nts voltage / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

nts voltage

previous 1 2 3 next

Move Thread

LAN_403

J. Aaron Holmes

Sun Jun 10 2007, 04:05AM

Registered Member #477 Joined: Tue Jun 20 2006, 11:51PM
Location: Seattle, WA
Posts: 546

Ok. Armed with fresh evidence, I'm going to re-assert the bogusness (or at least severe inaccuracy) of both of the following methods of measuring NST voltages:

1) putting 120V into the HV winding and measuring the voltage on the LV winding
2) putting 1V (or other extremely small V) into the LV winding and measuring the voltage on the HV winding

Subject transformer was a 15/30 NST by Transco. With my DMM and HV probe, I measure 15,231V no-load. I'm going to accept that as "correct".

To test method #1 above, I applied line voltage (119V from my wall socket) to the 15kV leads on my NST. I then measured .810V across the 120V leads. This suggests a transformation ratio of 119V/.810V => 146.9:1. If this were correct, then it would mean that applying 120V to the 120V winding should produce about 17.6kV. That's about 15% high (again, taking my 15.2kV measurement with the DMM as correct).

To test method #2 above, I applied 1.00V from my variac to the 120V leads on my NST. I then measured 104V across the 15kV leads. This suggests a transformation ratio of 104V/1.00V => 104:1. If this were correct, then it would mean that applying 120V to the 120V winding should produce 12.4kV. That's almost 19% low. Terrible! I'd actually expected this trick to be closer, but it's by far the worst!

So, again, because you aren't magnetizing the cores fully with these simple tricks, your accuracy is going to SUCK!

The voltage divider methods suggested above sound serviceable, if you have the parts and want to spend time building one. But to me, an HV probe or another HV transformer in reverse seem by far the more convenient. Who has just *one* NST anyway?

Cheers,
Aaron, N7OE

Ryan Arlis

Sun Jun 10 2007, 04:52AM

Registered Member #577 Joined: Sun Mar 11 2007, 11:48PM
Location:
Posts: 26

Im surprised that none of you suggested a VOM!!!

J. Aaron Holmes

Sun Jun 10 2007, 05:04AM

Registered Member #477 Joined: Tue Jun 20 2006, 11:51PM
Location: Seattle, WA
Posts: 546

Ryan Arlis wrote ...

Im surprised that none of you suggested a VOM!!!

I suggested using a DMM.

DMM and VOM are often used interchangeably. VOM generally implies the ol' needle-type units, though, whereas the "D" in "DMM" means "digital", which is what many (if not most) of us are probably actually using. That's a guess, though. I'm sure a few people will chime in and tell us how they're still using their dad's old Heathkit VOM from the middle of the last century

I can't remember what I paid for my Fluke 40kV (DC) probe. It was a non-trivlal sum, though. I suppose the assumption was that, if you asked this question, you probably didn't have one and were looking for more homebrew-type techniques.

EDIT: In fact, really everybody has thus far suggested using a meter (DMM or VOM), just indirectly (see all instances of the word "measure"). The average meter can't be subjected to more than 600V or so without frying, though. When you talk about measuring HV, the meter is usually implied, and discussion will center on how to effectively extend its range, which is what everybody so far has been talking about.

Regards,
Aaron, N7OE

DrZoidberg

Sun Jun 10 2007, 07:29AM

Registered Member #350 Joined: Mon Mar 27 2006, 05:14PM
Location:
Posts: 106

I also think that applying line voltage to the secondary is the best method to find the ratio.
And the reason why the voltage you measure at the primary is too low is not because of poor magnetization. It's because the secondary coil has a resistance over which there is a voltage drop. But you can compensate for that.
The impedance of the secondary can be calculated as
Z = (R^2 * X^2)^0.5
X is the reactance of the coil which is
X = 2 * pi * f * L
f is the frequency, L is the inductance

Only the voltage that drops over X is stepped down. The voltage that drops over R is lost.
But it's not correct to say that Vr + Vx = Vtotal. That's wrong.
As you can derive from above equations
Vtotal = (Vr^2 * Vx^2)^0.5

So to calculate the transformation ratio you need to get Vx - i.e. the voltage that drops over the reactance X of the secondary coil - and devide it by the voltage you measured at the primary.

For that you first measure R, i.e. you simply measure the resistance of the secondary with your DMM.
Next you connect the secondary to line voltage and measure the current.
Then you can calculate the inductance Z.
Z=V/I = 110V / I

Then you can insert Z and R into the equation
Z = (R^2 * X^2)^0.5
to get X

And then the voltage that drops over X is simply
Vx = I * X.

The transformation ratio is then
Vx / Vprimary

Of course that still doesn't tell you the output voltage of the nst.
Because this time you have to factor in the resistance of the primary coil.
If you connect the primary to 110V only the voltage that drops over Xprimary actually gets transformed.
Xprimary = Xsecondary / (turns ratio)^2

I think I will leave out the rest of the calculation.

sparky

Sun Jun 10 2007, 08:03AM

Registered Member #530 Joined: Sat Feb 17 2007, 07:56AM
Location: Victoria BC, Canada
Posts: 178

Dear Mr. Holmes,
I have been using my reverse voltage transformer method for years, and to all a great success. So I ask you to try again... you could try 240V if you like then simply divide the output voltage by a factor of 2. What ever works for you.... inaccurate? Bogus? Not a chance...pros use this method too when stuck without a high voltage probe.

