Discussion on the design of the T2 & other circuitry

[luvdunhill]

I didn't play with it too much, I tried 3 or 4 metal films (PRP, Dale, Panasonic, and another) and 1 CC and the difference was 4 or 5 fold after the first cap. Again, this is with a 22VAC transformer, so things may be different here.

[Linear]

What do you think of this design?

http://www.basaudio....heets/hvreg.pdf

II really like the basaudio thing. Sure wish i had seen it earlier. The T1 current source

feeding the gate of T2 makes all the difference. But it also needs the low voltage supply

for the opamp, which i did on the T2 with a current source and zener.

Will study more tomorrow. Thanks marc.

Still i like the hybrid approach that shuts down the current source in the absense

of a load. And adding that to the basaudio thing probably turns it into an oscillator

like the problems i was having.

Regarding the basaudio:

Run for your lives!!! It has TWO TL431's in it! Sprenger and Wright are infected with 431 fever!!!!

Seriously, though, the current source they use is what I suggested in posts #24 and #25. I then added T1 and sent it by email to KG as "dwpass4". HONEST, I had never seen the Basaudio cct before! (So, KG, you did see the T1 current source idea earlier!)

Also, as I previously suggested, they use a MOSFET Cascode in their shut regulator element. And I thought that was my original idea . Well, at least mine doesn't need a low voltage supply or an opamp, but those components may make theirs a better shunt (maybe?).

Damn, I re-invented their circuit! I HATE IT when that happens! I'll try to be more original in the future!

Linear

[spritzer]

Thanks for your comments, spritzer.

Yes, I had already thought about doing a mod on the SRA-14s to give it a balanced input. I even went to a lot of trouble to get complete schematics and a service manual for it (from Japan). Then, when it came time to "put it under the knife", I just couldn't do it! I just couldn't cut the chassis to pieces on a relatively rare bit of Stax history.

So, for balanced, I'll have to build a T2!

P.S. Actually, the 14s schematic is significantly different than the SRM1 Mk2. For one, it runs on higher voltage rails.......I'll send you a PM with more details.

Linear

You can always fit mini XLR's which do fit perfectly in the stock Stax RCA holes...

The SRA-14S is one of the few amps that I haven't bought and set to KG for analysis so it would be great to see the schematic.

[guzziguy]

Yes i have tried a few versions of that too. That was the one time i got

my ass kicked, first time in a while. For some reason getting hit with

voltages like that does not have the same effect it used to...

I'm going to have to try again.

Viagra would be a lot less painful.

Edited April 1, 2011 by guzziguy

[kevin gilmore]

i thought we already had the schematic for the sra14s... will look soon.

[spritzer]

Nope, we didn't have that one. Could be a variation of the SRM-3 circuit (i.e the SRM-Xh with a full size PSU) so it will be interesting to see.

[Linear]

I have tried every combination of capacitor of all kinds, and resistor and capacitor of all kinds

in parallel with the power diodes with no success. Now if i go back to really lousy diodes

and EI transformers, then the caps actually do stuff. But with the fast recovery diodes the

last remaining bit of noise is very hard to get rid of...........

Could you give us a quick (few lines) primer on the trade-offs with diodes and transformers?

EI transformers? Are they the standard ones with the E shaped laminated core? Are they better for noise? Why use toroids then? Because of their better regulation and lower losses?

What about diodes? Anything special about the DSEP12-12B? Will any fast recovery diode work as well? How do "lousy diodes" help with noise? Should we all be using lousy diodes and EI transformers to minimize noise?

I'm confused and need advice from the Doctor! I hope that the Doctor is in!

[kevin gilmore]

Not too many companies want to make EI power supply transformers these days.

Actually it seems that these days, no one wants to make anything unless they

can make a million of them. And then not even then (semicoductors i like)

Torroids are much more efficient and smaller for the same amount of power

delivered. Torroids radiate much less magnetic field. Torroids buzz less.

(well some of them anyway)

It may also have to do with the fact that for power torroid transformers, you can

wind the primary on one area of the transformer, and the secondaries on

the other side and avoid the hastle with interwinding capacitances and

other stuff that seems to get UL in a tizzy. There are some things about

the good old days i would rather have back.

The fast recovery diodes are at least 10 times (probably more like 25) faster

than the old high voltage diodes.

Once again, trying to find 1kv to 2kv diodes is getting harder and harder and harder.

So i use what i can get my hands on, with lead times of less than a year.

The SiC things might be really good. i'll let you know once i get some.

Fact is that if you build stuff the good old way with the good old parts, you are

nowhere near the ultra low noise levels that i can get now.

sucks don't it.

Edited April 2, 2011 by kevin gilmore

[Linear]

Now if i go back to really lousy diodes and EI transformers, then the caps actually do stuff. But with the fast recovery diodes the

last remaining bit of noise is very hard to get rid of.

