I've noticed a few designs lately, including a 1949 article on a "Peerless" amp, using cathode followers coupled to the grids of the output triodes. A more recent example was the "2001 2A3" amp project on the angela instruments site.
My question is: what are the benefits? and are they worth the additional tube & signal path?
Also, isn't any benefit in the JE Labs version offset by the fact that it's RC coupled to the ouput grid?
My question is: what are the benefits? and are they worth the additional tube & signal path?
Also, isn't any benefit in the JE Labs version offset by the fact that it's RC coupled to the ouput grid?
Thanks Dhaen,
Yes, I thought that might be one of the justifications for it. BUT...
a) Is drawing grid current even a good idea with tubes not specifically designed for it??? ie. My fragile old globe 71A's don't look like they have too hefty of a grid.
b) In a design where the CF is rc-coupled to the output grid, how does the follower supply current through the coupling capacitor? Doesn't that block the dc current in the plate-cathode circuit?
Yes, I thought that might be one of the justifications for it. BUT...
a) Is drawing grid current even a good idea with tubes not specifically designed for it??? ie. My fragile old globe 71A's don't look like they have too hefty of a grid.
b) In a design where the CF is rc-coupled to the output grid, how does the follower supply current through the coupling capacitor? Doesn't that block the dc current in the plate-cathode circuit?
Hi Joel,
a) Of course you sould be careful, but what can happen? For any grid deformation to take place, the grid must get very hot. Calculate how much grid power is being produced, it'll be milliwatts for music. You'll probably have to set up a test circuit to establish the dynamic resistance.
b) As for as AC coupling; of course it doesn't work for A2, as the bias point moves through rectification. Even though some fairly respected circuit designers suggest it slides a little into A2, I believe it's rubbish.
There can other advantages though. If the valve being driven has high input capacitance, or to avoid parasitic oscillations.
Cheers,
a) Of course you sould be careful, but what can happen? For any grid deformation to take place, the grid must get very hot. Calculate how much grid power is being produced, it'll be milliwatts for music. You'll probably have to set up a test circuit to establish the dynamic resistance.
b) As for as AC coupling; of course it doesn't work for A2, as the bias point moves through rectification. Even though some fairly respected circuit designers suggest it slides a little into A2, I believe it's rubbish.
There can other advantages though. If the valve being driven has high input capacitance, or to avoid parasitic oscillations.
Cheers,
dhaen said:There can other advantages though. If the valve being driven has high input capacitance, or to avoid parasitic oscillations.
Thanks for the input again Dhaen!
That's really what I was looking for, as both of those conditions describe the situation in the PSE thread - paralleled tubes have more of a tendency toward parasitics, and DHT's have a high input capacitance too, if I'm not mistaken.
![idea :idea: :idea:](https://files.diyaudio.com/forums/images/smilies/idea.gif)
Joel
and DHT's have a high input capacitance too, if I'm not mistaken.
I don't know if they have any higher input capacitance than IHT's.
Generally the higher the mu, the higher the capacitance, for a given size of valve.
Anyway, what's a down to earth DIY'er like you want with DHT's?
Are you looking for some micro-details?
Cheers,
JOEL,JOEL.
Hi,
Just can't leave you two alone,can I?
What this has to do with cathode followers is beyond me.
Let me guess:you saw a circuit where the CF was driving an output tube that was biased positive,hence drawing grid current?
Why would one use cathode followers that have no gain,even detract from the overall gain:
First of all not all CF topologies are the same but generally these are used to:
Linearize the signal from the gainstage preceding them.
(they rely on 100% feedback)
Provide low output Z to drive the next stage effectively without loss.(other than the mu -1 gain)
They can swing a very respectable amount of voltage and provide healthy current.
Drive a bunch of paralelled tubes in an output stage with very wide bandwidth to combat the combined Miller effect of the tubes.
Effectively buffer the previous stage from the rest of the world.
There must be a few other applications but I presume this will keep you all busy for a while.
Joel,do you understand how a coupling cap works?
It blocks off DC voltage,NOT current nor AC voltage.
If it would nothing would work,would it??
Also you can still direct couple even when the next stage is drawing grid current,it just becomes that much harder to drive.
Ciao,![Wink ;) ;)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
Hi,
Just can't leave you two alone,can I?
a) Is drawing grid current even a good idea with tubes not specifically designed for it???
What this has to do with cathode followers is beyond me.
Let me guess:you saw a circuit where the CF was driving an output tube that was biased positive,hence drawing grid current?
Why would one use cathode followers that have no gain,even detract from the overall gain:
First of all not all CF topologies are the same but generally these are used to:
Linearize the signal from the gainstage preceding them.
(they rely on 100% feedback)
Provide low output Z to drive the next stage effectively without loss.(other than the mu -1 gain)
They can swing a very respectable amount of voltage and provide healthy current.
Drive a bunch of paralelled tubes in an output stage with very wide bandwidth to combat the combined Miller effect of the tubes.
Effectively buffer the previous stage from the rest of the world.
There must be a few other applications but I presume this will keep you all busy for a while.
b) In a design where the CF is rc-coupled to the output grid, how does the follower supply current through the coupling capacitor? Doesn't that block the dc current in the plate-cathode circuit?
