Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Pulse transformer output measured vs. calculated?

Status
Not open for further replies.

Swend

Full Member level 4
Joined
May 14, 2019
Messages
237
Helped
10
Reputation
20
Reaction score
9
Trophy points
18
Location
Roskilde, Denmark
Activity points
2,304
Hi friends!

I have a 1000 ohm resistor connected to the secondary winding of my pulse transformer, and I'm measuring voltage and current which is all fine. But then I want to e.g. calculate the voltage by applying ohms law - then the calculated voltage is not even in the same range as the measured voltage, so there must be a problem somewhere, either my calculation, measurement and/or expectations? What do you think?

Figure_1.png

X-axis is nano-seconds...
 

While calculating Voltages and Currents with Pulse Transformer, how did you model it ?? Because Ohm's Law cannot be applied to Transient Domain in which is your case.
I mean, you cannot find Voltages and Currents by simply inserting Ohm's law ,instead you have to know Model of the Pulse Transformer and you have to use State Variable Technique to find Transient and Private ( Steady-State) Solution of Integro-Differential Equation of the circuit.
You may use a Transient Simulator to predict the response..
 
  • Like
Reactions: Swend

    Swend

    Points: 2
    Helpful Answer Positive Rating
While calculating Voltages and Currents with Pulse Transformer, how did you model it ?? Because Ohm's Law cannot be applied to Transient Domain in which is your case.
I mean, you cannot find Voltages and Currents by simply inserting Ohm's law ,instead you have to know Model of the Pulse Transformer and you have to use State Variable Technique to find Transient and Private ( Steady-State) Solution of Integro-Differential Equation of the circuit.
You may use a Transient Simulator to predict the response..

Actually it's a physical circuit, not a simulation. And of-course I was wrong applying ohm's law. Basically I just wanted to check if calculated would yield something close to measured in order to (perhaps) know if my measurements are way off, so how do you suggest I go about that?

And what do you mean by "model" of the pulse transformer?
 

Actually it's a physical circuit, not a simulation. And of-course I was wrong applying ohm's law. Basically I just wanted to check if calculated would yield something close to measured in order to (perhaps) know if my measurements are way off, so how do you suggest I go about that?
I suggest you extract a accurate model of the Pulse Transformer ( it might be simple or complex) then use a simulator because calculating Voltages and Currents of such a circuit will be very difficult even more erroneous.I believe the manufacturer of that Pulse Transformer has a more or less accurate model, request it..Then you can compare measurements and simulations to crosscheck the failing point..
 

I suggest you extract a accurate model of the Pulse Transformer ( it might be simple or complex) then use a simulator because calculating Voltages and Currents of such a circuit will be very difficult even more erroneous.I believe the manufacturer of that Pulse Transformer has a more or less accurate model, request it..Then you can compare measurements and simulations to crosscheck the failing point..

I hear what you are saying, but there are some issues with that. First is that what's on the primary side cannot be simulated, and secondly I'm the manufacturer of the pulse transformer which a simple air-core transformer.

But what if I measure the inductance of each winding and use LTSpice to couple two coils together, then apply the physical primary side waveform to the simulated primary? would that yield a usable simulation result on the secondary side?
 

I hear what you are saying, but there are some issues with that. First is that what's on the primary side cannot be simulated, and secondly I'm the manufacturer of the pulse transformer which a simple air-core transformer.
But what if I measure the inductance of each winding and use LTSpice to couple two coils together, then apply the physical primary side waveform to the simulated primary? would that yield a usable simulation result on the secondary side?

OK, if you are able to measure primary and secondary side of the transformer independently, you can use a "Lossy Transformer Model" that can be found in the literature and use it in LTSpice simulations.
 

Air core xfmr's are notorious for very high magnetising current loading the driving voltage ( i.e. pulling it down), and not so great coupling to the sec ...
 

Air core xfmr's are notorious for very high magnetising current loading the driving voltage ( i.e. pulling it down), and not so great coupling to the sec ...

Yes, could that be the reason that the voltage drops on both primary and secondary when I add more winding on secondary? I'm a little bit puzzled by that, could it have something to with the fact that I connect one side of the primary and one side of the secondary to 0V (GND), which I had to do because I don't have diff-probe yet?

Also I would like to have a non-air core, but I really can't find any suitable power cores that can do 1MHz.
 

I have a 1000 ohm resistor connected to the secondary winding of my pulse transformer, and I'm measuring voltage and current which is all fine.
How do you measure it? Sure that the probes don't affect the circuit? The question isn't clear at all.
 

How do you measure it?

For voltage, a resistive 1000:1 voltage divider 100M/100K. For current, a Rogowski coil.

Sure that the probes don't affect the circuit?

