Yorkshire Boatanchor event

lead boatanchors are best.

Moderator: Paul Barker

Post Reply
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Yorkshire Boatanchor event

Post by Paul Barker »

The event was a resounding success. I'll be linking a web page for it when built.

I'd like to use my area here to discuss with Steve Bench and you dave where to go from here in development of my 212 power supply.

On the face of it that may seem a stupid question.

I used Steve's resistorless capacitorless power supply method with the driver amp (ml4:PX25). This I found to be a massive improvement to that stage, so I knew that it would be great to develop such a supply for the 212 also.

However my first attempt ended in tears and I had to take a standard supply to the event.

What I built was 4 572b's in paralle and 20 75c1 references 75v/60mA. I put 150 ohm grid stoppers on the 572b's but no method of balancing them which was the problem.

With the driver supply also LR shunt smoothed, when placed the output stage supply on, the two supplies played ping pong, one lit the other went out and vice versa. In short they don't do sharing. This light show was oon followed by a Taylor 572b out of the quad imitating the bar of an electric fire, just before it changed it's mind and decided rather than be a heater it would prefer to be a lightning display.

Clearly I need to have a balance arrangement or better still use a single valve. The gm100 would walk it, but I have recon that the ty4-400 is up for it if I keep to the now lower operating point of 900v 100mA. The 212 sounds fine at that, nobody noticed any lack of power or anything. Run at lower ratings a power valve gives the lie that it is more powerful than another valve which would be flat out at that rating.

My biggest obstacles are now interaction between the supplies and my inept pathetic useless inability to wind any chokes with sufficient inductance capable of the current required.

For the first one, the answer has to be to further shunt regulate the main supply. I need more current than tap off's from the vr's would allow, but I could drop the main rail with voltage consuming vacuum tube ccs loads for instance.

Steve I'll do some drawings at some point but can you see what I'm talking about? Mainly I like the simplicity and absence of the need for caps anywhere of your method of biasing the shunt valve in A2 with the vr's.

Well done for an excellent design, everything is more musical base is better but most of all air and hf extension is greatly improved.

Dave, can you help me on or off line to develop a choke input choke fit for 500mA? I don't care how high the dc is. I have some large m6 scrapless lams, back of e is 180mm bobbins allow 63mm stack. I have no formal traing in inductor design, haven't any resources for it either, so I am clueless. I can't cope with a design that requires an economic perfect layer system which maximises truns for the space, becvause these bobbins are far too large for my small winding machine. I just have to bastard wind on the laithe for this one.

Do you need any more data like weight of core for instance? length of magnetic path? When dc current and air gaps are involved I have no data available to me.

ps Steve, remember Jonathan Noble from the vintage?
Thanks he's over in UK at the moment, so he came to the event. His little globe 45 amp on James D's dipoles was exquisite.

Paul
Guest

Post by Guest »

Hi Paul,

I'm neither Steve nor Dave, so if you don't want me here just say - will delete the post and shut up.

I have been thinking about the same problem - shunt reg for 304tl. I'm a big fan of Steve's solution and here is what I have had in my mind:

1) shunt element - 450th (pitty there is no 750th). Looks nice at 1kV: moderate current, gm approx 6umho giving (idealy) Rout=160 Ohm. Low capacitance is just a bonus keeping the efficiency high at HF. Maybe not so nice at 1.5kV due to 200mA plate current but in no compromise version should work ok.

2) Making strings of dozens of low volt VR's may be less than optimal: the internal resistances + inductances are summed up + problems with lighting them up (however the latter can be remedied by 200k resistors across each VR). There are EU VR's for higher voltages: STV280/40 for instance (although I rememebr seeing once something for 1kV) .From the datasheets one such tube should have lower internal resistance than 4x 75c1 (240Ohm vs. 340Ohm). There are also STV280/80 which are more expensive but with yet lower impedance. Both are quite common here in Germany.

Just my 2c,
Jarek

Just my 2c,
jarek
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

Jarek thanks for coming in.

I have already built the supply for the px25 part using two stages of 572b and compared to solen caps it's a revelation.

Thanks for the tips on using vr's. I'd better look out for those higher voltage ones then. I didn't however have a problem getting the 20 75/60's to light because of the riple before they are lit. I musn't use anything that in the end would be troublesome though do I?

Thinking about it I was foolish to think 4 572b's would share current happily.

