measuring OT's primary impedance

[booj]

I was just purusing a radio ameture's site <http://www.radioremembered.org/outimp.htm> about measuring the turns ratio, sqaring that and coming up with an impedance ratio. He came up with good numbers, but when I measured my '68 Bandmaster OT, my calculations came up with less than a K for the primary between the red and brown lead.
Here's my numbers: measured at 120 HZ

Red to Blue primary: 7.350HY
Red to Brn primary: 7.185HY
Blk to Grn secondary 48mhy

The impedance ratio is 7.185/.048? That would be about 150. Then the turns ratio = sq. rt. of 150, or 12.
When I hooked up my sig. generator to the Red/Brn leads I came up with 534mv. into the primary, and 34.2mv. at the Blk/Grn leads, indicating a turns ratio of 534/34.2, or 15.6.
So, lets just say 15. 15X15 = 225 (turns ratio squared = impedance ratio)
With that ratio, the reflected impedance to the primary from a 4 ohm load would be 225 X 4, or 900 ohms.
Where are my thousands of ohms of impedance at?

[tubeswell]

Reading the DC resistance on the primary like that has no relationship to reading the DC resistance on the secondary.

You have to put like a 1VAC source across the secondary and then (carefully) measure the VAC across the primary, and the ratio of the two is your impedance ratio (which you then square root to get your turns ratio)

[Enzo]

LOoks more like he measured the inductance instead of resistance, but the transformer doesn't run on inductance either.

You need to determine turns ratio, not inductance or resistance. And you do that by comparing voltages in and out.

ALL you need is an AC voltmenter, a signal source, and the transformer itself. Oh, and a calculator too, I suppose.

[Chuck H]

You can determine the ratio just like Tubes said. But I wanted to note that it helps to have an AC signal that can supply at least a little current. Otherwise the OT will load it down and give you an inaccurate read. I tried to use a little keyboard output once and got a bad read. So I switched to an AC wall wart type transformer and that worked fine.

Chuck

[Steve Conner]

OT primary impedance is always quoted between tube plates, which is four times the value you get between the center tap and one plate.

The method is fairly sound, since the magnetizing inductance, which is what the OP was measuring, is proportional to the square of the number of turns in the winding you're measuring it from, just the same as the impedance ratio is.

The only drawback is that magnetizing inductance is non-linear. You get a different answer depending on how much you magnetize the core in the process of measuring it. If your inductance meter always uses, say, 1 volt at 60Hz, it'll magnetize the core a lot more when connected to the speaker winding than when connected to the primary.

Hence this method is less accurate than what Enzo and others suggested, to feed a voltage into one winding and measure the step-up ratio. As long as the magnetizing current isn't stupidly high, its only effect is a small loading on the voltage source you're using, and any volt drop caused by this gets cancelled out when you measure both voltages and calculate the ratio.

[booj]

Thankyou for your feedback, guys. I did a voltage measurement with my sig. gen. which has an internal resistance of about 500 - 680 ohms. I thought that would drive the primary (plate side) strongly enough, but I'll just go with a variac and measure the ratio agin using higher voltages. Get back with you.

[Steve Conner]

The sig gen will work just as well as the variac, as long as you measure the voltage ratio like I explained, rather than using the inductance method.

[booj]

Thanks Steve. I tried it again with a variac. I did my calculations a bit differently.
Unloaded voltage from variac: 12.22v
After connecting to 4 ohm tap: 12.00v (not much loading)
voltage at red-brn lead: 188v
Since I know that the tranny puts out 40 w into 4 ohms, or about 12.6v, I know that the output for 12 volts would be 12x12/4, or 36 watts. I just calculated the current neccessary to produce 36 watts at 188 volts, or .191A. So R or Imp. = 188/.191, or 984 ohms.
The Bandmaster uses 2 6L6's.
What...am...I...missing...

[Enzo]

But it is not a resistor, it is a transformer.

Take it step by step.

What turns ratio did you calculate for the transformer?

From the turns ratio, calculate the impedance ratio.

What load impedance did you specify?

Take that impedance and transform it through the impedance ratio you just calculated.

[Steve Conner]

You're not missing anything, rather you're seeing something that most people miss.

984 ohms, or 900 or any of the other figures you posted, is a perfectly reasonable load impedance for a 6L6's plate to see, and indeed, 0.191A is a reasonable estimate for the current drawn from B+. (It'll be nearer 0.300 because the tubes are only about 60% efficient.)

It's just that by convention, plate load impedances in push-pull amps are quoted across the entire OT primary, between the two plates. Since impedance varies as the square of the turns ratio, and the center tap is halfway along the primary, the plate-plate impedance is four times the value you measured. And 984 times 4 is about 4k, which is a reasonable value for a 6L6 OT.

To sum up, the impedance from center tap to one plate is what matters to the actual physical operation of the tubes. When each tube takes its turn to work in the Class-B push-pull circuit, it will be pulling on a reflected speaker load that looks like a 984 ohm resistor to B+. But when shopping for OTs you have to multiply this by 4.

[Grim]

don't forget the transformer also has regulation, so the turns ratio won't work out exactly as you would think

a 10:1 transformer may well have 10:1.1 or even 10:1.2 depending on the size.

[Bruce / Mission Amps]

I've done this hundreds of times with my Variac or my 12vac .5amp wall wart with clip leads on the output.

As mentioned about three or four times above, you are measuring it wrong.

If I found an unknown OT, .... when 12vac is applied to the 'X' secondary (using the Variac or wall wart) it delivered 188vac across the primary, (Brown to blue, not from one side to the middle) I would be 100% convinced it was about 2000 Ohms across the primary with an 8 ohm speaker and 1000 ohms with a 4 ohm speaker load, or any ratio thereof.
No good for two 6L6s

Doing your test correctly, across blue to brown (the actual entire primary) and I'll assume yours would have actually read closer to 376vac from brown to blue:

376vac/12vac = 31.33
31.33^2 = 982
That means 982:1
So,
922 times a 4 ohm load = 3927 ohms.
That is very close to the correct primary Zed for pair of 6L6s in this amp.

[booj]

Thanks much! I got hung up on just using the red-brn wires because that's what is connected to one of the two tubes. And if you double the voltage reading, you square the ratio and multiply by 4.

[Robert M. Martinelli]

Hi,
to complement all of the above I would like to add the "mathematics" rationale for matching impedance - a perfect match happens when the maximum power is being transferred from one stage to another with the minimum loss, and this happens when the source and the load impedance are the same, the exact relationship for calculating an OT' s ratio being : ratio = Sqrroot( Zin/Zout ), or if you prefer, (squared ratio)= Zin/Zout.

Once you have determined the actual OT ratio with the "voltage" method, you can easily figure out things simply reversing the above equation : Zout=Zin/(squared ratio)
Zin=Zout*(squared ratio)

As already said, any OT has an inherent ratio, not an inherent impedance, it opposes, or "acts like" a certain impedance on one side depending on its ratio and the characteristic impedance you connect on the other winding.

My .001 ( Euro ) cent

Best regards

Bob

[Enzo]

I compare it to a set of gears. The gears always have a speed ratio between them, but they have no inherent speed of their own.