The art of Sen
- Thread starter Bemused
- Start date
martin clark
pinko bodger
10...9...8...7...cut the blue wire..no the red.. 3..2..no the blue
You are too late, Mr. Bon<BANG>
(Tony - I am not mocking. That's very tidy compared with the innards of my CD2...)
You are too late, Mr. Bon<BANG>
(Tony - I am not mocking. That's very tidy compared with the innards of my CD2...)
James Evans
Bedroom Bodger
Pardon, my french, but holy f*ck!
"Oh that goes in there.
Then that goes in there.
Then that goes in there.
Then that goes in there. & then it's over. Oh, what a hell of a show
but what I want to know:
what exactly do you do for an encore? 'Cos this is Hardcore"
"Oh that goes in there.
Then that goes in there.
Then that goes in there.
Then that goes in there. & then it's over. Oh, what a hell of a show
but what I want to know:
what exactly do you do for an encore? 'Cos this is Hardcore"
Sound or smoke?Just waiting for battery charge and some courage
To be honest a bit of both.
Sound of course must be improbable due to the low component count of Sen, we have after all learned that improved audio requires more components and larger complexity. not forgetting a vast array of exotic capacitors
First pre tests
As I bug wired the quads I could juice them up before fitting to the board and got about 40mA on each.
I then fitted the quads to the board and powered the Sen.
Initial current was around 38mA each and rose over about 4 minutes to 42.1mA on each at which point the quads were quite warm.
Next I connected the AVCC supplies to the little Vref boards. they produced exactly 1/2 AVCC, so I connected the Vrefs to the Sen.
So just battery supplies and Vrefs connected and checked output across Riv and it was a nice fat zero mV.
As I bug wired the quads I could juice them up before fitting to the board and got about 40mA on each.
I then fitted the quads to the board and powered the Sen.
Initial current was around 38mA each and rose over about 4 minutes to 42.1mA on each at which point the quads were quite warm.
Next I connected the AVCC supplies to the little Vref boards. they produced exactly 1/2 AVCC, so I connected the Vrefs to the Sen.
So just battery supplies and Vrefs connected and checked output across Riv and it was a nice fat zero mV.
Next I split the dual mono Buffalos.
Currently a whole Buffalo feeds each channel of a single Legato.
I split the left hand Buffalo so that the right hand half fed the left of the Legato and the left hand half fed the left stage of the Sen.
I split the right hand Buffalo so that the left hand half fed the right of the Legato and the right hand half fed the right stage of the Sen.
Buffalo is set for all DACs in phase not the anti phase way TPA do dual mono.
The Sen had an extra pair of sockets fitted to its output so I have Legato and Sen running together at the moment, each has its own half of a dual mono Buffalo.
The Buffalo is wired directly to my HackerNAPs so I just swap the power amp interconnects from Legato sockets to Sen sockets.
Note to self. Switch OFF amps before swapping
Currently a whole Buffalo feeds each channel of a single Legato.
I split the left hand Buffalo so that the right hand half fed the left of the Legato and the left hand half fed the left stage of the Sen.
I split the right hand Buffalo so that the left hand half fed the right of the Legato and the right hand half fed the right stage of the Sen.
Buffalo is set for all DACs in phase not the anti phase way TPA do dual mono.
The Sen had an extra pair of sockets fitted to its output so I have Legato and Sen running together at the moment, each has its own half of a dual mono Buffalo.
The Buffalo is wired directly to my HackerNAPs so I just swap the power amp interconnects from Legato sockets to Sen sockets.
Note to self. Switch OFF amps before swapping
Thanks but oops
I am there,
Forgot to say it works.
This is just one Sen board and half a Buffalo, well perhaps quarter of a Buffalo as I am only using DACa output and Analog ground not the Sabres balanced outputs ie DACa & DACb.
There is no BAL/SE converter, I am just using one half of a Sen per channel.
Really this should not be working with so few components
Batteries have no filters, they are wired direct to the Sen.
For now charging is just by a bench power supply, set to 22.4 volts and the current clamp set at about 100 / 150mA, the clamp is not so accurate and varies a little.
I am there,
Forgot to say it works.
This is just one Sen board and half a Buffalo, well perhaps quarter of a Buffalo as I am only using DACa output and Analog ground not the Sabres balanced outputs ie DACa & DACb.
There is no BAL/SE converter, I am just using one half of a Sen per channel.
Really this should not be working with so few components
Batteries have no filters, they are wired direct to the Sen.
For now charging is just by a bench power supply, set to 22.4 volts and the current clamp set at about 100 / 150mA, the clamp is not so accurate and varies a little.
Suffolk Tony
Aim low, achieve your goals, avoid disappointment.
Oh wow! That's truly amazing Tony. I'm lost with what you've done but no doubt it'll all appear clear at some stage. Does it sound good?
Tony, just want to express my admiration for the inspirational work you are doing. You often make remarks as to how modest your understanding is, but you are clearly very capable of managing the complexities of such involved projects.
