Mic Input Transformer Inductance

[Murdock]

I'm currently in the planing stages of a V241 clone.
On the hunt for a suitable input transformer I stumbled uppon some Sennheiser TB 432 "Broadband" mic input transformers with a ratio of 1:15. They  were quite cheap and the turns ratio is close to the originals of the V241 which was 1:18. So I bought them.
But I just found a paper from Sennheiser from 1971 where the TB 432 is pictured along some details.
The thing that puzzles me is the stated primary inductance of only 1,7H.
Isn't that way to low? I've read Ians paper "Audio Transformer Inductance" and the rule of thumb seems to be:

"For little loading at 20Hz, the inductance of a winding in Henries
should be approximately its specified winding impedance divided by 100."

So the specified primary winding impedance is 200 divided by 100 is 2H. Is that right? 

But than I found this post from abbey:

"The determining factor here is the inductance of the windings.
A microphone of nominal impedance 200 ohms should see a load not lower than 5x 200 = 1k
In order to qualify as an e.g. 200 ohms primary, the impedance at the lowest frequency of interest must be 1k.
Let's assume you're shooting for studio performance, you want 20Hz.
So you want L = 1000/2pi.20 = 8H"

So what's right now?
When looking at specs of other mic input transformers almost all of them araThe input transformer of the V72 preamp has an primary inductance of 25H...

I tested the transformer with an old RFT Signal generator with variable Rg of 20, 200 and 600 and an oscilloscope. I used Rg 200 and different input levels at a frequency of around 30hz. The max input voltage before the wave distorted on the output was around 1V RMS which is quite alot higher than most mics can output...

 

[Murdock]

Also when calculating the inductive reactance I calculated the following numbers for a frequency of 20Hz:

primary: 2*pi*20Hz*1,7H = 213 Ohm
secondary: 2*pi*20Hz*382,5H = 48.066 Ohm

Both are very close to the impedance rating. Source impedance of 200 Ohm into 1:15 ratio (1:225 impedance ratio) is 45.000 Ohm.

I don't know if that is good or not but I found it interesting.

 

[abbey road d enfer]

Also when calculating the inductive reactance I calculated the following numbers for a frequency of 20Hz:

primary: 2*pi*20Hz*1,7H = 213 Ohm

Which implies that the response at 20 Hz will be 3dB down; that is not a bad performance, since in many instances you will need a high-pass filter in order to prevent plosives. But frequency response is not everything. This is probably a source of LF distortion.
Please note that, in respect to your LF saturation measurements, visualizing with an oscilloscope will make THD of less than 5% go unnoticeable.

I would think these Sennheiser xfmrs are not much different than the Beyers (in fact I believe at one time they were made by the same company, that was later taken over by Neutrik). They are acceptable, but not excellent as a good Sowter or Jensen. Studer used them extensively, but always with a good dose of clever circuitry to alleviate their inherent weaknesses.
I don't know what you expect from them, but you can make pretty decent stuff with them, within limits.

 

[Murdock]

3dB down at 20Hz? The published frequency chart doesn't look like this... The source resistance was stated as 200 ohms and secondary load 500k/10pf. It's only 1dB down at 20Hz.
Here is the pdf. Starts at page 75.
http://lcweb2.loc.gov/master/mbrs/recording_preservation/manuals/Sennheiser%20Micro-Revue%2070-71.pdf

But you are right. An oscilloscope is probably not the best tool to look at distortion.
But also at lower levels the bass response is higher than the top response... Or something in my measurement procedure is awefully wrong.

This is not one of these tiny peanut miniature transformers like the Beyers.
It's a large and heavy transformer. Lenght is 52mm and width 40mm.
Was also quite expensive, about 185$ in todays currency...

Here is a picture:

 

[ruffrecords]

Transformers are complicated things. Usually the inductance varies with frequency and fortunately it tends to increase as the frequency lowers. Saying the primary inductance is 1.8H is quite meaningless if the frequency of measurement is not stated.

For minimum loading of a 200 ohm source you really want the primary impedance to be at least five times this i.e.1000 ohms so using my rule you would be looking for an inductance of 10 Henries at 20Hz. It is quite possible that a transformer that measures and inductance of 1.8H at 1KHz, for example, will have an inductance of 10H or more at 20Hz

For some reason, transformer manufacturers are reluctant to provide detailed information about inductance measurements but if they publish a curve using a 200 ohm source and it is flat down to 20Hz then it is a pretty safe bet the low frequency inductance is a lot higher than 1.8H at 20Hz.

Cheers

Ian

 

[abbey road d enfer]

Transformers are complicated things. Usually the inductance varies with frequency and fortunately it tends to increase as the frequency lowers. 

Thank you for reminding that. I totally forgot to caveat the OP about this.