The ty4/400 is only just in the running if I keep to a sensibly low operating point for the 212. Actually I find it quite powerful enough on ariels at 900v 100mA or thereabouts, the power supply is for both channels so I just want to consume a maximum total including the output valves 500mA.

Then comes the problem with chokes.

I had another thought about that and the problem of supply voltage variation say when I take the amp somewhere else. What about a variable series pass reg prior to the shunt reg, turn it up sufficient at the start for the vr's in the shunt reg to light, once lit, turn down the series reg to a satisfactory position. Could we then manage without inductors also? Any views on that?
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

Post by jarek »

Then comes the problem with chokes.
The higher the gm of the shunt element the less Hy's you would need to achieve the same filtering efficiency (disregarding the fact that the stroed energy is also lower), but at the high voltages the choices are limited. 450th with Rout=200Ohm, say, is equivalent to 8u cap at 100Hz. Not much but also not bad, given the quality. Pitty 304th cannot be imlpemented here. But 5H/500mA should not be very hard to get, especially that Rdc is our friend here.

BTW, one more idea for 1kV shunt: 805. At 1kV, +10V gm=5umho.
I had another thought about that and the problem of supply voltage variation say when I take the amp somewhere else. What about a variable series pass reg prior to the shunt reg, turn it up sufficient at the start for the vr's in the shunt reg to light, once lit, turn down the series reg to a satisfactory position. Could we then manage without inductors also? Any views on that?
The ingenious feature of Steve's design is that it also works at DC, stabilizing line voltage variation (you juts have to see within which limits so not to overrun the shunt element). Rdc acts as a balast element then, so high Rdc is good here!

I'm suspicious of series regs sonics wise, but I was also thinking along the same lines: CCS instead of a choke, However this seems too complicated for my taste (especially given the currents) - enough play with shunt tube!

Best,
jarek
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

Yes, I just spent all evening on phone to Vry, he put me right off the prefix series reg idea, though I do have some 13e1's which could do it.

Like you say rdc is not a problem at all.

Did I say how good the capless supply is in the driver part? Recovery from transients is instant, gives the impression that a 1.6 watt stage is 10 times as powerful, and how dirty those caps must have been?

pics of Saturdays get together at

http://homepage.mac.com/scress1958/WADF ... bum21.html
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

I have to say it's a big mistake phoning Vry because when you eventually get off the phone your head is full of all sorts of ideas thrown out by a mathematician. I only just ordered at great expense Mundorf Silver Golds to hopefully be the final tweak for the Ariels and Vry is telling me to build a small amp with an air cap and an air cored output transformer to drive the tweeters! I'd like to tell him where to stick his ideas, but he's right, only two days late with the information!
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

the one you describe as "Believe it or Not. This Amp Had a Discman Feeding it and Sounded Magnificent" is Mark's Amity under continuous development in Aurora direction.

"This is One Evil Looking Beast" is Jonothan Noble's globe 45 amp with trick outpout transformers. It was the most involving sound. Jonothan is my old buddy from Vintage Tubes Talk, he developed this sound over many years has crossed many hurdles to arrive at the place where he can just modestly say; "I go for the romantic sound". The output transformers are important but there are other tricks...

"The Phantom Amp Tweaker Strikes Again!" Will is that your amp?

"Loved the Polished Wood and Brass" Steve's original px4 build circa 6 or 7 years old now. This was the first px4 amp I heard and forced me to sell all my 300b's (about 8 I think there were).

He also brought his new much loved 1 watt MP41 amp, but spent a lot of time under the bonnet.
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

sorry guys last post was meant for another forum, woops.
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

Steve I'll do some drawings at some point but can you see what I'm talking about? Mainly I like the simplicity and absence of the need for caps anywhere of your method of biasing the shunt valve in A2 with the vr's.
Gee. That'll teach me not to look at the forum all day.... look away for a second and lotsa posts.
The filtering in that method is simply L"R" filtering using the series L and effectively the "R" of the shunting tube. Indeed it gets rid of any problems with C elements (there are none). For the most effective filtering, the embedded resistance within the inductor is useful. Dave's pretty dang good about coming up with the right R and the right L in a part, so have at it with respect to what you need.

Steve
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

Good day Steve,

I think I'll go dual mono so that I'm not shuntiung so much current. I could then use one 572b within it's ratings at say 100mA for the first stage of filtering and then use the ty4-400 for the second stage to shunt 200mA. The output stage would probably run at 130mA, the other stages 50mA.

Dave,

So my first pair of chokes need to be 500mA capable at choke input. That is the difficult one. Plenty of dcr is not really a problem, (I can adjust the HT secondary voltage empirically using an over voltage transformer and variac, then wind the correct power transformers later) is something like 30h possible?

Paul
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

Post by jarek »

Yo Paul,

Vry's mathematician? He has never told me that...The idea of air cores is just great - worth separating the freq. into channels, IMHO (look at my plan on the other forum). Air caps - the aviable sizes good only for RIAA pre. Take a look at Ebay on Russian vacuum caps - they have copper structure. Indeed cool for RIAA!

On the shunt: Rdc is a *good* thing here - you add more ballast. Sure you loose volts + produce heat, but you go no-compromise anyway, or?

The tube: it does conduct almost all of the signal to ground, so if I didn't like the tube on the output I would not use it for shunting...hence my idea of 450th.

Choke: 30h/500mA...WOW! Just to save a bit of Dave's work (where is he by the way??) here is a rough formula for getting turns:

n=L*I/(B_0*A)

B_0 - questient op. point of the core (like 0.8-1T), A - core area in m^2 (use SI units!!!)

This is the *lower bound* for the number of turns, but with high currents it becomes more and more accurate. At least you get a rough estimation on what's going on. Say you use 40cm^2, B=1T =>3750 turns at least...not so bad, just remember Rdc is your friend here.

-jarek
Last edited by jarek on Tue Jun 07, 2005 11:18 am, edited 1 time in total.
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

What Vry means is that because you are only inetrested in tweeter frequencies the air cap coupling inside the amp (or perhaps a pair) has sufficient pass point. You can of course tune if you get air variable caps. Maybe a vacuum cap and an air variable in parallel if more required than is available from air variable alone.

I knoiw the final shunt valve plays the music, but when using the 572b (even Chinease ones) the balloon px25 which already sounds great sounds greater by quite some margine. So though that shunt valve may affect the sound, it is already a lot better than the alternative capacitor. It's a win win situation, but I can see that the best still is the exhotic valve.

Probably at the operating point I might be looking at the best choice is the gm100, but sensibility holds me back, though I have a pair in the loft. I have 4 or so ty4 400's so I'll give that valve a try, sure it'll sound good.
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

I could have discussed this in that secret place where you get hit by the mob if you tell anyone it exists, but I know Steve and Dave, key witnesses lurk here, it seems you too, so we have a quorum. This is a public place afaik. Anyone that isn't visiting is missing out, and quite honestly isn't missed.
IslandPink
Posts: 56
Joined: Thu May 12, 2005 8:01 am
Location: Denbigh, North Wales

Lurk

Post by IslandPink »

I'm here, lurking .
So you were up to much more than I thought in the last few weeks !
MJ
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

Post by jarek »

What Vry means is that because you are only inetrested in tweeter frequencies the air cap coupling inside the amp (or perhaps a pair) has sufficient pass point. You can of course tune if you get air variable caps. Maybe a vacuum cap and an air variable in parallel if more required than is available from air variable alone.
Yes, due to aviable sizes aircaps would be good only for coupling. But anyway why cap couple if you can use aircore trannies anyway? No iron, no saturation, no problem 8) And for PS final caps they are too small anyway - we want to filter something just not to have IM products comming from huge 50Hz component. here Steve's shunt rulez.
Probably at the operating point I might be looking at the best choice is the gm100
Some people do have a masochistic tendency to use this tube at any cost :wink: but I'd take a closer look at 805...it is still good at 1200V +8V, 100mA, gm=5m. I couldnot find any curves for ty4-400. What's its gm at your given point? That's the figure of merit.

-jarek
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

Hi again Jarek,

at 90mA gm is stated to be 5, mu 25. I don't have transfer curves to predict what it might be at my operating point.

It has the benefit of the extra dissipation 350w if used below 100mcs with ambient cooling conditional upon really good heat sinking connectors, particularly anode. It suits better than the 833 from the point of view that it goes into a2 at less current / voltage and heating is less of a challenge (slightly) it also fits my pocket because I hold them in stock, though I have a pair of 805's, I reconed the 572b looked more apropriate than 805's with the 160 watt dissipation, but were I to use those I'd probably need two per channel on the final stage. I would however put an ameter on each cathode and a variable resistance on the anodes to match them up.

