• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Anyone build a Minimal Reactance PSU?

Hi,

I recently added a Minimal Reactance psu by Dave Davenport has made for the Raleigh Audio Line Stage. I recently added his psu to my RALS and it has made a huge difference.
Here is an early version of http://www.raleighaudio.com/chapter_5.htm the design. the current design is more advanced with hybrid shunt reg. I.e. lm431 and 6n1p...(Something like this has been published by John Broskie as well and is therefore in the public domain...)

The MR psu is state of the art WITHOUT the cost of a choke or big expensive caps like BG of Elna Cerafine. It really is a fresh and clever way do make a psu for a low current valve amp...like a preampflier..and I am suprised I have seen very little on the web on similar concepts...apart from Allen Wright's superreg and Emile Sprenger's version of it...but that is waaay more complex and expensive.

A number of people are going to build an Aikido and have asked me to design or help with a psu for the Aikido.

Obviously I can't steel Dave's design and was hoping someone could "supply" us with his or her version of a MR psu?

Regards,
Bas
 

mach1

Banned
2004-09-13 4:06 am
Brisvegas
Bas,

I have used one on my 1LE3 pre for some time. It uses an aerovox 2uF oil cap - pimm CCS - Broskie 5687 / 431 shunt reg followed by individual Pimm CCS' to each plate. There is one additional sangamo 0.25uF oil cap just before the plate CCS' for stability. Total psu capacitance is 2.25uF.

Very nice indeed - very fast and transparent with a super solid bottom end.

I believe Gary Pimm had been using fairly small value caps / chokes before CCS-shunt reg-CCS psu's for some time before the minimal reactance article was published However, the minimal reactance psu took this concept to the limit and provided a sound theoretical base for its implementation.

A great thing about these psu's is that you don't need a choke at all and can afford to use really good quality caps.

Has anybody used one for a power amp yet ?
 
I don't see the point of using a vacuum tube rectifier, as seen in the schematic http://www.raleighaudio.com/ps_schematic.htm
as the mosfets would "filter" out any "tube magic" from the tube rectifier. Also the power on time delay circuit takes care of avoiding full B+ on preamp tubes not up to heat yet. So I'd use some good silicon rectifier diodes in place of the tube rectifier.

Something else to keep in mind is that the power disippated in the pass mosfets will be less if the ripple is allowed to be high before the mosfets.
 
as the mosfets would "filter" out any "tube magic" from the tube rectifier

That is what I thought as well..but listening to it ...you can even detect changes in the caps that are placed before the CCS.

But I would agree that it is subtle, and while one is doing the cost effective thing...something like BYV95D or something with RRSF and a snubber would be a cheaper solution...particularly in the long term.
 
Of course the source impedance of this type of supply is high with its attendant limitations unless a shunt regulator or capacitor is added at its output to, for instance, limit crosstalk between different stages supplied from it. I'd probably dump the mosfet and use a tube current source, anyway:)

Also, I find many references to a DN2470 & DN2540, but hardly any to a DN2450 on the web.
 

mach1

Banned
2004-09-13 4:06 am
Brisvegas
wa2ise

On the contrary, I believe a minimal reactance psu constructed using Gary Pimm's latest CCS design and a decent shunt reg (preferably with a tube carrying out ac reg duties) lets all the natural magic intrinsic in a tube shine through unhindered.

If you want to give it a go try without going all the way, try using IXCP10M45S current regulators in cascode mode - in my opinion they are every bit as good (if not better than) the DN2540 in this application and are also better able to dissipate heat. Plus you can actually buy the things (the DN2540 has now been discontinued and was always extremely difficult to source).

I totally agree with Bas that the character of caps in the psu are quite audible in this topology. I have found that even the type of cathode bypass cap used in the shunt reg is critical to the overall sound.

BTW Bas, I thought one of the main attributes of the Aikido topology was its excellent psu rejection qualities - perhaps the earlier version of the minimal reactance psu using a resistor as the shunt element would suffice? If you use the 10M45S as a CCS perhaps there may be sufficient differentiation from the published circuit to allay your conscience.
 
Hey Douglas! Long time no see. Are you banned from AudioAsylum or something?

It is a neat exercise, but it has not been my preferred sol'n.
Regards,
Douglas
Ofcourse..so many people so many tastes..

I've got to admit that I tried making my own version of Dave Davenport's CCS...and was not too happy either. Then I got his real board with the IXS parts and tube shunt reg. It was a definite step up in transparency and everythin. Anyway..I'm not trying to convince you....just my experience.
:D
 
IXCP10M45S

Good suggestion..the DN2540 is still available..but you need to buy in quantities of 50 not very practical.

