Constant current instead of fixed bias?

[markr14850]

f you are talking about replacing R10 and R20, go for it.

I've never tried it, but I've read that running both a constant current source on the plates and a constant voltage on the cathodes (the LED) doesn't work so well.

Might want to try one or the other first before combining them.

[skyboltone]

I was planning just the power tubes by the way. Sort of like a HI WATT setup only SS.

I'm a bit confused...

Are you interested in a low impedance drive for the power tube grids - like "MOSFET Follies", or "PowerDrive"?

Or are you considering cathode biasing the output tubes via LED?

Or something else?

Option number two. And forget the HiWatt connection, I was thinking something else. But, like I sorta said, I need to smarten up a bit about it. The amps that I've seen with cathode LED bias take a board with a big string of Reds to get to bias voltage. It then becomes more like fixed bias because it's no longer dependent on voltage drop pushing ~ current across a fixed resistance. Even with ordinary fixed bias it's anything but fixed unless it's supply is totally separate and robust. I have experimented with regulated screen voltage (like the Leslie amps) in the past because some tubes can run awfully close to triode mode unless pushed very hard. 807s don't like that. Regardless of these conventional means, bias and screen voltages can flop around quite a bit when a tube is clipping. Having observed all that, I still suspect that this "loose" relationship is probably integral to "guitar amp" tone. But another instance; Leslie speaker amps clip just fine!
One thing I wonder though, perhaps you can help. Why not just set bias voltage with a zener string sourcing a transistor? What advantage do LEDs have over a zener other than that they drop just about exactly what you want in a pre amp bias arrangement?

[TheGimp]

LEDs are not constant current, they are constant voltage. So, I was wrong to state they would not work well for class AB. They should work fine.

The combination of CC anode load and LED constant voltage cathode bias will produce a very linear (low distortion) output.

[TheGimp]

Zeners are noisy, especially around 6.8V where the technology transitions from true zener to avalanch.

The dynamic impedance of the parallel LED strings is very low and this contributes to rapid overlaod recovery.

Sy has a good write-up on the RLD amp here,

http://syclotron.com/?page_id=3

[skyboltone]

LEDs are not constant current, they are constant voltage.

The combination of CC anode load and LED constant voltage cathode bias will produce a very linear (low distortion) output.

I think I just had a light bulb moment so to speak. So what I thought:
So if the LEDs nail the cathode voltage then the cathode current MUST be nailed too. Wrong!

What I think you are saying and I just grocked:
The LEDs only give a constant DC bias. The AC current is still dependent on the grid swing and it's voltage drop across the anode resistor. So that's why you need the fancy CC anode load to arrive at that linearity across all kinds of drive levels. Sounds boring in a guitar amp eh?

Did I pass?

[Ian444]

The amps that I've seen with cathode LED bias take a board with a big string of Reds to get to bias voltage. It then becomes more like fixed bias because it's no longer dependent on voltage drop pushing ~ current across a fixed resistance. Even with ordinary fixed bias it's anything but fixed unless it's supply is totally separate and robust. I have experimented with regulated screen voltage (like the Leslie amps) in the past because some tubes can run awfully close to triode mode unless pushed very hard. 807s don't like that. Regardless of these conventional means, bias and screen voltages can flop around quite a bit when a tube is clipping. Having observed all that, I still suspect that this "loose" relationship is probably integral to "guitar amp" tone. But another instance; Leslie speaker amps clip just fine!
One thing I wonder though, perhaps you can help. Why not just set bias voltage with a zener string sourcing a transistor? What advantage do LEDs have over a zener other than that they drop just about exactly what you want in a pre amp bias arrangement?

A string of LED's in the cathode of the output tube is very close to fixed bias. I have only tried it in several hifi amps and it kicks ass compared to self-bias, it gives extended freq response, big bottom end, slam, attack, fast, higher top end, drums sound great, better dynamics. It's like night and day. Dunno if this is good or bad for a guitar amp. One of the amps that was running LED bias got converted to fixed bias and I could not tell any difference. Even then though, many hifi people like the sound of self-bias too. Something else I did on a couple of amps was add mosfet source followers (ZVN0545 with source resistor instead of CCS) between the phase splitter outputs and output tube grids, that gave some extra clarity and was a startling improvement in a hifi amp.

Not all LED's are the same, they have different current vs voltage curves, which means some will give a bit more cathode feedback than others, much the same as the difference between a 1 ohm or 10 ohm cathode resistor for measuring bias in a fixed bias amp. For example, a 10 ohm cathode resistor with 150mA peak across it gives a 1.5V p-p across it, subtracting from the grid drive voltage. Nice bit of local feedback for a hifi amp.

>Why not just set bias voltage with a zener string sourcing a transistor? What advantage do LEDs have over a zener other than that they drop just about exactly what you want in a pre amp bias arrangement?

LED's are really quiet apparently. In hifi, I never liked an LED in the preamp, but others love it. Guess it's down to the individual ears.