Impedance matching

[antknee]

I understand the theory of impedance matching but the practicality of it i'm still struggling with.

I want to put the signal from a signal generator into a transformer and step up the voltage.

The system would be:

Sig gen (output 10V AC, 50 ohms, 0.5amps, 100KHz) ---> transformer primary (10uH, impeadance 6 ohms at 100KHz ) --> transfromer secondary (1500uH) ----> R_load

How do i get the 10V across the transformer primary?

I tried putting a 50 ohm resistor in series with the signal generator output but then the voltage just appeared across the resistor. I tried putting the 50 ohm resistor in parallel with the transformer primary and the voltage was tiny.

What is the simplest way to impedance match the signal generator with the transformer?

Thanks.

Antknee.

 

[Hero999]

You need an amplifier to get 10V across the primary.

Your idea won't work as the signal generator can't output enough power.

Assuming it's 10V RMS (not peak or peak to peak) the maximum power transfer is 10²/(2×50) = 1W

 

[antknee]

So the events would be something like below (the numbers wouldn't be exact or perhaps even close) ???

Sig gen @ 10V ---> amp @ 100V ----> transformer primary @ 3V ----> Transfromer secondary @ 300V

 

[Hero999]

I forgot to say that the voltage will drop when a load is connected because the output of the signal generator looks like a voltage source with a resistor connected in series.

The maximum power is transferred when the load impedance is equal to the source impedance which is 50R and the voltage drops to half the initial value which is 5V.

I made a mistake before when I said the maximum power transfer is 1W, it's 5²/50 = 0.5W

 

[antknee]

I think I get that!

The sig gen can be modelled with 10V AC and 50 ohms in series. I did try the circuit in LTSpice but didn't know what to do with the output impedance, so it wasn't much help.

I get Z source = Z load for best power transfer but I don't understand why the best value would be 5V?

Thanks,

Antknee.

 

[antknee]

Oh I see. I meant +10V to -10V. Perhaps you thought it was +5 to -5. The terminology is confusing to me sometimes.

Cheers.

Antknee.

 

[Hero999]

That's 10V peak, the principle is still the same, if the load impedance is equal to the source, the voltage across the load will be halved.

Picture a potential divider, it's simple Ohm's law.

 

[antknee]

I have dug up some theorem which explains the halving, understanding it would take too long, so i'll just take the potential divider as the way to think of it. Thanks!

 

[Warpspeed]

o/k, your signal generator is fifty ohms, and to keep the magnetizing current of the transformer anything like acceptable, (say ten percent), the primary inductive reactance will need to be at least 500 ohms.
It will still then draw 50mA even with no load on the secondary.

At 100 Khz, the primary inductance must be at least 796uH minimum,
The primary must operate well below resonance, which means few turns (to reduce stray capacitance) on a high permeability core.
The core needs to have low loss at 100 Khz, and operate well below saturation flux density.

Given your figures, the secondary has 150 times the inductance of the primary, or maybe 120mH.
That will be quite a challenge getting that much inductance and keeping the secondary self resonant frequency well above 100 Khz.

But if you do manage it, and the primary to secondary coupling is o/k, the output voltage should be around 120 volts rms.

I wish you luck doing this, the transformer design, particularly the secondary winding will be a real challenge.

 

[antknee]

Hi warpspeed,

So it is possible just difficult. I will knock this avenue on its head. I think sig gen --> amp --> transformer is the best route. I do have an amp that i'm going to use but i'm waiting on the power supply, in the mean time I thought I would investigate this. Transformers look so simple but aren't!

Thanks.

 

[Space Varmint]

You need an amplifier to get 10V across the primary.

Your idea won't work as the signal generator can't output enough power.

Assuming it's 10V RMS (not peak or peak to peak) the maximum power transfer is 10²/(2×50) = 1W

That's what it sounds like to me exactly. Everything Hero said. Might I suggest a Darlington amplifier. What that will do is provide a stiff current source. I will not load down the signal generator and it will step down the impedance to your transformer.

edit*

Try 100K for R5 & R6 and about 470 for R7.

 

[antknee]

Hi Space Varmint. I have an audio amp I'm going to use. I bought it as a kit because the chip on it has a good gain bandwidth product (LM3886). I can't do much with it while i'm waiting for a bought power supply. I think Hero has hit the nail on the head too.

Thanks.

 

[Space Varmint]

Hi Space Varmint. I have an audio amp I'm going to use. I bought it as a kit because the chip on it has a good gain bandwidth product (LM3886). I can't do much with it while i'm waiting for a bought power supply. I think Hero has hit the nail on the head too.

Thanks.

Oh, you need to speak with Audioguru. But I don't believe you need a transformer. Let me cut a piece of schematic out that I use for a related chip that works quite well. Be right back...

 

[Space Varmint]

This circuit I used for a very simple regenerative receiver and I had plenty of audio. The diode is where I rectified the rf output for the audio. You can get data sheet here: LM3886 Datasheet catalog

 

[Warpspeed]

Hi warpspeed,

So it is possible just difficult. I will knock this avenue on its head. I think sig gen --> amp --> transformer is the best route. I do have an amp that i'm going to use but i'm waiting on the power supply, in the mean time I thought I would investigate this. Transformers look so simple but aren't!

Thanks.

An amplifier is not going to help you, because the transformer itself is the problem. It is just not capable of working at 10v input.

Here is a thought experiment for you Antknee.

Suppose I have a small six volt to sixty volt transformer, (ten to one step up ratio). It works fine at six volts input.

But I want to run it at 100 volts input, and get 1,000 volts out of it.
When I do that, it draws huge input current, looks like a short, and blows a fuse.

So maybe what I need is a power amplifier that will give me 100v at maybe five hundred amps of current so it can drive my little transformer.
That is just going to blow up that transformer.

Your problem is not impedance matching, or having enough input drive power, but the transformer itself is just not capable of working at the voltage and frequency you want it to work at.

There are a couple of ways to do this, depending on what it is for, but the transformer you already have is just not suitable.

 

[antknee]

This transformer isn't up to the job, it was an interesting exercise but not something i will pursue. While reading the above comment the film Jaws came to mind. I was trying to catch a big shark with a little fishing rod. It was never destined to work, I needed to blow the shark up. Or a much bigger fishing rod!

Thanks for your comments.

 

[Hero999]

This is for your piezo driver project isn't it?

Why not keep it to one thread that way your requirements will be more clear and we won't have to repeat ourselves?

 

[antknee]

Hi Hero,

Yes, I only have one project, it is to make a piezo driver. I did stick to one thread originally but the project has a number of different areas so I began posting those areas when interest in that thread waned. I posted about transformers and impedance matching separately yesterday and they both ended up being about the same thing. Apologies if you repeated yourself.

 

[antknee]

I will return to the thread below tomorrow. I should finally have my LM3886 amp working tomorrow so I will be able to test the 4-5 transformers I have and see if I get anywhere. I don't think they will be any good, so I'll make the transformer mentioned in that thread and post what goes wrong there...!

https://www.electro-tech-online.com/threads/check-my-transformer-spec.105788/