Joel,do you understand how a coupling cap works?
It blocks off DC voltage,NOT current nor AC voltage.
If it would nothing would work,would it??
Also you can still direct couple even when the next stage is drawing grid current,it just becomes that much harder to drive.
Ciao,
dhaen said:I don't know if they have any higher input capacitance than IHT's. Generally the higher the mu, the higher the capacitance, for a given size of valve.
Guess I could just look it up - but I was feeling lazy.
Anyway, what's a down to earth DIY'er like you want with DHT's?
Are you looking for some micro-details?![]()
Haha.
Well, I guess I just thought it'd be fun to build a circuit or two with some really old tubes. They were all outside my budget except used 71a's. End result is in the "my current 71a amp" post. And luckily it sounds good too.
I think the real motivation was just to make a "DHT SET" amp, that everyone's been raving about, for a quarter of the money they spent.
REDUNDANT PERHAPS?
Hello,
For a rhetorical question it did get quite a few answers.
Ciao, (-5 minutes)
Hello,
My question is: what are the benefits? and are they worth the additional tube & signal path?
For a rhetorical question it did get quite a few answers.
Ciao, (-5 minutes)
Re: JOEL,JOEL.
Do you? (see below)
Frank, I'm no engineer, but I'm sure dc cannot flow through a capacitor. Were you really saying that???
fdegrove said:Joel,do you understand how a coupling cap works?
![redhot :redhot: :redhot:](https://files.diyaudio.com/forums/images/smilies/redhot.gif)
It blocks off DC voltage,NOT current nor AC voltage.
If it would nothing would work,would it??
Frank, I'm no engineer, but I'm sure dc cannot flow through a capacitor. Were you really saying that???
NOT QUITE.
Hi,
I know,I know I can be a major PITA,but:
Not necessarily:
What the problem is is that this high mu amplifies it's input capacitance therefore limiting and rolling off the high frequency range.
Ciao,![Wink ;) ;)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
Hi,
I know,I know I can be a major PITA,but:
Generally the higher the mu, the higher the capacitance, for a given size of valve.
Not necessarily:
What the problem is is that this high mu amplifies it's input capacitance therefore limiting and rolling off the high frequency range.
Ciao,
DIFFERENT WAVES?
Joel,
If it passes the AC signal voltage,why should it then be unable to pass AC current?
What they mean by it's capability to deliver current is that when high current is called for the CF will be able to put it out even when the impedance gets on the low side.
This is due to it's low output impedance as opposed to say a run off the mill anode follower.
Of course you will choose a CF with low mu and low Rp (natural state of affairs) and the more this tube can dissipate the more it will be able to supply this current.
It is not that complicated really.
Ciao,![Wink ;) ;)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
Joel,
If it passes the AC signal voltage,why should it then be unable to pass AC current?
What they mean by it's capability to deliver current is that when high current is called for the CF will be able to put it out even when the impedance gets on the low side.
This is due to it's low output impedance as opposed to say a run off the mill anode follower.
Of course you will choose a CF with low mu and low Rp (natural state of affairs) and the more this tube can dissipate the more it will be able to supply this current.
It is not that complicated really.
Ciao,
NO PROBLEM
Joel,
As I said,no problem better to ask then to make mistakes like this one:
Actually this is NOT correct.
I had it correctly in my mind but somehow managed to miswrite it
completely.
It should say: unity gain -1.
This means that this stage will have some insertion loss.
Sorry,my mistake,![Eek! :eek: :eek:](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
Joel,
As I said,no problem better to ask then to make mistakes like this one:
(other than the mu -1 gain)
Actually this is NOT correct.
I had it correctly in my mind but somehow managed to miswrite it
completely.
It should say: unity gain -1.
This means that this stage will have some insertion loss.
Sorry,my mistake,
I don't think you will notice the difference. If you're driving a low mu output valve, you'll probably want lots of peak-to-peak voltage, so the Va (Vp) and Ia (Ip) will be important.Of course you will choose a CF with low mu and low Rp (natural state of affairs)
Power pentodes are good for this.
None of the triode vs pentode arguments apply to cathode followers...
YOU AGAIN?
Hi,
True.
The lower the output impedance the easier it gets when the though gets going...
I never tried it,but penthodes are high Ri compared to triodes and are noisier as well.
Usually I go for White cathode followers using something like a 12BH7A.
I never had a situation where they couldn't handle it.
A penthode CF may be useful for driving a 845 for instance.
(admit it John,that was what you had in mind.)
Cheers,
Hi,
I don't think you will notice the difference. If you're driving a low mu output valve, you'll probably want lots of peak-to-peak voltage, so the Va (Vp) and Ia (Ip) will be important.
True.
The lower the output impedance the easier it gets when the though gets going...
Power pentodes are good for this.
I never tried it,but penthodes are high Ri compared to triodes and are noisier as well.
Usually I go for White cathode followers using something like a 12BH7A.
I never had a situation where they couldn't handle it.
A penthode CF may be useful for driving a 845 for instance.
(admit it John,that was what you had in mind.)
Cheers,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Amplifiers
- Tubes / Valves
- Cathode follower driver stage