>Even though most of you seem to recommend batteries, i will probably build four separate shunt-regs and use each of one winding each.
I have tried separate secondaries (toroids) and I had serious noise issues. I tried a couple of different regulators, but no go. The noise does not seem to have anything to do with the quality of the regulator. Be warned.
I might try with R-cores, or separate toroids - one day.
For now I have absolutely no reason to downgrade from batteries. Especially after listening to the SEN directly into the F5X
Have you tried one tx with two secondaries per channel, i. e. two toroids with four secs for Sen I/V stereo in balanced version?
What I tried was a custom build (for another project) audio grade toroid with multiple secondaries. This is clearly not optimal, but not that different from one toroid per channel with dual secondaries.
I did buy 4 tiny R-cores to try one last time with regulated supplies for the SEN when I find the time. They did not cost much so I have no intention of experimenting anymore with multiple secondaries.
I never had problems (that I could hear at least) with multiple secondaries in other applications.
Higher priority is trying CEN V18 as my battery packs are still going strong after 1 year of heavy use
I've just got my SEN up and running with Buffalo II and have a virtually hum free background without recourse to batteries/chargers. I didn't have to resort to common mode chokes either. I used 4 separate R cores for each 18 volt rail. These centre tapped transformers were full wave rectified using just the two diodes and choke regulated using 1.5H 200mA Hammond chokes. This was followed by two 100ohm/6800uF R-C filters and then a Salas regulator for each rail. I had used this same supply setup very successfuly with the I/V stage of my Legato so it was just a matter of disconnecting the earths from the supplies and upping the voltage to 18 volts. Here are some picies:
The Salas regs for the SEN are the four to the bottom right of the picture. The transformers, chokes, rectifiers and filtering are off board at the moment but will be housed in a separate case:
I see no good reason to regress to battery supplies which will be noisier, off higher impedance and of varying voltage and therefore create varyiations in output voltage offset and ultimately more expensive in the long term. The little very low level hum I am getting is because of the choke regulated supply for the digital stages of th Buffalo which is just over an inch below the SEN boards. Increasing this distance removes this minor hum altogether.
John
The Salas regs for the SEN are the four to the bottom right of the picture. The transformers, chokes, rectifiers and filtering are off board at the moment but will be housed in a separate case:
I see no good reason to regress to battery supplies which will be noisier, off higher impedance and of varying voltage and therefore create varyiations in output voltage offset and ultimately more expensive in the long term. The little very low level hum I am getting is because of the choke regulated supply for the digital stages of th Buffalo which is just over an inch below the SEN boards. Increasing this distance removes this minor hum altogether.
John
> I see no good reason to regress to battery supplies which will be noisier, off higher impedance
> and of varying voltage and therefore create varyiations in output voltage offset and ultimately more expensive in the long term.
That I do not agree, especially you are using regulators with your mains supply.
To have a fair comparison, you could also use regulator after the battery.
It is NOT my experience that batteries are noisier than mains (before the regulator).
The output impedance, if any, has no consequence in a circuit which only draws constant current.
I personally will continue to use batteries with this circuit, and others.
But everyone to his own taste.
Patrick
> and of varying voltage and therefore create varyiations in output voltage offset and ultimately more expensive in the long term.
That I do not agree, especially you are using regulators with your mains supply.
To have a fair comparison, you could also use regulator after the battery.
It is NOT my experience that batteries are noisier than mains (before the regulator).
The output impedance, if any, has no consequence in a circuit which only draws constant current.
I personally will continue to use batteries with this circuit, and others.
But everyone to his own taste.
Patrick
Agree. Batteries, based on recent measurements, are usually much quieter than I had concluded from seeing some discredited data online, that seem to have a life of its own. And Patrick and others avoid contact problems by using ones with welded leads, rather than battery holders with those particularly troublesome little clips.
I think the main problem with batteries is a psychological one: we are aware that they are running down. It reminds me of the time I in essence hypnotized a rather susceptible person, who was sitting in an old barber's chair, to feel that he was slowly sinking toward the center of the earth
Thanks NicMac, maybe the toroids are not the optimal choice cause the hi capacitance between pri and sec.
Great John! I thinked the same thing. Four R core txs with Salas shunt ( I prefer the 1.2r version to BiB one). I don't like the chokes, cause dinamic suffers. Have you tried the ps without the chokes? I think that the very high PSRR and the low output impedance are sufficient to keep out noise...I hope.
Finally
Great John! I thinked the same thing. Four R core txs with Salas shunt ( I prefer the 1.2r version to BiB one). I don't like the chokes, cause dinamic suffers. Have you tried the ps without the chokes? I think that the very high PSRR and the low output impedance are sufficient to keep out noise...I hope.
Finally
Well it is a significant step up from my Legato I/V stage which I had been very happy with. I can't identify any downsides. Bass lines are both more even and easy to follow, vocals are a great deal clearer and complex passages remain open and clear. More real instruments and a less synthetic delivery all round. After a day with it I can't identify any shortcomings compared to the Legato.
