Assessing a PP transformer to use as SE

[systemtruck]

Pardon this question, which surely some of you well seasoned folks have had to begrudgingly entertain for the newbies once in a while.

The reason I’m asking is because I have my hands on some pretty beefy output transformers that are difficult to find specs / info about. They were made by HP in the dawn of their development.

I’m also asking because david at Cinemag mentioned to me recently that one of their transformers may not even bat an eye at some DC current because they’re so robust. This leads a new guy like me to think, hey, if these other HP transformers I have are quite literally 2X the physical size and weight of those Cinemag’s, or more, maybe they can handle some DC current.

Is any of this sounding like it has some mathematical backing?

I know that inductance is key for low frequency signal. I assume that gapping transformers makes up for lack of strong inductance. So im guessing here that if a transformer isn’t gapped but has massive inductance, maybe they’ll pass DC without too much loss of bass. But how much?

A completely different take on this is a quote I found somewhere…

“”” P-P transformers can work in a SE fassion if you measure your idle current of the tube and feed the current through the center tap as well as an equal current that goes to the other end of the 'unused' winding and you put a load resistor that draws the same amount of idle current as the tube. “””

 

[Murdock]

There are far more experienced people here who will hopefully help you but I have to correct you on one thing.

I know that inductance is key for low frequency signal. I assume that gapping transformers makes up for lack of strong inductance. So im guessing here that if a transformer isn’t gapped but has massive inductance, maybe they’ll pass DC without too much loss of bass. But how much?

Gapping transformers is done so the DC current does not saturate the transformer. Gapping a transformer drastically decreases its inductance. Therefore SE transformers are typically larger than their PP counterparts to make up for the loss in inductance. As larger core and more windings lead to more inductance.

Some PP transformers still have a small gap build in to handle small current imbalance of the push pull tubes. So if the current of the SE circuit is small enough, some PP transformers could be used in SE configuration.

Do you know out of which HP units these transformers are? Maybe you can find schematics and get data from there.
Post some pics and maybe DC readings of the windings.

 

[emrr]

“”” P-P transformers can work in a SE fassion if you measure your idle current of the tube and feed the current through the center tap as well as an equal current that goes to the other end of the 'unused' winding and you put a load resistor that draws the same amount of idle current as the tube. “””

That can be seen in RDH3 and other ancient texts, never have I encountered it. It’s still a PP output stage with no signal feeding one side, hard to imagine what conditions would make it a favorable arrangement.

 

[systemtruck]

The HP transformer part number is 9120-0037 and it was in a handful of old oscillators.
Not much in the way of specs can be found, except size and then some cross part numbers..
https://www.parttarget.com/5950-00-...37.html/-C397F95D-AA8E-41A7-AF73-328350F3E2D1
https://mil-nsn-parts.com/parts-catalog/nsn/5950000784083-5950-00-078-4083-transformer-audio-f.html

Attached is a pic someone else took of one with the bells off, and there’s a ruler for size reference.
Below that is a pic of one of mine still installed in a unit.

My application would be for a parallel 12SN7, so single ended.

These come from a standard 6v6 PP config, to 600R line out. I’m assuming primary is in the 6-7k area. I could shunt the secondary some to lower the primary to what a 12SN7 parallel wants, though I’m not sure what that is yet. And since it’s a triode, not pentode, I guess there is some flexibility in the primary being high anyways.

But two 12SN7’s will kick out up to 20mA DC. I could do an inductor + cap parafeed setup but the inductor cost makes it a moot point. Just looking to make use of good parts I already have!

I was confusing saturation with inductance. Mixing them together. Now I understand that gapping helps reduce saturation, and that inductance suffers from gapping so has to be made up for with added size. So I guess I’m separately wondering if size itself also helps avoid DC saturation. This is the impression I get when someone experienced says thatsome given non gapped transformer won’t be affected by some DC current… that the size of it can absorb DC and gapping isn’t needed. So I’m wondering is 20mA is way too high of a number to throw at any non gapped transformer.

This applies to some other transformers I have too, so I’m looking to understand a little more about it. Thanks!

