Zen -> Cen -> Sen, evolution of a minimalistic IV Converter

> This particular circuit (Zen, etc.) is sensitive to the output impedance of the floating supply.

I am sorry, but the SEN / CEN IV are completely different circuit principle than the ZEN IV, as already explained in the article.

> Unless the batteries are heavily bypassed, the source impedance will have a negative effect on the performance of the circuit.
> The circuit requires that the floating supply cancel out all differential mode signals and the only way to do that is to make sure it is a dead short for AC signals.

Really ?
BTW, have you actually built and tested one ?


Cheers,
Patrick

I'm sorry, I was in fact talking about your circuit, which is inspired by the one by Dr. Leach. I'm sorry you apparently do not understand what I'm talking about.

Your article does not discuss batteries vs. regulators for the floating supply, or at least I can't find it.

Did you submit your article to Dr. Leach for his feedback before publishing it? I would consider it rude not to do so myself. He complains about people ripping off his circuit in the article, so it's obviously a sore point for him.
 
Last edited:
I posted the graph to simply show the discharge curve of NIMH and how radically it changes with lower currents and I fail to see how a higher capacity battery will behave any differently at such low currents. myself I use much higher current single cell lifepo4 cells that put the above graph to shame as far as current, but NIMH are no slouches either. basing your argument on zero experience and decades old research is pretty meaningless i'm afraid.

I wonder when people who have no experience of using quality batteries in their gear will stop quoting that article to back up their present day arguments? its getting pretty tiresome, the number of times its been covered is right up there. I dont have the gear to test or I would be answering the call to update it.

also notice the voltage scale on the side, this is not a big battery

AC and batteries and switching supplies all have pros and cons, areas they do well and others where each is a pita, but low current low noise line level gear is most definitely an area where batteries excel

"decades old research"? are you serious? The copyright on that web page is 2004.

If you don't have the gear to test it yourself, then you have no basis to say your way is better or worse than anyone else's.

I'm sorry, but the facts about batteries do not support their use in circuits that are sensitive to power supply impedance and absolute voltage.
 
Inspired yes, as I promptly acknowledged in the first section of the article. Copied certainly not. In fact, the use of JFET cures one of the deficits of the Leach circuit, namely the base current.

And I am certainly not aware that, even in academic publications of which I have had quite a few a decade or so back, one has to submit one's article to the author of a previous publication for approval, when the later publication is a further development of the former, and that the former article has been properly referenced to and acknowledged.

If you have any solid experimental evidence of the detrimental effect of the internal resistance of the power supply in this particular circuit, I am sure all of us would be glad to discover and learn from you.


Best regards,
Patrick
 
Last edited:
Inspired yes, as I promptly acknowledged in the first section of the article. Copied certainly not. In fact, the use of JFET cures one of the deficits of the Leach circuit, namely the base current.

And I am certainly not aware that, even in academic publications of which I have quite a few, one has to submit one's article to the author of a previous publication for approval, when the later publication is a further development of the former, and that the former article has been properly referenced to and acknowledged.

If you have any solid experimental evidence of the detrimental effect of the internal resistance of the power supply in this particular circuit, I am sure all of us would be glad to discover and learn from you.


Best regards,
Patrick

You published in a commercial magazine, not an academic one. You're selling evaluation boards from the article also. So, this is different than an academic environment. I never said you had to submit anything to anyone. I merely stated that I thought it would be rude not to do so given the tone at the beginning of Dr. Leach's article in particular. That is my opinion obviously.

Even so, a collaboration with Dr. Leach could have possibly achieved an even better design. I don't see a huge difference between your circuit and Dr. Leach's.

If I wanted to pursue the battery issue further, I would contact Dr. Leach about it myself. I'm not that interested. I'm convinced by the article I cited above and that's good enough for me.
 
As I do not have a working circuit with me to do real measurement, the least I can do is a Spice simulation.

The attachments show two standard SEN IV circuit variants with 2SK170s.
The first one has ideal current and voltage sources, so that the distortion is only due to the circuit itself.
The 2nd is the same, except that a 1R series resistor has been added to the voltage source.
This is as bad as a battery power supply without any decoupling capacitor can ever get.
The corresponding distortion spectra are shown and can be easily compared.

I also attached the Spice model I am using, so that anyone can repeat the results for themselves.


Patrick
 

Attachments

  • Standard SEN IV w Ideal PSU.png
    Standard SEN IV w Ideal PSU.png
    27.4 KB · Views: 676
  • THD Standard SEN Ideal PSU.png
    THD Standard SEN Ideal PSU.png
    42.5 KB · Views: 669
  • Standard SEN IV w PSU 1R.png
    Standard SEN IV w PSU 1R.png
    28 KB · Views: 667
  • THD Standard SEN PSU 1R.png
    THD Standard SEN PSU 1R.png
    43.9 KB · Views: 661
  • Spice Model SEN IV 2SK170 PSU 1R.zip
    1.3 KB · Views: 107
Last edited:
diyAudio Member RIP
Joined 2005
Power supply rejection of Sen and cascode Sen

This is helpful to assess the appropriate levels of concern. In each case the low-frequency attentuation is noted, and the point at which things are up by 3dB, after which they rise (that is, attenuation falls) at 6dB/octave out to VHF.