John
It may be taste but in retrospect it may also be unfair of me to glorify mains supply to the SEN to the detrement of battery supplies without comparable experience of the battery supply with equivalent regulation. I've tried regulated battery supplies in the past but not in constant current supplies which lend themselves to shunt regulation as this one does. I tried them in headphone amps and phono stages with little success compared with well designed mains supplies. The noise batteries add to proceedings varies with the current sourced from them and that might be the root cause of my lack of success.
I am actually an advocate of as complete a separation of DC supplies as possible which is why I use choke regulation for every stage of my system apart from the power amp output stage, where I have to make do with choke filters. I even use choke and shunt regulation on my Squeezebox Touch and Router as they are on the same spur as my music system. Obviously the best way to make sure each power rail doesn't interact through the mains is to not connect to the mains at all and use batteries, so I will try a comparison in this application and report back my findings. Has anyone else compared the two options of powering the SEN? I'm presuming you have Patrick? How did they compare for you?
John
I have always found choke regulation i.e. near constant and non zero current through the choke sited before the first PSU caps to be beneficial. It is essential (but relatively simple)to properly snub the choke to stop ringing at diode switch off and to have enough constant current draw to keep the choke in conduction. I haven't ever perceived any problems with dynamics with this implementation. When used in this way chokes are an additional energy storage device that augment the PSU caps and restrict the passage of interference and noise both in and out of the PSU, rather like a linear regulator tries to.
Perhaps you should try full choke regulation. If you have a Buffalo it is most effective and noticable on the Avcc feed PSU.
John
I think this is as close as it can be to what I found. I don't know why it is better, but my assumption is simply simplicity.
With the SEN connected to the F5X I only have three FETs and no caps between DAC and speakers. Can't be (much) more simple than this and the difference is audible - again
I'm using Qspeed LQA03TC600 silicon diodes (which are very low storage and quiet switchers) and two 220nf caps across the chokes with the centre connection of these going straight to earth. This dissipates the switching spikes from the diodes across the internal impedance of the choke and controls their frequency as well (keeps it low).
I'm wondering, are you suggesting that the choke increases the noise or do you want to know if the result is noise free without it? I suspect the latter.I could remove the four of them and see although it is a lot of trouble as the rectified voltage will exceed the voltage rating of the first set of caps and changing the caps will damage the boards. BTW they are 72-0-72 volt transformers! They were what I had handy.
John
I'll probably try Schottkys, since they should give minimum noise.
As you guess, it was because i wanted to see if it was noise-free without the chokes. Still, i was thinking about either use one 2x100mH CM-choke before each regulator, or maybe even two. With a small cap in between them.
Anyway, now i at least have something to go after. So thanks alot!
Nic has already reported in detail his experiences with mains supply and batteries.
I don't have much to add.
If you use a tranformer with regulator, or simply an L-C, you should use the same on the battery.
Then the difference is significant, especially in such floating supply circuits.
4x R-Core is probably as close as you can get to the complete isolation of the 4 IV circuits as you can to batteries.
The difference is in imaging.
Patrick
I don't have much to add.
If you use a tranformer with regulator, or simply an L-C, you should use the same on the battery.
Then the difference is significant, especially in such floating supply circuits.
4x R-Core is probably as close as you can get to the complete isolation of the 4 IV circuits as you can to batteries.
The difference is in imaging.
Patrick
I've just got my SEN up and running with Buffalo II and have a virtually hum free background without recourse to batteries/chargers. I didn't have to resort to common mode chokes either. I used 4 separate R cores for each 18 volt rail. These centre tapped transformers were full wave rectified using just the two diodes and choke regulated using 1.5H 200mA Hammond chokes. This was followed by two 100ohm/6800uF R-C filters and then a Salas regulator for each rail. I had used this same supply setup very successfuly with the I/V stage of my Legato so it was just a matter of disconnecting the earths from the supplies and upping the voltage to 18 volts. Here are some picies:
The Salas regs for the SEN are the four to the bottom right of the picture. The transformers, chokes, rectifiers and filtering are off board at the moment but will be housed in a separate case:
I see no good reason to regress to battery supplies which will be noisier, off higher impedance and of varying voltage and therefore create varyiations in output voltage offset and ultimately more expensive in the long term. The little very low level hum I am getting is because of the choke regulated supply for the digital stages of th Buffalo which is just over an inch below the SEN boards. Increasing this distance removes this minor hum altogether.
John
would you please tell us what are red caps you are using with Salas BIB
why you are using only two from four wires from BIB
Hi,
The Red caps are Shizuki 4.7uf 100v polpropelene film. They were NOS from Cricklewood electronics in the UK but I'm afraid I got their last remaining stock. The closest equivalent to these that can be made to fit and is available are the Wima MKP range, but the Shizukis sound a little better when driving analogue circuits.
I plan to use the four wires eventually but bear in mind that the SEN is a constant current draw (set by the Idss conbination) and so low impedance is of no great advantage. There is less to go wrong with a two wire circuit. It is also worth bearing in mind that the SEN presents a load of two series 220uF Electrolytic capacitors. Some series resistance needs to be introduced between the Salas reg and this effective 110uF approx 0.3 ohm ESR load in order to at least critically damp the resulting resonant circuit. I've since added a series 1 ohm resistor and could hear the improvement it made. It took a slight glassy sheen off the sound and improved the bass definition.
john