 

[PRR]

> the stated primary inductance of only 1,7H.
Isn't that way to low?
...Ians paper ... at 20Hz ... specified primary winding impedance is 200 divided by 100 is 2H.

Ummmm, if Ian wants 2H and you got 1.7H, how is that "way too low"?

It's only 15% "low". OTOH it suggests 24Hz response. How many sources, and listeners, can handle 24Hz?

And 1:15 (1:30!!) is a BIG step-up. Design must be adjusted, compromised, to reach such a step-up. You are saying you want to overwhelm tube noise, maybe more than capture deep bass.

OK, -3dB @ 24Hz is -1dB @ 50Hz, a "real" frequency. I doubt there are many speakers (or recording rooms!) accurate to 1dB @ 50Hz.

Very few of your directional mikes will try to reach 20Hz. (When I did the half-big pipe organ I used the omnis.)

Anyway you got a bass knob somewhere. 98% of the time I have bass-cut; but where I know I have deep bass and it will not embarrass the final mix, I'll add a bump before the cut.

Or when you get the job to record the BIG pipe organ, use the solid-state mike preamps; they can be flat to a few Hz.

 

[CJ]

that transformer should be fine for most mics, 

i would worry more about input level than inductance,  in other words, it might handle a SM57 just fine, but what about a large capsule condenser mic ?

it is hard to have a lot of primary inductance and a high turns ratio at the same time and still have good frequency response, the specs on that xfmr say 25 Hz to 20 KC which is fine for a mic input,

if you add more turns to get more primary inductance, then you will have 15 times the turns on the secondary side,  so lets say you wind 400 turns on the primary to get 8 Henries,  now you need 15 times the turns (1:15 ratio) on the secondary side which means 15 x 400 = 6000 turns.  now we are going to have a tough time with hi freq roll off, that is why most transformers never go past about a 1:10 ratio,

so you have to drop the pri turns with a 1:15 to get good response, maybe 150 turns on some mu metal, so you have 150 x 15 = 2250 secondary turns which is reasonable.

there are some exceptions to the rule, like the V76 1:30 input, the way they does this is with a huge core and few turns,  but it is a very expensive transformer, the core is expensive, the winding structure is labor intensive, and the triple layer mu can is almost as expensive as the transformer,  so figure about 400 bucks.  If you could get them, but poor ol Ollie has moved on so the only option would be a used one with funky silk insulation around the wire which is gonna short out.

another problem with the 1:15 ratio is if you use a small core and have 1.7 Henries of inductance, this means low turns on a small core which means low input level.

so if you want the option to use all mics in all situations, (close mic on a Marshall cab) then you will need a transformer with more inductance  (bigger core and more wire)

Sennheiser makes some good stuff, give that transformer a try and see how it works.

If you need more bass, and higher input level, try a lower ratio Jensen, Sowter or Lundahl, 

I stuck a Jensen 1:8 input on a V76 instead of the 1:30 monster and it sounds just as good if not better than the 1:30.  Of course that circuit has a lot of gain to make up for the lower ratio.

sometimes a transformer will say "mic or line input,"  , a line input should probably have about 8 Henries to reproduce everything you throw at it, and the additional turns needed for the 8 Henries will also allow it to handle a  higher level as the more turns per core, the higher the flux level you can handle,

Core loss is also something to consider, a large 1:30 transformer that will handle a U47 might not be the best choice for a SM57. The reason is core loss. The bigger the core, the more core loss. Core loss is usually speced as some figure vs pounds.  so the more poundage, the higher the loss. so a small core might be the ticket for a delicate source like a dynamic mic. This is why a lot of Beyer mic inputs are pretty small, they are not going cheap, they are trying to save signal by matching the core to the mic.

 

[Murdock]

Wow, thanks alot guys! Alot of good information.

Transformers are complicated things. Usually the inductance varies with frequency and fortunately it tends to increase as the frequency lowers. Saying the primary inductance is 1.8H is quite meaningless if the frequency of measurement is not stated.

Ok, thats good to know. The 1.7H is probably taken at 1Khz then. 

> the stated primary inductance of only 1,7H.
Isn't that way to low?
...Ians paper ... at 20Hz ... specified primary winding impedance is 200 divided by 100 is 2H.

Ummmm, if Ian wants 2H and you got 1.7H, how is that "way too low"?

I was more puzzled by abbey's post which stated a minimum of 8H for 20hz in a 200 ohm primary...

