Fewest gain stages.....

[imaradiostar]

Would paralleling get you lower output impedance?
Sorry, but I'm not familiar with using pentodes in preamps. How would the output impedance of a single LND150 stage compare with a 12ax7 stage in the first place?

russ

OK, I'm not an EE but this is how I understand it.

The output impedance is a function of the plate (or drain) resistor. The gain is a function of the device being able to "pull down" the plate resistor.

In other words, a high transconductance low plate impedance device (like an ECC88) can create a large voltage change even when used with a smaller than usual plate resistor. A 12ax7 isn't a terribly high transconductance device (but it still has a high plate resistance) so it demands a high plate resistance and impedance to make a lot of gain.

From that standpoint an LND150 is like a perfect tube. Up to its power dissipation limit you can use as small a drain resistor as you want to achieve a lower output impedance. It has super high transconductance- a tiny voltage change at the gate creates a huge drain-source current change.

that said- what's wrong with a 47k to 100k plate or drain resistor? Why do you need a lower impedance?

jamie

[Russ]

that said- what's wrong with a 47k to 100k plate or drain resistor? Why do you need a lower impedance?
jamie

Well going to 47k sounds good to. My reason for lower output impedance is that in the case of using an LND150 as a clean boost stage to overdrive a tube stage after it, it's my understanding that lower output impedance equates to lower source impedance, which will allow the overdriven tube stage to be overdriven more before going into blocking distortion. I'm basically trying to limit blocking dist. Maybe this works, or maybe only up to a point, not sure.

russ

[imaradiostar]

Ah, I see. If your goal is to gently draw grid current then why not directly couple (no cap) a mosfet source follower directly to the grid of a triode? If I understand things correctly then the grid stopper between the source and the grid could function as a sort of "softness" control for the clipping- the lower the value of the grid stopper, the more current the source follower could "push" into the grid, softening the transition into grid current draw. It might be an interesting experiment to wire something like that up with a pot between the source and the following grid.

jamie

[J M Fahey]

Well, the original poster left us in the middle of 2007, never to come back, so in a way this is like beating a dead horse.
Anyway, I *love* Rube Goldberg type designs , and find them very funny.
Just as an example:

For the full explanation:
Rube Goldberg: Self-Operating Napkin
Best of all: they work !!! (he *was* an Engineer after all)
He would have loved the 42 LED biasing scheme, although probably would have added some photodiodes in front of them, so as to not waste their light and, for example, bias an earlier stage with that voltage.

[Steve Conner]

Hey, they're not Rube Goldberg machines, they're Heath Robinson contraptions.

You can parallel two of any amplifying device for a 3db noise reduction. But I'd expect a LND150 to be quieter than a tube anyway. They don't have as much gm as you think, about 3mA/v if I remember right, which isn't that much more than a 12ax7.

No matter what the source impedance, if it's AC coupled there will be blocking distortion. If you have a low impedance source, you can make the grid leak resistor of the following stage smaller and that will swamp it, especially if you add a series grid stopper too. If you get rid of the bias shift/ blocking behaviour completely the amp will sound boring.

Or you can just use another LND150 for the next stage, they are MOSFETs so can't draw any grid current. Of course, that is an oversimplification, if you put more than about 15-20v on the gate you can blow it. The LND150 may contain a Zener to protect the gate, which would give you some "grid" current.

[imaradiostar]

Hey, they're not Rube Goldberg machines, they're Heath Robinson contraptions.

You can parallel two of any amplifying device for a 3db noise reduction. But I'd expect a LND150 to be quieter than a tube anyway. They don't have as much gm as you think, about 3mA/v if I remember right, which isn't that much more than a 12ax7.

No matter what the source impedance, if it's AC coupled there will be blocking distortion. If you have a low impedance source, you can make the grid leak resistor of the following stage smaller and that will swamp it, especially if you add a series grid stopper too. If you get rid of the bias shift/ blocking behaviour completely the amp will sound boring.

Or you can just use another LND150 for the next stage, they are MOSFETs so can't draw any grid current. Of course, that is an oversimplification, if you put more than about 15-20v on the gate you can blow it. The LND150 may contain a Zener to protect the gate, which would give you some "grid" current.

You're right, the transconductance isn't that different. Which value on the LND 150 corresponds to plate resistance on a triode? Whatever the case, it sure seems like the mosfet has way more gain into a lower resistance drain load than the comparable 12ax7 plate load.

New thread time?

jamie

[J M Fahey]

New thread time?

Excellent idea.
Won't answer here to avoid "upping" this one, see you in the new LD150 (or HV Mosfets) topic.

[mooreamps]

What are the fewest gain stages that can be used to build a "practical" (ie, something strong enough to be heard in a busy nightclub) amp? Will solid state require more gain stages? Or is the pure, single ended class A tube amp the simplest?

6U8 > SRPP > Volume > one knob tone stack > KT-88...

-g