I am pretty sure I do not agree with the R4 in your sim, which is the main contributor of the distortion.
In the real PCM1704, the output current does not go through the 1k internal impedance, and the internal impedance only have a secondary effect on the accuracy of the current source. If you want to sim this properly, you need to put in the entire equivalent circuit of the R2R DAC.
In real life, the distortion of the cascode SEN is unmeasurable on my AP system 1. We have access to a Rhodes & Schwarz UPV, but that means shipping the entire circuitry across continents. Postage and insurance are too much to justify generating numbers for the sake of it. We like the sound (for ourselves), which is all I care.
Rechargeable drill batteries are just NiCd, unless you use the newer Lithium ones. I have only changed my NiMH for my preamp once in 6 years. So for me batteries are no issue. And the beauty of the SEN, as described in the article, is that it draws constant current in a separate loop as the signal. So impedance of the power supply has next to no effect at all.
Just my biased opinion, without saying,
Patrick
In the real PCM1704, the output current does not go through the 1k internal impedance, and the internal impedance only have a secondary effect on the accuracy of the current source. If you want to sim this properly, you need to put in the entire equivalent circuit of the R2R DAC.
In real life, the distortion of the cascode SEN is unmeasurable on my AP system 1. We have access to a Rhodes & Schwarz UPV, but that means shipping the entire circuitry across continents. Postage and insurance are too much to justify generating numbers for the sake of it. We like the sound (for ourselves), which is all I care.
Rechargeable drill batteries are just NiCd, unless you use the newer Lithium ones. I have only changed my NiMH for my preamp once in 6 years. So for me batteries are no issue. And the beauty of the SEN, as described in the article, is that it draws constant current in a separate loop as the signal. So impedance of the power supply has next to no effect at all.
Just my biased opinion, without saying,
Patrick
Last edited:
Regal, Patrick, heb1001, thanks for your replies.
Here are my quick measurements, sorry for the noise and hum. From left to right: Cen (2SK170 and 2SJ74), Sen (paralleled 2SK170's), Sen (paralleled BF862's).
Cen has vanishingly low distortion, it clearly beats the Zen I had before.
Somehow I didn't have any luck with the Sen. With 2SK170's its not that bad, but with BF862's the harmonics above 2nd rise remarkably. Maybe the parts aren't matched tight enough, or I've done something wrong. Could somebody help me with this?
I would happily buy a kit with matched fets, if there is any available. Maybe it would work better.
Peter
Here are my quick measurements, sorry for the noise and hum. From left to right: Cen (2SK170 and 2SJ74), Sen (paralleled 2SK170's), Sen (paralleled BF862's).
Cen has vanishingly low distortion, it clearly beats the Zen I had before.
Somehow I didn't have any luck with the Sen. With 2SK170's its not that bad, but with BF862's the harmonics above 2nd rise remarkably. Maybe the parts aren't matched tight enough, or I've done something wrong. Could somebody help me with this?
I would happily buy a kit with matched fets, if there is any available. Maybe it would work better.
Peter
Attachments
Last edited:
Your measurements do not totally disagree with ours.
CEN is push-pull, and have some even harmonic cancellation.
But SEN has 2nd harmonics below -90dB, and what I call a health distribution down to 5th to give good sound.
I personally would be very happy with it as is.
If you want to chase numbers, then try cascoded SEN.
We can only support SEN V18 right now with matched 2SK369Vs.
I am not sure whether it will give you even better numbers than paralleled 2SK170s for you application.
Suggest you try to get 2SJ74 / 2SK170 matched sets from Spencer.
Do some listening first ?
Patrick
CEN is push-pull, and have some even harmonic cancellation.
But SEN has 2nd harmonics below -90dB, and what I call a health distribution down to 5th to give good sound.
I personally would be very happy with it as is.
If you want to chase numbers, then try cascoded SEN.
We can only support SEN V18 right now with matched 2SK369Vs.
I am not sure whether it will give you even better numbers than paralleled 2SK170s for you application.
Suggest you try to get 2SJ74 / 2SK170 matched sets from Spencer.
Do some listening first ?
Patrick
CeeVee in SEN heaven
First and foremost...a huge thank you to Patrick.
This is a gorgeous IV ...never had such crystal clear, dynamic sound..across whole spectrum...great defined bass, voices...highlevel details...the whole package.
