Aikido

[shaun]

Hi (all)Simon

I thought this might be a good place to talk aikido.

My line stage is using JB’s design and started its short life with 6n1p and 5687 B+220v 430r cathode resistors but I’ve swapped the 5687 for 6n6p which are much more my cup of tea. By the way Simon I’m using a series reg PSU which works pretty well and is pretty low in terms of output impedance so if you've not tried that yet it could be worth thinking about for later.

The thing with me is I always seem to get things back to front. I guess most people start with a good understanding of how things work and then build but for me I build something that sounds interesting and get the understanding through the build process. Odd I know but when i see how it goes together it makes it ''real'' to me (shall i do the 211 now or later ). Ok so the penny dropped a bit on this topology and the thought process is starting. So a discussion on the ins and outs would be nice for me and a good way for me to put the fragments of my understanding together.

What I have so far

Can we do this in bits?

1st stage

1) The Aikido input is (give or take a foot ) working pretty much as a Mu follower variant. I.e. a common cathode amplifier with an active load. But the difference is the trade off between gain and flexibility. The big penalty penalty is a reduced gain of just under half Mu.

2) The nice part of this arrangement is that the load line is effectively a straight line. so within the min/max operating conditions of the selected valve and fixed voltage its possible to cherry pick the most linear OP possible @half B+for the valve used. If I’ve got that right its pretty cool I’d say and leaves enough scope for experimentation.

One thing I’m not to sure about is if it’s ok to use different values of cathode resistor for input and output? I think it probably is as long as the resistors are equal on the top and bottom of each valve.

have i got that right


i do get the noise reduction thing but a bit at a time works fine for me.

Yes Simon the line stage is sounding pretty nice in my set up and it’s made me interested enough to learn more.

Hearing Nick’s Aikido into Steve’s 300a was enough to get me reading.

Nick you stated in your aikido thread that you found that running the valves less hard reduced 2nd harmonic distortion. That’s interesting. Any more thoughts on that.

So I’ve tried to make a start on the getting to grips with it all thang. But I need all the help I can get.

If I’ve got it totally wrong (would not be the first time) please say IM HEAR TO LEARN AND HELP if i can

Take care

Shaun

[Nick]

Sounds like you have it understood well. At least as well as I do, so maybe that says less.

Nick you stated in your aikido thread that you found that running the valves less hard reduced 2nd harmonic distortion. That’s interesting. Any more thoughts on that.

Well, what I did think, that because I was using a ecc40 before the first stage, I was passing a lot more signal than expected into the first stage, so by increasing the cathode voltage I was giving the stage more headroom.

But what make me doubt that, is that the clipping is gentle. I should try with a standalone aikido. I will see if I can get to that over the next week or so, and see if the same reduction in distortion happens with a normal two stage version.

[simon]

Hi Shaun,

Sounds like you're doing okay. You understand it as well as I do anyway; don't think I'm the person to be increasing your knowledge . I'm more of a practical person so I tend to build before I really understand. It's not ideal, but at least this way I get to build stuff .

One of the nice things about the Aikido is that once built it seems pretty tolerant of messing with, whether that's changing valves or cathode resistors. And it sounds pretty good too.

I'm interested in regulated PSUs but haven't tried one yet (lack of time and opportunity). Any pointers? I'd like to build a Steve Bench type experimenters amp with variable HT, but I'm at an early stage of reading at the moment.

Cheers,
Simon.

[shaun]

hi Nick SImon

Nick as always your input would be most welcome

''I'm interested in regulated PSUs but haven't tried one yet (lack of time and opportunity). Any pointers? I'd like to build a Steve Bench type experimenters amp with variable HT, but I'm at an early stage of reading at the moment''.


