DCR and Impedance - modern audio transformers

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
It's economics, surely to large degree. 

Note I didn't say there was a flaw in functionality, on the face of it.  There's a difference in method, the majority of modern offerings are not 'same as'.  It leads in to the same morass as any discussion about the realism of clones. 

Some data I did write down, not a lot directly apples to apples, but enough to make general trend observations and supported by all the other measurements I've taken but not written down. All of these are averages of multiple units, so not unconfirmed single outliers.

I've noted various things at 120Hz, L, sometimes ACΩ, sometimes DCR, quoted Z.  In general, higher L and DCR equates to lighter ACΩ loading.    None of it says anything definite about bandwidth or frequency linearity. 

One example, it's extremely rare to see a modern mic/line to grid transformer with a secondary DCR over 3K, in old stuff it's normal to see 4K-7K. 

some outputs of various types

Sowter 8344E - Fairchild signal output 60K:600
32h, 90k ACΩ, 510 DCR - you would expect 5-10x these numbers in an older transformer
1040mh, 320 ACΩ, 5R5 DCR
- not the H is still a fraction of the lowest below, the lowly Peerless 15095. 

UTC LS-50 (SE)
15K 645h 328k ACΩ, : 500 20.3h 13k9 ACΩ

UTC LS-51 (PP)
30K 2487h 856k ACΩ : 500 34.57h 13k9 ACΩ

Gates SA-70
15k 477h 186k ACΩ : 500 16.38h 11k ACΩ

Collins 6Q-1
15k 287h 156kACΩ : 500 13.58h 9k9 ACΩ

Peerless 15095
179h 55k ACΩ : 9.5h 4k75 ACΩ

Various RCA single ended DC coupled outputs 30-50K : 250-500
240-250H : 3-5H5


Some interstages note the Hammond 850Q test, source and load Z of 40K with full bandwidth

Sowter 9830 BA-6 interstage DCR 1680Ω, recall approx 200H  +36 rating!  Why?!?!

Hammond 835  10K/40K:” +/- 0.5db 50-15K +15dbm, 0mA unbal DC, 980:982H

Hammond 850Q  10K/40K:” +/- 0.5db 20-20K +15dbm, 0mA unbal DC, 1400-1685:1400-1685H (40K:40K) 480K AC Ω 5K DCR
looks the same connected either side to ground, 40K source and load yields
+/-1dB 10.9-29K4Hz rising 34K9 peak +1.5dB
+/-0.5dB 16.8-26K5Hz

RCA 86-A limiter interstage 5K2ct:, 2029Hct:1023H

UTC A-18 grey badged era
1070 : 2948 DCR
246H : 1301H
321K :  greater than 1M AC impedance at 120Hz

RCA 76 console program amp interstage
258H : 1605H
875 DCR : 3K08 DCR


a couple 600:600

Edcor WSM600/600
10H3 3k6 ACΩ

WE 111C
51H 19K ACΩ

Hammond 850G
12H9 10K ACΩ

UTC A-20
23H  14K AC Ω
 
CJ has done dissections that show a variety of techniques used in winding

The methods to interleaving and different ways of controlling distributed capacitance is really interesting

I imagine it would not be economically feasible to replicate some of those techniques today
 
shabtek said:
CJ has done dissections that show a variety of techniques used in winding

The methods to interleaving and different ways of controlling distributed capacitance is really interesting

I imagine it would not be economically feasible to replicate some of those techniques today
CNC winding machines can do it economically.
 
@EmRR

I largely agree with your list. I have found Sowter output transformers to be low on inductance and DCR. Their input transformers are generally fine in this respect.

Regarding the 600:600 types, 10H inductance of the Edcor seems to me to be quite sufficient inductance at 120Hz representing a 7K5 load on a 600 ohm source. The WE 111C at 51H that you compare it with seems excessive to me. Assuming a similar core material and size, the additional inductance requires more turns of thinner wire and hence the increase in dcr. However, leakage inductance and capacitance both increase  with more turns but maybe that did not matter back in the day when the top end limit was 15KHz. So although I agree there are differences between vintage an modern version, I can see no good reason for the higher values of dcr and inductance they provide.

Certainly for mic input transformers, dcr directly affects noise performance so should be kept as low as possible so it is better to select a core with a high Al so fewer turns are required to reach the required inductance.

Cheers

Ian


 
As Doug and Ian said, the Sowter replacement for the Fairchild output does indeed have much lower inductance (and DCR) than vintage.

While this may not be much of an issue in a perfect balanced circuit, in real life there is invariably some imbalance in DC currents through the primary. 

There are a couple of Cinemag output transformers I've used that are better in that regard and also, while Tom Reichenbach was around, I was also able to specify a small gap at times to allow for some imbalance.  I don't know if David is still open to these slight customizations, maybe so?

 
On the WE 111C, it's a roughly 3lb toroidal type.  Not unlike a clay brick.  CJ's taken one apart and posted about it.  The one's I've looked at are -0.8dB at 1Hz and 40kHz when loaded with 600Ω.  10K load, the bottom looks the same and the top down -0.1 at 44kHz. 

The 1980's Jensen (they may still make it) repeat coil that matches and fits the footprint of a UTC A-20 is 45H and looks like 31KΩ at 120Hz, unloaded.

ruffrecords said:
For input transformers, excessive dcr is bad for noise

I don't see this as a practical problem in any tube preamp I've ever heard.  Tube and source noise always swamps anything from a transformer in my experience.  You can take any high DCR secondary into a modern active unity gain DI, turn that up with a modern transformerless preamp, and it'll be more quiet than the noise floor of any tube I've ever heard. 

All the American classics are at 3K DCR or above, and it doesn't stop use of ribbon mics either. 
Current Hammond 850N (40K) is 5K DCR.

Then something about some vintage outputs I've brought up repeatedly, not clear on the purpose.  Numerous RCA, Langevin, and Western Electric 150-600Ω output windings with DCR as high as 80%.  An output isolation theory?  Semi-self-loading?  What? 
 
Winston O'Boogie said:
There are a couple of Cinemag output transformers I've used that are better in that regard and also, while Tom Reichenbach was around, I was also able to specify a small gap at times to allow for some imbalance.  I don't know if David is still open to these slight customizations, maybe so?

David will make them any way you want them, in my experience.

As for Cinemag and vintage transformer repros, they still make the UA transformers from the same blueprints they were originally specc’ed from. And with their S-217-D repro, they went to great lengths to stay true to the original, down to custom building their own tooling (very costly) for making the original L-12 lams, the machinery for which was scrapped long ago. Same alloy as the original. I have one of those which I’ll measure up to compare to the original numbers that folks have posted here over the years. I’d be surprised if it weren’t the same.

Mainly making the point that Cinemag is one you could group with Hammond and Crimson who still offer some vintage style iron made like vintage iron, rather than approximating performance using modern techniques (or not even fully approximating performance).
 
EmRR said:
Then something about some vintage outputs I've brought up repeatedly, not clear on the purpose.  Numerous RCA, Langevin, and Western Electric 150-600Ω output windings with DCR as high as 80%.  An output isolation theory?  Semi-self-loading?  What?

Doug, I don’t know about the theory with those 150:600 inputs, but I thought PRR made a convincing case in that M-1433 thread that I linked on page 1 that high DCR on outputs is more forgiving of impedance mismatches downstream, thus answering this same question you’ve posed many times regarding the high DCR of the older stuff. I’d think the same would apply to your question. See the linked thread for more.
 
Thank's for the replies - seems to be rather complex - but maybe there is just a lack of knowlegde on my side. I'll try to explain one of my experiences with "different behaviour" at lower DCR. During calibration of a U73b clone I fed a Sowter T188 copy with original input pad with a 600R generator output and a 600T-Pad attenuator. It's a 0.5+0.5:1CT. The Sowter has a (measured) 200R DCR - so 100R on 0.5 Input. The second "primary" goes to the sidechain. The T188 originals range around 600 DCR. The input pad has 2 non inductive serial 10k resistors and one 4k across the 0.5 input. Even with the 600T attenuator fully open and at several dB levels resulted an intolerable loss on the signal (on various freq.) between output of generator and output of the T188. Calibration impossible. No other issues on the secondary side. With a double L-Pad attenuator of 10k each instead of the 600T the loss was gone. So about the same procedure as protecting a GZ34 from low DCR. I wonder whether this type of remedy would affect any parameter in theory. Else it's rather easy to adopt original DCR by resistor pads. Since the signal was choked before - does low DCR have an influence on impedance matching apart from impedance ratings ?? I have to add, that I believe to have no audible disadvantage and frequency response measures quite flat over the full unit - but have not been able to compare with the original T188 behaviour yet.         
 
I don't think your problems have anything to do with the Sowter input transformer. The two 10K series resistors in combination with the parallel 4K and 5K resistors make a 20dB attenuator so the signal level will be low at the transformer input.

Cheers

Ian
 
Sc said:
Thank's for the replies - seems to be rather complex
No, it's not. Just Ohm's law.

I fed a Sowter T188 copy with original input pad with a 600R generator output and a 600T-Pad attenuator. It's a 0.5+0.5:1CT. The Sowter has a (measured) 200R DCR - so 100R on 0.5 Input.
Actual Z is much higher than DCR.

The input pad has 2 non inductive serial 10k resistors and one 4k across the 0.5 input. Even with the 600T attenuator fully open and at several dB levels resulted an intolerable loss on the signal (on various freq.) between output of generator and output of the T188.
As Ian said, too much attenuation there.

With a double L-Pad attenuator of 10k each instead of the 600T the loss was gone.
Your description is, to say the least, ambiguous. I read it as if there were two attenuators, one loading the other. If that's the case, no wonder it does not do.

So about the same procedure as protecting a GZ34 from low DCR.
I don't know how these two subjects are related, except that xfmrs are involved.

I wonder whether this type of remedy would affect any parameter in theory. Else it's rather easy to adopt original DCR by resistor pads.
Do you mean adding resistors? That can only attenuate. Resistors resist current.

Since the signal was choked before - does low DCR have an influence on impedance matching apart from impedance ratings ??
Strange question. Perfect impedance matching implies knowing and controlling teh DCR of all windings, so yes, it has an influence, but it is generally relatively easy to master.

I have to add, that I believe to have no audible disadvantage and frequency response measures quite flat over the full unit - but have not been able to compare with the original T188 behaviour yet.       
In audio, impedance matching has almost zero effect on frequency response. Actually, impedance matching has been abandoned long ago, to the profit of bridging.
 
Following the comment on matching versus bridging:

Back in the early days many companies offered input transformers in two forms, open secondary or loaded secondary.  One type for input stage volume controls and/or fixed input load Z, another for source defined input Z and/or downstream volume control. 

The loaded secondary types I have are flat with matched loading, and have rather severe resonant treble peaks as soon as you move away from bridging.  In one I recall +3dB with a 50% increase in resistor loading, and something north of +12dB when open. 

The unloaded types that correspond start losing treble with less than a bridging secondary load, meant to feed a grid with no resistive load. 
 
I coincidentally just stumbled upon what may be an example of the phenomenon Doug was describing:

Sowter 1290 listed as replacement for Triad HS-50.

It lists a primary inductance of 97H and a primary DCR of 176 ohms.

On CJ's analysis of an actual HS-50, he measured a primary inductance of 525H at 20 Hz, with a primary DCR of 1,238 ohms.

Quite a difference!
 
soapfoot said:
I coincidentally just stumbled upon what may be an example of the phenomenon Doug was describing:

Sowter 1290 listed as replacement for Triad HS-50.

It lists a primary inductance of 97H and a primary DCR of 176 ohms.

On CJ's analysis of an actual HS-50, he measured a primary inductance of 525H at 20 Hz, with a primary DCR of 1,238 ohms.

Quite a difference!
It's quite surprizing from Sowter.
I guess they have designed this xfmr from analyzing teh circuit, without actually evaluating a genuine one.
 
abbey road d enfer said:
It's quite surprizing from Sowter.
I guess they have designed this xfmr from analyzing teh circuit, without actually evaluating a genuine one.

Exactly my suspicion, looking at the specs in their 'vintage replacement' offerings.  Pretty much all of them are 'different than'.  All of the correct info is reasonably available, always has been.
 
soapfoot said:
I coincidentally just stumbled upon what may be an example of the phenomenon Doug was describing:

Sowter 1290 listed as replacement for Triad HS-50.

It lists a primary inductance of 97H and a primary DCR of 176 ohms.

On CJ's analysis of an actual HS-50, he measured a primary inductance of 525H at 20 Hz, with a primary DCR of 1,238 ohms.

Quite a difference!
No difference at all because the Sowter reading was most likely made at 100Hz or 120Hz which is the frequency used by common inductance measurement devices. It is very well known that inductance rises significantly at low frequencies and can easily be several times the 120Hz value.

The problem here is really that all manufacturers, both modern and vintage, fail to properly specify their transformers.

Cheers

Ian
 
ruffrecords said:
No difference at all because the Sowter reading was most likely made at 100Hz or 120Hz which is the frequency used by common inductance measurement devices. It is very well known that inductance rises significantly at low frequencies and can easily be several times the 120Hz value.

The problem here is really that all manufacturers, both modern and vintage, fail to properly specify their transformers.

Cheers

Ian

But Ian....the DCR!!!!  That's plenty a clue!

Sowter 1290 listed as replacement for Triad HS-50.

It lists a primary DCR of 176 ohms.

On CJ's analysis of an actual HS-50, he measured a primary DCR of 1,238 ohms.
 
EmRR said:
But Ian....the DCR!!!!  That's plenty a clue!

Sowter 1290 listed as replacement for Triad HS-50.

It lists a primary DCR of 176 ohms.

On CJ's analysis of an actual HS-50, he measured a primary DCR of 1,238 ohms.
I don't know why everyone is so hung up abut dcr. Lower dcr is generally a good thing and leads to better performance. However, a near 10:1 difference is very hard to explain and even harder to believe. I suspect CJ's measurement to be in error. In fact if you look at what the transformer is supposed to do I think it must be.

This is a 5:1 step down designed to be driven from a source impedance of 2K ohms. Now the primary dcr appears in series with the input so you want it to be significantly lower than the source impedance. A few hundred ohms would be OK but over 1K is just going to lead to unnecessary insertion loss. Bottom line is the Sowter is an improvement on the originals not a clone (and they do not claim it is a clone).

Cheers

Ian
 
ruffrecords said:
I don't know why everyone is so hung up abut dcr. Lower dcr is generally a good thing and leads to better performance. However, a near 10:1 difference is very hard to explain and even harder to believe. I suspect CJ's measurement to be in error. In fact if you look at what the transformer is supposed to do I think it must be.

This is about recognizing the differences, not what is 'better'. 

CJ's measurement is exactly what I would expect to see in a transformer of that type, and I would think the Sowter faulty at a glance.  Triad follows the general 10% of impedance rule as do most, the Sowter is nowhere near it.  I've got hundreds of old transformers here to back this data, though it's not my job to do so, it's easy enough for anyone curious to have a look at, with relatively few examples on hand. 

Look at the Hammond DCR specs in the 850 series.  They represent much more closely what older transformers ALL look like.  Look at the Telefunken V series transformer specs for corroboration, generally even higher in resistance, with many multi-chamber bobbins. 

Whether related or not, higher DCR appears to go hand in hand with higher inductance, and lighter loading at the reference AC frequencies.  That lighter loading is easily the WHY of ribbon mics sounding one way with vintage transformers, another with modern transformers. 

Hell, even the current Jensen JT-110K mic input (1:81) is 3K DCR for a measured secondary Z of 13K.    Their 1:10 is 2K5 DCR.  Their 10K primary tube line output is 740 DCR.  Those #s are closer to old than I expected to see. 
 
I am confused. On the one hand we were discussing a Sowter output transformer and next you are talking about the differences in input transformers and the effect they have on the sound of ribbon mics.

Let us be clear. I am not disputing there are differences between modern and vintage output transformers. I agree dcr and inductance are related but the relationship is complicated by the AL of the core and the thickness of the wire used but other than that there is no advantage per se in high dcr.

When it comes to input transformers you say yourself you are surprised how close some modern ones are to vintage types. This is my experience also. For example the Sowter 1:10 transformer I use in my tube mic pres has a 495H secondary inductance measured at 100Hz.

Of course there are plenty of crap transformer manufacturers out there but Sowter, Cinemag, Jensen and Carnhill are exceptions.

Cheers

Ian
 
Back
Top