MJ Statistical Regulator PCBs

What people are working on at the moment
User avatar
Mike H
Amstrad Tower of Power
Posts: 20157
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

#196 Re: MJ Statistical Regulator PCBs

Post by Mike H »

Been watching this sort of, on and off - embarrassing silly question time - what's it for? :oops:

Not seen the book
 
"No matter how fast light travels it finds that the darkness has always got there first, and is waiting for it."
RhythMick
Old Hand
Posts: 747
Joined: Fri Jun 23, 2017 12:24 pm

#197 Re: MJ Statistical Regulator PCBs

Post by RhythMick »

At the risk of jumping in and getting it wrong, MJ describes using a string of 5.6V zeners to achieve a very low noise high voltage low current regulator. I just read the thread from the start and most seem to be using it for screen grids, though not exclusively.

Thinking back to the stats part of my degree, if the noise generated by a 5.6V zener is normally distributed then adding the noise should result in a normal distribution with a very low standard deviation, is the noise is reduced. I can't remember the formula and I haven't read the MJ book for a bit but I remember looking it up from my notes when I last read it. I had to blow the dust off first mind...
User avatar
Nick
Site Admin
Posts: 15708
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#198 Re: MJ Statistical Regulator PCBs

Post by Nick »

Yep, the simple way to look at it is adding together two equal noise sources increase the noise by 3dB, but the voltage increase by 6dB, so the signal to noise ratio is increased by 3dB each time the number of diodes is doubled. I think he also determined that the 5.6v zeners were the quietest ones as well.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
RhythMick
Old Hand
Posts: 747
Joined: Fri Jun 23, 2017 12:24 pm

#199 Re: MJ Statistical Regulator PCBs

Post by RhythMick »

Nick wrote: Sun Dec 27, 2020 1:29 pm Yep, the simple way to look at it is adding together two equal noise sources increase the noise by 3dB, but the voltage increase by 6dB, so the signal to noise ratio is increased by 3dB each time the number of diodes is doubled. I think he also determined that the 5.6v zeners were the quietest ones as well.
Nice way to view it, thanks. Saves all those nasty stats formulae...
RhythMick
Old Hand
Posts: 747
Joined: Fri Jun 23, 2017 12:24 pm

#200 Re: MJ Statistical Regulator PCBs

Post by RhythMick »

I realise I'm late to the party here, but if there are a couple of kits available I'd like them please.
mach1
User
Posts: 5
Joined: Tue Apr 12, 2016 1:21 am

#201 Re: MJ Statistical Regulator PCBs

Post by mach1 »

This is a very interesting article which looks ar getting the best out of conventional zener strings

https://publicwork.wordpress.com/2017/1 ... -ccs-80db/

If a similar ccs was applied to the zener string on the statistical regulator it is possible that further gains in noise reduction may be forthcoming.
User avatar
jack
Thermionic Monk Status
Posts: 5493
Joined: Wed Dec 29, 2010 8:58 pm
Location: ɐılɐɹʇsnɐ oʇ ƃuıʌoɯ ƃuıɹǝpısuoɔ
Contact:

#202 Re: MJ Statistical Regulator PCBs

Post by jack »

mach1 wrote: Tue Jan 05, 2021 2:12 am This is a very interesting article which looks ar getting the best out of conventional zener strings

https://publicwork.wordpress.com/2017/1 ... -ccs-80db/

If a similar ccs was applied to the zener string on the statistical regulator it is possible that further gains in noise reduction may be forthcoming.
This is, IMHO, a poor article. The MJSR is a statistical regulator, i.e. it uses large numbers of low noise true Zeners to achieve noise cancellation - it already has an improved cascoded CCS driving them.

The referenced article uses a single Zener with a very simple CCS so no statistical improvement at all. Further, there are no real-world measurements - the whole article relies on simulations which in turn rely on the noise model of the Zener diode in use. The noise models used are generally generic and voltage independent, so I would have little faith in simulations giving accurate noise figures for a physical implementation of the circuit. MJ measured loads of Zeners and ended up making very specific device selections.

The second FET is simply a standard bypass device to increase the current available - this could be added to the MJSR if required. However, the MJSR is specifically designed for optimal performance with exceptionally low noise at low currents, typically for grid/screen drive.
Last edited by jack on Tue Jan 05, 2021 10:30 am, edited 1 time in total.
Vivitur ingenio, caetera mortis erunt
User avatar
Nick
Site Admin
Posts: 15708
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#203 Re: MJ Statistical Regulator PCBs

Post by Nick »

The second FET is simply a standard bypass device to increase the current available - this could be added to the MJSR if required. However, the MJSR is specifically designed for optimal performance with exceptionally low noise at low currents, typically for grid/screen drive.
Yep, I agree with Jack, I would worry about any circuit that didn’t automatically have the gate protection diode in place. Also, the mosfet is not "perfect" by any means, it will only provide regulation based on its gm. That's why the next step is to add an error corrector and maybe some low pass filtering on the reference voltage. But I do use a similar circuit in several places.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
mach1
User
Posts: 5
Joined: Tue Apr 12, 2016 1:21 am

#204 Re: MJ Statistical Regulator PCBs

Post by mach1 »

Points taken.I only skimmed the article and didn‘t realise it was only a sim.
Post Reply