Single ended series vs. parallel feed

[soundguy]

lets assume, in the vacuum of my fantasy, you could get a pp and se transformers that were similar enough to compare (maybe you can from lundahl). 

this assumed

anyone have any real experience comparing the not so theoretical difference between a series single ended output stage with similar parafeed stage using a plate choke (not a css)?  Trying to figure which would be best for the best low end response.  I mean, you DO have that cap to massage things in the parafeed. 

Assuming all transformers are the same (which they obviously arent, but....) what is the real advantage to a parafeed circuit compared to the series?

Been searching for days to try to find a good resource on parafeed circuits and coming up with very little that is inspiring.  The lundahl 1620 is available with a big and little gap, so I can take it either way, I just cant intelligently weigh the difference.  One you get the grunt of single ended series, the other you get to add a choke (plus) and control a cap value and construction (plus).  Is there something Im missing?

 

[emrr]

I'm confused about several things.  I have never encountered the term "series single ended"; am guessing you mean DC coupled vs. parafeed, but maybe I'm wrong.  What are you trying to build?    I'm also not sure what about SE vs. PP we are comparing here; using a SE output with a PP parafeed circuit?   

Some SE transformers are designed with orientation in mind, and may exhibit imbalance when used otherwise.  I have seen problems with some center tapped input transformers when used SE, with same units looking perfect when connected PP.  Many UTC's look better connected PP, even when SE is the normal app. 

I once repaired a PP 6F6 Collins broadcast amp by converting to parafeed operation.    The PP output transformer functioned, but it was noisy; a static type sound.  By taking the DC load off through addition of chokes, the noise went away.  In that case, I added no coupling caps; just floated the output primary. 

This looks like some interesting reading, though you've probably seen it.

http://www.siteswithstyle.com/VoltSecond/Parafeed_fun/Parafeed_fun.html

 

[Gordy]

The real difference is that a parallel-feed allows for *proper* transformer design. By contrast having an anode line transformer means DC flowing in the primary and hence a complete and utter compromise in the transformer design. It seems as if the advocates of anode line transformers actually enjoy the spray of complex harmonics that they create (...that were never in the original signal).

Using a competently designed parallel-feed transformer (such as the Sowter 8995 / http://www.sowter.co.uk/single-ended-output-transformers.php / http://www.sowter.co.uk/Parafeed-transformers.htm) brings all-round performance benefits. Also, as you have noted, you can manipulate the bass frequency response via the output coupling capacitor in a parallel-feed circuit.

I suggest not using CCS loads in power stages, as by definition they are also current limiters. Reserve them for line stages that are reasonably lightly loaded and they will sparkle. Instead consider a semiconductor anode load on top of the vacuum tube configured as a mu-follower power stage. This is generally cheaper than an anode line choke and capable of excellent performance. A 2A3 loaded with something like a IRF820 should give a very clean 5W. However there is a limit to the output that can be extracted this way owing to available FET types.

Good luck.

 

[soundguy]

am guessing you mean DC coupled vs. parafeed, but maybe I'm wrong. 

yes, DC coupled versus parafeed.

It seems as if the advocates of anode line transformers actually enjoy the spray of complex harmonics that they create (...that were never in the original signal).

for the longest time, this seemed to be the primary reason to me, to go with a single ended circuit.  More or less what is peaking my curiosity about experimenting with parafeed.

Ive looked at those sowter transformers and am slightly confused.  Seems that the primary reason to use parafeed is to use ANY regular ungapped transformer with the addition of anode choke, not really seeing what could be so special about those sowters, am I overlooking something?

 

[Gus]

With parafeed you need to think about the energy that goes to
the choke and to the transformer.  You also need to think about the inductance of the choke vs the inductance of the transformer and both inductance values for the load line.

There is a good PDF on the web about the output design of parafeed with chokes
 http://web.archive.org/web/20021012050843/http://www.bottlehead.com/valve/ramseypg1.html.

  Nice tube microphones often have a a resistor for the choke with a cap to transformer.
http://greygum.net/sbench/sbench/outstru.html

 

[Gordy]

Seems that the primary reason to use parafeed is to use ANY regular ungapped transformer

Yes, I think that you can. But why settle for 'ANY regular' performance?

