Single ended series vs. parallel feed

[CJ]

use the resistor. it sounds fine.


cap sounds fine too, just use a Big one.
cj

 

[emrr]

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?

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.

My only observation there is that they are being specific about the usage.  It strikes me that winding type, power, and impedances would not really be used for any other purpose.  I'm not sure you would find that set of parameters, and it not be intended for parafeed use.  The mix of core materials is non-standard, and enhance the product.  Cost is negligible compared to the relevant DC coupled SE type in Sowter's line.  Low performance is vastly better, but that's a given.  It's a hard apple to find a comparative orange. 

 

[tardishead]

It seems as if the advocates of anode line transformers actually enjoy the spray of complex harmonics that they create

We are looking for differences between anode line transformers and parallel feed - somebody already mentioned the creation of odd harmonics in DC coupled transformer outputs - can anyone else shed any light on this. There is nothing really on the web.
Presumably PF topology would not be compromised and be theoretically better freq response. What about the bottom end in a PF stage - does the capacitor always somehow compromise the low end??

 

[lassoharp]

It seems as if the advocates of anode line transformers actually enjoy the spray of complex harmonics that they create

We are looking for differences between anode line transformers and parallel feed - somebody already mentioned the creation of odd harmonics in DC coupled transformer outputs - can anyone else shed any light on this. There is nothing really on the web.
Presumably PF topology would not be compromised and be theoretically better freq response. What about the bottom end in a PF stage - does the capacitor always somehow compromise the low end??

I'm not 100% on the impedance(choke) loaded plate case, but during parallel feed with standard resistive plate loading the cap will be resonating with the transformer primary.  I don't think the cap itself is degrading or limiting the low end(outside of normal roll off behavior) if that's what you mean by compromise. If you're referring to the resonance bump as part of the compromise then just up the cap value vs pri inductance until you're too low to notice. The transformer is probably going to be where the compromise is in either case(DC direct or PF). In the case of a PF voltage amp, use a choke to get a bigger voltage swing off the plate and omit the transformer and get max bass.

 

[PRR]

> during parallel feed with standard resistive plate loading the cap will be resonating with the transformer primary.

Same for transformer, with or without choke.

The problem with resistive feed is that it is 6% efficient at best. Choke/trans feed can do 50%. Real cases don't reach 50% or 6%, but there's still ~~8X difference which means crummy performance.

 

[abbey road d enfer]

somebody already mentioned the creation of odd harmonics in DC coupled transformer outputs - can anyone else shed any light on this. There is nothing really on the web.

Operating in one quadrant of the B-H curve does not provide natural compensation of 2nd-order harmonics. Same principle as SE vs. push_pull. Symmetry provides natural 2nd-order cancellation.

What about the bottom end in a PF stage - does the capacitor always somehow compromise the low end??

As I said earlier, it may improve the LF response. By careful tuning of the resonance, one can achieve a better response than the usual overdamped high-pass.

 

[analag]

This is why when building power amps I wind my own. I would go DC single end on C core and play with the gap by simply tightening clamps. That's just me, though.

 

[lassoharp]

> during parallel feed with standard resistive plate loading the cap will be resonating with the transformer primary.

Same for transformer, with or without choke.

The problem with resistive feed is that it is 6% efficient at best. Choke/trans feed can do 50%. Real cases don't reach 50% or 6%, but there's still ~~8X difference which means crummy performance.

Taking this one step further, in the general case of choke loading, what is the proper way to treat that entire network in terms of figuring out phase shift, filtering, and resonance?  I can see the output cap in series with the transformer primary as one calculation, but with the choke you now have a coil and the power supply decoupling cap as another series resonance.  So, I see two LC branches in parallel and I can't seem to find or remember the rules/method for combining parallel resonances - is the net resonance a simple matter of combining the two branches?

 

[PRR]

No.

 

[CJ]

the choke takes the pwr supply cap out of the equation, hopefully, you don't want to hear the cap,

so one end of the choke is stable, the other end is movin and groovin,

 

[lassoharp]

the choke takes the pwr supply cap out of the equation, hopefully, you don't want to hear the cap,

so one end of the choke is stable, the other end is movin and groovin,

Right.  Make sure the PS cap value w/choke inductance keeps resonance a non factor.  What I wasn't sure about was the output cap sandwiched in between two inductances - the OT and the choke.

So lets say for sake of example the choke is 100H, the OC is 2uF, and the OT pri is 100H - are you going to get 2 humps at 11Hz that add together?  I'm assuming this is why it's suggested to stagger the inductance values between choke and OT pri.

 

[CJ]

i am looking for a lost link that had all this explained with cool  color graphs and charts filled with

resonance figures for different cap/Henry combinations, but it seems to have been overwritten on my last WIN98 install.

i think i have the most versions of WIN98 ever installed on one computer,

the sectors look like a jig saw puzzle!   

meanwhile,

searching...


ok, this is not it, but this article looks better than the one i was searching!>


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

 

[lassoharp]

Good article.  Thanks CJ.

 

[abbey road d enfer]

Right.  Make sure the PS cap value w/choke inductance keeps resonance a non factor.  What I wasn't sure about was the output cap sandwiched in between two inductances - the OT and the choke.
So lets say for sake of example the choke is 100H, the OC is 2uF, and the OT pri is 100H - are you going to get 2 humps at 11Hz that add together?  I'm assuming this is why it's suggested to stagger the inductance values between choke and OT pri.

Things are not that simple. As I mentioned earlier, the circuit has to be tuned. The attached file shows the effect of varying the value of the coupling cap, all the rest being the same. 50k simulates bottle output Z. Signal is measured at primary of OT, secondary not shown but 50k res simulates reflected load (600 ohms with 9:1 xfmr). Blue curve is with 2uF, green with 100n. The results are surprising, aren't they? The smaller cap give extended LF response over the larger...
On the left, the 2uF cap gives an overdamped transfer function, and the 100nF gives a kind-of-Chebyshev curve. One would be tempted to use the one that has the most extension, but the steep roll-off is considered by some as unacceptable. There's also headroom issues there.

 

[lassoharp]

Nice. Thanks for sharing.

"Surprising" - On the face of it - yes.  First guess would probably have gone to the .1 for an earlier roll-off.

 

[CJ]

from a practical standpoint, when you sit at the bench to tinker a parafeed, the first thing you notice

is there are not that many  parts that you have to mess around with,

usually the limitations such as voltage ratings, stray C, & size determine the choice of parts in a parafeed,

how big a cap can you get that will be non polarized and handle 200 VDC?

20 uf would probably require a separate can like you see on the bottom of a Fender amp,

and the output transformer pri  inductance is never going to be so huge that you need

2,000 H. in a choke to get good response,  and how much high end would a 2,000 H roll off?

you would have to have a 5 or 6 chamber bobbin, to get good high end,

so we are somewhat stuck with values of probably 8 uf max and 800 H. max,

so kind of a bummer, you read the article, then sit down to try out all the wonderful cominations,

and you find out that there really are not that many. copy the old Telefunken and you are good.

so you can grab a 1,2 4, 8 uf, and grab so chokes, and see what matches up to your output the best.

one thing you don't have to worry about is a speaker on the end, that really messes things up.