 

[Murdock]

20mA is a lot for SE… can I ask for what circuit it is?
You can try but I would say keep it for a project where it fits. 6v6 PP would fit to RCA BA-6A or other Vari-Mu compressors.
What you can do is check the lams from outside to see how their are stacked. If it looks like in that picture it has a gap. If that gap is big enough to handle 20mA you have to try.
If it has two lines it’s not gapped.

 

[systemtruck]

I think it looks non gapped. I see two lines. Unless I’m looking for the wrong lines. Pic below.

I may try some of my good PP transformers in parafeed mode with what parts I have on hand. I can toss a decent cap between plate and transformer and see how it runs. To the best of my understanding thus far of parafeed, the big inductor between B+ and plate is to filter supply ripple noise. If that’s all it does, maybe I can still get good audio through the setup using just a cap, and if it works well I can then check around for some decent inductors to filter supply noise.
I did however find an interesting article about adding supply ripple to the cathode so that it cancels out, and you don’t need a a giant expensive inductor for anything.

https://www.tubecad.com/2014/09/blog0308.htm

View attachment IMG_0392.jpeg

 

[CJ]

Should work, try it,

Use the whole primary, safe off the CT.

HP was the Cadillac of the electronics industry back in the day. Everything was over engineered for the utmost reliability. At least that's what my electronics teacher told me back in 71,

 

[systemtruck]

Should work, try it,

Use the whole primary, safe off the CT.

HP was the Cadillac of the electronics industry back in the day. Everything was over engineered for the utmost reliability. At least that's what my electronics teacher told me back in 71,

Are you referring to the illustration in that article about adding ripple to cathode from B+? You think I should give that a go?

I’ve attached the image below. It does seem relatively straight forward to calculate for one tube… C2=C1x(Mu).
I suppose one 12SN7 has a Mu of 20, so C2 should be 20X the value of C1.

But what about the context of two parallel 12SN7’s? Maybe C2 = 40X C1 now?

Yeah these HP outputs seem way oversized, comparing them to what I have that I know is nice. I have done some solid listening tests after recap and the HP output sounds great.

I also have a pair of brand new Cinemag 9600T’s burning a hole in my parts bins. Those are 15K:600, where as the HP’s are probably lower (PP 6V6). Maybe what I’ll do is set up this whole parafeed circuit and try swapping the HP and Cinemag. I’m really not sure what impedance a parallel SN7 pair wants to see because I’m not even sure what a single SN7 wants to see.

 

[k brown]

https://www.misterpp.com/

...sorry.

 

[Cqwet Dbdfte]

How much output do you expect from this setup?
Using a regulated power supply easily gets rid of ripple to sub millivolts.
12SN7 is not a prime candidate for an output tube.
A PP circuit with one tubes input grounded would yield a SE output, with some (active) bias control the non-active tube can set the balancing current for the PP transformer.

 

[systemtruck]

Happy holidays everyone!
So I’m putting together a test build of a Gates SA-134 along with a number of other preamps that have less DC current (so I can use other OPT’s for those). This one uses parallel SN7’s. I’ll stick with that for now.
That’s a great point about a regulated supply. I have a PACO B-12 on hand which is regulated, but not the cleanest signal.. spec’d at 300mV ripple though I think I measured much lower before. I also have an external supply I built that is somewhere in the ballpark of 125mV, though not regulated. But I could build one that’s regulated very tightly and I suppose not have to deal with chokes or ripple-canceling caps as discussed.
I also can add some more stages of filtering to my unregulated diy supply and see how low I can get ripple. The voltage is already pretty high so a couple steps down can be afforded.
If ripple filtering truly is the only purpose of the plate choke in a parafeed setup, then for my project it seems very reasonable to address ripple from the source instead of trying to filter each unit.

 

[systemtruck]

I read the PACO B-12 specs completely wrong. Was reading the regulation accuracy.
The ripple is 3 mV.
Seems serviceable for this situation, and then maybe it can be beat with a custom supply.

 

[Wavelength]

Systemtruck,
First a little surprised HP would use 6V6 pp for a line output at 600ohms, that seems silly unless they are driving 5000 feet of cable.
Second... ok yes this will work but not too well. Peerless use to include max offset dc current when specing their PP outputs. Remember a PP output is basically 2x sides in parallel with the CT going to AC ground. While yes full primary is best for SE on a PP but base is going to be minimal and 15K is a bit low for 6SN7/12SN7. You would actually be better off using both halves of the 12SN7 to drive the damn thing.