I'm sure that it does, but I use the same measurement technique on the primary, so I'm hoping that any measurement errors are the same on both sides, as I'm satisfied with just a approximate primary:secondary average power ratio.

The question isn't clear at all.

I have made measurements using one approach (see post #1) and I would like to use an alternative approach and compare results. So if both results are in the same ball-park I would have more confidence in the first result. Or if there is another way to do it, I'm all ears.
 

calculate the voltage by applying ohms law -

In response to a voltage pulse, the current increases linearly; that is clearly seen in the graph.

In both the plots, I see considerable high frequency ringing. Both voltage and current plots show the same effect (also same frequency).

I presume you are measuring the voltage on the primary side and the current on the secondary side using the same (or identical) probe.

Your probe is not responding well to the square pulse. You can use Ohm's law for the 1K resistor only (not for the pulse transformer).
 

There is no square pulse anywhere, what you are seeing is the pulse.

I am not quite sure. How are you generating the pulse?

The waveform seen is typical ringing process (it hit a bell and it produces sound that slowly dies in intensity).

I still suspect that the primary cause is a square (or rectangular) pulse that hits a resonant circuit. You catch that in the scope but you cannot see the original pulse.

The current curve is on a slope but why? But that is the typical response of a voltage pulse.

Is your probe on a DC coupled mode? It looks you are using AC coupling.
 

For voltage, a resistive 1000:1 voltage divider 100M/100K. For current, a Rogowski coil.
Is it a similar divider as in your other thread?

divider.jpg

If so, I doubt that the voltage waveform isn't reproduced correctly. Similarly, what kind of Rogowski coil and integrator circuit are you using? Did you calibrate voltage and current probes with a test generator?
 

If so, I doubt that the voltage waveform isn't reproduced correctly....

Once you use a voltage divider, your bandwidth is also getting scaled down in the same ratio, right? Similar to gain-bandwidth product

Plus, frequency compensation should be done with this divider in place, because the big resistor adds more capacitance. Sometimes the small adj cap will be at its end and external compensation has to be done.

I too think that there are problems with the measurements.
 

With an estimated capacitance of 1 pF for the 100 Mohm resistor, you get a zero at 1600 Hz. A single 1 nF compensation capacitor can at least improve the transfer function.
 

I am not quite sure. How are you generating the pulse?

It's a commercial thing, if I were to discuss that in public, it would become public domain and then it wouldn't be commercial anymore, I'm sure you can appreciate that, I'm not trying to be a smarty-pants. But what I can say is that if you use a resistor instead of a reactive component, the ringing will gradually diminish with increasing resistance, until about 1.5Kohm when the pulse becomes a smooth exponential decay. But a smooth exponential decay does not produce the desired effect, it has to ring as much as possible as long as it does not become pure AC. So that's why I'm pretty sure there is no square anywhere to be seen.

- - - Updated - - -

Is it a similar divider as in your other thread?

If so, I doubt that the voltage waveform isn't reproduced correctly.

Yes. And I will be checking it as you said, always meant to get around that but other stuff that I needed to learn was piling up so I kind of repressed it and just put it on the todo list.

Similarly, what kind of Rogowski coil and integrator circuit are you using? Did you calibrate voltage and current probes with a test generator?

Well similarly that was also a steep learning curve, after reading a dozen of scientific papers on rogowski coils I decided to roll my own. And I still have doubts about my calculations, as I found that there are no two papers that agree or use the same calculation methods, there is always a small (but significant) difference in the way it is being calculated. But I was content with two identical coils, because then I could at least measure a relative difference, which was my first objective. And only when that objective was reached (which was several months ago) I would review my calculations and have it calibrated against a known current, but it's still on the todo list.

So I use my own home-rolled coil, connect it to the scope probe x10, dump the trace to a file and then run it through a script which does the integration and other stuff. I found that it would be too cumbersome to make an active integrator and it would still be error prone as I have nowhere to calibrate it.
 

No, I do not wish to know the internal details. I just want to know if you substitute a resistor instead of the pulse transformer, what kind of waveform you will get? say you use a 1k or 10k resistor instead of the transformer primary.

But you need to make sure that your measuring instrument is not contributing to the ringing.

Better to use a series resistor of 1k and 1R and measure the voltage across the 1R to reduce the applied voltage.
 

No, I do not wish to know the internal details. I just want to know if you substitute a resistor instead of the pulse transformer, what kind of waveform you will get? say you use a 1k or 10k resistor instead of the transformer primary

if you use a resistor instead of a reactive component e.g. the transformer, the ringing will gradually diminish with increasing resistance, until about 1.5Kohm when the pulse becomes a smooth exponential decay.
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top