I mentioned the Sowters because they have interleaved windings, a combination of core materials, wide bandwidth and (in my subjective experience) very fine performance.

However it's your design, and your choice!

 

[soundguy]

thanks for the links gus.

Im still confused about sowter's marketing on that parafeed transformer.  Its gotta be just marketing, right?  Is there something specific to a parafeed transformer that Im missing?  You either have a big gap to allow for the DC or you dont or am I missing something?  I dont doubt they sound good, just not understanding what makes for a parafeed transformer...  Leaning towards lundahl though.

 

[Gordy]

am I missing something?  

Yes, you are.

understanding what makes for a parafeed transformer...

A parallel-feed transformer is one that does not have an air-gap and is correctly dimensioned for your particular application.
However that does not automatically make it a good transformer. A good transformer takes experience in design, appropriate materials and skill in manufacture.

If you choose to believe that Sowter (and similar) are all about marketing then naturally that is your prerogative.
I wish you every success.

 

[abbey road d enfer]

A parafeed transformer is somewhat easier to design than a SE, because you don't have to care about the reduction of magnetic permittivity (reduction of inductance) and the displacement of the B/H curve (2nd-harmonic distortion). A gapless core has a much higher AL (inductance per sqrt_turn), so typically, you will probably need 3-5 times less turns for the same inductance (LF performance), which will also reduce stray capacitance, which is good for HF response.
But in fact the load has just been shed on the design of the choke: it must withstand DC current and stay linear, must have low stray capacitance, because, once it is in parallels with the transformer, all the non-linearities affect the signal in the same manner as they would in an SE configuration. The only benefit of parafeed is that the design of the choke does not need to take into account the coupling factor. In other words, you don't need to sandwich the choke, so you can use the multi-bobbin construction to reduce capacitance.
But IMO, a well designed SE can be made as good as a parafeed for less money.

 

[emrr]

I don't know what the f$&% you guys are doing with the term SE; I think someone other than me is confused.  There seem to be two different muddled sideways conversations going on here.  The newbies certainly have no f'in idea what this conversation is about.

SE = single-ended.  SE does not mean 'DC coupled'.
PP = push-pull
DC coupled versus parafeed; pick SE or PP.

There are SE transformers that do and don't take DC, likewise there are PP transformers that meet the 'no DC' descriptive, though they are much less common.  I have never ever EVER heard SE used to describe anything that might be used in a PP circuit, DC coupled or not. 

Dave, are you interested in DC coupled versus parafeed for PP or for SE circuits?  I assume you are building a PP power amp, from some of your other recent threads, along with the few transformers mentioned. 

 

[abbey road d enfer]

OK, it seemed that there was some kind of consensus, in this thread at least, that SE meant a transformer of which the primary is connected to the anode on one side and to B+ on the other side. I suggest we call that TOWTPICTTAOOSATBOTOS to simplify .
But you're right, parafeed is also SE, with its two variants:
1) A transformer of which the primary is AC coupled to a choke that is connected to the anode, which may be called choke-parafeed
and
2) A transformer of which the primary is AC coupled to a resistor that is connected to the anode, which may be called resistor-parafeed
In the end, to (over)simplify:
TOWTPICTTAOOSATBOTOS = DC in tranny, gap needed, more turns, more stray cap, more difficult, expensive
Parafeed = no DC in tranny, no gap, higher AL, less turns, less stray cap, easier, cheaper. But added cost for choke-fed or less performance for resistor-fed.

Who cares about DC coupling in a transformer output?

 

[lassoharp]

You may have already been there but Magnequest has a bit of info on parafeed and plate chokes.

Also would be curious to know if there's a particular non-parafeed design that's coming up significantly short on the low end response that prompted the interest in the parafeed alternative.

 

[soundguy]

Holy ****.  And people wondered why I stopped posting here.

1. Im building a single ended amplfier.  As such,

2. Im considering using either a gapped transformer in a DC coupled circuit or a standard push pull transformer in a parafeed circuit and wanted to know

specifically

if

3. anyone had any experience building both and could comment on the theoretical difference in bass response with all things being as equal as possible between the two very different circuits.