I have a PT with an unknown ratio made by Westinghouse, just got it today from a junk dealer - and tonight I did a reverse feed with 120V into the high side. I got a reading of 0.798 V on primary--- so I believe this transformer to be a 150:1 PT ---oh yeah and just to be on the safe side I found a code stamp on the case - guess what, I pulled up the specs on this transformer and success
I have a Westinghouse 18000V - 120V PT. ;)

Bored Chemist

Sun Jun 10 2007, 09:00AM

Registered Member #193 Joined: Fri Feb 17 2006, 07:04AM
Location: sheffield
Posts: 1022

I'm surprised that nobody suggested measuring the voltage with a sphere gap. OK, it's not the last word in accuracy but it's cheap and robust.

Anyway, in the light of Dr Zoidberg's post I wonder if Mr Holmes would be kind enough to let us know the DC resistances of the 2 coils of the transformer he tested so we can see how much better the results get when the DC resistance is taken into account.

J. Aaron Holmes

Sun Jun 10 2007, 01:18PM

Registered Member #477 Joined: Tue Jun 20 2006, 11:51PM
Location: Seattle, WA
Posts: 546

sparky wrote ...

I have a PT with an unknown ratio made by Westinghouse, just got it today from a junk dealer - and tonight I did a reverse feed with 120V into the high side. I got a reading of 0.798 V on primary--- so I believe this transformer to be a 150:1 PT ---oh yeah and just to be on the safe side I found a code stamp on the case - guess what, I pulled up the specs on this transformer and success
I have a Westinghouse 18000V - 120V PT. ;)

The question is about NSTs, not PTs. I'll be happy to try a PT and reply back, although that would be somewhat OT

) You may get better accuracy with a PT because you don't have such a high-DC-resistance secondary dragging you down. I have no reason to believe that repeating my NST experiment will produce different results, so I stand by my comments about this approach being no good for NSTs (at least the "simple" version; see DrZoidberg's post)

DrZoidberg wrote ...

I also think that applying line voltage to the secondary is the best method to find the ratio.
And the reason why the voltage you measure at the primary is too low is not because of poor magnetization. It's because the secondary coil has a resistance over which there is a voltage drop. But you can compensate for that.
<snip>

Hmmm..."voltage drop" is a semi-unintuitive concept here; isn't the voltage drop across the secondary always whatever voltage you apply? Maybe I misunderstand, but factoring the DC resistance component out like you're doing seems very strange to me. I'd really like to see this approach tried.

EDIT: Thinking about this over my morning coffee just now, I guess it makes a little more sense to me. I'll have to try it tonight and see what it produces. Not sure if the same phenomenon would be responsible for the dreadful accuracy of the "1V-on-primary" approach, though. Hmmm...

Cheers,
Aaron, N7OE

DrZoidberg

Sun Jun 10 2007, 03:12PM

Registered Member #350 Joined: Mon Mar 27 2006, 05:14PM
Location:
Posts: 106

The primary has a much smaller resistance but then again it also has a much smaller inductance. Could be that the relation between dc resistance and inductance is better there. I don't know.

I used those equations some time ago for small 0.1 - 0.5 VA transformers and it seemed to work quite well. They were 230V to 10V and their dc resistance is very high so it needs to be factored in. That's also why their winding ratio is smaller then what you would expect. A 230 - 10V transformer might have a winding ratio of 1:20, 1:18 or even 1:15 depending on its resistance to compensate for the voltage drop.

You can think about it the following way.
Let's say you have a transformer with superconducting coils. When you apply a voltage to the primary the magnetic field in the core will start to increase. This induces a voltage in the primary and the secondary that is proportional to the rate at which the field increases. The field strength will now increase exactly at the rate necessary to induce a voltage in the primary that is equal to the one you applied. Thats why, if the secondary has the same number of windings as the primary, the voltage induced in the secondary will be equal to the one applied at the primary.
But if the primary has a dc resistance things look different.

Now the rate at which the field strength increases will only be so high that the voltage induced in the primary will be equal to the voltage applied, _minus_ the voltage drop over the dc resistance of the coil, which depends on the current.
So therefore V = Vx + Vr
The equation is not quadratic this time because I was assuming you apply a dc voltage to the coil like e.g. you do with a flyback (I know you don't call it dc because it's only applied for a fraction of a second, but for that small time it's a constant voltage and that's what matters here). When you apply a sine wave ac you get a phase shift and then you have to use the quadratic form V = (Vx^2 + Vr^2)^0.5

Hazmatt_(The Underdog)

Sun Jun 10 2007, 06:51PM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

You guys do this the absolute hardest way, you know that right?!

120V from wall ---> 12 VAC Center tapped transformer, so you have 6VAC or 12VAC, take the 6VAC first ----> Primary of your NST.
Measure the secondary bushing of the NST with your meter set to 600V or higher. Your probes go to the bushing and the case.

Write the ratio down, multipy by input voltage (120), DONE.

We're using a step down transformer in this case because you may not have a variac, some people don't. You could use a variac too set below 10VAC.

DrZoidberg

Sun Jun 10 2007, 07:31PM

Registered Member #350 Joined: Mon Mar 27 2006, 05:14PM
Location:
Posts: 106

You're right. That would be easier of course.

Now that I think about it you only need to factor in the dc resistance when you are using the nst in reverse.
When you connect the voltage to the primary you also lose voltage at the dc resistance of the primary but the percentage you lose is independant of the voltage. So if you e.g. lose 10%. That's 11V at 110V and 1.1V at 11V input voltage.
So if the winding ratio is 1:100 and the voltage drop at the resistance of the primary is 10% without any load at the secondary you have 9,900V output at 110V input voltage. And 990V out at 11V in. So it doesn't matter.

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob