Cookbook Solution for driving 600 Ohms from a Cathode Follower?

[velo]

Say we're coming out of triode cathode follower with a 1uF coupling capacitor into 1M resistor as the starting point. I think the cathode would be sitting on top of 470 + 10k (with feedback to grid between those). What is best way to drive very low distortion 600 ohm input impedance of next stage? ...without a solid state part?

[EDIT: Skip the 1M as it isn't really important.]

 

[Rusan]

Without knowing the particulars of your circuit and application, off the top of my head I'd maybe use a SSM2142, THAT 1606/1646, or LME49860.

 

[gyraf]

What does your 1M do?
How low distortion, and why?

Note that feedback cathode to grid does not compute, would be positive feedback if so. You probably have impedance bootstrapping? Or across multiple stages?

 

[ruffrecords]

Driving a 600 ohm load with a cathode follower is not easy. Yes, if you choose the right tube, the output impedance can be a lot less than 600 ohms but, as I have said many times before, output impedance is no indicator of drive capability. Just to drive 0dBu into 600 ohms requires nearly 1.3mA rms current or a peak current of over 1.8mA. This current has to come from the quiescent current in the tube. So if the tube is biased for 2mA you might just about be able to drive 0dBu into 600 ohms (if it can swing down to 0.2mA and up to 3.8mA). But of course this solution has no headroom. So you need to raise the quiescent current. If you raise it to 10mA you might be able to achieve up to +14dBu into a 600 ohm load. Of course there is a lot more to it than that but you need to get the basics right first.

Cheers

Ian

 

[velo]

Driving a 600 ohm load with a cathode follower is not easy. Yes, if you choose the right tube, the output impedance can be a lot less than 600 ohms but, as I have said many times before, output impedance is no indicator of drive capability. Just to drive 0dBu into 600 ohms requires nearly 1.3mA rms current or a peak current of over 1.8mA. This current has to come from the quiescent current in the tube. So if the tube is biased for 2mA you might just about be able to drive 0dBu into 600 ohms (if it can swing down to 0.2mA and up to 3.8mA). But of course this solution has no headroom. So you need to raise the quiescent current. If you raise it to 10mA you might be able to achieve up to +14dBu into a 600 ohm load. Of course there is a lot more to it than that but you need to get the basics right first.

Cheers

Ian

I was challenged to come up with a solution. So what I am describing here, I may not know all the constraints but I think that some attenuation might be made up after the following stage. So my first thought was to place a resistor in series with the 600 ohm stage--a passive EQ. But that'd be something like 10k-600ohm resistor and the loss would be around -12dBv I think.

 

[velo]

I was challenged to come up with a solution. So what I am describing here, I may not know all the constraints but I think that some attenuation might be made up after the following stage. So my first thought was to place a resistor in series with the 600 ohm stage--a passive EQ. But that'd be something like 10k-600ohm resistor and the loss would be around -12dBv I think.

Could something like this drive it? Maybe I lied about solid state.

https://www.tubecad.com/2011/08/blog0212.htm

 

[velo]

What does your 1M do?
How low distortion, and why?

Note that feedback cathode to grid does not compute, would be positive feedback if so. You probably have impedance bootstrapping? Or across multiple stages?

I believe low distortion vocals chain. I had similar questions. I was presented with a partial sketch of the stage so I am not sure. I think the 1M was just to make SPICE happy for modeling prior to considering design alternatives to couple to next stage.

 

[abbey road d enfer]

I was challenged to come up with a solution.

Are you obliged to answer this challenge?
It's like asking to tract a mobile home with a moped.
In order to drive decently ( i.e. at least +18dBu) a 600 ohms load, you need at least 35-40mA quiescent current with a class-A stage. You would need extensive modification of the circuit.
I would decline and propose a different approach, using a step-down transformer, probably a push-pull or SRPP topology, or even a solid-state solution.

 

[k brown]

Can't just raise the input impedance of the downstream gear it'll connect to?

 

[velo]

Can't just raise the input impedance of the downstream gear it'll connect to?

How? One of my "outside the box" proposals was simply to add a large resistor in series as if Z + R = higher Z. The dB loss would be large but could make up for that. However, that doesn't address the minimum current needs that Ian was pointing out. I still have a lot to learn, coming from a mostly digital background.

I ask how in non-obvious ways. He does not want to use a transformer or an op amp, but I think a single transistor or FET would be ok.

 

[velo]

Are you obliged to answer this challenge?
It's like asking to tract a mobile home with a moped.
In order to drive decently ( i.e. at least +18dBu) a 600 ohms load, you need at least 35-40mA quiescent current with a class-A stage. You would need extensive modification of the circuit.
I would decline and propose a different approach, using a step-down transformer, probably a push-pull or SRPP topology, or even a solid-state solution.