I will definatelly try the transformer but I was just wondering about the stated 1.7H primary inductance and the associated frequency response which did not add up.
But as I now know that inductance is frequency depended it makes more sense. 

it is hard to have a lot of primary inductance and a high turns ratio at the same time and still have good frequency response, the specs on that xfmr say 25 Hz to 20 KC which is fine for a mic input,

if you add more turns to get more primary inductance, then you will have 15 times the turns on the secondary side,  so lets say you wind 400 turns on the primary to get 8 Henries,  now you need 15 times the turns (1:15 ratio) on the secondary side which means 15 x 400 = 6000 turns.  now we are going to have a tough time with hi freq roll off, that is why most transformers never go past about a 1:10 ratio,

The primary DC resistance is only 16 Ohms and the secondary about 1600 Ohms. Which seems quite low in comparison to some other higher ratio transformers I have here. The core probably isn't that small. The case is 42x50mm big. Quite bigger than most of the high ratio input transformers from Jensen, Cinemag and the like.

I will definatelly try it.

Thanks for your time and great information!

 

[abbey road d enfer]

I was more puzzled by abbey's post which stated a minimum of 8H for 20hz in a 200 ohm primary...

Which is not very different than Ian's calculation who rounds it at 10H...

 

[PRR]

> it is hard to have a .....high turns ratio ...and still have good frequency response....

Not so much the ratio as the high-side impedance.

On the 1:30 ratio, 200r on primary reflects over as 180K.

Any lump of conductor has capacitance. Something the size of a mike transformer, say 100pFd. 100pFd across a 180K circuit is a low-pass at 10KHz! Getting the C down to 50pF would make this 20KHz, but is hard. (Note the specs say only 10pFd added C from the tube grid; this tends to rule-out triodes.)

There will also be stray inductance, forming a resonance. With good design this may be 1/1000th of the bass inductance, so will also be around 20KHz. So a 2nd-order resonance with Q about 1, more or less.

But to extend the bass inductance will tend to increase the leakage inductance and lower the resonant frequency.

Generally if you think the Old Guys were wrong, you come to find they knew just what they were doing.

 

[abbey road d enfer]

Quote CJ> it is hard to have a .....high turns ratio ...and still have good frequency response....

Quote PRR Not so much the ratio as the high-side impedance.

As an illustration, check the specs for moving-coil cartridge xfmrs, such a s
http://www.jensen-transformers.com/wp-content/uploads/2014/10/jt-34k-dx.pdf
Ratio of 1:36, response up to 180kHz (that's if you managed to completely cancel capacitive loading... ), but actual secondary Z about 5k.

 

[PRR]

> to 180kHz (that's if you managed to completely cancel capacitive loading... ), but actual secondary Z about 5k.

Right; but the C is not real critical. Matchbook math says it can support 200pFd. They actually want you to add 100pFd (and some R), while driving a MM phono input which may have 100pFd in it. However the transformer seems to have ~200pFd in it. And the Q is low. Too much to figure just now. But does show that lower-Z is easier.

 

[Murdock]

Little update.
I bought a little LCR meter ( XJW01 AUTO LCR) and made some measurements.
At 100Hz the transformer has an inductance of about 3H. But 100Hz is the lowest frequency it can measure at...
I found a datasheet from the Sennheiser TB 421 series, which is, I think, a predecessor of the TB 432.
The inductance values are the same and it also states the frequency (50Hz) and amplitude (25mV) at which the 1,7H was measured. It says it is equal or more than 1,7H at the stated frequency and amplitude.

But I wanted to find a way to measure the inductance at lower frequencies. So I found a post from CJ where he explains a way where you have to divide the signal voltage by the signal amperage to get the input Z and then dividing that by 2*p*F.
The thing is, the inductance varies with signal amplitude... at around 500mV I get the highest inductance. At 1V it's a bit less but at 25mV it's alot less (about 1,5H less).

I read that transformers with Iron core behave that way but I know that the core is made out of Mumetal.
I then also traced the signal with a scope and found that the frequency response from 31,5Hz to 15Khz is the flattest when applying about 500mV.
At 25mV I get a big bump at the high frequencies...
Is that normal? Or am I missing something here?

 

[abbey road d enfer]

So I found a post from CJ where he explains a way where you have to divide the signal voltage by the signal amperage to get the input Z and then dividing that by 2*p*F.
The thing is, the inductance varies with signal amplitude... at around 500mV I get the highest inductance. At 1V it's a bit less but at 25mV it's alot less (about 1,5H less).

Please describe your measurement set-up.

I read that transformers with Iron core behave that way but I know that the core is made out of Mumetal.

They all behave that way. Only coreless inductors are perfectly linear; ferrite cores are less prone to this effect, though.

At 25mV I get a big bump at the high frequencies...

I would suspect you are overloading the meter with the higher level, which tends to linearize all measurements.

 

[Murdock]

Please describe your measurement set-up.
They all behave that way. Only coreless inductors are perfectly linear; ferrite cores are less prone to this effect, though.
I would suspect you are overloading the meter with the higher level, which tends to linearize all measurements.

I have an old RFZ signal generator with 31,5Hz being the lowest and 15Khz being the highest frequency it can generate. I can switch between 20, 200 and 600 Ohms impedance and vary the level where as 3V rms seems to be the highest level.
I connect it to the primary of the transformer and measure the voltage and amperage at the primary with a cheap digital Multimeter.  And then I calculate according to CJ's post (reply #16) in this thread:
https://groupdiy.com/index.php?topic=61436.0

Formular: L=R/2pi.F where R=V/I

Is that the right way to calculate inductance?

At 31,5Hz and 0,5V it gives me an inductance of 5,13H.
With 1V I get 4,72H
The same frequency at 0,025V (25mV) gives me an inductance of 3,71H.
As this is a mic input transformer the 25mV reading is probably alot more valid and meaningful then the 0,5V, right?

EDIT: how would I calculate the -3dB cut off point with these values?

 

[abbey road d enfer]

Formular: L=R/2pi.F where R=V/I

Is that the right way to calculate inductance?

That is certainly correct, but I would suspect the accuracy of measurements done at 15kHz. Most mavometers have an optimal frequency range of about 1kHz.

At 31,5Hz and 0,5V it gives me an inductance of 5,13H.
With 1V I get 4,72H
The same frequency at 0,025V (25mV) gives me an inductance of 3,71H.

Which means the current you measured was  .034mA; I very much doubt the accuracy of your meter there, particularly in AC mode. Anyway these are not unlikely results; it is known that magnetic permittivity decreases at very low level (Barkhausen effect) and ay high level (saturation).

 

[PRR]

> I read that transformers with Iron core behave that way but I know that the core is made out of Mumetal.

Iron -or- Nickel.

Mu-metal is a Nickel-Iron alloy. No surprise.

Look at the chart at bottom of this page.  http://www.mu-metal.com/technical-data.html 

Turn head 45 degrees.

Mu-metal runs permeability over 100,000 for flux from 100 Gauss to like 6.000 Gauss, lower outside this range. Or from ~~10 Gauss to ~~3000 Gauss we will get about 1:10 swing of inductance.

All normal for "iron core". Paying for Mu buys you more inductance in a given size (reflects back to capacitance and shielding concerns) but still wonky iron. The usual goal is to make your *minimum* inductance over the working range, and design so more inductance at some levels does not do any harm.

 

[Murdock]

Got another question regarding input transformers.
I know that the winding capacitance and leakage inductance affect high frequency response and also relate to each other. Less winding capacitance means more leakage inductance and vice versa.
But is there something like a preference of what should be higher/lower than the other?
I have two mic input transformer with high step up ratio.
One has a primary leakage inductance of  around 260 μH but winding capacity of a few nF. And the other one has a leakage inductance of about 2 mH but a winding capacity in the range of a few pF.

The one with the lower leakage inductance is flat from 30Hz to 15KHz, secondary unloaded.
The other one with the higher leakage inductance has a rise at high frequency, secondary also unloaded. Probably needs a zobel network to be flat.
Is this related to the leakage inductance and winding capacity ratio?

 

[ruffrecords]

Got another question regarding input transformers.
I know that the winding capacitance and leakage inductance affect high frequency response and also relate to each other. Less winding capacitance means more leakage inductance and vice versa.

As with all things transformer related it is not quite as simple as that. Leakage inductance is a measure of the degree of coupling between primary and secondary. If the coupling is 100% the leakage inductance is zero. One way to get very good coupling is to bifilar wind the primary and secondary. Assume for the moment a 1:1 transformer so primary and secondary have the same number of turns. If you wind both windings at the same time and keep the wires right next to each other, the coupling between the winding will be near perfect and the leakage inductance very low. However, because the the two winding wires are very close to each other, the capacitance between windings is quite high. So in this case you rule of low leakage tends to mean high capacitance does hold true.

if instead to wind the primary on the upper half of the bobbin and the secondary on the bottom half, the capacitance between winding is much lower but because the windings are physically separated the coupling is not so good. So inter winding capacitance goes down and leakage inductance goes up. So again the rule holds

However, wind this transformer the same way on a toroid and the inter winding capacitance will remain the same but the leakage inductance will drop dramatically because nearly all the flux is contained within the core. In this case the rule no longer holds.

And so far all we have talked about is the capacitance between windings, we have not considered the capacitance across each winding. Which is pretty much the same in all the above cases.

There are all sorts of winding techniques aimed at reducing one or the other parameter or in somehow optimising them both for an application - input transformers have different design goals to output transformers.

Bill Whitlock's paper on audio transformers is a must read:

http://www.jensen-transformers.com/wp-content/uploads/2014/09/Audio-Transformers-Chapter.pdf

Cheers

Ian