to even be safer but I think 35 watts dissipation is enough for new production 6550/KT-88's. Using 25 watt dissipation would just lower the ma ranges making the tubes run even cooler.
 
harddriver said:
I have not heard this myself I do believe 6550 must be run as any other vaccum tube within their bias operating range specific to that tube referring to the - bias voltage on pin 5 in the amplifier circuit in relation to the output transformer impedance too operate properly.

Now Marshall transformers are actually spec'd out impedancewise for EL-34 that's why Marshall was worried the 6550's might have runaway bias/instability issues that's why they used the 150K bias splitters as a happy medium to ward off any issues. From what I gather those runaway bias issues never really became an issue for whatever reason. I personally have never had an issue running 6550's in my Splawn modded 79 Marshall an Scott installed 220K bias splitters back in 2003 when he installed one of his boards when he rebuilt the amp he knew it was running 6550's. That amp is running a Drake Power transformers and a Mercury Magnetics Output transformer.

NOS GE6550's are rated for 42 Watts dissipation. For newer production 6550's/KT-88 I would use 35 watts dissipation to be on the safe side. Now I guess you could use the EL34 watts dissipation of 25 watts (if that is what you are referring to) to even be safer but I think 35 watts dissipation is enough for new production 6550/KT-88's. Using 25 watt dissipation would just lower the ma ranges making the tubes run even cooler.
Click to expand...

On my 6550 2203 I ran them at 25 watts to be safer. I think it was a little cold but it still sounded ok. Running them a little hotter would be better though.

The 6550s were cleaner and more umph. With a 2203 JMP it made it more like acdc to me. Neither is better but thinking there will be a dramatic change isnt true. I do understand that its easier to find el34s though, that makes sense.

PS. My 2203 JMP came to me with el34s. The amp shipped with 6550s. Whoever the previous tech was didnt update anything. He installed el34s and biased it as hot as he could but when I checked it was running at 13. But the bias splitters were still stock 150k.
 
harddriver said:
EL34 Marshall's run 220K bias spillter(bias grid resistors). 6550 Marshall's from the factory used 150K bias splitter resistors(bias grid resistors) The fear of runaway bias never really became an issue. If you are a stickler for technicalities 6550 amps should run 82K bias splitters and .1uf PI coupling caps to pass the same bass frequencies as the 220K/.022uf configuration does. With all that being said I have been running my Splawn modded 79 Marshall with the 220K/.022uf bias splitters with 6550's with absolutely no runaway bias or any other issues for that matter since 2003. Use the 150K bias splitters if you want some piece of mind as it is what Marshall used in all the USA 6550 amps.

I also used the 82K/.1uf configuration in a 4 stage Marshall build and it worked and sounded fine...just like a Marshall. I would go with the 150K that way that resistor/cap combination is close enough you don't need to swap the .022uf coupling caps out. If you went lower to the 82K's then it would need at least .047uf or .1uf to mover the same amount of bass frequencies.

As for the bias supply circuit a stock factory EL34 Marshall is 27K to the diode then 15K then 47K before the bias pot.
A 6550 Marshall is 15K to the diode then a 15K then a 47K before the bias pot.

Marshall used the 8 ohm tap for EL34's and the 4 ohm tap for 6550's but really it's a personal preference. 8 ohm tap is slightly cleaner sounding and 4 ohm tap is a little rattier sounding.

If the 10uf bias supply capacitors are still original then I would change them out. If they have been replaced recently I would not worry about them unless you want to change them while you are lifting the board for the other work for peace of mind.
Click to expand...

Im curious about the PI cap values. I never looked into this in context.

EL34 - 220k splitter - .022 caps
6550 - 150k - still has the .022 caps right?
 
You must log in or register to reply here.
Back
Top