Help with Maxwell-Wien Bridge

Do you have a function generator and an oscilloscope? That's about all you need.

You'll need to measure that thing at a low frequency, a few Hz maybe.

Does the Bible give any guidance on the issue? That's the first place to look.

Maybe fool around with the R values to get a bigger C?

What do you do with the other side of the transformer? Open circiut?

George H.

Yes. Describe the technique.

Do I have any concern about the core not acting the same a 10Hz vs

The C has to be a value to equal the reactance of the Inductor being measured. I don't know good values for the R's, I pick 10K out of the air.

Yes.

Mikek

Set up a fungen to make a sine wave. Connect it to the inductor through a resistor, 50K roughly. Put a scope across the inductor. Make sure there's no DC offset... add a big series cap if there's any doubt.

Twiddle the frequency and find the 3 dB point, and do the math. If you plot the frequency response, you can separate out the inductor L and copper resistance.

A square wave can be interesting, too. Fungen, L, small R to ground, so the scope displays inductor current. The current should be a triangle wave, and the slope lets you calculate L. Any saturation effects should be visible. For a very low frequency square wave, the L/R time constant will be visible.

Lots of variations.

Yeah. Apply a voltage to it and record the current from 0 to E/5000.

I've never used this bridge, but if you believe the numbers on the wiki page then making R1 and R4 1k ohm would make C 100 times bigger... and R2 100 times smaller. (But you mostly care about the L value, I think.)

George H.

Oh, I have done that, although the 50k doesn't work for me. I dropped half my voltage at 920Hz and the series resistor was 5.94Mega ohms. This calculates to 1,029 Henries, this is more in the range I was expecting for this transformer although when I rang it I got 2000 to

Would you describe that better, I'm not seeing where the R breaks out.

If you graph that frequency response, a pure inductor would have the voltage trend to zero as the frequency goes down. If there's series R inside the inductor, you'll get a non-zero plateau at low frequencies, where Xl is approaching zero and the resistors make a voltage divider.

At high frequencies, core loss and distributed capacitance will start another rolloff, maybe even a peak if the Q is there. Gotta consider scope probe C, of course.

Yes, this transformer has -3db points at 80 Hz and 3000Hz. I suspect most of the high frequency rolloff is from the self capacitance.

Yep, I had major problems with probe C to start. My first test I measured 42.5 Henries, then I pulled another different brand transformer a I got the same answer 42.5 H. (uncomfortable feeling) The third different transformer measured much lower but I got 1/2 voltage at two different points. After mulling awhile, I checked my scope probe and it was on X1 with about 100pf. ARRGH! Rather than using a X10 probe with 15pf capacitance, I have a high impedance, very low C input amplifier that I used for the new tests. With that I measured the 1,029 Henries. But I want to try to duplicate the number with another method. Hence the Maxwell-Wien Bridge.

Mikek

Ahh, have you tried reducing the resistors yet? I'd try

George H.

Given a primary with 1000H + 5K ohms, the inherent L/R tau is 0.2 secs, w=5, so Fc~~1 Hz. So with a low-Z primary drive and high-Z secondary load, it should be down 3 dB around 1 Hz.

John, I'm sure you know more about this than I do, But the manufacturer graph on the spec sheet shows the response as I said. Now, maybe you are suggesting my 1000 Henry number is wrong and it could be, that's why I'm working to verify it. To get 1000H takes a lot of turns and with that comes self capacitance, wouldn't the parallel capacitance cause the high frequency to drop. Here's the manufacturers graph.

Are you nuts? One THOUSAND Henries? I seriously doubt it. However, a (much) lower test frequency would be in order; say 10Hz. The idea is that all arms to be roughly the same impedance AND have (stray) capacitance of small effect. So, make R=69.1K RN55C 0.1% (62.83K+Rx=5.27K) (ASS-u-ME-ing 1000 Henries). Also ditch the variable cap and make Rs=62.6K RN55C 0.1% in series with a 10K rheostat preset to middle (about 5K). If brave, and know Rx~5K but >5K, use Rs=66.5K RN55C 0.1% in series with a 5K rheostat preset to middle.

Check!!! Low freq is for better results; also balance all arms.

DANG! Tawt I taw a puddy tat! Thought the "O" series was Triad. Must be an OLD series long gone..could not find it by any current xfmr maker. Well, it _has_ been a few daze since the thermally stabilized, vacuum encapsulated FETs..

Bill Whitlock of Jensen Transformers wrote an interesting Handbook for Sound Engineers available at

He warns on Page 26:

"Do not use an ohmmeter to check winding resistances unless you are able to later demagnetize the part. Ordinary ohmmeters, especially on low-ohm ranges, can weakly magnetize the core. If an ohmmeter simply must be used, use the highest ohm range (where the test current is least)."

He discusses demagnetizing on page 29.

If your inductor is really 1,000 H, the bandwidth will be very low due to stray capacitance if it is operated into an open circuit, as you have already noted.

However, the transformer must be operated into its rated load impedance. For example, the Jensen JT-11P-1 has a recommended load of 10k as shown on page 27. The low frequency response is -3dB at 200 millihertz as shown on the same page.

This implies a large inductance, which would normally resonate with the stray capacitance at a low frequency. However, the high frequency response extends to 100 KHz at -3dB. This illustrates the need for proper loading.

A transformer with 1,000 H primary inductance is not unreasonable. He discusses a 4:1 input transformer with 300 H primary inductance on page

Unfortunately he does not talk much about stray winding capacitance or how to measure it.

The mfr tested it with high impedances, which wrecks the LF response. HF response, too.