Another go at PP -> IT -> PP

What people are working on at the moment
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#1 Another go at PP -> IT -> PP

Post by Max N »

I have been having another go at a push-pull amp, specifically using a push-pull first stage, interstage coupled to a push-pull output stage.
Or, as per the thread title, PP -> IT -> PP

I know I have been down this road a few times before, for example:
211 push pull thread
300B push pull thread
and a closely related thread
measuring transformers

In my previous attempts, I haven't been able to produce a system using this topology that satisfies me over the long term. I think I now understand at least some of the reasons for this:
1. Every push-pull stage cancels the even harmonics and sums the odd harmonics. Even harmonics are pleasing to the ear, odd harmonics are displeasing. Therefore, it will not be easy to make an amp which consists of multiple push-pull stages sound good.
2. Good interstage transformers are not easy to design and produce. However good they are, they will need to be driven by a low impedance to give their best.

These reasons may seem obvious, but I think actually doing the experiments and looking at the results has given me a better appreciation. Getting up and running with a sound card (actually a USB audio interface) and Audiotest software has also been a big step forwards in understanding what is going on

You may ask 'why persevere with this topology?'. A couple of reasons. Firstly, more and more DACs now feature balanced outputs, and it seems perverse to go balanced, single-ended, balanced. Secondly, the alternative (going single-ended for at least some of the stages) doesn't really work for me either. I find it sounds great with most music, but not all. Looking back, the best version of my 211 amp ended up with a single-ended 6sn7 gain stage, followed by a phase-splitter, then a push-pull 45 driver stage, IT coupled to the push-pull 211 output stage. The 6sn7 gain stage produced almost pure second harmonic, the PP 45s at a suitable operating point produced low odd harmonics and gave a reasonably low driving impedance for the very good IT transformers. But it still didn't really hit the spot. Also, in its final incarnation, it wasn't so much a boat anchor as a barn door of an amp.

Hopefully this rambling post gives an introduction to where I had got to. Looking back through the photos on my phone, I am surprised to see that I first started assembling some of the building blocks for another go at PP -> IT -> PP in February of 2019. I always intended to post what I was up to on Audio-Talk. Better late than never!
RhythMick
Old Hand
Posts: 747
Joined: Fri Jun 23, 2017 12:24 pm

#2 Re: Another go at PP -> IT -> PP

Post by RhythMick »

Interested to hear how you get on and your conclusions. I started with SE, then went PP -IT-PP and have settled on SE with choke-cap coupling. Of course as we learn and understand more our ability to design and construct a good sounding amp improves so whatever we've done most recently is more likely to be the best sounding... To a point.

I may try PP again. In the meantime I have some very nice Lundahl PP interstages and output transformers. Depending on where you are you would be welcome to borrow them for your experiments.
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#3 Re: Another go at PP -> IT -> PP

Post by IslandPink »

Max, I understand exactly where you are with this. The class A PP output stage drives speakers really well - the bass and timing can be excellent.
Did you ever try the 'Harmonic Equaliser' trick with the resistor in the ultrapath cap from the cathodes ? It would have to be an 'Amity' or 'Aurora' style circuit for this.
When i tried it, it certainly helped.
The data that Gary Pimm or Lynn Olson posted, showed some reduction in 3rd & 5th harmonic and some more 2nd-harmonic.

It's a shame you didn't get the result you wanted from the SE front end though. I think I had a CCS-loaded 76 in there, at the end, which would not have had much distortion at all.. !
One of the issues can be transparency of the phase-splitter. Which ones were you using ?
Some of them would not be driven sufficiently transparently with 10ma or less from a 6SN7. The small Sowter ones are an exception, but i wouldn't go below 10ma driver.
"Once you find out ... the Circumstances ; then you can go out"
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#4 Re: Another go at PP -> IT -> PP

Post by Max N »

Hi Mick
Which Lundahls do you have exactly?
As it happens I'm currently using a pair of LL1663 outputs for my experiments. I have thought that they may be a weak link, so may invest in something better one day. You've also reminded me that I have a pair of LL1671AM PP interstages somewhere which I should really measure and have a listen to. I think I originally bought them when I was thinking of 300B PP to drive 211 PP - I was going to try them in 2+2:1+1 step down. But I came to my senses so they are in a drawer somewhere.
I am roughly between Oxford & Reading - I think you are further north? Maybe we can meet up at an Owston one day, when the world is back to normal.
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#5 Re: Another go at PP -> IT -> PP

Post by Max N »

Hi Mark
I haven't tried the Harmonic Equaliser yet. I think MJ gave me the impression (in the book, not in person!) that it only worked properly if your speaker was an almost purely resistive load or some such. I may have remembered that wrong. I think there was another reason as well which will come back to me I'm sure......

On the phase-splitter, I wasn't very clear, it was 6sn7 input dc-coupled to a 6sn7 cathodyne phase-splitter, cap-coupled to the 45s. The only slightly unusual thing was that because I had 750 volts available I could dc-couple and still have very nice operating points for everything. It measured very well.
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#6 Re: Another go at PP -> IT -> PP

Post by Max N »

I'm still probably not being totally clear.
Where I ended up was
6sn7 SE input -> dc coupled -> 6sn7 cathodyne phase-splitter -> cap-coupled -> PP 45s -> IT-coupled -> PP 211s

Four stages of madness. No wonder it was big!
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#7 Re: Another go at PP -> IT -> PP

Post by IslandPink »

Been a busy day, & need sleep ! ( finished work now )
More to say - but -
Thanks, I made assumptions, missed option of valve phase splitter.
I do think it's significant that when Olson/Pimm worked up the 'Amity' design about 20 years ago, they rejected valve phase splitters. is there anything on the Nutshell HiFi site about this, in the Amity pages ? It will likely come down to the symmetry of the amplitude/phase and how that propagates.
Lundahl - I used LL1620 first, then bought some from Bud Purvine that were better. A pair of the cheaper ones ( still good ) that Paul has/had, then some really nice ones with Litz wire secondaries, that i have . I have a kit forming now in order to make the 'Tabor' amp using the same transformers, but not sure when I'll get round to that ! ( the input phase splitters I like are the Sowter 3575 )
"Once you find out ... the Circumstances ; then you can go out"
User avatar
izzy wizzy
Old Hand
Posts: 1496
Joined: Fri Nov 02, 2007 7:02 pm
Location: Auckland NZ
Contact:

#8 Re: Another go at PP -> IT -> PP

Post by izzy wizzy »

Somewhere there is a test that Lynn and Gary did of the Lundahl 1676 as a phase splitter. It measured incredibly well. I'll see if I can find it.

I was a fan of the Sowter 3575 and used it up front in my line stage which is also PP. Swapping out the ll1676 on the GM70 amps for them was a big step backwards. I then replaced the line stage pair for lundahl 1544a. I have quite a few 3575 left over now.

I've also been reading that different types of phase splitter all sound different and for those that like the SE sound, the unlikely (to me) candidate was the floating paraphrase. Personally I don't see how any of them work perfectly symmetrically.

I keep reading about pp cancelling 2nd as though it does it perfectly. Given the phase splitters aren't perfect and the gain devices aren't either the might be some cancellation but there must be some left. Been reading some Nelson Pass articles on distortion and apparently we like considerable 3rd harmonic as well. It's the higher odd harmonics that are the most objectionable.

Given we all have our own sonic preferences, I can say, for me, I've never had such good sounds at home. I find the involvement such that I lose all track of time and am not focused on what I should do next other than get them into smaller chassis. But it took a lot of work to get my amps to this point.

Looking back, that was a bit of a rant, sorry. I look forward to watching your journey in this thread.

Cheers,
Stephen
Last edited by izzy wizzy on Wed Dec 23, 2020 9:40 am, edited 1 time in total.
User avatar
Nick
Site Admin
Posts: 15708
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#9 Re: Another go at PP -> IT -> PP

Post by Nick »

All I can add is that I find it helpful to prevent losing touch is to try and work out just what you dislike in a design sonically and try and devise a plan to address what you hear. If you can correlate measurements with that all well and good, but work out what you are trying to fix first. I would agree that valve phase splitters are problematic, look at the work Radford did maybe with high frequency balance.

