• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

cascode vs asymmetrical cathode coupled

http://www.diyaudio.com/forums/showthread.php?s=&threadid=120643
split from post #16
Sheldon said:


Which triode characteristic does the cascode kill? The one triode characteristic I'm aware of, that the cascode arrangement kills, is the Miller capacitance. I don't think the plate curves don't all of a sudden become pentode like.

Sheldon

Hi Sheldon,
the influence of the plate voltage in the plate current
for example is lost.
Yes, the plate curves become pentode like.
The bottom tube sets the plate current of a cascode
arrangement, in a pentode the g2 current sets the
plate current.
Similar to the base current setting the collector current in a
bipolar transistor. ;)

Kind regards,
Darius
 
Ok, I'll buy that part. However, I think that the distortion profile will not be pentode like, in that the grid curves will not bunch up going from 0 to negative. So maybe it's something in between. Are there any examples of cascode curves published?

In my preamp (a version of the Bench amp), the second gain stage is a triode. I would expect that this would dominate the distortion profile, as the cascode should be quite linear.

Sheldon
 
Sheldon said:
Are there any examples of cascode curves published?


Yup, some datasheets for tubes designed to be run that way actually have curves. Look at 6BK7 and 6BQ7, for example...
 

Attachments

  • cascode.gif
    cascode.gif
    42.8 KB · Views: 866
#2

Originally #2 posted by Sheldon
... However, I think that the distortion profile will not be pentode like, in that the grid curves will not bunch up going from 0 to negative. So maybe it's something in between. Are there any examples of cascode curves published?

In my preamp (a version of the Bench amp), the second gain stage is a triode. I would expect that this would dominate the distortion profile, as the cascode should be quite linear.

Sheldon

Hi Sheldon,
I am sure it will be pentode like because in a cascode
the lower triode works in short circuit and not with µ gain. :(

The output tube should dominate the distrotion profile.
But this does not mean that the preveous stages have
no influence in it.
I want an all triode line. :)

Kind regards,
Darius
 
Sehr interresant!

Below 150V the plate curves are triode like, above that, pentode like. Well Darius, maybe we can eat our cake and have it too.

Sheldon

edit: I see that you were busy posting before I responded to Pete's neat little graph. Nice site Darius. I'll have to stare at it a while before I comprehend some fraction of it. For sure, corrected some assumptions I held and learned some new stuff today about cascodes.
 
#7 #8

Originally # 7 posted by Sheldon
Sehr interresant!

Below 150V the plate curves are triode like, above that, pentode like. ...

Hello Sheldon,
interesting discussion, thanks.
Pelase note that the anode of the cascode can not provide
any anode current at voltages below the cathode voltage
of the upper section. In Pentode curves the g2 voltage is fixt,
in this graph it comes from a high R voltage divider.
The curves below say 150V are mainly given by these resistors.
Thus there is nothing triode like in this cascode arrangement.

#8 Thank you very much, Pete.

Kind regards,
Darius :)
 
Re: #7 #8

oldeurope said:


Hello Sheldon,
interesting discussion, thanks.
Pelase note that the anode of the cascode can not provide
any anode current at voltages below the cathode voltage
of the upper section. In Pentode curves the g2 voltage is fixt,
in this graph it comes from a high R voltage divider.
The curves below say 150V are mainly given by these resistors.
Thus there is nothing triode like in this cascode arrangement.

Darius :)

I think I understand what you are saying. But let me make sure. The region below about 150V definitely looks triode like (the influence of plate voltage on current). The upper grid is set at 125V according to the figure. Are you saying that the high value resistors are functioning as a type of grid leak biasing, and that is causing the effect of plate voltage on current? Does it follow that in this region this configuration is not actually behaving as a cascode?

Sheldon
 
Re: #10

oldeurope said:
think about how the curves would look like with a
stabilized 125VDC supply at the grid of the upper triode.;)

I wasn't trying to be clever Darius, it was an honest question. I'm nowhere near your level on this stuff, but I like to learn. I think your answer is yes then? If so, then a lower impedance supply for the grid would shift the curves to look more like a typical pentode?

Sheldon
 
Dunworryboudit. The cascode is a two stage cascade: grounded cathode --> grounded grid. Loading down the grounded cathode input stage limits the resulting Cmiller. That's how I treat 'em for design purposes. Do the loadline for the second (grounded grid) stage, and figure out what the plate load for the input stage is. That gives enough open loop gain margin for a decent amount of gNFB in one stage. Unlike the high-u audio triodes, you don't get all that Cmiller to interact with the volume pot.

6BQ7 Loadlines

Here I used a 6BQ7 to make a cascoded LTP phase splitter/gain stage. The 6BQ7 characteristics aren't real good for audio, not as bad as some types, but not good either. Furthermore, the type doesn't have any pretenses of being an audio type at all. The 6BQ7 tends to be microphonic since you don't care if you get ringing at ~2.0KHz when the signal frequency is 100+MHz.

So how did this sound? Really good, actually! Of course, that's operating as a differential. How it works as a SE stage, who knows? It is promising enough to give it a try.
 
#16

Originally #16 posted by Miles Prower
Dunworryboudit. The cascode is a two stage cascade: grounded cathode --> grounded grid. Loading down the grounded cathode input stage limits the resulting Cmiller. That's how I treat 'em for design purposes. Do the loadline for the second (grounded grid) stage, and figure out what the plate load for the input stage is. That gives enough open loop gain margin for a decent amount of gNFB in one stage. Unlike the high-u audio triodes, you don't get all that Cmiller to interact with the volume pot....

Hello Miles Prower,
I agree, but the loading down of the grounded cathode stage gives this cascode
a pentode characteristic. You don't get the triode loadlines shown in your link #16.
What you get is something "transistor like" shown in #3 . This why I don't like the
cascode topology.
The asymmetrical cathode coupled is: cathode follower --> grounded grid.
In this topology you'll get the triode loadlines.

Kind regards,
Darius
 
Re: #16

oldeurope said:
This why I don't like the cascode topology.

Best way to find out: try it! That's what I did when I couldn't find out anything about the sonics of either a SE cascode or the cascode LTP. The cascode LTP worked out just fine, and it's definitely a do-it-again circuit. The main problem is a rather thin output swing for the Vpp.


The asymmetrical cathode coupled is: cathode follower --> grounded grid.
In this topology you'll get the triode loadlines.

That's a differential. It'll help with the Cmiller problem, but will cost you half your voltage gain.