Ribbon Mic Transformer size question

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

brukker

Member
Joined
Dec 26, 2009
Messages
7
Hi everyone,

I've been thinking of building my own ribbon mic, and winding my own transformer, and thanks to the tonnes of stuff on here and lurking for a good few months, I'm pretty much set to begin sourcing the parts, which I'm trying to do locally.

I have a question,... most mic trannies are rather smallish, and I'm wondering if there would be any major advantage/disadvantage to using a larger core?

i can source a c-core locally, but it is considerably larger than your average mic tranny so I thought i would ask as I am contemplating rewinding it to match my ribbon-motor-to-be. I'm only really building this for myself so the weight and size of the final mic doesnt really concern me so long as the mic sounds good.

How would using a larger than average c-core affect the output of a ribbon mic?

I'm guessing shielding would be more of an issue right?

Also I'll be experimenting with both ciggie pack foil and some aluminium leaf I've managed to find, incase that makes any difference other than re-calculating for each.

Cheers,

brukker.
 
A large core means more copper per winding, which means higher resistance. And higher dc resistance means higher noise.

Since levels are relatively low usually, you can use a small core without having to worry about distortion too much.

That said, I'd recommend building with what you have and try to improve on that in the next version.
 
every  core needs a little amount of power to run it,

the bigger the core, the more power. this power comes from your mic signal.

this is why all your good mic inputs are rather small, look at the Beyer stuff,

you want the core as small as possible but with enuff metal so that it does not distort the vocals.

 
Hi,
Pretty much the same project here except I ever built the motor.

If your core is still resonably small, I'd second Rossi's recommandation to try with what you have...

A large core will means more copper per winding, which means higher resistance.

And I'd say that on the other hand, a larger core would mean more winding space, allowing for the use of heavier wire gauge, wich could compensate the resistance increase due to wire lengh.

as fo shielding, if you can find a core supplier locally, perhaps he could also help you finding shielding material...

Laurent.

 
and just so we are on the same page, most mic inputs are 80% Ni, so the math is the same.

Nickel cores will have a lot less loss than 29gaM6. (grain orient)

a bigger core means less turns needed to get the same inductance,

but the increased wire length means more stray C,

so you trade insertion loss and a little high end for reduced distortion,

or you give in a little to distortion and allow less loss amd better response,

see what a nightmare these darn transformers are? you can not have your cake and swallow it too.

??? ::) 

you know what is weird, everytime I type a word, the guy says the same word at the same time on the radio.

this time it was "orient."

at the exact moment i typed orient, dangerous dan on kfjc mentioned a band called orient express.
look at the bottom track>

29fczmg.jpg


you might say, "gee cj, what an awareness you have, nobody else notices that crap"

not really, after about 1000 times of this stuff, you start to take notice,

does this happen to anybody else out there?  :-\
 
For the same/similar core material:

The larger cores allow fewer turns for the same inductance or saturation -and- the larger windows on larger cores allow fatter wire. So even though the mean length of turn goes up, and the total length may go up, the resistance generally goes down. (Otherwise KVA transformers would be hopelessly lossy; instead they are more efficient than small iron.)

As CJ's oriented brain says: stray capacitance rises. This really hurts high-Z performance. It may not be an issue for 0.2:200 impedance.

The killer problem: shielding! A larger core will intercept more room-hum, and will need a larger can to shunt hum around it. Good magnetic shield-stuff is expensive.

(Also you don't want large lumps near microphones because they skew the response and directionality. We finally got magnets smaller than old Alnico, and you want to put a doorbell tranny in its place?)

There are two large groups of "iron". When you need POWER, you use a silicon steel. It can handle a fair amount of power per pound, and it is not real expensive per pound. But as you note, it is so inexpensive that they don't make it in small sizes.

What is the power of a ribbon mike? 10mV into 1K? That is a TENTH of a MICROwatt! If I can get 5 Watts per pound of sil-steel, then I need much less than a micro-pound of core. That's invisible.

And as CJ says, iron eddy-loss is a shunt resistance, and excess iron is excess loss. On modern si-steel with "normal" power density, the loss may be under 2%. However when you have a million times more iron than you need, it may matter.

All of this was understood by 1930. Telephones and especially vacuum-tube audio amplifiers needed good inductance and low loss at VERY low power level. What was needed was an iron with much higher inductance per turn. Power-handling was often not a problem. The answer was RadioMetal and Permalloy: juiced-up iron/nickel metals with magic heat-treatment. You get 5 or 10 times more inductance; or rather, you get your desired inductance with a much-much smaller core. The smaller winding has less stray capacitance, the smaller core is easier to shield. The stuff is expensive, but you don't need much as long as you are talking low-low-LOW power level.

Dat's how it should be done.

But yeah. Take your power-core. Figure 300 turns secondary, 10 primary. Each gets half the available window area. Compute your maximum wire gauge, compute your copper resistance. This should be much less than the nominal impedance: you want your "200" winding to be near 20 ohms. If it is much less, increase the number of turns: I think that reduces the leverage of iron eddy-loss, which I suspect will be an issue. Wind it up, feed from/to appropriate impedances, and measure frequency response. "Appropriate impedance" load would be a mike preamp, so now you can quickly tell if the thing picks up so much room-hum that shielding will be problematic.

You should interleave primary and secondary, but at these low impedances a simple S-P-S interleave will probably give more high-end than a ribbon mike.

10 turns in a power core window suggests 3/16" copper rod. Which won't want to bend around. You will want to be ribbon or parallel strands of say #16. Thin enough to wrap, paralleled enough to use the available space for low copper loss.

> increased wire length means more stray C

For simple systems: it is the exposed area of the winding, not "length". So a 1"x1" bobbin will have a fairly consistent stray C, a few hundred pFd, no matter what length/diameter you do. Of course as you interleave, each leave adds more inter-winding capacitance. And long/thin wire goes with high impedances where C really sucks.
 
wow, cheers for the amazing amount of info everyone!

needless to say i'll be rereading this over and over, like i have all the other ribbon mic threads.

Well, I guess I'll keep shopping around for a more regular mic sized core,.... for the time being I have a few salvaged EI's and toroids so I'll play around with those until it comes time to step things up ( how pun-tastic  :D ).

I have been looking into sourcing high nickel iron, and being in Japan its quite easy to find top grade manufacturers,.. but I havent checked on minimum orders yet, which may be quite high. definately on my to-do list though.

Cheers and thanks again for the input guys,


brukker.
 
I was a bit too general with my statements: Of course a larger bobbin allows thicker wire. But at a certain point it's hard to find thicker wire that's still flexible enough to bend around the bobbin. Very very Low DCR on the primary winding is crucial, so you use very thick wire. But anything thicker than about .8 mm is hardly manageable. Paralleling two wires can help and will produce a flatter winding than one very thick wire. But you see that a large core will make things more complicated.

Here's an older thread about my experiments with a power core (from a 9V-wallwart)

http://www.groupdiy.com/index.php?topic=19046.0

What gives better results is rewinding a Chinese mic transformer. I rewound some ribbon trannies for active ribbon operation (ca. 1:100 + electronic impedance conversion). I haven't tried cores from discarded China condenser mics, but I'm sure they'd perform much better than a power tranny core. A dual bobbin humbucking arrangement is a good idea, as it will require less shielding. Most Chinese mic trannies are constructed that way.

 
yeah I went through your mcgyver thread atleast 3 times in the past month, it made for very interesting reading.

The core i had in mind would have come from an amplifier this guy was looking to off load so not a power tranny, but still on the chubby side of large for sure. So I wanted to ask before I put any money down as I'm a cheap bastard  ;D

I did think of just picking up a chinese mic and fiddling with it, but then even those "cheap" chinese mics are no longer exactly cheap once they get shipped into Japan. I reckon if i spent the same amount on a core and on winding, I stand a decent chance of producing something half decent.

Besides, its just more fun trying to do this from scratch, and I never thought I would be getting into tranny building but strangely i'm enjoying learning about it. there's something incredibly masochistic about this DIY stuff.

PRR's comment about "expensive material but small amount" got me thinking as I passed a friend's jewelry shop in town today,....

apart from price, how stupid an idea is a mostly nickel core, does a core have to have iron?
 
> how stupid an idea is a mostly nickel core, does a core have to have iron?

All the pure metals can be made better.

Plain soft iron "works", but resistance is low enough that eddy loss is an issue. Strangely, a cup of Silicon (sand, a natural impurity) helps a lot. (And "steel" is cheaper than pure iron; nobody takes ALL the carbon out, anyway it improves the stuff.)

Soft or Silicon iron can be better magnetic if you roll and anneal it just-right. That's why a handful of nails, or a mile of painted iron wire, is not so good a core as transformer-stuff.

Pure Nickel has much less magnetic property than the irons/steels. (And pure nickel is hard to find; a US 5c coin has as much copper as they can get away with and still be nickel-color.) I think all the transformer Nickels are magic-recipe alloys and treatments. Unless you want to spend your life in metallurgical alchemy, you buy this stuff.
 
cheers PRR, guess its time for me to go shopping and get down to building.

Thanks alot to everyone that replied to this thread, its been most educational.



brukker.
 

Latest posts

Back
Top