help winding an auto transformer

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

saxtim

Well-known member
Joined
Jun 4, 2004
Messages
88
Location
Melbourne, Australia
I'm hoping someone can help me with winding an autotransformer. I had some tips ages ago at the old place, but the thread is long gone. It's for a U87ai mic clone, the schematic can be found here:

http://www.omnipressor.com/MicSchLib/u87ai.gif

The inductor in question is in the DC-DC converter.

I've spent the last couple 6 months on and off on this project and I've now designed/etched/stuffed the PCBs, soldered everything up and it's more or less ready to go bar the autotransformer (and a suitable capsule - another story).

I've got a toriod core with an AL value of 2770nH. It's 16mm in diamater. I've also got some thin magnet wire (I think .25mm from memory).

Bo Hansen has penciled in the values for the inductor on the schematic, but I'm not sure of the units - it looks like micro henries? I know it was mentioned sometime ago at the old place when this was orignally being discussed but I can't remember.

Anyway I have some questions about winding etc:

1. Will it actually matter that much if I don't get the same inductance values? I remember Jacob mentioning that I should just trying winding something like x+2x turns and seeing what voltages were generated. Any suggestions what number of turns to start with?

2. Does it matter if the individual windings contact each other, or should each turn round the core be separated slightly from the last?

3. Its just one long continous winding around the core I need right? I don't need a second winding over the top do I?

4. How do I make the connection for RT? With GE and GN, because they're at the start/end there is extra magnet wire to play with to solder connections to. But for RT, I was fiddling around today and I had great difficulty trying to solder a wire onto the actual loops on the coil, the magnet wire seemed to not like be soldered to and my connections would simply break off if they were knocked - is there a trick for this?

thanks

tim

Thanks

tim
 
You could likely do it as you mentioned - a single layer. It's not so critical as it operates at a single frequency.

There are a couple ways to solder to magnet wire. Some magnet wire insulation will melt under the heat of a soldering iron - practice first. If you have this sort of wire, take your iron and apply some solder to it, then let it sit for 20 or so seconds on the magnet wire while moving the iron back and forth. At some point the insulation might break down allowing the solder to stick to the wire and not the insulation. If the solder comes off easily it's just wrapped around the insulation and not making contact.

If you don't have that sort of wire, use a lighter or torch to lightly heat the wire so the insulation breaks down but not enough to make the wire glow. Now clean the wire off with methanol to get the insulation off. You should be able to solder to it ok after you clean it.

If it's thick wire you can use sandpaper but I don't like this approach.

The values are 220uH for the small section and 1400uH for the large section. Looks like about 10 turns for the little section and 20 turns for the large section should be about right, for a total of 30 turns.

Funny you should mention this because I'm building a U89 circuit, which includes this power supply, probably next week. I just etched the boards a few days ago.
 
Oh yea, what I do when winding and need a centre tap is I leave about 12" of wire going out in a loop (6" each direction), then twist the loop together so it's like one wire sticking out about 5-6". I don't try splicing on the wound area - you're likely to make a short circuit there. After the coil is done, remove the insulation from where you need it in that length (often it's only an inch or so from the coil) and trim off that length.
 
Thanks. Well I tried putting the inductor into the circuit and it's not working (I'm just trying to get the correct voltages +\- 60V at this stage, not trying for audio).

I sorta supsect my inductor, as I've never wound one before, however I've got some voltage measurments. I'm not really where to start looking other than that.

I should note that I'm testing this circuit without the capsule attached and without the output transformer. I'm pretty sure I can leave the capsule off without issue, but I'm not as sure about leaving the output transformer out of the circuit - I think that's ok too though at the moment.

Head amplifier and DC board are on separate boards.

I'm getting around 45V in from phantom.

At T2 (head amplifier) I've got:

E36.4 B37.2 C 43.8

This means I've got 36.4 going to the DC converter board (from E of T2)

On the DC board:
T9
E 10.7 B 11.3 C18.3
T10
D 11.2 S 10.5 G 10.5

(there's no short that I can find between S and G -should these be at different voltages?)

After r10 and before r103 I'm getting 33 volts (should be getting 60)

I'm getting 0 volts between r104 and r13 (where I should be getting - 60V)

Any suggestions?


Funny you should mention this because I'm building a U89 circuit, which includes this power supply, probably next week. I just etched the boards a few days ago.

Are your boards based on the Neumann ones, or are they your own design? Whould you mind sharing them - if yes personal message me. I designed my own - I'm sure they're correct in regards to everything attaching to what it should etc, however I don't know how critical actual layout is for the DC board seeing as it operates in the RF area.

thanks

tim
 
Tim,

First of all, make sure that your oscillator - that the autotransformer is a part of - is actually running, and at the approximately correct frequency. Possibly the adjustment of frequency is acheived by "primary" inductance of the autotrafo, i.e. the lower half of the inductor.

Once you've got the oscillator running, add turns to the "secondary", the upper half of the inductor, untill you have the right voltages.

Jakob E.
 
First of all, make sure that your oscillator - that the autotransformer is a part of - is actually running, and at the approximately correct frequency. Possibly the adjustment of frequency is acheived by "primary" inductance of the autotrafo, i.e. the lower half of the inductor.

Jacob - thanks, however I don't know how to check if this thing is oscillating. I don't have a scope - is there another way to check?

thanks

tim
 
I can't find AC voltage anywhere, so I guess that means it's not oscillating.

I went throught it again and check all the component values and diode orientations - everything is correct as far as I can see. Where should I start looking to get the oscillator going?

thanks

tim
 
It looks like you've got about 230 uA flowing there. You should have about zero volts on the gate of the JFET. Make sure you got the pinout of the FET right. It almost looks backwards from the voltage measurements but that's pretty much a guess. There should also be 0 volts (DC) on all terminals of the inductor - and from your measurements it looks like it is ok in that regard. If you don't have a scope, to see if you have oscillation, the best bet I can think is just look at the -60 volt output. If you have oscillation it'll show something. First, though, is find out why the DC voltages are wrong. I would expect on T10 to get 0 on G and maybe 3 to 5 volts on S.

I just finished the head amp of my U89-like circuit, and it works ok. The DC-DC converter board is my project for this coming weekend. I just did a pretty simple layout - it runs at under a MHz so the layout should not be that critical.

-Dale
 
It almost looks backwards from the voltage measurements but that's pretty much a guess.

Dale, I should've checked the data sheets :oops: Both the BC182B and BF245B had different pinouts than listed at the bottom of the schematic.

Now I've got the following voltages:

T2 (not sure if this is important or not - it's on the head amp board, but it feeds supply to the DC board)

C45.1 B41.4 E41.1

T9
C 35.2 B34.0 E24.5

T10
G0.1 S 3.5 D34.5

I'm getting 37.1 volts between r10 and r103 and -1.3 between r104 and r13 - still way off, but it must be oscillating now as I'm getting the negative voltage.

I guess it's now down to tuning the inductor - first though I want to make sure I've got the primary/secondary part around the right way.

The first section, 10 turns is that from GE to RT or from GN to RT? I think it's from GE to RT and then RT to GN is 20 turns. Or have I got this backwards?

How should I go about tuning the inductor - should I just add windings to the secondary, or as Jakob mentioned:

is actually running, and at the approximately correct frequency. Possibly the adjustment of frequency is acheived by "primary" inductance of the autotrafo, i.e. the lower half of the inductor.

should I also be fiddling with this as well? I have no idea of the approximate frequency that it should run at (or how to measure it)


Thanks

tim[/quote]
 
If it's oscillating and getting a bit of voltage then that's a good thing. Try increasing the value of C25 (330pF) and see if voltages pick up. This capacitor sets the gain of the oscillator. I think you have the relationship right - the 10 turns is between 3 and 2, the 20 turn coil should be between 2 and 1. You can try increasing or decreasing the number of turns a bit and see if you get more or less voltage. I am getting the right voltages but the oscillator is cutting in and out instead of stabilizing at the right value. Mine oscillates at about 500 kHz, though I'm not sure if this is right.
 