Simon
i found Steve's regulator primer is really useful when i started to try to get my head around how things work. seems we have something in common as far a building is concerned. i don't let an incomplete understanding get in the way of the fun of building things. i began to view steves primer as a swimming pool, i had a good look through and jumped in where i felt comfortable i built a basic series reg with no frills. it works well and it's simple to do. the thing is that with not to much work the extra bits can be added later.i always try to KISS every time.

i try to do a passive supply for my projects first to get things up and running and then do a reg when all seems well. i've built three so far. nicks shunt regulated supply for the WAD phono2, a Morgan Jones type series reg for my 76-45 and one of Steve's for the Aikido line stage. i'd have to say that to my ears the differences in each case have been quite obvious. the passive supplies always seem to sound a bit ''squashed and fuzzy by comparison whilst the series regs seem to have a more even ''wideband sound'' (think Island pink got that spot on). of course you may not like the results you get in your setup but i'd be interested to know what you find. i think you'll find plenty of help here

just one thing

with steve's reg i changed the pass valve from a EL509 pentode with a nasty top cap to a 6080. i just could not get the 509 to run quiet

take care

shaun

[Nick]

Nick as always your input would be most welcome

Well, as before you have most of what I understand.

What started out though as a simple "build a regulator for the 300b", got a bit strange, there were lots of emails between myself and Mark (IslandPink), the body of what I was finding was 1. regulators had a sound, and 2. to my ears, a "better" regulator in terms of measurement wasn't the better one in terms of sound.

There are still a lot of ideas there that I am unsure of, I am happy to talk of them, but it will come out as a general ramble, as I still haven't got a handle on what is actually going on.

In theory, series regs should be better, and in the case of the 45 amp (driver stage for the 211), A 12e1 shunt reg, certainly sounded better than a similar series reg, even though the series reg had a measurably lower source impedance, and lower ripple because of that.

This is the basis of my confusion, I went from the shunt on the 45, to try and replicate this for a 300b, but the extra current requirement defeated me, I had a thereticall A2 reg using a 572b, but it never worked well enough to get ripple down low enough.

This is one of my feelings (see I said it would be a ramble), I suspect that once B+ noise gets below a certain value, extra microdynamics starts to appear in the amplifier, and I believe that this is one of the reason that regulated supplies can sound better (see next for the other)

But there is another side to this, after having several drunk and sober conversations with Andrew at Whittam, I have come to believe, that the conventional CLC type of supply, contains the seeds of its own failure. Once the system contains enough C's and L's its starts to become very hard to model or (for me at least) to visuliase its operation. I think we are so used to thinking about our circuits in a single direction, power starts at the transformer, moves towards the amplifer, and is smoothed in the process. But this is just a partial view of whats going on. The electrons have no concept of where the amp is, they have no "arrow" from the mains transformer to the output transformer. So (after a small sideline involving prime numbers) I started to think any supply that let two chokes "see" each other was potentially flawed. And of couse the output transformer looks just like a choke. So any supply, that lets the L in the supply, "see" the L in the output transformer has a problem.

The extreme of this was the muitiple choke stages in the original LCR phono stage It modeled and measured as if the system was becoming chaotic.

Sometimes I wonder if some of the liking for low C, high L, high R supplies in may SET designed, is the effect of the humps it creates in the dynamic response of the system, but I just don't know about that for sure.

So it seemed to me that a regulator would provide a solution to keeping the L's apart. in the same way a CCS isolates a simple shunt reg from the upstream power supply.

But then you have the problem of where the current loops created by the signals in the amplifiers have to go in a series reg. A shunt reg provides a simple path (well simple enough, I still think some goes via the error amplifier). Thats why I think that a better measuring supply, with higher feedback in the regulator may not sound as good. In fact what I found, what the "better" supply sounded worst in much the same way as that which can happen when you apply feedback around a amplifier.

What I also found, was by forceing the regulator to be poorer, by introducing a small resistor in the output could reduce some of these problems. In fact it seemed that just adding a one ohm resistor could change the sound for the better.

And this is the point where the cap that may follow the regulator comes in, I think, that the effective inductance of the cap, compaired to the inductance of the supply forms a potential divider, and this determines where the current goes.

This (and the idea that phase shifts were bad things) got me to the multipel parallal series caps I had a Whittam.

But to add more confusion, the 6c33c amp has allowed me (by removing the output transformer from the output stage) to try and measure the way the impedance of the regulator changes over the frequency band. But what thats starting ot show, is the supply before the regulator is important. A shunt reg, makes the load seen by the supply more constant, but a series reg makes the supply see a more changeable load than a normal supply, so now I am trying to understand just what the requiremenst the series reg places on its supply, and how that affects the operation of the regulator, and the supply conditions that the amplifier sees because of that.