Dan
PS love the photos of your setup. They remind me of some fantastic lego creation, but from grown ups!
Dan
PS love the photos of your setup. They remind me of some fantastic lego creation, but from grown ups!
This is of no importance for a hard-core diyer
Look at the apparent complexity of the SEN and you can see that it simply can't sound any good.
BTW. The twisting of wires is a piece of art!
Well thanks for the encouragement you make me blush.
Paul and Nic have helped me to speed along a lot, so thanks guys.
First my system, Touch, dual mono Buffalo-III, Single Legato 3.1, HackerNAPs, Thiel CS1.6.
All mostly modified of course.
What it did before Sen, great pace and involvement, quite Naim like. Good image width and stunning depth often causing the listener to look behind them. Non HiFi people comment that the sound is everywhere.
First of the sound was a little coarse but this was mainly down to a gain / level matching problem and perhaps the Sen needing a little burn in.
Volume turned out to be way too high to start off with.
I convinced myself the implementation was too poor, I know its not good. I have great long leads from the Sabre outputs to the Sen and lots of other little things that should be better, but this is a development rig and principally has to allow rapid changes and strip downs. Probably at the forefront of my mind was the fact that the Riv resistor needs to be a lot better than a 1% metal film no name.
After a half dozen tracks it became obvious the volume was to high and reducing volume gave the same involvement but more clarity, I am sure you have all been here before.
Gain was further reduced over a few tracks more. I / we settled on 3dB more than previously used. This is with 150R I/V conversion resistor, also comparing to a Legato with non standard gain. (I halved the Legato)
After a while the excessive gain took toll on our ears so we thought to leave it a while. I have often found system improvements lead to less volume, perhaps gain is still a little high.
At the end of the session we both believed the Sen out did the Legato, not a night and day difference but quite subtle. The dual mono Buffalo / single Legato is after all a good combination.
Most noticeable was vocal clarity.
The Blessings Back from Managua is one of our audition tracks and neither of us can decipher all the words but with Sen we got a few more.
I had serious misgivings about throwing four of the Sabres converters out of each channel, previously we thought (Suffolk Tony also) stereo to dual mono was a step change. Dual mono with single Legato of course, the TPA method is not so good for SE users. Anyway away they went and we are certainly no worse and probably a little better. This makes me think the dual mono Buffalo / single Legato is more about classic dual mono architecture than the extra converters.
And of course tossing out the Sabre balanced output system seems bad as we have completely lost all the even harmonic / common mode noise rejection that we had for free but that's assuming there there was significant noise to be canceled. Maybe it's a trade off against increased common mode from the Sabre and less from the simplicity of the Sen.
I am convinced there is more to come with Sen, and will be moving onward with some low cost changes but need a while to settle down with the new sound first.
But early on the list I will be halving the load on the TPA AVCC regulator, I can now simple replug and half the load, probably need some proper low noise Riv resistors. Further along I need to get shut of the TPS AVCC regulator, setup the other Sen board.
And not forgetting half of each Buffalo is still running a powered Legato, may be some problem here.
Tony
PS You will never guess what happens when Sens battery runs out
No it did not go flat already, I forced it to check the HackerNAPs would be safe.
Paul and Nic have helped me to speed along a lot, so thanks guys.
First my system, Touch, dual mono Buffalo-III, Single Legato 3.1, HackerNAPs, Thiel CS1.6.
All mostly modified of course.
What it did before Sen, great pace and involvement, quite Naim like. Good image width and stunning depth often causing the listener to look behind them. Non HiFi people comment that the sound is everywhere.
First of the sound was a little coarse but this was mainly down to a gain / level matching problem and perhaps the Sen needing a little burn in.
Volume turned out to be way too high to start off with.
I convinced myself the implementation was too poor, I know its not good. I have great long leads from the Sabre outputs to the Sen and lots of other little things that should be better, but this is a development rig and principally has to allow rapid changes and strip downs. Probably at the forefront of my mind was the fact that the Riv resistor needs to be a lot better than a 1% metal film no name.
After a half dozen tracks it became obvious the volume was to high and reducing volume gave the same involvement but more clarity, I am sure you have all been here before.
Gain was further reduced over a few tracks more. I / we settled on 3dB more than previously used. This is with 150R I/V conversion resistor, also comparing to a Legato with non standard gain. (I halved the Legato)
After a while the excessive gain took toll on our ears so we thought to leave it a while. I have often found system improvements lead to less volume, perhaps gain is still a little high.
At the end of the session we both believed the Sen out did the Legato, not a night and day difference but quite subtle. The dual mono Buffalo / single Legato is after all a good combination.
Most noticeable was vocal clarity.
The Blessings Back from Managua is one of our audition tracks and neither of us can decipher all the words but with Sen we got a few more.
I had serious misgivings about throwing four of the Sabres converters out of each channel, previously we thought (Suffolk Tony also) stereo to dual mono was a step change. Dual mono with single Legato of course, the TPA method is not so good for SE users. Anyway away they went and we are certainly no worse and probably a little better. This makes me think the dual mono Buffalo / single Legato is more about classic dual mono architecture than the extra converters.