I have a bunch of 9600T that David built for me recently and using them with a 6SN7/12SN7 would have poor top end and of course best to run them parallel feed and not direct.
I built a boat load of USB DACS with B7/15K in high nickel parallel feed with a 220H loading inductor and a 6SN7 didn't fair well with that at 10ma. You need something with higher transconductance and lower Rp like 6922, WE417/5842A, 5687 or others.

 

[systemtruck]

Systemtruck,
First a little surprised HP would use 6V6 pp for a line output at 600ohms, that seems silly unless they are driving 5000 feet of cable.
Second... ok yes this will work but not too well. Peerless use to include max offset dc current when specing their PP outputs. Remember a PP output is basically 2x sides in parallel with the CT going to AC ground. While yes full primary is best for SE on a PP but base is going to be minimal and 15K is a bit low for 6SN7/12SN7. You would actually be better off using both halves of the 12SN7 to drive the damn thing.

I have a bunch of 9600T that David built for me recently and using them with a 6SN7/12SN7 would have poor top end and of course best to run them parallel feed and not direct.
I built a boat load of USB DACS with B7/15K in high nickel parallel feed with a 220H loading inductor and a 6SN7 didn't fair well with that at 10ma. You need something with higher transconductance and lower Rp like 6922, WE417/5842A, 5687 or others.

I’m surprised to hear that 15K is low for a pair of parallel 12SN7’s. It’s not been easy finding specs on the 12SN7, such as a load resistance that is recommended. Probably since they aren’t typically power tubes.

I’ll probably give it a shot still, as I’d described with a parafeed transformer loading of a pair of 12SN7’s, despite the possibility of a “poor top end”. I’m coincidentally going for a very dark sounding system. So, maybe that works out just fine.
I have a set of CM-27101’s that are gapped, but only for 10mA. I believe the SA-134 output will require 20mA. But I could at least try them. It’s a 30K load.

But again I had been assuming that a pair of 12SN7’s was going to want a load in the ballpark of 6K or even less. So learning that 15K is not high enough is surprising.

I could also hook up the 9600T to provide 30k, since I could use just half the secondary and load it with 600R.

 

[CJ]

people used to alter the freq range a bit on those HP 200's, stock they went to 600 K Hz, so if you give it a bump you could start your own bootleg AM radio station with the 6AV5's pumping out the juice,
so yeah, you can drive a truck with those things, they are nice for testing transformers,

the HP used two output transformers, one for the low freqs and one for the high,

here is a guy who turned one into a SE amp>

https://audiokarma.org/forums/index.php?threads/hp-200cd-6av5-se-amp.376120/

 

[systemtruck]

people used to alter the freq range a bit on those HP 200's, stock they went to 600 K Hz, so if you give it a bump you could start your own bootleg AM radio station with the 6AV5's pumping out the juice,
so yeah, you can drive a truck with those things, they are nice for testing transformers,

the HP used two output transformers, one for the low freqs and one for the high,

here is a guy who turned one into a SE amp>

https://audiokarma.org/forums/index.php?threads/hp-200cd-6av5-se-amp.376120/

Yeah HP seems to have a lot of nice transformers in their stuff. I have a pair of simple interfacing units, I think the 353A, that has a high freq and low freq transformer kinda like the 200CD does, though I think these are 1:1. I want to utilize them in the master section of a mixer I’m working on.

Anyways that’s a cool project there. Was excited at first that maybe he’d implemented some HP push pull transformers as SE, but it looks like he just installed some Edcor SE transformers and otherwise the rest was all HP parts.

 

[Wavelength]

ST,
Rp is like 7K for the 6SN7/12SN7 at 10ma, so the drive or damping and distortion would be off for 15K. In general more like 21K would be best (3x rule).
As I said I tried this with a B7/15K to 500 and was not impressed with the results using the 6SN7 loaded with a 220H choke and parallel feed with a 3.3uF film cap. The B7 was loaded with high nickel and wound with Nippon %99.999 pure copper. It worked a lot better with higher transconductance tubes.