And so it goes on. See, I said it would be a ramble...

[richardcooper2k]

wow, very interesting nick. and explained in a way i could get the gist of.

[Nick]

Glad it made some sort of sense, I still don't have it fixed in my mind, but I do feel that there is something in there that is correct in some way.

[simon]

Hi Shaun,

Seems like I'm following in your footsteps . I was thinking about starting with the first error amplifier circuit on the third page as a starter. Interesting what you say about not being able to get the EL509 quiet. I haven't checked the 6080 out yet, SB refers to them in cascode - is this how you have them? I don't have any EL509s or 6080s so from that point of view I can use whatever's appropriate/convenient for the pass tube. I do have some ECL82s, the pentode section I believe is equivalent to a 6BM8 which SB uses.

Interesting stuff Nick, plenty of food for thought. Does your hypothesis of two chokes seeing each other extend to the PSU? What I find interesting about the SB stuff is that he only uses a 5H 50R choke. The regulation does away with much of the L of a passive supply, of which I have a lot in my easy-peasy lash-up phono PSU. Certainly for passive PSUs there seems to be different schools of thought - some prefer 2 chokes, some a single choke. Which is why we have to try these things for ourselves to find out what's best for us.

[Cressy Snr]

I don't have any EL509s or 6080s.

I have plenty of these Simon. I have about 10 6080s and the three EL509s which survived the OTL. You are quite welcome to come and pick up a couple of each for experimentation.

Steve.

[Nick]

Does your hypothesis of two chokes seeing each other extend to the PSU?

Thats exactly where it started. Every time I try and model a two stage choke supply, I find that the interactions go both ways, its not a simple linear progression that we may like to think of it as.

[simon]

Thanks Steve, that's very kind. When I'm ready I'll pop over.

[shaun]

hi all

yes simon i had many a happy hour trying to find where the noise in my psu was coming from i even did a passive by way of comparison and it was quiet enough so i ended up swopping out the EL509 for a 6080 and the noise was gone. at the time i put it down to two things

1) the 509 had a top cap arrangement.fine but the lead to the top cap picked up noise no matter what i did in the way of screening.

2 i did not know what i was doing

so you may do better. i also have a 509 you would be most welcome to please


Nick (this is a ramble)

thats an interesting thought on making the reg worse improving the sound. any thoughts on why that might be

something else that's not to clear from reading up on regs is how it works in relation to a two stage amplifier. is the usual resistor decouplig capacitor arangement between stages needed. if we have a fas supply that can sinkas well as source current i'm struggling to understand the need for a big cap in there. i have tried with and without cap and it does not affect the sound to much.

any thought

take care

shaun

[Nick]

if we have a fas supply that can sinkas well as source current i'm struggling to understand the need for a big cap in there

Thats part of the problem with series regs, they can only sink very very little current, the asymmety is why a cap is needed.

I feel (but haven't proved anything) the problem with the "better regulator" is the lower output impedance means the signal path is through the regulator, but when the regulator is "poorer" its higher impedance, means that more of the signal passes through the cap. Tjis is also where my multi cap array comes into the picture, its aimed at lowering the esr of the cap, meaning that the regulator can be better, and the cap still pass the bulk of the signal current.

I think the same effect happens when a extra ohm or so is added to the output of the reg, it allows less of the signal to go through the several paths through the regulator to ground,

[simon]

Thanks for the offer Shaun. I'll see how I get on with Steve's top cap-less JJs - they may work a treat, who knows?

I had a very quick think about the series reg at the weekend. There's not a lot in it really, the only things I'm missing are the EL509/6080 and maybe a cap or two. Think I've got a 105V VR tube somewhere I can use. I cut a piece of sheet aluminium at the weekend and thanks to Steve for his kind offer of the 509 and 6080 I think I'll put an octal socket in as well as the 9 pin jobbie for the 509. Easier than adding one later... I just need to think how I can supply separate heater windings for the pass tube and the error amplifier. I have a 7.5V tx I could bodge with dropping resistors I guess, but it's not very sophisticated. Probably get me going though.

[Nick]

I would check the heater ratings Shaun, most valves that are aimed at pass tube use have very good heater/cathode insulation, I think you will be able to get by with one heater supply elevated just above the error amp cathode.