I'm playing around with some ideas for a "spud" cathode follower output amp using a 21LR8 with the pentode side triode strapped (for the moment). The smaller triode is used as the driver in an ordinary common cathode setup.
The design isn't done yet (may never be...) but one really weird thing is that the distortion I'm measuring across an 8 ohm resistor on the secondary of the output transformer is consistently about 2X lower than the distortion coming from the driver stage. For example 1.5 % THD in, 0.8 THD out at 1 watt. Is this really possible or is something off in my measurements?
The design isn't done yet (may never be...) but one really weird thing is that the distortion I'm measuring across an 8 ohm resistor on the secondary of the output transformer is consistently about 2X lower than the distortion coming from the driver stage. For example 1.5 % THD in, 0.8 THD out at 1 watt. Is this really possible or is something off in my measurements?
A power cathode follower will not have the low distortion of a well-designed small-signal CF, so some cancellation may occur. However, it is likely that only 2nd will cancel; higher orders may get worse.
Thanks for the info. It turns out that this particular magic only happens over a narrow range of operating conditions, and not the ones I want to use. As soon as I biased the driver to where it could swing 200V+ to drive more than a watt the distortion went to 4 - 5% in and out.
I need to get a lower impedance OP tran to lower the drive requirements. I've ordered a few but it will take a couple of months to get them so this one will need to go back on the shelf for awhile.
All I've learned so far is that Cathode Follower output stages are great for freq. response, even with cheap-ish transformers. But the distortion problems just get kicked down the road to the driver.
I need to get a lower impedance OP tran to lower the drive requirements. I've ordered a few but it will take a couple of months to get them so this one will need to go back on the shelf for awhile.
All I've learned so far is that Cathode Follower output stages are great for freq. response, even with cheap-ish transformers. But the distortion problems just get kicked down the road to the driver.
4...5 % THD is very typical to triode SE stages, and still most DIYers do not want to build or have anything else (=>better).
How are you measuring distortion?
Get a copy of a spectrum analyzer (FFT) package and use it with your PC to perform FFTs of the single tone output. This will show you a plot of the distribution of the distortion components, which is much more important than absolute THD distortion levels.
Many of these packages will also perform multi-tone IM distortion testing as well.
Get a copy of a spectrum analyzer (FFT) package and use it with your PC to perform FFTs of the single tone output. This will show you a plot of the distribution of the distortion components, which is much more important than absolute THD distortion levels.
Many of these packages will also perform multi-tone IM distortion testing as well.
This may be of interest: http://www.diyaudio.com/forums/tubes-valves/288372-idea-driver-cf-output-stage.html
I'm using the frequency (FFT) view in a Picoscope. I don't have the numbers but a pretty good picture of it in my head. For the 4-5% case the second is the highest, but the third comes in not far under it. After that they drop off, the "odds" all seem higher than the "evens" but it's a bit hard to make out. With a 1khz signal and a 10khz window I still see things all the way to the 9th. I didn't hook it to a speaker (so not to disturb my wife and daughter at night) but my eyeball judgement was that it would sound pretty ugly, not "nice triode" type distortion. But I'm driving the snot out of the driver triode to get anything reasonable out of the cathode follower output stage.
- Status
- Not open for further replies.