BTW Bas, I thought one of the main attributes of the Aikido topology was its excellent psu rejection qualities - perhaps the earlier version of the minimal reactance psu using a resistor as the shunt element would suffice? If you use the 10M45S as a CCS perhaps there may be sufficient differentiation from the published circuit to allay your conscience.

I think you are right...I'll just go ahead and ask Dave if we can use that as a psu for the aikido.
 
Hey Bas,

I think the performance is improved with good CCS. I have tried with the same CCS several PS design, up to a 25 Hy, 816 rectified, 100 uF of motor-run version.

Let me tell you that the L-C/Hg rectified one was my distinct preference!

I tried this same one with 6AX4 damper diodes( found a few volts increase in the un-reg ), and liked it quite well. Then on to the 816's...

I am not banned from AA but I may resign. Banned from the DIY/Tube section for telling the world that MQ-Mike had his head up his backside when it came to understanding how the transformers work( remember the PP, class A loading discussion? ). I topped it with then spreading the news that I could also produce reproductions of his cash-cow Peerless designs. So I was booted.

I am done with it, and the idea of making audio work my primary means of support. I now work at designing HEV's( and this pays far better than the audio ever did ).

regards,
Douglas
 
Hype and hyperbole.

A conventional series regulator with the drain as input and source as output (and made using a conventional enhancement-mode mosfet) presents itself as a high slope-resistance and therfore a constant current sink to the smallish capacitor following the rectifier. It has the added advantage of a low output impedance, unlike the MR psu.

The idea of using a minimal size capacitor is good because anything that widens the conduction time of the diodes will generally lower the VA's demanded from the transformer.

That said, IMHO I think the MR psu offers absolutely nothing over a conventional but chokeless series regulated supply.
 
Cyclotron,

I fail to see the hype and hyperbole? Ok..maybe the state of the art statement? Maybe over the top..but in my experience...Dave's CCS with the shunt reg sounds very good.

And once again..it's not like I'm preaching the gospel here....but I appreciate that Dave writes about his journey and tells one what he is doing. I would seriously like to see what you prefer...so make a post and show me? I'll build it and compare it, so I can learn.

Cheers,
Bas
 
Hi Bas, interesting kind of power supply. I’m thinking to try something similar to feed the heaters of Direct heating tubes. I’ve often listened that series regulators provide bad sound in DHTs and that ccs are good. I’ve try ccs power supply for not direct heating tubes and I found this kind of feeding really good. No hummm and no “cloudiness”…
But a ccs is not a regulator and in that way you can’t fix the voltage rock stable and if you try to use a three pin regulator in front of the ccs you’ll probably loos most of the advantage of unregulated power supply.

With a minimal reactance psu I’ll be able to set easily the output voltage without nasty series elements
What do you think about this idea?

Mark
 
Hi Mark,

if you try to use a three pin regulator in front of the ccs you’ll probably loos most of the advantage of unregulated power supply.

From what I READ..this seems to be true. Otherwize Guido Tent would not be selling his heater solution. ;-)

With a minimal reactance psu I’ll be able to set easily the output voltage without nasty series elements
I'm not too sure...the early MR was really a CCS with resistor soaking up the extra current..the latter versions use shunt regs which lower the impedance (I believe) and stabilize voltage...apparently a DHT heater does not like to be fed from a low impedance supply (which I think is the reason for the popularity of current regulated supplies over voltage regulated ones)

Regards,
Bas
 
Bas look at the schematic, I think that I was not clear in my first post....

I don't want to use any shunt regulator like zener or so but only a trimmer to gnd. When I used lm317 or lm350 I found that the voltage drop across them was not always the same so I'll ever need to adjust the voltage in front of them.
In the way you can see in the attached, this kind of setting will be much easier.

Mark
 

Attachments

  • vreg.jpg
    vreg.jpg
    18.4 KB · Views: 558

mark_titano

Member
2004-10-19 10:10 pm
Sure, I hope that too.

What I thought is that if you set the current in the first ccs higher that needed, due to the constant current over the second ccs the only way for that current to flow is in the direction of gnd.
If there is a resistor in this path then you’ll be able to set the voltage between ccs 1 and 2 changing its value. So you’ll have a voltage regulator using only another ccs and a trimmer…but maybe I’m wrong and we’ll wait for experts opinions ;-)

Mark
 
drifting away from the original subject??

I have not build it yet ... but I will use the first IC to set a fixed voltage (2,5V higher than the heater nominal voltage) at the input of the second IC who actually feeds the current.

Btw BAS if you will end up with a standard PCB for this minimal reactance PSU I would really like to give it a try.

Ciao