I have a question that may have a real simple answer, so please forgive me if it sounds dumb. I see many amplifier designs in which one half of a dual triode is hooked up as a common cathode voltage gain stage and then is buffered by the second half as a cathode follower. Since both halves run at about the same bias, usually pretty shallow, why doesn't the cathode follower clip like crazy?
Thanks in advance!
Thanks in advance!
The cathode follower can withstand a mucho bigger input signal before clipping, due to the feedback..
AS much as this is going to sound wacky....I can show it mathically if need be... The max input signal before clipping is approximately the difference between the DC Plate voltage and the DC Grid voltage.... Since the Plate is at AC ground potential and it's voltage is essentially locked down...Then you can view the input upside down.. SO think of the input signal being applied between Grid and Plate not Grid and Cathode....Since Grid and Cathode "follow" or track each other, therefore there is virtually no movement between Grid and Cathode due to the almost 100% feedback...
To understand followers a bit better...take the plate curves graph and then derive the cathode curves from them by hand on the graph....
A resistor loaded follower durring large signal analysis is not very linear.... This is primarily due to the wide variation in gm when signal is applied.... A CCS in the tail tends to keep the gm fairly constant, thus a more linear follower...
Chris
AS much as this is going to sound wacky....I can show it mathically if need be... The max input signal before clipping is approximately the difference between the DC Plate voltage and the DC Grid voltage.... Since the Plate is at AC ground potential and it's voltage is essentially locked down...Then you can view the input upside down.. SO think of the input signal being applied between Grid and Plate not Grid and Cathode....Since Grid and Cathode "follow" or track each other, therefore there is virtually no movement between Grid and Cathode due to the almost 100% feedback...
To understand followers a bit better...take the plate curves graph and then derive the cathode curves from them by hand on the graph....
A resistor loaded follower durring large signal analysis is not very linear.... This is primarily due to the wide variation in gm when signal is applied.... A CCS in the tail tends to keep the gm fairly constant, thus a more linear follower...
Chris
Thank you many times! I see now why it's called a cathode follower--the cathode follows the grid. For a resistive loaded follower (where the load resistor is beneath the bias resistor), how does the distortion compare to the distortion of the preceding resistor loaded common cathode stage? That is, if both loads are the same, does the distortion cancel?
No, they're very different and won't cancel.
The distortion of the cathode follower will be lower by a factor of the amplification in common-cathode. For example, let's say you have a tube in common-cathode that gives a gain of 50 with 2% THD with a certain load resistance. As a cathode follower with the same load, the distortion will be 2/50 = 1/25 = 0.04%. This can be understood by looking at the cathode degeneration as a form of feedback.
The distortion of the cathode follower will be lower by a factor of the amplification in common-cathode. For example, let's say you have a tube in common-cathode that gives a gain of 50 with 2% THD with a certain load resistance. As a cathode follower with the same load, the distortion will be 2/50 = 1/25 = 0.04%. This can be understood by looking at the cathode degeneration as a form of feedback.
This is great stuff. While I am on a question and answer roll, another question. I just added a cf (rc coupled) between a 12ax7/12at7 and a trioded el34. I did this in order to be able to drop the grid to ground resistor of the second tube to be able to try a trioded 6l6, which was unstable with the original value (470k). Anyway, the sound improved considerably--even with the el34 still in place. The grid to ground of the 6bl7 (the cf) is 1m, while that of the second stage (el34, 6l6, 6v6, etc.) is now 10K. The amp sounds clearer, more natural and more dynamic. I think the 6bl7 is loaded with 15k. The question is why this sounds better. Is it 1) having a better impedance ratio between the two stages, 2) having more current available for the grid of the output, 3) stabilizing the grid of the output, or 4) adding another degree of freedom for tube rolling?
The trioded EL34 has a whoppingly high Miller capacitance. The 12AX7 has a disastrously high output impedance. Put them together and you have rolled off highs and treble distortion. I've been playing with 12AX7s recently after abandoning them years ago as an inferior tube. They're not, but they're VERY finicky and almost demand a cathode follower as a load.
I opened the amp back up to see what I did. Beneath the cathode is a 670 ohm bias resistor in series with a 5.5K load resistor to the ground. The grid of the 6bl7 is referenced to the junction of the two with a 1M ohm resistor. The 5.5K drops 45 volts for 8 mA current. I have about 250 volts to play with. Should the load resistor be dropping closer to half the available B+?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.