6550 vs EL34 swap?

[bassybeats]

Just swapped my JJ EL34L's for some EH6550's and am very pleased with the sound. A lot warmer tone, deeper bas as well.

My question is do I have to swap the feed resistors from the stock ones? They are 220k at the moment but some say 150k is better for 6550's? My bias falls within spec at 50ma (64ish % of a 35w tube) and the plate voltage is 448v. The amp was a build from a Triode electronics layout, using Magnetic Components transformers which look like they can handle the extra current required by the 6550's.

Hope someone has done something similar before. I've read a few posts of similar things but people are arguing either way so didn't get a definite yes or no for my situation.

Cheers

[soundguruman]

the smaller the resistor, the less AC voltage swing, so installing the 150K limits the amount 6550 can be driven at the grid (or over-driven) by the preamp.

[demonufo]

the smaller the resistor, the less AC voltage swing, so installing the 150K limits the amount 6550 can be driven at the grid (or over-driven) by the preamp.

but also, and more importantly, the smaller the resistor, the more negative bias is applied to the tube. In a standard marshall circuit, 220K is often not giving you enough.

[joey]

I would check with EH's Datasheet for the optimum grid load for PP ab1fixed biased pentode operation. The values that are chosen have really nothing to do with trying to limit drive voltage/signal swing to the output tube grids, as 6550's/KT-88's are not very input sensitive valves.

The reason they are typically suggested to be a lower value for KT88/6550's than your typical 220K grid load is because those valves may have leakier grids(grids that will be more prone to attract electrons boiling off the cathode, before they can make it to the plate), and too high a value may permit unnecessary, and unwanted grid current to flow during normal, and quiescent operation, which will really muck up fixed biased Q operating point, resulting it what some have termed "Bias Runaway". It is very much the same concept as the maximum grid load spec'd on triode datasheets. If you notice some valves like 12AT7's for example have leakier grids , and specify no more of a grid load (including any possible grid stopper) than 1M, before unwanted current may flow. This phenomenon is the whole reason Gridleak Biasing can even be possible

If you notice the maximum permissible value is typically larger for Cathode bias operation on any given valves datasheet.

[demonufo]

I would check with EH's Datasheet for the optimum grid load for PP ab1fixed biased pentode operation. The values that are chosen have really nothing to do with trying to limit drive voltage/signal swing to the output tube grids, as 6550's/KT-88's are not very input sensitive valves.

The reason they are typically suggested to be a lower value for KT88/6550's than your typical 220K grid load is because those valves may have leakier grids(grids that will be more prone to attract electrons boiling off the cathode, before they can make it to the plate), and too high a value may permit unnecessary, and unwanted grid current to flow during normal, and quiescent operation, which will really muck up fixed biased Q operating point, resulting it what some have termed "Bias Runaway". It is very much the same concept as the maximum grid load spec'd on triode datasheets. If you notice some valves like 12AT7's for example have leakier grids , and specify no more of a grid load (including any possible grid stopper) than 1M, before unwanted current may flow. This phenomenon is the whole reason Gridleak Biasing can even be possible

If you notice the maximum permissible value is typically larger for Cathode bias operation on any given valves datasheet.

Aha. (Soaking it all up...)

So is there a link there between grid leakage, and why 12AT7's are NOT happy running as cathode followers in Marshalls?
I'm also wondering if that may have been an issue with the bad batch of Tung-Sol 12AX7's that appeared a few years back...

Of course, I'm probably way off base here

[soundguruman]

All I can say about the maximum possible values is that most amp manufacturers ignore them. Especially bogys.
The point being that the way to make an amp sound good is to overdrive the output tubes until they are screaming for mercy. And this is not accomplished by following the maximum values, rather by disregarding them.
And so if you are trying to make tubes last longer by biasing them cooler, you are also making the amp sound worse.
The hotter the output runs, the better the amp sounds. The cooler the amp runs, the worse it sounds.
Amp crafting includes the ability to compromise between short tube life (better sound), and longevity (crossover distortion). Unfortunately, amp builders just keep rehashing the same old design, and fail to create new and better compromises.

[joey]

All I can say about the maximum possible values is that most amp manufacturers ignore them. Especially bogys.
The point being that the way to make an amp sound good is to overdrive the output tubes until they are screaming for mercy. And this is not accomplished by following the maximum values, rather by disregarding them.
And so if you are trying to make tubes last longer by biasing them cooler, you are also making the amp sound worse.
The hotter the output runs, the better the amp sounds. The cooler the amp runs, the worse it sounds.
Amp crafting includes the ability to compromise between short tube life (better sound), and longevity (crossover distortion). Unfortunately, amp builders just keep rehashing the same old design, and fail to create new and better compromises.

I think this is a very single-minded viewpoint, which isn't necessarily true at all.

[joey]

I would check with EH's Datasheet for the optimum grid load for PP ab1fixed biased pentode operation. The values that are chosen have really nothing to do with trying to limit drive voltage/signal swing to the output tube grids, as 6550's/KT-88's are not very input sensitive valves.

The reason they are typically suggested to be a lower value for KT88/6550's than your typical 220K grid load is because those valves may have leakier grids(grids that will be more prone to attract electrons boiling off the cathode, before they can make it to the plate), and too high a value may permit unnecessary, and unwanted grid current to flow during normal, and quiescent operation, which will really muck up fixed biased Q operating point, resulting it what some have termed "Bias Runaway". It is very much the same concept as the maximum grid load spec'd on triode datasheets. If you notice some valves like 12AT7's for example have leakier grids , and specify no more of a grid load (including any possible grid stopper) than 1M, before unwanted current may flow. This phenomenon is the whole reason Gridleak Biasing can even be possible

If you notice the maximum permissible value is typically larger for Cathode bias operation on any given valves datasheet.

Aha. (Soaking it all up...)

So is there a link there between grid leakage, and why 12AT7's are NOT happy running as cathode followers in Marshalls?
I'm also wondering if that may have been an issue with the bad batch of Tung-Sol 12AX7's that appeared a few years back...

Of course, I'm probably way off base here

No, 12AT7's are not happy running as CF's for a variety of reasons, and the Grid leak maximum quoted doesn't hold up when we are talking about AC, or What we are tricking AC to believe its seeing such as in bootstrapped scenarios.

AT7's don't like the CF position because 1) the voltage at the cathode for CF's in most guitar amps can far exceede (double the maximum rating) the maximum heater to cathode differential. They would work perfectly fine for lower voltage DC coupled applications, or as AC coupled followers, or if you elevated their heater reference by 90V or so.

The second reason is that they have a bad reputation for not sounding all that great when overdriven, and the Cf of most marshall type amps does get overdriven.

[bassybeats]

Wow thats a lot to soak up!

I will try and find the spec sheet for the EH's. I'd rather bias on the hotter side as I don't tend to run my amps for hours and hours at a time, so I doubt i'd notice tube life that much, and if in doubt I have a big stash of spares anyway.

As it stands my Bias pot is full wound down. So I am at 50Mv and can't go lower even if I wanted to. So would installing the 150k give me a different range to work with?

I really love the sound of these and If I can make it run better/more efficient then I will try to do that. But am I doing any harm to any of the components/valves by leaving it as is? (except tube life shortening).

Cheers