mic transformers

[timtimtim]

Another theory question. A transformer is often used twixt mic and preamp to give voltage gain. Why limit the transformer ratio to 4:1? Why not use say 32:1 giving 30dB lower noise? To avoid loading the mic too much use high value biassing resistors on the opamp such as 470k or 1M. 1M and a 32:1 traffo would give about 1k load on the mic. Why is this not done?

 

[Minion]

Well when you use transformers to create gain you also increase the Impedance going to the next stage and high impedance circuits are inherantly more noisy , and there is only so much gain you can get from a transformer (maybe 25db) before the signal gets loaded down....

Here is a Shure white paper on audio transformers which might help ...

http://www.kvglabs.com/PDF/xformer.pdf

I"m sure MSH will correct me as I rarely use audio transformers and my knowlege is limited in this area .......


Cheers

 

[timtimtim]

Right I can see that having a high impedance at the opamp input will generate more thermal noise, but surely there will still be an advantage because, no there wouldn't would there because doubling the ratio will double the voltage but quadruple the impedance so doubling the thermal noise. OK I understand that. But are there any other reasons why high ratio transformers are no good, to do with frequency response maybe? That link you posted said transformers can give up to 25dB gain and no more, but didn't say why.

 

[Minion]

With a Transformer like that it takes a Low impedance (low voltage high current ) signal and turns it into a High impedance (high Voltage low current) signal ..... after a certain turns ratio the current on the output of the transformer is so low that the signal gets loaded down in the next stage ....

You are basicly useing the transformer to convert low voltage high current into high voltage low current and eventualy you run out of current to drive the next stage .....

I"m sure MSH will correct me as I rarely use audio transformers and my knowlege is limited in this area .......

 

[mshilarious]

Thermal noise is a function of DCR of the windings (and source). You will step-up the thermal noise of the transformer primary + source and that gets added to the thermal noise of the secondary in parallel to its load, so you need an appropriate transformer if you expect a noise benefit. I'm not a transformer design expert by any means, so other practical considerations might escape me.

My guess is it's just not really necessary, and it gives you a larger amount of minimum gain, which is usually a hassle because it will require a pad for loud signals (or a transformer bypass switch).

With a 1:4 and a proper circuit you can hit the theoretical minimum noise without too much fuss.

See the note at the Jensen site:

http://www.jensen-transformers.com/mic_in.html

To determine the proper ratio transformer for your application, you will need to know the Voltage Noise and Current Noise of the active amplifier device to which the secondary of the transformer will be connected. If you divide the Voltage Noise by the Current Noise of the device you will get a value in Ohms. This value is the Optimum Source Impedance for best noise performance using that particular amplifier device. Select the transformer whose secondary impedance most closely matches the calculated value.

Example: An AD797 has an input Noise Voltage of 1.2nV per root Hertz Max. @ 1kHz and an input Current Noise of 2.0pA per root Hertz. This gives an Optimum Source Impedance of 600 Ohms. The proper choice of transformers for this device would be either the JT-16-A or the JT-16-B.

For INA217 that would be 1625, for 5534 it's 8750.

 

[Marik]

Let me start with a lame statement that any design is just a set of compromises. The transformer is even more so.

The most wide spread fallacy about the transformers is the idea that their voltage gain is free. Wrong!!! Here is a little write up on the noise considerations in transformers I posted on the groupdiy forum not long ago. That was about ribbon microphone transformers, but the same rules can be applied to any transformer. Since I wrote it myself, I don't have to ask anybody's permission just to paste it :

Some more on how winding DCRs affect noise. Any resistance is a noise and the problem with high step up transformers is the primary resistance gets multiplied by the turn ratio square.

Say, a typical Chinese ribbon transformer has Pri DCR=0.3 Ohm, Sec DCR=90 Ohm, and ratio 1:55. The primary translated into the output will be already 0.3 x 55^=907.5 Ohm. Add to this the secondary DCR and we get 997.5 Ohm of pure noise.
Another element of the noise increase is lower output due to the series losses (think of it as a resistive divider). You can see with 1.5K typical input impedance of preamp the losses can be quire considerable.

Now let's take a low DCR transformer with Pri DCR=0.002 Ohm, and Sec DCR=6 Ohm (for 1:55 ratio). 0.002 x 55^=6.05 Ohm+6=12.05 Ohm, which also adds very little to the series losses.

But we forgot that the ribbon itself also has resistance (noise). Typical DCR (or for practical purposes, impedance) of a large ribbon of 1.8 um is about 0.15 Ohm. Translated into output it becomes 0.15 x 55^=453.75 Ohm, which leaves us with total resistance in Chinese transformer 453.75+997.5=~1.45 KOhm, and with low DCR transformer 453.75+12.05=465.8 Ohm.

Another consideration is output impedance. In the modern equipment the rule of thumb is the input impedance of next stage should be 5 times highr the output impedance of preceding stage. Most of the preamps are 1.5K in, so the output stage should provide 300 Ohm.

Say, you have a tube microphone with a tube's plate impedance of say, 12K. We take a nice off shelf transformer of 6.5:1. Our output impedance will be 12,000:6.5^=284.02. Close enough.

Don't forget, if you are using say, 6.5:1 then it is actually STEP DOWN... meaning with 32:1 you will get very low output impedance (which we won't need), and also you will lose about 14dB of gain.

The next consideration is input impedance. For example if the source impedance is 0.2 Ohm (ribbon microphone) we don't need high inductance to get a good bass response. When we are going to 10K range, then even we will need much turns, which translates into high DCR (look at the first writeup about DCR). On top of that with special tricks (which greatly increase the price) we will need to fight high capacitance, leakage, etc.etc.etc.

It is an art in itself to make a good and highly optimized for a given application audio transformer!

Best, M

 

[timtimtim]

Thanks for all that info. I'll certainly have to think very carefully before using a transformer. Luckily I have no intention of ever using a dynamic or a ribbon mis so hopefully I won't have to. Incidentaly I have gone through a whole life of fiddling with electronics without being aware that resistance generates thermal noise! I knew that high impedance input stages were noisier than low impedance, but I always assumed that low resistor values shunted thermal carriers generated in the semiconductor junction down to ground. We live and learn.

 

[mshilarious]

It's not the input impedance that is the issue, unless you are measuring unterminated noise (which is a silly thing to do). It's the source impedance, which does shunt the thermal noise of the input impedance (so it's actually parallel resistance taking into account any series resistance). But once you are on the other side of the transistor whatever noise you brought to the gate is now part of the signal, along with the added semiconductor noise.

 

[timtimtim]

Yes, but I didn't realise resistors created noise at all, I thought only semiconductors generated noise, and the source impedance somehow shunted some of those carriers down to ground (how I thought it sucked them out of the silicon I don't know). Anyway I know better now thanks to you.