Mark, I know you and Nick were conspiring to find out what I was up to, which made it all the more fun keeping it a surprise! Got yer!
dave slagle
Posts: 2022
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

paul wrote: Do you need any more data like weight of core for instance? length of magnetic path? When dc current and air gaps are involved I have no data available to me.
I shouldn't scrapless lame are easy. I'll have to convert to The equivalent US lam size, but that’s not an issue. So you have an EI-180, that puts the tongue width at 60mm, and your 63mm bobbin suggests a 60X60mm cross section. This would be a square stack of US EI-237 (if that exists 2.375 inch tongue) the Length of the magnetic path should be 14.22 inches (31.6cm) and if that checks out everything else below should be close.

Now punching the old calculator buttons.

3000 turns of #23 will net you a DCR of ~65 ohms
A .1 inch gap will give you about 15hy with a Bdc of 7500 at 500ma.
Bac at 100hz will be 2.18 gauss per volt (so 1000VAC will net you ~2200gauss)

Knowing that you still have dome flux available I'd shrink the gap a bit to get to say 10KG to buy some more inductance and going to a .07 gap will up your Bdc to 10,500 and up your inductance to 20hy.

The 60 ohm DCR will dissipate 15W but that shouldn't be an issue given the size to lose that amount of heat. (you will be running the copper at 1000 circ mils per amp)

A somewhat braver soul might go for #24 wire which will net you the following.

3800 turns DCR 100R
.1 inch gap gives you a Bdc of 9300 and 24hy of inductance
Bac @ 100hz becomes 1.7 G per volt. At 25W of dissipation it might start getting toasty but you should still be fine.

Going to #25

4700T dcr 160

.12 gap nets 32hy at 10KG Bdc with 1.3G per volt AC.


The above numbers should be close, and I strongly suggest two things.

Play with the gap in circuit to confirm proper behavior of the device.

Get ahold of some bondable magnet wire to assure quiet operation of your device.

dave


PS,

let me know if you want this to be an entire (paul barker) forum so the various ideas will be kept separateyet together (much like steve has) then you will get the keys ot moderate and be responsible for keeping your own yard clean :-)
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

Post by jarek »

at 90mA gm is stated to be 5,
gives ideal Rout=200Ohm, which is equivalent to 8uF cap on 100Hz. Given the quality it's a good tradeoff I'd say.
It has the benefit of the extra dissipation 350w if used below 100mcs...
This is a good point. As I see it, for the safety of operation shunt should be able to withstand the full voltage + current at no load condition comfortably.

Best,
jarek
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

Dave,

yes please you are such a kind sole, I'll have my own place please?

That is most useful information, we have a saying a friend in need is a friend in deed.

Jarek and Steve included of course, and Mark lurking Uncle Tom Cobbly and All.

Jarek, you too are a mate.

you too are a big help, and also it's nice to have someone to talk to about it when Dave was busy probably in his day job.

Dave I' spent a whole evening getting my head filled by vry. If you've ever had an evening on the phone with vry you'll know that leaves you with all your plans in shreds and a whole lot of new ones.

The long and short of it is that I may be wanting you to bespoke wind my mumetal double c's for something or other (tweeter amp, grid chokes or some such) and maybe more mountainously my huge double c core 50% Nickel's as output's for one project or another.

The big ones have damage to the corner of one core where it is cut, but discussing it with volts at vsac he didn't think it would matter that much.

Anyhow, the smaller cores I could post but the big ones would be best waiting for someone over there to take in his luggage or anyone here going there, then post to you from inside your country.

If anyone here's of anything let me know.

I might be able to get away with a 45 globe out of the mumetals, whe you get the cores you could advise Dave.

I have resigned myself to the fact that transformer winding is best left to your ilk.

let me know if any transport arrangements present themselves.

This is a lifetime thing so no big rush in deciding designing or transporting.

Paul
shinebox
Posts: 88
Joined: Fri Jan 28, 2005 9:44 am

guilty as charged...

Post by shinebox »

Ooops... sorry Paul, didn't mean to do your head in so much! Good chatting to you even if you regrets it...

Btw, I'm not remotely a mathematician... those guys (proper ones, at any rate) are far, far more scary.

The air-core may have been an extreme idea; I've only managed to get mine running on the bench with a signal gen so far, they arespecialised job for 3k:0.5 ohms for ribbon direct drive and the ribbons aren't ready yet, but the txs measure well. So I think getting Dr Dave to wind those mumetal C-cores for a tweeter transformer is the way forward. They should be just about big enough for a treble amp. Criminal not to use them for something.

One of the most beautiful sounds I ever heard from a valve amp was with my Sowter mumetal opts on EI on a clipped up lashup. I think with Intact's winding techniques, the superior double C cores and your amp/ tuning skills; well, that would be something to behold.

Just don't underestimate the power you need for a tweeter amp - also, depending on how you crossover, you may need even more power due to something called crest factor, but I don't want to scare you any more just yet. :twisted:

shines aka da vry
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

I don't mean it really Chris, just having a laugh as I'm sure you know. There is so much going on otherwise I'd be glad to almost put it to a vote, we'll see what Dave thinks of the cores first.

I suppose vis a vis not underestimating the power needs of tweeters, the answer is to see what power they need using lesser transformers first and then decide on the output valve.

Must get hold of Peter about my transformers from him. Just talking to you has again reminded me. And must get onto that other manufacturer and ask what they want me to do with ahem you know what.
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

Post by jarek »

you too are a big help, and also it's nice to have someone to talk to about it when Dave was busy probably in his day job.
I think he rather sleeps at that time. Anyway, thanks for the invitation!

-jarek
dave slagle
Posts: 2022
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

paul wrote:The long and short of it is that I may be wanting you to bespoke wind my mumetal double c's for something or other (tweeter amp, grid chokes or some such) and maybe more mountainously my huge double c core 50% Nickel's as output's for one project or another.
I feel honored :-)
The big ones have damage to the corner of one core where it is cut, but discussing it with volts at vsac he didn't think it would matter that much.
I agree, in the whole scheme of things it shouldn't matter. Hell call it a custom gap and make it a feature.

Anyhow, the smaller cores I could post but the big ones would be best waiting for someone over there to take in his luggage or anyone here going there, then post to you from inside your country.
Or you could just hang out with a heavy box and a few quid outside Heathrow.

Seriously, I don't need the cores however bobbins would be nice! If you get me the core dimensions and we can come up with bobbins, then we can get a design done, and a first go at a wind. You can test it just as easily as I can and since you will be testing it in circuit your results will be more appropriate. If we cannot come up with bobbins, Pieter would be your best bet since he layer winds. We can come up with a suggested design for the bobbin to make his life easy or just tell him what you need. It would actually be very cool to approach the same design both layer and "random wound" to see if any differences show up in testing and listening.


dave
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

Post by jarek »

One more shunt tube idea: T150-1. I remember seeing those babes in Warsaw and was impressed by the looks - cool tantalum plate! Should be ok up to 1.2kV, but high current rather.

-jarek
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

Post by jarek »

well, gm is about 3.5m only, but still those tantalum plates...

jarek
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

CLR

Post by jarek »

Paul,

re clr shunt supply, I was having the same thoughts: is C input better here due to its worse load characteristics? Is using HV diodes better than MV due to their higher R? I eventually came up with a negative answers to both questions: in both cases the equivalent series resistance is nonlinear. As for using resistors to make up a ballast, I think it's simply better to wind appropriate chokes - you avoid crappy resistive wire in da path + get high impedance as a bonus.

-jarek
Paul Barker
Posts: 59
Joined: Sun Feb 20, 2005 10:20 pm
Location: UK

Post by Paul Barker »

Thanks Jarek for the thoughts.

Over in England the rectifiers are oxide coated U19's roughly speaking 250mA 2k5 max secondary, 30 second preheat.

Do you think it best to parallel then or series them to cope with the extra current above 250 mA? In parallel one might hog the work. What do you think?

Thanks

Paul
jarek
Posts: 62
Joined: Wed Feb 02, 2005 10:02 am

Post by jarek »

Paul,

you connect in series if you need higher PIV; you connect in parallel if you need higher current.

Best,
jarek
Suepemponge
Posts: 3
Joined: Tue Feb 21, 2012 5:09 pm

Yorkshire Boatanchor event

Post by Suepemponge »

Whereabouts in Yorkshire are you thinking?

It would be worth getting in touch with kaya and/or chrisv40 who are coordinators for Retrobike rides in Yorkshire and the North East.

I coordinate rides in the North West area but Id be happy to help out if needed.

Also have you any dates in mind...?
Post Reply