Just to clarify. (I'm a bit foggy today, I guess!)

If we built a PS with and EI style transformer connected to a 1n4007 bridge, would it have lower switching noise that a torroid /ultra fast diode setup? (Assuming the same capacitor values, etc)

[Linear]

Well, there has been a lot of “talking” on this thread, but not too much bench testing of the ideas. I think that KG is the only one to help out in that area. So, I decided to do my share.

I’ve built a breadboard of what I egotistically call the kgdwhps1, and I think that I have some interesting results. The circuit incorporates KG’s hybrid idea (post #19), and my series element (post#24 on the left), but with a bipolar pass device with an ixys FET current source driving its base. (I suggested this exact configuration as “dwpass6” to KG in an email on March 31.) This pass element only has the isolated package MD2310FX on the heatsink Also, I’m using a re-designed version of the “active battery” section (in the blue box) that I call “dwsr”. It doesn’t use any HV transistors (more details on the dwsr later). See the attached schematic. I’ve also attached a photo of my bench setup.

This thing is designed to deliver 250V @ 50 mA. It can run as a shunt, hybrid, or series PS, just by changing the values of Rs, and Rh. I’m using a Hammond 229B230 “semi-toroidal” 12VA EI type transformer. These are widely available and only cost about $12. It would take 10 of them to build a full T2 supply, and they could be mounted on the PCB, thereby spreading the transformer heat around. I know they exhibit poor load regulation (20%), but the regulator can handle that. Tell me again why I should order custom toroids? Unfortunately, my ixys diodes are on order. I thought I had some 1n4937 (200ns) diodes around here somewhere, but I can’t find them. So, I’m using “double O seven’s”. I’m not sure how bad (or good) the initial switching noise is with this arrangement (compared to DSEP12-12B’s and a toroid). Comments are welcome (see posts #82, 83, and 84).

RESULTS

It works! I had to put 50 pF caps across the CB junctions of Q2, and Q1/Q3, to stabilize and set a dominant pole for the two feedback loops. Once that was done, I tried it with no load, dead short circuit (47 mA) and with a 25 mA load. All OK. Hybrid mode also works fine. Obviously, I need to do much more testing, and probably add a couple of zeners for protection. Could try MOSFETS (eg fqpf8n80c), which would hopefully be stable. Looks like a TL431 (Basaudio style) in the series element is NOT stable with the dual feedback loops (KG’s tested it). I’m probably wasting my time testing noise until I get my ixys diodes, right?

HEATSINK

I have been looking at the heatsink issue for a while (see DIY T2 Clone thread), and so I decided to try a piece (3” long) of the “super large” one I have. It’s cross-section is 4” x 5” with 19 fins! You can see it in the photo. I ran “kgdwhps1” in shunt mode, pumping 11.2 W into it. The ambient was 74 F, and that 3” piece of heatsink stabilized at 114 F. That means an 18” length could handle 67.2 WATTS! That’s just ONE side of a T2 type chassis, so it’s 134 W for a full chassis. KG, your 150 watt chassis is reality!

Stay tuned. More to come.

Linear

[kevin gilmore]

Those are the same as the krell heatsinks from the ksa series,

anodized black, in 9 inch lengths, and turned sideways.

[Linear]

Think I'm going to build a pure shunt PS for my T2 (the one that isn't built yet!) This "kgdwhps" seems to be working really well and the heatsinks really "eat the heat".

For transformers, why not use 10 Hammond 229B230's, a 229B56 to get 60V for the 560V rail, and a 229B24 for the +/- 12V. Make it modular, using 12 identical PCBs. The PCB layout could take any of the 229 Series transformers and could then be stuffed to create any voltage. (Triad also second source the 229 Series, as their FP line.) As I said before, this would really spread the heat generated by the transformers around the PS chassis.

So what's wrong with using these "semi-toroids"? Let's hear all the advantages of using toriods. Check out what you get for $12 in the pic below.

[cetoole]

Isn't that Hammond just a split bobbin? I don't see how that is related to a toroid at all. I would imagine the benefits would be extremely low interwinding capacitance and very low bandwidth.

[Linear]

Isn't that Hammond just a split bobbin? I don't see how that is related to a toroid at all. I would imagine the benefits would be extremely low interwinding capacitance and very low bandwidth.

Yes, the Hammond 229 series is a split bobbin. (They call it a "semi-toroid" because it has a few characteristics in common with toroids.) I want to compare the 229 against true toroids, which would have better regulation, lower self-magnetizing current, etc.....

In other words, are true toroids that much better that we can rule out using the 299 series?

[Linear]

"Bug Fix" on the ".kgdwhps1". Latest version attached: "kgdwhps2"

Added diode in series with series element to avoid reverse current through 2n3904 and Rs.