I think it would be important to know what frequency the original oscillates at. That the way you'd be sure to be in the right frequency range, not getting too much side effects..

Jakob E.
 
If I have a 'scope probe hooked up to the circuit, it oscillates at maybe a 5V p-p output, but as soon as I disconnect the probe it oscillates normally, and I get about +/- 55V output. The 'Q' of the inductor needs to be quite good, I tried some powdered iron cores I had lying around and they kinda sucked - they didn't oscillate.

-Dale
 
If I have a 'scope probe hooked up to the circuit, it oscillates at maybe a 5V p-p output, but as soon as I disconnect the probe it oscillates normally, and I get about +/- 55V output. The 'Q' of the inductor needs to be quite good, I tried some powdered iron cores I had lying around and they kinda sucked - they didn't oscillate.

Cool! I'm going to fiddle some more with mine, probably not till the weekend though. Any tips on getting a good 'q' - is this mostly material dependent, or does the way you wind it effect it as well? What type of core did you end up using? I really don't know much about inductors - I wonder if the core I'm using is ok - it's an epcos 'ring core' Farnell (Australia) part number 3110266 . It's epoxy coated.

Dale, have you got any way to post pictures? I wouldn't mind seeing how your inductor actually looks - I stil have some reservations on the way I've wound mine.

thanks

tim
 
Getting a good "Q" value inductor core also depends on you being able to specify the working frequency..

think it would be important to know what frequency the original oscillates at. That the way you'd be sure to be in the right frequency range, not getting too much side effects..

I'm sure there are plenty of people here that have or have access to a U87ai - would anyone be willing to measure the frequency of osciallation in the DC-DC converter circuit?

tim
 
And if you are measuring it, here's a tip - measure it from the centre tap on the inductor, not the top. Measuring it from the top loads it down and things don't work right after that.

The working frequency appears to be somewhere in between 200 kHz and 2 MHz, that should be good enough for design.

I don't have a digital camera at home, but my inductor is a ferrite rod that's about 1.5 cm long, and about 5mm in diameter, and I have about 100 turns tapped at 35 turns, scramble wound. I got the core out of a burnt up PC power supply. I don't know the AL value of the core or anything, I just guessed at the number of turns. I might try doing a neater version of the inductor when I get a chance, and when I'm happy with it I'll glue it down to the board. Right now it oscillates at about 1 MHz, but I think that's a bit high. Probably a neater winding will bring the oscillating frequency down to where it should be.

-Dale
 
I picked up a ferrite rod today and tried scramble winding a similiar number of turns 35/100 - and I've now got +/-47.5V Much more healthy looking :grin: I'll try winding a couple more and adding some more turns to the secondary to see if I can get it up closer to +/- 60V.

Dale - just out of interest, have you tried you're U89 circuit yet? What sort of capsule are you using? Are you building it into an existing microphone body?

tim
 
I built it into a cheap Chinese mic body ('Apex 430'). The capsule is the original (yuck) Chinese capsule, but I am working on building a capsule, it'll be a lot like the KK47 in appearance and hole pattern. One thing to remember is that the output of this power supply is very wimpy - even the load from a meter drags down the output, so getting 47.5 volts with a meter might be pretty close to what it's supposed to be. I get almost exactly that, but if I use the oscilloscope (with a 10 meg input impedance) it's closer to 60. If you have no meter attached you're likely getting 55 or 60 volts.

The U89 circuit just powers up and worked, though the jury is still out on the sound - I'll need a real capsule to judge. I'm using a Sowter transformer for the output. One thing that's nice with it is that it doesn't die for ten seconds or so on a puff of air, which I found the U87 circuit does.

-Dale
 
Back
Top