I figured when I posted this that the odds of finding someone with that experience were slim, but it was worth a shot anyway.  In the end I suspect Im going to wind up building both to reach any kind of authoritative conclusion.

 

[abbey road d enfer]

2. Im considering using either a gapped transformer in a DC coupled circuit or a standard push pull transformer in a parafeed circuit and wanted to know specifically if
3. anyone had any experience building both and could comment on the theoretical difference in bass response with all things being as equal as possible between the two very different circuits.

The latter, because of the additional cap, will exhibit a 2nd-order high-pass characteristic, the former just 1st-order, so all things being equal, the overall response of the whole amplifier will have one order more. The Q of this 2d-order filter may be tuned by selecting the coupling cap and the damping. Transient behaviour may benefit from proper tuning.

 

[emrr]

I know why you stopped posting; we talked about it.  I'm trying to get clarity on this conversation, since post #1 says 'PP', and it's been ignored until I got loud.  Maybe it's a typo.

lets assume, in the vacuum of my fantasy, you could get a pp and se transformers that were similar enough to compare (maybe you can from lundahl). 

PP =  two plates fed from either resistive B+, a center tapped choke, or a center tapped transformer, each pushing opposite polarity audio, or at minimum present to create current balance in a winding and cancel some degree of distortion. 

I feel like I'm reinventing the wheel here, between asking Dave for clarity, and addressing the other posters.  As I said, maybe PP is a typo.  Appears to be.  So....we're talking exclusively about the differences between direct coupled transformer B+ and parafeed in a single ended circuit.  Thank you; I was not clear on that; fixating on the first sentence of the first post. 

To the 1st question:

Why would you use a PP transformer in a SE circuit, with parafeed?  That I've never seen.  Why not use any typical SE transformer that's not intended for DC current, and use parafeed?  What am I missing?    I still suspect we are using differing confused terminology here.  As I mentioned earlier, I have seen significant response problems with some PP transformers when connected SE, due to orientation of winding capacitance I assume.  Nothing to do with parafeed. 

Or, are you going way way way back, and considering the sort of circuit that is SE, but has a dummy PP tube as the final, simply so DC can be balanced in the output transformer?  I don't think that's what you are talking about, since parafeed does away with the need to do such a thing. 

I think you will have to build, and make the sonic judgement call yourself.  I know how you feel about the minutia of sonic variation.  I have no doubt you could get both types to generate identical response charts.  You free up the transformer by going parafeed, but then you have to worry about the inductor's response.  The transformer is immediately capable of better bass response.  Maybe the idea of "all things being equal" is a bit impossible in the first place.  Conventional wisdom seems to dictate that parafeed simply saves you size and cost in the output, but it's certainly more complex than that.  A bit I've only seen RCA use in a few cases; an inductor feeding a SE transformer which is DC coupled to the plate.  And, also the same with AC coupling and resistive B+.    Have I experimented with all in a tight A/B manner?  No.  With either, you get the advantage of greater voltage swing at the plate compared to resistive coupling, which is good.  Every tube voltage amp RCA ever built was direct coupled through the output transformer.  They went to PP circuits in the BA-20 series to get size and cost down, and I feel that really is the primary reason.  There are no examples of resistive B+ with AC coupling on their output stages.  The portable OP-6 is a rare example of a parafeed SE output, and I am unclear as to the reasoning there, since the transformer can is the same size as the DC coupled stuff, and all the values and ratings work out about the same as the DC coupled types and circuits.  I need a loaner of that piece to diagnose further, and I intend someday to check out that output stage using their standard transformers for comparison. 

 

[soundguy]

its not a typo.  The lundahl 1620 is available with and without a gap, so you can use the same-ish part for single ended or push pull operation.  In an attempt to make a worthwhile comparison of the two circuit topologies I tried to find a part that would be close enough for some kind of logical comparison.  Now obviously, pushing a ton of DC through a transformer and completely bypassing DC from a transformer is probably not a comparison worth making, but Im curious, so there you go.  I figured this at least would be a better comparison than two completely different transformers in two different circuits...

