Forums

Transformer Turns Question

Started by Chris February 14, 2016
Hi guys,

I need to count the number of turns of wire on this relay core I have. 
Trouble is, there are way too many to do so manually. The DC resistance 
is over 500 ohms and the wire gauge is *very* tiny; there must be a 
couple of thousand turns at least. I've thought of an ingenious way of 
counting them and I'd like to know if it will work in practice.

What I plan to do is wind say 10 turns of wire around the outside of this 
core, which is effectively just a solenoid of course. Then I feed this 
outer winding with say - for example - 10mV of AC at say 10kHz. I 
determine the voltage output level from the 'secondary' with an 
oscilloscope. From the ratio of p-p input voltage to p-p output voltage I 
can calculate the unknown number of turns on the core solenoid from the 
knowns: the input & output voltages and the primary no. of turns.

What do you think? Will this work in practice?
On Mon, 15 Feb 2016 00:05:06 +0000, Chris wrote:

> Hi guys, > > I need to count the number of turns of wire on this relay core I have. > Trouble is, there are way too many to do so manually. The DC resistance > is over 500 ohms and the wire gauge is *very* tiny; there must be a > couple of thousand turns at least. I've thought of an ingenious way of > counting them and I'd like to know if it will work in practice. > > What I plan to do is wind say 10 turns of wire around the outside of > this core, which is effectively just a solenoid of course. Then I feed > this outer winding with say - for example - 10mV of AC at say 10kHz. I > determine the voltage output level from the 'secondary' with an > oscilloscope. From the ratio of p-p input voltage to p-p output voltage > I can calculate the unknown number of turns on the core solenoid from > the knowns: the input & output voltages and the primary no. of turns. > > What do you think? Will this work in practice?
I think you'll have so much leakage inductance that the number you get will be very rough. Other methods: * estimate the volume of wire, measure it's diameter, get the length, then estimate the turns from the area. * Weigh the thing, estimate the weight due to copper, measure the diameter, get the length, ditto. -- www.wescottdesign.com
On Sun, 14 Feb 2016 18:11:11 -0600, Tim Wescott wrote:

> On Mon, 15 Feb 2016 00:05:06 +0000, Chris wrote: > >> Hi guys, >> >> I need to count the number of turns of wire on this relay core I have. >> Trouble is, there are way too many to do so manually. The DC resistance >> is over 500 ohms and the wire gauge is *very* tiny; there must be a >> couple of thousand turns at least. I've thought of an ingenious way of >> counting them and I'd like to know if it will work in practice. >> >> What I plan to do is wind say 10 turns of wire around the outside of >> this core, which is effectively just a solenoid of course. Then I feed >> this outer winding with say - for example - 10mV of AC at say 10kHz. I >> determine the voltage output level from the 'secondary' with an >> oscilloscope. From the ratio of p-p input voltage to p-p output voltage >> I can calculate the unknown number of turns on the core solenoid from >> the knowns: the input & output voltages and the primary no. of turns. >> >> What do you think? Will this work in practice? > > I think you'll have so much leakage inductance that the number you get > will be very rough. > > Other methods: > > * estimate the volume of wire, measure it's diameter, get the length, > then estimate the turns from the area. > > * Weigh the thing, estimate the weight due to copper, measure the > diameter, get the length, ditto.
D'oh. Measure the DC resistance, measure the wire diameter, use the conductivity of copper or a copper wire table to estimate the length ... -- www.wescottdesign.com
On Sun, 14 Feb 2016 18:11:11 -0600, Tim Wescott wrote:

> I think you'll have so much leakage inductance that the number you get > will be very rough.
Whoa! "leakage inductance"?? Where's that coming from already? Concentrically-wound transformers are a perfectly valid topology AFAIAA.
On Sun, 14 Feb 2016 18:12:40 -0600, Tim Wescott wrote:

>> * estimate the volume of wire, measure it's diameter, get the length, >> then estimate the turns from the area. >> >> * Weigh the thing, estimate the weight due to copper, measure the >> diameter, get the length, ditto. > > D'oh. Measure the DC resistance, measure the wire diameter, use the > conductivity of copper or a copper wire table to estimate the length ...
Not so simple. The windings aren't single layer; they're stacked about 1/4" thick which really fouls up the C=pi*D part of the calculation.
On Mon, 15 Feb 2016 00:05:06 -0000 (UTC), Chris <cbx@noreply.com>
wrote:

>Hi guys, > >I need to count the number of turns of wire on this relay core I have. >Trouble is, there are way too many to do so manually. The DC resistance >is over 500 ohms and the wire gauge is *very* tiny; there must be a >couple of thousand turns at least. I've thought of an ingenious way of >counting them and I'd like to know if it will work in practice. > >What I plan to do is wind say 10 turns of wire around the outside of this >core, which is effectively just a solenoid of course. Then I feed this >outer winding with say - for example - 10mV of AC at say 10kHz. I >determine the voltage output level from the 'secondary' with an >oscilloscope. From the ratio of p-p input voltage to p-p output voltage I >can calculate the unknown number of turns on the core solenoid from the >knowns: the input & output voltages and the primary no. of turns. > >What do you think? Will this work in practice?
That can work. It might be better to drive the relay coil from a signal generator and measure the added winding voltage as the secondary. For more accuracy, measure the DC resistance and the inductance of the relay coil, pick a good frequency, and do the math to account for the copper loss. Or Spice it. Don't expect an exact turns count, but you should be able to get into the ballpark. If you "close" the relay (push the armature down) there will be less flux leakage error and you might get within 1%. -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
On Sun, 14 Feb 2016 16:19:48 -0800, John Larkin wrote:
> > That can work. It might be better to drive the relay coil from a signal > generator and measure the added winding voltage as the secondary.
Certainly *safer* I guess!
> For more accuracy, measure the DC resistance and the inductance of the > relay coil, pick a good frequency, and do the math to account for the > copper loss. Or Spice it.
Not sure what you mean by "a good frequency" - how do I determine that? I also thought maybe pulling the iron core out (if possible) and replacing it with a ferrite rod in case saturation might be an issue?
On 2/14/2016 6:45 PM, Chris wrote:
> On Sun, 14 Feb 2016 16:19:48 -0800, John Larkin wrote: >> >> That can work. It might be better to drive the relay coil from a signal >> generator and measure the added winding voltage as the secondary. > > Certainly *safer* I guess! > >> For more accuracy, measure the DC resistance and the inductance of the >> relay coil, pick a good frequency, and do the math to account for the >> copper loss. Or Spice it. > > Not sure what you mean by "a good frequency" - how do I determine that? > I also thought maybe pulling the iron core out (if possible) and > replacing it with a ferrite rod in case saturation might be an issue? >
I would add more than ten turns. Why? I don't know, I just feel you would have more accuracy with 100 or 200 turns. Mikek
On Mon, 15 Feb 2016 00:45:34 -0000 (UTC), Chris <cbx@noreply.com>
wrote:

>On Sun, 14 Feb 2016 16:19:48 -0800, John Larkin wrote: >> >> That can work. It might be better to drive the relay coil from a signal >> generator and measure the added winding voltage as the secondary. > >Certainly *safer* I guess! > >> For more accuracy, measure the DC resistance and the inductance of the >> relay coil, pick a good frequency, and do the math to account for the >> copper loss. Or Spice it. > >Not sure what you mean by "a good frequency" - how do I determine that? >I also thought maybe pulling the iron core out (if possible) and >replacing it with a ferrite rod in case saturation might be an issue?
Well, you need to do a little math and maybe some measurement. The frequency should be high enough that the voltage drop in the inductance is above the voltage drop in the copper resistance. And stay below the self-resonant frequency. A nonlaminated iron core in a DC relay will have a lot of loss at high frequencies, too. If you are willing to disassemble the relay, you could use a better, closed core. Or unwind and count! Just curious, why do you want to know the number of turns? -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Tim Wescott wrote:
> > I think you'll have so much leakage inductance that the number you get > will be very rough.
If you drive it from both sides, and compare the ratios, maybe you could compensate for that? --