There may (actually, probably will) be additional "fixes". So, don't generate your BOMs just yet! This means you, Justin! (By the way, I really like your chassis work.)

Linear

[Linear]

Attached is "kgdwhps3". I haven't tested this yet, but if it can be made stable, I think it would be the best topology for the new Hybrid PS. (I'll get it on the bench as soon as my fqpf8n80c's arrive from Mouser.)

Compared to version 2, the HV NPNs are gone, replaced by the MOSFETs. The ixys10m90s is gone, replaced by the Supertex LND150 (available at Mouser). This is a TO-92 device, off the heatsink. Only the isolated package MOSFETS are on the heatsink, making this easy to build. For higher power and voltage (I'm testing these ideas at 250V @ 50mA), we might add more '8n80c's in parallel, and more LND150's to share to load.

Thanks to luvdunhill and the Basaudio guys for suggesting the LND150 (only $0.69!). It's a handy little HV current source. Check it out.

Linear

Update: Has been tested and has stability problems. See next (#92) post.

Edited April 17, 2011 by Linear

[Linear]

Time for an update on the "kgdwhps3" schematic shown in the previous post (#91). My MOSFETS finally arrived, and I got it on the bench.....

Bottom line: It is VERY difficult to make stable with the MOSFET output on the active battery. Using the FET increases the gain of the circuit dramatically, and this is a problem. I cut the gain down by adding a resistor (51K) across the gate and source of the fqpf8N80c, but it still takes lots of capacitance to tame it. So, right now the MOSFET version is a work in progress.

MEANWHILE, the bipolar version "kgdwhps2" (post #90) seems to work perfectly, with only 50 pF compensation caps. I've taken it up to 50 mA with 50 uF on the output and it's fine. I haven't measured any noise yet because my "Plexiglas blast shields, Teflon caps, and 5 digit floating noise analyzer" are still on back order! (as recommended by KG for diode noise measurements. )

Regarding the DWSR block that I re-designed, I will have more information very soon.

Linear

[luvdunhill]

Maybe Kevin should investigate using a Music Rail in the PSU:

http://www.bybeelabs.com/

[ujamerstand]

"These things (Music Rails™) are stupid good. I have no idea how they work, but because they work so well, I actually don't even care!

wow! Wow! and WOW!"

Edited July 5, 2011 by ujamerstand

[kevin gilmore]

The amount of bullshit Bybee can produce makes Ray look like a 2 year old.

Sales of quantum purifiers must be slipping.

My guess based on the wiring diagram and the extra parts for higher voltages

is that the thing is a power fet of opposite polarity plus a couple of diodes

and small capacitors.

Similar to the noise reducing capacitance multipliers used in the Melos units.

So for positive voltages, a p-channel fet, for negative voltages a n-channel

fet.

Price not specified. Bet they are $200 to $300 each.

Pathetic.

[luvdunhill]

Ah, thats a good possibility. Maybe a magic series resistor as well for good measure. Ah, physicists today.. Don't they have better things to be doing?

[kevin gilmore]

Actually unlike the quantum purifiers which do jack shit, this

thing actually does something. And that something could be very

bad.

Notice the .5 to 2.1 volts of input to output differential.

If you put these things on a dynahi, dynafet, or B22 you will

absolutely kill the effect of the dual tracking power supply,

deliberately offset the rails, and probably force the servo

to clip in one or the other direction.

Now do i believe the 45db of noise reduction... Nope, in testing

a few of the possible circuits, maybe 10db at most, and only

when you put such a device between the unregulated voltages

and the rest of everything else. At which point with a well

regulated power supply you can't tell the difference between

the thing in and out of circuit.

But wait there is more.

Lets say you are dumb enough to put this on say a kgsshv power

supply, or something (anything) else that is 450 Volts.

Do a quick calculation. The R1+R2 load resistor eat up 7 watts

(which would be 14 watts total for a dual supply) plus the

power in the zeners, say 18 watts total.

33% more power all going directly into heat including the

extra power the supply has to generate to light the thing up.

In fact a number of tube preamps and small tube otl amplifiers

with tube diode power supplys, are unlikely to be able to

supply the additional power to light this stupid thing up.

Yep, another quality product from jack bybee.

But i notice a new difference too, bybeelabs.com and bybeetech.com

Each selling different products. I wonder what that is all about.

[kevin gilmore]

I'm thinking that this is probably pretty close.

http://gilmore.chem.northwestern.edu/bybeething.pdf

[Kerry]

I just found this thread. Late to the party again.

I've got some stuff to post here, but I just got back from a business trip and still need to catch up with life.

[Milosz]

CAUTION: If you use a Quantum Purifier along with any Machina Dynamica product or service you risk creating a Ponzi-Barnum condensate, which may result in a singularity.....