Advise on interstages...

[Max N]

Sorry guys, coming to this thread a bit late.
I'll put up some more detailed info about what I had at Owston in a separate thread to keep this one clean, but to address the specific questions here:

Nick, the interstage transformers I was using have their origin in a 211PP project that Neal and I started. They were specced by Andy Grove and Neal got them wound by Morite. They are PP 1+1:1+1 and were designed to be driven by a 5687, but similar valves should obviously work. They are bi-filar wound, so bandwidth should be good (I'm hoping to measure them asap and will post the results). The bi-filar winding means that there is a limit to the voltage between primary and secondary, so you need to keep the HT on the driver valve sensible, and you probably don't want to negatively bias the secondary (ie you may be limited to cathode bias on the output valves). Depends how far you are happy to push things, but you're obviously relying on the enamel to isolate primary from secondary.
The main downside was the interminably long time it took Morite to wind them, but that's understandable since it was such a small order.
I don't want to drop Neal in it, but if there was a lot of interest he might be persuaded to organise another batch.......

To answer Neal's original question:
Neal, if you already have a phase-splitter then surely you need a PP:PP interstage? Why not try the Morite ones since you already have a pair.
I'm thinking you could use a 5687 as your long-tailed pair, and load the 5687 with the primary of the interstage. Is there any reason why the 5687 can't be both the phase-splitter and the IT driver at the same time? The limit might be, if you're DC-coupled from the input stage to the long-tailed pair, then the voltage on the primary might end up a bit too high for the Morite ones. Nick, or anyone else, do you think this would work? Two halves of a 5687 as a long-tailed phase-splitter, CCS tail, loaded with the IT primary. I suppose normally you would tweak the loads of the long-tail pair to get perfect balance, which you wouldn't be able to do, but the IT will balance for you anyway. Neal, if you don't fancy this then I will give it a try and see how it works out. It might take me a while though, work has gone mad at the moment.

Neal, if you do want to get rid of the phase-splitter and try SE:PP interstage, then you already have a pair of interstages in the WAD 300BPP amp (do you still have two of those?) You could whip the ITs out of one of those amps and try that in your 2A3 PP? My gut feel is that SE:PP interstages are harder to get right than PP:PP interstages - but I'll be better placed to comment after I've done some measurements on the various ITs I have here - I also have some Lundahl PP:PP I could check.

Max

[Max N]

Andrew I, the Lundahl LL1676s can be excellent, I would stick with them if I were you. I did a lot of testing on these a while ago - don't wire them up as per the data sheet. You need to use the primaries as secondaries, and the secondaries as primaries. And you need to parellel what are now the primaries.
Connect 7 to 9 and use as input +
Connect 6 to 10 and use as input -
Connect 1 to 4 as output CT
Connect 2 as output +
Connect 5 as output -

Don't load them with a variable impedance as K&K do, experiment to find the best load and then buy/build a volume control to give that load (so no shunt volume controls)

Used like this they are flat from 3Hz to 50kHz

This seems to be a common theme for Lundahls, the data sheet sometimes not specifying the best connection.

[andrew Ivimey]

Thanks for all that stuff Max. I will check how I have 1676 wired ......................... Yep, I got them that way - spot on!

This morning I got an email from Mr Purvine. O-Netics transformers are not 'off the shelf' I'm afraid. Eight weeks seems the normal time for order/shipping.

But he also sent me the latest Karna information pack as a .pdf that I'm happy to email to anyone - except that if you order from him, I will get my transformers even later!

[Neal]

Max, yes that may work, I don't know... AG's idea was to drive the 5687 with a Concertina splitter. I could use my existing longtail pair directly as you say...hmmmmmm

I don't want to touch my 300B! I sold the other to Richard on the WD forum a while back.

[IslandPink]

Andrew - I'd like to see that Karna information if you can
markjeffs at btinternet dot com

Max - very interesting comments . I found the LL1676 real pigs to wear in and later changed to Sowter 3575's - but maybe your extra understanding here is what's needed .

Bud Purvine is making some chokes for me currently and some of the tech e-mails he as sent me are mighty interesting - about small signal throughput and 'latency' . It's got James's appetite whetted for a new PP project !

Mark