Eggs may have been sucked in this post by the way. Sorry.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
izzy wizzy
Old Hand
Posts: 1496
Joined: Fri Nov 02, 2007 7:02 pm
Location: Auckland NZ
Contact:

#10 Re: Another go at PP -> IT -> PP

Post by izzy wizzy »

PS. Forgot to say, I really like the 1671 as finals driver. The am version should be even better if amorphous is your thing.
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#11 Re: Another go at PP -> IT -> PP

Post by Max N »

Thanks everyone, really good input there.
Ref phase splitters, the intention this time is to to stay PP/balanced all the way thus avoiding phase splitters. But I will say in passing that I am a big fan of the Jensen JT-11-P1 used as a phase splitter. Data sheet here:
https://www.jensen-transformers.com/wp- ... -11p-1.pdf
I have found it to be sonically transparent. They come up on Ebay packaged inside ISO-MAX boxes:
https://www.ebay.co.uk/itm/JENSEN-ISO-M ... SwXGRcKUJP
Ref output transformers, I think that Bud has retired, so I'm still looking for a recommendation for something better than LL1663. Cathode feedback seems to be becoming popular, has anyone tried that in PP?
Stephen, good points about harmonic distortion, I shall have another read through your thread. I will dust off the LL1671s when (if) I finally get back to the 211s.
Nick, wise words there from you as always. I do still find myself getting lost on occasion, but not nearly as much as I used to! I have found that access to better measurements has made a great difference for me. Sometimes I can pin down what I don't like about the sound, but sometimes it is still just a feeling.
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#12 Re: Another go at PP -> IT -> PP

Post by Max N »

I've been trying to figure out the best way to explain where I've got to and how I got here. I think I'll just do a brain dump.
OK, back to February 2017, I was starting to get my audio mojo back after a few years playing with telescopes. I had built up some of Andrew's VCCS filament modules to heat the 45 drivers. It occurred to me that the amount of heat the modules generated was more to do with the current than the voltage, and that led me to think about a quad of 6b4g's that I had bought. 1 amp filaments made them 'ideal' candidates for DC filament supplies?
6b4g heaters.jpg
Sure enough, the modules seemed to run cooler feeding the 6b4g's than the 45's
User avatar
IslandPink
Amstrad Tower of Power
Posts: 10041
Joined: Tue May 29, 2007 7:01 pm
Location: Denbigh, N.Wales

#13 Re: Another go at PP -> IT -> PP

Post by IslandPink »

I think that Bud has retired, so I'm still looking for a recommendation for something better than LL1663.
I would find out what Sowter can do for you. My gripe with the Lundahls is the kapton interlayers, unless they have changed to paper or something else in any of their recent units.

Just an aside on the 3575 input phase splitters, I thought about this and for a two stage amp using DHT's, the input levels would have to be significantly higher than what I had ( using 6N6P's drivers ) , the 3575's would be likely a bit overloaded, so 1676's would make more sense.
"Once you find out ... the Circumstances ; then you can go out"
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#14 Re: Another go at PP -> IT -> PP

Post by Max N »

I think back then I had already decided to have another go at PP all the way. But I realised I had a lot (too many) options, and I figured the only way to decide the best way forward was to build everything and listen and take measurements.
So I took some left over plywood and cut it into handy sizes that would fit on my work bench. And I started to build up 'modules' that I could arrange into combinations along the bench. Most of the modules were designed to drive IT transformers, so I was looking at ways to get low impedance.
Potential candidates for the first stage included parallel triodes and high gm triodes and pentodes such as E55L. Started with a pair of Hammond transformers to cover a range of HT:
hammond traffos.jpg
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#15 Re: Another go at PP -> IT -> PP

Post by Max N »

Added a few of Pete Millett's modular sockets, bench PSU for the heaters. Notional first stage ready to go:
high gm driver valves.jpg
Some of you may have seen this module in the 'Measuring Transformers' thread..
Post Reply