And of course tossing out the Sabre balanced output system seems bad as we have completely lost all the even harmonic / common mode noise rejection that we had for free but that's assuming there there was significant noise to be canceled. Maybe it's a trade off against increased common mode from the Sabre and less from the simplicity of the Sen.
I am convinced there is more to come with Sen, and will be moving onward with some low cost changes but need a while to settle down with the new sound first.
But early on the list I will be halving the load on the TPA AVCC regulator, I can now simple replug and half the load, probably need some proper low noise Riv resistors. Further along I need to get shut of the TPS AVCC regulator, setup the other Sen board.
And not forgetting half of each Buffalo is still running a powered Legato, may be some problem here.
Tony
PS You will never guess what happens when Sens battery runs out
No it did not go flat already, I forced it to check the HackerNAPs would be safe.
This is of no importance for a hard-core diyer
Look at the apparent complexity of the SEN and you can see that it simply can't sound any good.
Quite right and its fun too.
The good ones are done with a vice and a drill, but I ran out and you can see what is hand twisted.
Perhaps if you see the Bom it would look clearer Tony, I have rambled a lot along the way and in parallel we are trying for a neater Buffalo implementation.
Current BOM
From Xen
1 x Sen only evaluation kit.
This is two boards, I have only used one so far.
Four quad T092 heat sinks.
I have only used two so far.
Sixteen Toshiba N Jfet 2SK369V
I have only used eight so far.
Perhaps some guys could share a Sen kit?
From elsewhere.
Four Nichicon Muse ES bipolar 470uF 16v capacitors.
Four Panasonic ECPU SMT ECPU 16Vdc 1uF, SMD 1210. RS727-0376
Two Panasonic ECHU SMD,3.3nF 50V 2% SMD 1206 RS414-7254
Two 680k metal film 1% resistors.
Two 470k metal film 1% resistors.
Two 150R metal film 1% resistors.
Four Vishay thick film 1% SMD 0805 resistors RS679-0807
For the pair of Vref boards.
A little bit of strip board.
Four 1K0 metal film 1% resistors.
Two Nichicon Muse ES bipolar 220uF 16v capacitors.
Some thermally conductive glue, also electrically insulating RS155-8320
Enough glue to to lots and lots and lots.
Some terminal headers / turned pin board connectors to personal choice.
Yes using Legato internal BAL/SE
HackerNAP is a reworked Naim amplifier, its a pinkfish thing.
I halved the gain on the Legato BAL/SE stage for sure, I changed the Legato I/V resistor but sure I changed it back to stock.
So Sen should be 3dB quieter.
Volume does not sound the same level when we are listening to Sen and use around 3 to 5dB more gain. We started of with towards 12 dB extra gain and hence the pain. Not a Sen problem.
As Sen has a different presentation which I cannot describe it has not been easy to match levels.
Sen is settling in quickly now and we are starting to listen at lower levels, the first levels had way to much gain (as in I turned the knob up too much)
Legato is powered and fed from the other half of the Sabre when Sen is running yes.
Probably not good, yes, that is why I mentioned it.
Wont be on for long though Sen is starting to sing, Legato was left on for AB testing or rapid fallback should Sen have smoked out.
Legato is connected to output 1-4 using DACa and DACb
Sen is on output 5-8 using DACa only.
DACb is open circuit.
This seemed the most basic approach and lets me perhaps learn a little as the conversion begins to go more Sen
Yes there is not a lot of difference at the moment, Sen is subtly more detailed.My balanced out SEN sounds very different to Legato. Even if you prefered the latter- which would be strange to me, but we're all different- you shouldn't be struggling to differentiate them IMO
Your comment is good news, but I was sure Sen would get better once the implementation improves.
Just spent some time cutting the grass and thinking about Sen.
I believe I goofed somewhat.
Maybe with the half a Sen board a channel approach or maybe the grounding.
DAC analog ground, Vref ground and Vout ground are all the same point.
I think this results in DACa output passing through the upper fets, though the big capacitor and on through Riv to ground. Regardless of Lower fet CCS, batteries connected or not.
Needs more thought...
I believe I goofed somewhat.
Maybe with the half a Sen board a channel approach or maybe the grounding.
DAC analog ground, Vref ground and Vout ground are all the same point.
I think this results in DACa output passing through the upper fets, though the big capacitor and on through Riv to ground. Regardless of Lower fet CCS, batteries connected or not.
Needs more thought...
What is the input impedance of your pre/power amp?
The Sen has a highish output impedance.
Did you check that you have the phases of the ES9018 right?
Like Paul, I am surprised that you find it similar to the Legato. I found it very very different right from the first power on.
The Sen has a highish output impedance.
Did you check that you have the phases of the ES9018 right?
Like Paul, I am surprised that you find it similar to the Legato. I found it very very different right from the first power on.