 

[Marik]

Dave,

You mention 1620, so are you after a power amplifier? For some reason while going through this thread I thought you were talking about a mic pre...

I'd give a call to Kevin Carter asking about it. As a Lundahl distributor he knows the application of that transformer very well. I think he also played a lot with both SE DC coupled and parafeed choke loaded topologies. My guess, he would suggest choke loaded parafeed.

http://www.kandkaudio.com/

Best, M

 

[emrr]

Yep, I'm misreading your purpose, and not reading far enough into the data sheet.  Without the data sheet saying so, I'd never trust the idea of using PP iron for SE, but there ya go, they sell it.  

Historically one sees use of parafeed at all voltage stages, but virtually never at the output stage, at least with power amps.   What sort of power are you going for?  

I suppose you may have to compare available power with SE versus same power in PP, with PP obviously having a lot of headroom in that comparison.  If you're not going for significant power output, then the comparison becomes more worthwhile.  

 

[soundguy]

have spoken to kevin briefly.

I am building a power amp however there is no reason that I can see why any conclusions I come to about this couldnt be assumed for line stages as well.

not comparing se to pp.  comparing single ended to single ended...  one circuit is dc coupled with a gapped transformer and one is parafeed with an ungapped transformer.  cant see the puropse AT ALL of doing a parafeed circuit with a gapped transformer.  In my mind, the whole reason why some guy came up with a parafeed stage was to eliminate a gapped transformer in the first place.  Again, all Im trying to do is verify the sonic advantages of a parafeed output stage versus a dc coupled output stage both obviously in a single ended design, both designs as identical as possible within limitations and what the difference is.  In my mind this is about as straight forward as you can go in comparing topologies, Im a little short to understand the cluster **** of communication this thread represents.  Perhaps thats my fault.

Im not concerned with power, its single ended, 3 or 5 watts, whatever.  What I am concerned with is how the two circuit topologies sound different. 

I thought this would be pretty clear to understand and fairly straight forward, Im a little befuddled at either my lack of ability to explain this or whatever.  But then again, I cant say Im exactly surprised since there is virtually NO information that I have been able to turn up which can marginally satisfy my curiosity on this subject.  I cant accept for a second that Im pioneering anything special here but I really am shocked that I cant find a single person out there to chat about this outside my usual core group of awesome people.  Not even the hi-fi guys I know can speak with any authority about this.  Seems a little weird to me as if there were ever two topologies to make a direct comparison with, to me it would be these two, but whatever. 

 

[emrr]

I think historically the parafeed approach was about saving money and space.  Particularly when you look at possible power for same size iron with each method.  It is historically interesting to me to note that RCA built several different interstage parafeed units that included the choke in the same can; so the whole thing is bigger due to combo-in-a-can approach, so why not direct coupled instead?  Like all of their voltage amp output stages.  The only examples of a final parafeed power stage I can think of all date to the early 1930's, and I don't spend any time to speak of studying recent hi-fi trends.  The closest recent trend I can think of is people building hi-fi systems with choke coupled output and cap coupled series hi-Z speakers with no step-down ratio.  The 1930's circuits I suppose were tuned with some resonant boost; easier and cheaper than big iron, and pre-dating use of NFB.  Today?    I'm just rambling and posing questions, as I find it interesting too. 

What does the choke contribute to the sound?  Does it roughly equate to the DC through a similar output, and come out in the wash?  Does the output sound improve vastly with parafeed, and the choke contribute a much more invisible fingerprint?

If you instead built parafeed with a transformer that is already gapped for the current used, you might still gain some performance and sonic improvement.  But to what end I can't say.  I don't delve into power amps much.    And I agree; why?  You want the best iron you can use for the job in the first place. 

Sorry to get bent out of shape, but the mention of PP wasn't clear without direct reference to the 1620 data sheet, and a description of build intent always helps.  I wasn't clear at all that you were building an SE amp until I prodded more, and I wasn't sure exactly what we were comparing; several things seemed possible from the first post.