I can...almost have to enjoy the music and forget looking for what else to work on....
SEN is running on 16 AAA enoloops per QUAD...sucking about 37mA bias and running at a shade under 50ºC each heatsink after 2 hours playing.
DC offset was trimmed to around 1 mA stable temp.
Running with SDTRANS - UGS - F5.
Testing tomorrow with my BJ2..
Gorgeous....build one !
First and foremost...a huge thank you to Patrick.
This is a gorgeous IV ...never had such crystal clear, dynamic sound..across whole spectrum...great defined bass, voices...highlevel details...the whole package.
I can...almost have to enjoy the music and forget looking for what else to work on....
SEN is running on 16 AAA enoloops per QUAD...sucking about 37mA bias and running at a shade under 50ºC each heatsink after 2 hours playing.
DC offset was trimmed to around 1 mA stable temp.
Running with SDTRANS - UGS - F5.
Testing tomorrow with my BJ2.
.Gorgeous....build one !
Attachments
Last edited:
CeeVee,
Many congratulations with a truely DIY build (including own PCB).
And glad you like the results.
If you would ever consider another build, I still advice you to move the IV away from the digital circuit.
Physical separations of 50mm or more is no issue as the DAC is current out.
Try it once to convince yourself.
50°C is OK for the sinks, as their main purpose is thermal tracking.
If you would want lower temperature, just increase sink thickness.
But I personally am quick happy with 50°C.
As to the Cascode SEN, it will be realised once we finish the ES9018 DAC.
So please be patient. It is also only a hobby for us, as it is for you.
Patrick
Many congratulations with a truely DIY build (including own PCB).
And glad you like the results.
If you would ever consider another build, I still advice you to move the IV away from the digital circuit.
Physical separations of 50mm or more is no issue as the DAC is current out.
Try it once to convince yourself.
50°C is OK for the sinks, as their main purpose is thermal tracking.
If you would want lower temperature, just increase sink thickness.
But I personally am quick happy with 50°C.
As to the Cascode SEN, it will be realised once we finish the ES9018 DAC.
So please be patient. It is also only a hobby for us, as it is for you.
Patrick
Thanks Patrick,
This is quick and dirty first try setup.
Even so it is highly rewarding.
This is proto...now that it works and certainly deserves my utmost effort for final version I'll design another PCB taking separation of digital and analog into account , with physical separate compartments in case.
...but now i can do this while listening to gorgeous music ....
...and keep working on my other main project
Do have one problem though...since i don't have any more 2SK369Vs ( this matched set was all i had ) I'll use 2SK372Vs in next build, heatsink must be a bit bigger though since power handling is smaller with these.
Last edited:
Which one for PCM1704, load?
Hi Patrick:
Between SEN and CEN, which one is better suited for PCM1704? What load (impedance) they present to PCM1704 respectively?
Regards
Hi Patrick:
Between SEN and CEN, which one is better suited for PCM1704? What load (impedance) they present to PCM1704 respectively?
Regards
Hi Patrick and all, just finished my Sen IV 18, and I am very satisfied by the sound . I can
confirm all the amazing rewiews made by the other members that made the IV conv.
It sounds VERY fantastic in all the areas ; dynamic, soundstage, bass , pace and rythm...
and very high spl levels without any form of distortion or compression. Very outstanting!
I supplied the SEN with four separated Salas shunt regs, 1.2r version. I used four r core
txs from Selectronics rated at 50VA, Hexfred diode bridges, Mundorf Lytics and none choke. I've not noise issue at all, the system is dark silent.
My preview IV conv was a Borbely EB 254 that sounded well , but the SEN is another thing. It wins low hands every comparison.
Thanks Patrick, NicMac and all the other members for their help.
confirm all the amazing rewiews made by the other members that made the IV conv.
It sounds VERY fantastic in all the areas ; dynamic, soundstage, bass , pace and rythm...
and very high spl levels without any form of distortion or compression. Very outstanting!
I supplied the SEN with four separated Salas shunt regs, 1.2r version. I used four r core
txs from Selectronics rated at 50VA, Hexfred diode bridges, Mundorf Lytics and none choke. I've not noise issue at all, the system is dark silent.
My preview IV conv was a Borbely EB 254 that sounded well , but the SEN is another thing. It wins low hands every comparison.
Thanks Patrick, NicMac and all the other members for their help.