I measre at the speaker terminal into 5R load. I think we are comparing apples to oranges at this point regarding THD. In my case it is a PP using the tubelab power drive method (FET drive with a CCS load) similar to what you are doing. I also have 9dB NFB for better dampening and reduced gain. I needed to use the FET since I wanted AB2 capability with a t7 driver. The t7 couldn't drive the low impedance required without losing gain or suffering blocking distortion. You can see the schematic and amp on the OPUS thread.
What caught my eye here was you are using a pretty strong drive tube with a FET follower. Something I considered strongly at one point, and am still interested in. I will be in the exact same place as you are today later this summer.
Here is a FFT plot with my EHkt88's. In this case thd is about 0.04% with the EH instead of the GEC's. The GEC's ran lower until I found I had a short in one of them. It was running 0.07% prior to raising the FET gate impedance. I still have some work to do as I have the other channel at <0.01% THD. I am focused on just the THD for the moment as I am still "voicing" the amp.
P.S. I will try to get some THD at the source for the FET follower(s) in the next few days so we can compare notes.
What caught my eye here was you are using a pretty strong drive tube with a FET follower. Something I considered strongly at one point, and am still interested in. I will be in the exact same place as you are today later this summer.
Here is a FFT plot with my EHkt88's. In this case thd is about 0.04% with the EH instead of the GEC's. The GEC's ran lower until I found I had a short in one of them. It was running 0.07% prior to raising the FET gate impedance. I still have some work to do as I have the other channel at <0.01% THD. I am focused on just the THD for the moment as I am still "voicing" the amp.
P.S. I will try to get some THD at the source for the FET follower(s) in the next few days so we can compare notes.
Attachments
Last edited:
Tom,
I think you may want to look at the bias current of the FET. As I said before I set up with 10mA initally based on a comment from George. I did some rudementary tests and found that at low powers, the bias didn't seem to have much effect, however, at higher powers, in particular close to grid current and above the THD was higher for lower bias currents and reduced at higher bias currents.
Again this is a different amp, tubes etc. at 5mA I measured 0.936% at 40 Watts. for 10mA I measured 0.828%, and finally at 15mA I measured 0.736%. I am going to try 20mA tomorrow.
I think you may want to look at the bias current of the FET. As I said before I set up with 10mA initally based on a comment from George. I did some rudementary tests and found that at low powers, the bias didn't seem to have much effect, however, at higher powers, in particular close to grid current and above the THD was higher for lower bias currents and reduced at higher bias currents.
Again this is a different amp, tubes etc. at 5mA I measured 0.936% at 40 Watts. for 10mA I measured 0.828%, and finally at 15mA I measured 0.736%. I am going to try 20mA tomorrow.
I measured THD+N for the following loads on the source follower: 22.6 kOhm resistor, 2 mA CCS, 5 mA, 12 mA and 22 mA CCS (IXYS 10M45). Results below. I find negligible difference in THD+N for the different loads, but note that the bias current of the 300B shifts slightly from setup to setup due to shift Vgs with shift in Id on the source follower. This change of operating point on the 300B impacts the THD much more than the change in source follower load. For an apples-to-apples comparison I adjusted the idle current in the 300B to 60 mA in all experiments before measuring the THD.
~Tom
~Tom
Attachments
Very interesting, Thanks Tom. I assume that the measurement is taken at 1 Watt output power. If so that would correlate with my measurements that at lower power the CCS bias doesn't have much effect. I also saw the same need to rebias at different currents due to Vgs.
Now if that data was taken at levels near A2 operation then our data differs.
P.S. Here is a plot at 1Watt, 4R load. 15mA bias on CCS. Using powerdrive in PP amp.
Now if that data was taken at levels near A2 operation then our data differs.
P.S. Here is a plot at 1Watt, 4R load. 15mA bias on CCS. Using powerdrive in PP amp.
Attachments
Last edited:
You are correct, Sir. I measured at 1.0 W output power (RMS, 8 ohm). I s'pose, I could remeasure at 6~7 W (close to max output power). I'll give that a whirl over the weekend.
Why the plunge in THD after 10 kHz in your plot? Are you using 44 kHz sampling? I assume the plunge is caused by the harmonics being cut off by the prefilter and sampler on the ADC you're using. Pretty impressive THD, though.
~Tom
Why the plunge in THD after 10 kHz in your plot? Are you using 44 kHz sampling? I assume the plunge is caused by the harmonics being cut off by the prefilter and sampler on the ADC you're using. Pretty impressive THD, though.
~Tom
I don't know why it cuts off. I havn't played enough with audio tester to figure it out completely. I sample at 96kHz, so I am thinking it is bug in the software (or operator, as both are quite buggy)
The Powerdrive has alot of possibilities. You are doing it in a SE and mine in PP. I am glad to see that both our data is matching so far.
The Powerdrive has alot of possibilities. You are doing it in a SE and mine in PP. I am glad to see that both our data is matching so far.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.