The cornering occurs a little lower in frequency when the PS noise is in the positive leg of the battery. But it's pretty decent, even without the cascode structures, and when fed from a lowish presumed DAC output resistance.
 

Attachments

  • simple sen PS rejection 07-31-12 001.jpg
    simple sen PS rejection 07-31-12 001.jpg
    254.1 KB · Views: 649
Cascoded SEN IV

Thanks Brad for the analyses and for the cascoded circuit.

Actually we are not quite ready to publish what we are working on, this one for some 10 months now.
We know it is super low distortion.
But we want to have a chance to do proper testing and A-B listening first before publishing.

We cannot possibly publish everything we have in the pipe ..... ;)


Patrick
 

Attachments

  • Cascoded SEN IV SMD V1a public.jpg
    Cascoded SEN IV SMD V1a public.jpg
    101.2 KB · Views: 623
Last edited:
diyAudio Member RIP
Joined 2005
Thanks Brad for the analyses and for the cascoded circuit.

Actually we are not quite ready to publish what we are working on, this one for some 10 months now.
We know it is super low distortion.
But we want to have a chance to do proper testing and A-B listening first before publishing.

We cannot possibly publish everything we have in the pipe ..... ;)


Patrick

Understood :) I just wanted to put a bit of perspective on the concerns from some quarters about power supply sensitivity, which can be seen to be fairly small even for the simplest topology.
 
As I do not have a working circuit with me to do real measurement, the least I can do is a Spice simulation.

The attachments show two standard SEN IV circuit variants with 2SK170s.
The first one has ideal current and voltage sources, so that the distortion is only due to the circuit itself.
The 2nd is the same, except that a 1R series resistor has been added to the voltage source.
This is as bad as a battery power supply without any decoupling capacitor can ever get.
The corresponding distortion spectra are shown and can be easily compared.

I also attached the Spice model I am using, so that anyone can repeat the results for themselves.


Patrick

Yes, I simulated your circuit up to 10 ohms for the power supply resistance with no significant ill effect. This particular circuit seems to be fairly resistance to power supply impedance, however the model does not take into account the fact that a chemical battery does not have a linear resistance with frequency. The battery becomes inductive above about 1kHz.
 
I mistakenly posted this in the GB thread -excuse me butting in as it's a bit outside the above current discussion. .....

I haven't seen any mention of this maybe "silly" idea but apart from the variation of output impedance, is there any particular reason why an LDR (Light Dependent Resistor, as per the LightSpeed Passive Vol Control) couldn't be used as the Riv in the Cen/Sen O/P stage?
It is relatively easy to "dial-up" a resistance from about 100R up to say, 1kR, 2kR etc that is typical for a 2mA O/P current dac chip (ie 1541A) - and matching in this range is fairly standard.

and Patrick's comment

..... In principle you can use R_iv for volume control, if this is the only source you have.
 
Inspired yes, as I promptly acknowledged in the first section of the article. Copied certainly not. In fact, the use of JFET cures one of the deficits of the Leach circuit, namely the base current.

And I am certainly not aware that, even in academic publications of which I have had quite a few a decade or so back, one has to submit one's article to the author of a previous publication for approval, when the later publication is a further development of the former, and that the former article has been properly referenced to and acknowledged.

If you have any solid experimental evidence of the detrimental effect of the internal resistance of the power supply in this particular circuit, I am sure all of us would be glad to discover and learn from you.


Best regards,
Patrick

I'm sorry, I didn't realize that Dr. Leach had passed away. His website makes it look like he's still alive.

Also, I confused your Sen and Cen circuits.
 
diyAudio Member RIP
Joined 2005
On the subject of noise in batteries, here's an NIST paper from 1995 measuring Ni-Cads:

http://tf.nist.gov/general/pdf/1133.pdf

The authors claim of order 56pV/square root Hz at 10kHz, and mention excess noise below 20Hz. That's mighty quiet.

I don't know if there has been followup for other more-recent-vintage rechargeables.

I'm going to buy some NiMH AAs and take a look at them with a decent preamp.
 
On the subject of noise in batteries, here's an NIST paper from 1995 measuring Ni-Cads:

http://tf.nist.gov/general/pdf/1133.pdf

The authors claim of order 56pV/square root Hz at 10kHz, and mention excess noise below 20Hz. That's mighty quiet.

I don't know if there has been followup for other more-recent-vintage rechargeables.

I'm going to buy some NiMH AAs and take a look at them with a decent preamp.

Yeah, I saw that. Other folks have said that batteries are some 40dB more quiet than a typical linear regulator.
 
Hi Patrick

I had presumed that the -3 dB low frequency limit of SEN was 1/(2 x pi x Riv x Ccoupling). So for - 3dB at ~2-3 Hz, a 330 R Riv and a 220u coupling cap were appropriate; or 150 R with 470 u. Is this correct? If not, what value for the coupling cap should I use with 150 R (smallest value possible please)?

Thanks

Paul N