I have a few unfinished projects over the years that I want to tidy up in 2018,
and the XCEN Balanced to Single-Ended Converter is one of them.
https://www.diyaudio.com/forums/analog-line-level/216557-xcen-balanced-single-converter.html
In 2012, we started work on an all-JFET XCEN. AC performance was fine (-93dB THD), and it sounded great.
But I was not totally happy with the 50mV start-up DC offset, even though it went down to < 10mV after a few minutes.
https://www.diyaudio.com/forums/ana...-balanced-single-converter-6.html#post3159210
The revised circuit still uses 2SK246/2SJ103 or their SMD equivalent, but now Y-grade and at much lower bias.
And the current sources are now BJT based and self-tracking, as in our Pioneer Super Linear Headphone Amplifier circuit.
DC offset has been reduced to < 10mV at start up, and stable to +/2mV over 30 minutes.
And a final bonus is that distortion improves to -105dB H2, and -110dB H3 at 1Vrms 1kHz output.
You can read about the topology changes and their reasoning in the article.
Finally, I am remembering bcarso on this occasion with some sadness.
He inspired me on numerous occasions through interesting and open discussion, also at the XCEN 2012 thread,
but is sadly no longer with us.
https://www.diyaudio.com/forums/in-memoriam/306454-rip-brad-wood.html
Patrick
.
and the XCEN Balanced to Single-Ended Converter is one of them.
https://www.diyaudio.com/forums/analog-line-level/216557-xcen-balanced-single-converter.html
In 2012, we started work on an all-JFET XCEN. AC performance was fine (-93dB THD), and it sounded great.
But I was not totally happy with the 50mV start-up DC offset, even though it went down to < 10mV after a few minutes.
https://www.diyaudio.com/forums/ana...-balanced-single-converter-6.html#post3159210
The revised circuit still uses 2SK246/2SJ103 or their SMD equivalent, but now Y-grade and at much lower bias.
And the current sources are now BJT based and self-tracking, as in our Pioneer Super Linear Headphone Amplifier circuit.
DC offset has been reduced to < 10mV at start up, and stable to +/2mV over 30 minutes.
And a final bonus is that distortion improves to -105dB H2, and -110dB H3 at 1Vrms 1kHz output.
You can read about the topology changes and their reasoning in the article.
Finally, I am remembering bcarso on this occasion with some sadness.
He inspired me on numerous occasions through interesting and open discussion, also at the XCEN 2012 thread,
but is sadly no longer with us.
https://www.diyaudio.com/forums/in-memoriam/306454-rip-brad-wood.html
Patrick
.
Attachments
> I was wondering if you were planning on selling the boards or kits?
Making PCBs available is relatively easy.
Important is that you can find 2SJ103Y or 2SJ106Y for group match.
2SK208Y is Mouser or DK active. So no problem with those.
If I have to supply matched 2SJ106s then it is a lot or work, and will take time.
> Also, how about single ended to balanced ones?
NECX and NESX Single-Ended to Balanced Converters
Patrick
Making PCBs available is relatively easy.
Important is that you can find 2SJ103Y or 2SJ106Y for group match.
2SK208Y is Mouser or DK active. So no problem with those.
If I have to supply matched 2SJ106s then it is a lot or work, and will take time.
> Also, how about single ended to balanced ones?
NECX and NESX Single-Ended to Balanced Converters
Patrick
Though very happy with the performance of the prototypes, I only have SMD devices for the front end.
So a pure SMD version will serve me better.
It also allows all the devices requiring thermal tracking to be coupled to a small heatsink.
The only change in the circuitry is replacing the TL431s with 6x NXP BZX84C3v3 in series.
The latter is known to be very low noise, further cleaned off by an additional RC low pass before the cascode gates.
All performances of the previous prototypes can be repeated.
And I declare myself happy with the final result.
A happy and healthy 2019 to all,
Patrick
.
So a pure SMD version will serve me better.
It also allows all the devices requiring thermal tracking to be coupled to a small heatsink.
The only change in the circuitry is replacing the TL431s with 6x NXP BZX84C3v3 in series.
The latter is known to be very low noise, further cleaned off by an additional RC low pass before the cascode gates.
All performances of the previous prototypes can be repeated.
And I declare myself happy with the final result.
A happy and healthy 2019 to all,
Patrick
.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.