Electronics > Projects, Designs, and Technical Stuff
Small signal audio transformer design for a vacum tube microphone
(1/5) > >>
SK_Caterpilar_SK:
So I have been planning to build the GIC microhpone for quite a while (here is the link to the page about this microphone http://www.gyraf.dk/gy_pd/g7/gic.htm ) and I already have the capsule and so on... Where I got a bit stuck is the transformer. Sadly I am still just a student and not a milionaire so I cant afford really any transformer that would be quality certified for this job (and before you tell me "you already bought the expensive capsule just pay up the transformer" well no i bought the cheapest possible double sided capsule on ebay which was 30 bucks and thats already expensive for me).

(before you dive deep into this I just want to warn you I am not native and my grammar will probably not be on the spot. Its also 3:40 AM at the time of writing this)

Since buy isnt an option here I went with the "well what can go wrong just wind one" option.

I am not an expert on this so dont burn me please. So I pretty much have the impedances and turns ratio with that.. It would be 15K to 600ohm (and no DC current would ever pass trought the transformer itself). For me to figure out how many turns I need on a given core with a given permeability I need to know what kind of inductivity I need to pass audio. I figured since I already have bought some 1:1 600 ohm transformers from china I should be able to just measure them and go with that value. It turned out about 1.2H (which sound alright to me). I have done some checking arround and I also found other microphone transformers that have the same 600 ohm output impedance which are about 1.8H on average.
I assumed it would be the right inductance to go with.

I was checking arround on TME and I found a pretty decent ferrite core (AL 8700nH or rellative permability 2700) and with this one I would need arround 130 turns of 0.1mm wire to achieve arround 1.8H (I just used a online calculator for this and I know it does not take layer thickness into account and neither the bobbin itself). Since I know the ratio should be 25:1 then i should just multiply 130 by 25 and bam I got my secondary turns ratio which is 3250 turns. At this point I think to myself great math checks out and I have now a number to go with. However I just had a slight thought of how about the primary tube side inductance. I know it will be a whole lot larger but I did not expect it to turn out to be 1200H. I am almost certain I messed up somewhere in my maths and 1200H just doesnt right. Then again I never ran across such a special purpose transformer in my life and didnt measure it so I dont really know.

Best I could figure is to check other tube transformers and so I did check my power transformer for a EL34 push pull amp(I know its wildly not the same as what I am trying to create). It has a 6k6 ohm winding and it turned out to be arround 590H so I regained some comfidence. I dont want to go ahead yet on this because I am a bit worried about this very high inductance and I wouldn want to wind and unwind 3250 turns of wire (nope to happening). I just want confirmation on this if I am doing this right, if the 1200H sound right for you at all. I have never come across anything with that high inductance ever before let alone in a transformer.

All I want to know is if the inductance of the primary 15K ohm winding (which is 1200H) is it sounds right to YOU. I will appreachiate any advice and opinions about what could be done if its right or if not.


Technical parameters I look forward to achieve :
Primary impedance : 15K
Secondary impedance: 600ohms
Bandwidth 20Hz-20kHz
Power: just signal

My results: (RM core AL=8700nH/relative permability 2700, not gapped, code: B65815E N30, link to datasheet: https://www.tme.eu/Document/4a01e423b117af0ac949e52794f2a2d2/rm_12.pdf )
Primary:
             Impedance: 15K
             Turns: 3250
             Wire: 0.1mm
             Inductance (estimated without accounting for layers and bobbin): arround 1200H(!)

Secondary:
             Impedance: 600ohms
             Turns: 130
             Wire: 0.1mm
             Inductance (estimated): 1.5H-2H

To anyone who is wondering why I did not buy the transformer that costs 60 dollars. I am sorry I am a student and with limited finances I will gladly go down the route of more effort, besides its gonna be a great adventure to turn that wire round 3250 times... and educational aswell :).

(also if someone could link me a PDF that would contain all the required math for this I would be gratefull. I have only used online calculator tools which are good only so far.)

Adam.
Cerebus:
Before you go any further I suggest you go back and check your understanding of the basics. You've got a turns ratio of 25:1 and an impedance ration of 25:1 - that's a problem as the impedance ratios of transformers go as the square of the turns ratio. So your 25:1 turns ratio has an impedance ratio of 625:1. To get a 13k:600 (25:1) impedance ratio you need a turns ratio of 5:1.

Remember that a transformer scales voltage and current in opposite directions simultaneously. So say you have a 1:2 turns ratio transformer and put 1V at 1A into it (an impedance of 1 ohm) on the primary, on the secondary you'd take out 2V at 500mA (an impedance of 4 ohms).

And yes, 1200H sounds way out of the ball park for a mic transformer. I've no idea what would be a reasonable inductance, but I know enough to know that 1200H is a large transformer. You shouldn't be aiming for any particular inductance here, but for a particular magnetic flux level in your core. You have to know that you're not saturating it, equally you need to know that your signal is going to use a reasonable amount of the magnetic flux available to you in that core. You further need to aim to be working on a nice linear part of the core's BH curve, and you need a core with low magnetic hysteresis.

I don't want to dissuade you, but designing low level signal transformers is not the easiest of tasks and I'd want to go and re-read all my materials on transformer design before I tried it. By all means go ahead, but I'd suggest that you have a few days reading of textbooks in front of you if you want to produce a competent design.
trobbins:
Also suggest you visit groupdiy.com and search on transformer teardowns and design.
SK_Caterpilar_SK:

--- Quote from: Cerebus on April 15, 2020, 02:18:20 am ---Before you go any further I suggest you go back and check your understanding............design.

--- End quote ---

Well man...That was the problem. I forgot to realise that impedance is squared. Thanks for opening my eyes. I did build three tube amplifiers with microwave transformers and there I did have to correctly estimate and calculate (and thankfully no rewinding was needed :) ). I dont have an idea how could have I missed that.

In that case the primary 15K ohm side would end up being 650 turns and the inductance of that would be 46H. That makes much more sense. Core saturation wont be an issue in my case. Its such low power and on top of that just by the shear size you can tell its not gonna happen. I can still fit the transformer into the body of the mike but its larger than any comercial transformer. But to be hones this is just guessing here. If you could link me to some theory I would be glad. The text books they gave us in school are useless beyond trying to make a astable oscilator.

I will quite possibly post updates on actually physically making the transformer here.

Many thanks, Adam.
sam[PS]:
Have you seen the video series Dave made few years ago with former Rhode microphone head engineer ? If you get into microphone design that's a must see.
Note what he says about transformer :
-they have lot mojo so good for the sale (other than that not so good)
-they are not needed
-most of the time they are detrimental to the quality/signal integrity

If you are on a budjet i'll suggest you to not use transformer at all. Instead inspired by his work at Rhode just consider using a simple resistor as equivalent load for your tube circuit's output folowed by a BD139/BD140 pair as a output buffer and balance the output by matching the output impedance with a resistor between the ground and the S- terminal. Not only that works but most of the time this solution will be better than a transformer. It will cost you just couple $ and easy to buil on a small veroboard. I do recon it kinda kill the mojo, but when you'll get the money for a nice jensen or carnhill transformer it's easy to swapp.

Just my 2cts...
Navigation
Message Index
Next page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod