How about "Just Dropped In (To See What Condition My Condition Was In)".
Kenny Rogers ca. 1967![]()
Is that the song in The Big Lebowski? Allways wondered where that came from
And was actually first recorded by Jerry Lee Lewis in 1967
So I got some more JFETs did some experiments at Pico's suggestion. Here are the latest results.
The 2SK170 measured at 9.1 ma. I tried various Idss 2SJ74's and settled on 6.0 ma. Going with lower values (I even tried 4.3 ma) seemed to have diminishing returns - I could not improve distortion or get the N-channel pot resistance much above 200 Ohms. I probably have a pretty mismatched pair of laterals.
I dialed the JFET source pot to get 2H THD about 7 db higher than 3H, while keeping overall THD around .05% (at 1 Watt). If I try to get the spread to 10db, THD creeps up to around .07%.
Anyway, the phase of 2H has now flipped, so I believe I am getting negative phase 2H - but need to verify with a filter.
So listen to Pico - he is the wizard
when it comes to this topology (there I said it - I promised I would - hehehe).
The 2SK170 measured at 9.1 ma. I tried various Idss 2SJ74's and settled on 6.0 ma. Going with lower values (I even tried 4.3 ma) seemed to have diminishing returns - I could not improve distortion or get the N-channel pot resistance much above 200 Ohms. I probably have a pretty mismatched pair of laterals.
I dialed the JFET source pot to get 2H THD about 7 db higher than 3H, while keeping overall THD around .05% (at 1 Watt). If I try to get the spread to 10db, THD creeps up to around .07%.
Anyway, the phase of 2H has now flipped, so I believe I am getting negative phase 2H - but need to verify with a filter.
So listen to Pico - he is the wizard
![wizard :wiz: :wiz:](https://files.diyaudio.com/forums/images/smilies/wizard.gif)
Attachments
Last edited:
...Try 5.4 and 8.4 with 5 ohms and 15 ohms source respectively...
Hi pico,
What is your rule of thumb, x mA different or a percentage?
Some more
Thanks Pico for your effort, it's hard to find this kind of data, maybe it's good idea to make thread that will contain this kind of measurements that we can't see in datasheets..
By the way if you have the time and will, I think that lot of members are interested in these kind of curves for smaller mosfet types like IRF610/9610 at low currents, for example 10mA to 100mA.
Anyway thanks a lot for the data, cheers.
Some more
My my - those Exicons are sissy
![yikes :yikes: :yikes:](https://files.diyaudio.com/forums/images/smilies/yikes.gif)
I better get some double-dies. I see that they are now offering a -S version which they claim are banded for better matching (actually in our case, deliberate mis-matching). I think the trick might be to select the proper bands to reduce the Idss delta between the N- and P-channel parts. This might help improve things a bit. See:
https://www.profusionplc.com/pdf/selected-mosfets.pdf
Last edited:
Thanks Pico for your effort, it's hard to find this kind of data, maybe it's good idea to make thread that will contain this kind of measurements that we can't see in datasheets..
By the way if you have the time and will, I think that lot of members are interested in these kind of curves for smaller mosfet types like IRF610/9610 at low currents, for example 10mA to 100mA.
Anyway thanks a lot for the data, cheers.
Ok I will do those too.
Yes exactly.
You want N Channel parts with the Vgs threshold the highest they have available and P channel parts with the lowest Vgs threshold available.
Perfect - that's what I was thinking. I believe their banding effort just happened recently - it would certainly make life easier and cheaper than buying batches of random parts!
I"m not sure Pico but to me it looks like the 20P20 has a lower transconductance than the 20N20, so for lower distortion you need a higher amplitude signal on the gate of the 20P20. However, the amplitude is too high on the 20P20 verses the 20N20 than it should be due to having too lower the 2SJ74"s drain resistor to get zero dc offset. Somewhere near an 18 percent higher amplitude signal on the 20P20 seems to match the transfer curves with the models I am using.
I think that is typical. Here are the Exicon curves from the datasheets. I think they mislabeled the Y-axis for the P-channel. Assuming it should have said transconductance and not drain current, the P-channel is lower for sure. The nature of the beasts, along with significantly different Idss. That is what makes this circuit so intriguing (and frustrating). ![Big grin :D :D](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
Attachments
Last edited:
So I got some more JFETs did some experiments at Pico's suggestion. Here are the latest results.
The 2SK170 measured at 9.1 ma. I tried various Idss 2SJ74's and settled on 6.0 ma. Going with lower values (I even tried 4.3 ma) seemed to have diminishing returns - I could not improve distortion or get the N-channel pot resistance much above 200 Ohms. I probably have a pretty mismatched pair of laterals.
I dialed the JFET source pot to get 2H THD about 7 db higher than 3H, while keeping overall THD around .05% (at 1 Watt). If I try to get the spread to 10db, THD creeps up to around .07%.
Anyway, the phase of 2H has now flipped, so I believe I am getting negative phase 2H - but need to verify with a filter.
So listen to Pico - he is the wizardwhen it comes to this topology (there I said it - I promised I would - hehehe).
I modified the other channel. It also had a 9.1 ma 2SK170. I replaced the matching 2SJ74 with one having an Idss of 5.6 (closest I had to a 3 ma difference). After adjustments, I got identical distortion results to the first channel. I am very close to Papa specs at this point. I think the only way to get any closer is to look at selecting laterals. For now, I am going to finish buttoning up the amp and let it play for a while.
![cheers :cheers: :cheers:](https://files.diyaudio.com/forums/images/smilies/cheers.gif)
I"m not sure Pico but to me it looks like the 20P20 has a lower transconductance than the 20N20, so for lower distortion you need a higher amplitude signal on the gate of the 20P20. However, the amplitude is too high on the 20P20 verses the 20N20 than it should be due to having too lower the 2SJ74"s drain resistor to get zero dc offset. Somewhere near an 18 percent higher amplitude signal on the 20P20 seems to match the transfer curves with the models I am using.
I am sure.
Think about it more. You're completely missing something.
- Home
- Amplifiers
- Pass Labs
- On A Hippie Trail, Head Full Of Zombie