Leaning toward a hybrid push-pull yes with solid-state.

 

[merlin]

Could something like this drive it? Maybe I lied about solid state.

You could do that, yes. It's similar to a White cathode follower, which we know can just about drive 600 ohms even with two triodes.

 

[gyraf]

..please specify what level you need to drive.

No, not "it would be psychologically easiest if it could drive +28dB into 600 Ohms" - but what you actually need.

Asking because mentioning driving passive eq indicates the existence of a later makeup gain amplifier, which could lift this burden much easier..

Perhaps this is a bit of a xy-problem? https://xyproblem.info/

/Jakob E.

 

[abbey road d enfer]

I ask how in non-obvious ways. He does not want to use a transformer or an op amp, but I think a single transistor or FET would be ok.

Is there any objective motivation for that?
I'm inclined to think your client is an idiot, and thus would recommend avoiding him as far as possible.
He's asking you to solve a problem (which may turn out not to be one) with a restricted set of tools and tunnel vision.

 

[velo]

..please specify what level you need to drive.

No, not "it would be psychologically easiest if it could drive +28dB into 600 Ohms" - but what you actually need.

Asking because mentioning driving passive eq indicates the existence of a later makeup gain amplifier, which could lift this burden much easier..

Perhaps this is a bit of a xy-problem? https://xyproblem.info/

/Jakob E.

Yes definitely an XY problem.

But nevermind that, since we're driving a passive RLC circuit yes (!) one of the primary debates is final makeup gain as an alternative. So boiling off the XY fog...

Going with a solid state part and jellybean circuit simulator I get maybe -6dBV from the whole thing and 441µA.

I'd need to look at reactance frequency spectrums for each filter setting and maybe use one of the SPICE tools. Also thinking if I am really feeling stubborn I could write a Python or Javascript program to compute theoretical frequency response given any combinations of passive filters inside the EQ.

For grins I wanted to know what the current would look like in a purely passive impedance matcher a.k.a. voltage divider compared to hybrid. I get 82µA.

 

[velo]

Is there any objective motivation for that?
I'm inclined to think your client is an idiot, and thus would recommend avoiding him as far as possible.
He's asking you to solve a problem (which may turn out not to be one) with a restricted set of tools and tunnel vision.

Objective is mostly to work within real estate and cost budgets from what I gather.

 

[abbey road d enfer]

YFor grins I wanted to know what the current would look like in a purely passive impedance matcher a.k.a. voltage divider compared to hybrid. I get 82µA.

My LTspice sim says different.
At 1kHz, with 1Vrms input, I get 0.53Vrms into 600 ohms, for a current of 0.88mA rms. This with a 12AX7, which is far from optimum.
A 12AT7 would deliver 0.66Vrms for 1.1mArms.
Of course, the 1uF output cap results in a -3dB LF point at nearly 200Hz.
Running the tube at a larger quiescent would result in significantly increased output capability, but not enough yet to approach usual standards.

 

[Eskiclown]

Could something like this drive it? Maybe I lied about solid state.

If your output stage has to be a cathode follower and a hint of solid state is acceptable for your client, John Broskie has another hybrid solution to "boost" the drive capability of a cathode follower : Solid-State Superchargers
(look at the "Supercharged Cathode Follower" schematic of the "Supercharged Totem-Pole Output Stages" section)

If you have power supply constraints, it has the advantage of not requiring an additional -12V power supply compared to the hybrid White cathode follower that you consider.

But since the solid state device is doing most of the job in this configuration, it may be tempting to replace this whole hybrid thing with a simple MOSFET source follower...

 

[velo]

My LTspice sim says different.
At 1kHz, with 1Vrms input, I get 0.53Vrms into 600 ohms, for a current of 0.88mA rms. This with a 12AX7, which is far from optimum.
A 12AT7 would deliver 0.66Vrms for 1.1mArms.
Of course, the 1uF output cap results in a -3dB LF point at nearly 200Hz.
Running the tube at a larger quiescent would result in significantly increased output capability, but not enough yet to approach usual standards.

Thanks! I think that fasltad Javascript app has limited thermionic simulation and I put in a low mu to force me more into worst case territory.

 

[velo]

If your output stage has to be a cathode follower and a hint of solid state is acceptable for your client, John Broskie has another hybrid solution to "boost" the drive capability of a cathode follower : Solid-State Superchargers
(look at the "Supercharged Cathode Follower" schematic of the "Supercharged Totem-Pole Output Stages" section)

If you have power supply constraints, it has the advantage of not requiring an additional -12V power supply compared to the hybrid White cathode follower that you consider.

But since the solid state device is doing most of the job in this configuration, it may be tempting to replace this whole hybrid thing with a simple MOSFET source follower...

https://tinyurl.com/27ayh3wz

Nice, thanks! With and without Turbo Boost.