Transformer Impedance

[mmuj]

Hi

my simple yet confusing question (at least to me) is, can audio transformer with 600 ohm primary and 600 ohm secondary winding be used in 50ohm primary 50ohm secondary circuit ?

little about back ground, I have to tx/rx data over 30 to 50 ohm coaxial cable using interface that is actually designed for 600 ohm line. I will change op amp output impedance by manipulating resistor & capacitor but do i need to change transformer as well ?

Actual transformer user is Midcom 82111 which is designed to reflect 600 ohm on primary(line side) with 470 ohm load on secondary (op amp side)

I asked similar question some time ago (search for "Modem coaxial cable interface") but i was diverted to do something else and now i am back to same problem so rather than cotinuing with the same thread i started new one

I will come to other aspect (not related to this problem) later



Thanks in advance

mmuj

 

[goldsmith]

Hey Gentleman!
Hi
See this famous formula:
(N1/N2)^2=R1/R2
don't forget that these transformers can not use in higher frequencies.
I hope that this become helpful for you.
Goldsmith

 

[mmuj]

Well i know these two facts

(N1/N2)^2 = R1/R2 shows how impedance is reflected from one side to the other. in my case as ratio is 1:1 so there is no impedance translation i.e. will remain same.

Now the second tip, "transformer can not use in higher frequencies". I already mentioned that it is audio frequency i.e. 20Hz to 20kHz. to be more precise it is 300Hz to 5kHz.

but how these two facts answer my question ? probably not


thanks anyway

 

[FvM]

Strictly spoken, the 600:600 ohm specification tells two things:
- the transformer has 1:1 windings ratio
- you can expect suitable behaviour in audio frequency range with 600 ohm circuits

It does not tell about the behaviour with 50 ohm circuits. If you look to the datasheet, the about 100 ohm primary and secondary windings resistance can't be ignored. In addition, you should expect a leakage inductance, that results in additional attentuation at higher frequencies. All in all, the transformer is badly suited for the intended purpose, I think.

 

[Raza]

(N1/N2)^2 = R1/R2 shows how impedance is reflected from one side to the other. in my case as ratio is 1:1 so there is no impedance translation i.e. will remain same.

What about the signal transfer ratio. Will the same signal strength be kept in case of 50Ω and 600Ω?

 

[mmuj]

Thanks for reply

I agree with FvM, high winding resistance can not be ignored. (simulation verify that as well)

what i have understood by now is that i need transformer with low winding resistance.

one point if some one can explain, i have noticed that almost all modem transformers have inductance in the range of 4H. what could be a reason of that?

do i need a transformer with high inductance for my interfacer as well ?

to answer raza's questions, intention is to keep same power level at the receiving end. as coaxial cable is more than 10km long so high voltage n low current should be better than reverse of it.

cheers,

mmuj

 

[FvM]

In your previous thread, we suggested to use a 600:50 ohm transformer with the original modem circuit. Now you told, that the modem circuit has been modified to support 50 ohm matching. You didn't tell the details, there may be doubts if you considered parameters like maximum OP output currents in the redesign. Thus I think, the original suggestion still holds. On the other hand, a rewinded audio transformer with about factor 3.5 less turns and respective higher wire gauge is feasible of course. I used to make these devices as a school boy.

 

[Raza]

a rewinded audio transformer with about factor 3.5 less turns and respective higher wire gauge is feasible of course. I used to make these devices as a school boy.

Support it.

 

[mmuj]

I have wound transformers with 3.5 turn ratio with available core (toroidal electrical steel core with 100 turns on primary and 29 turns on secondary with 55mH and 4.6mH inductances and have started playing with it.

meanwhile, if someone can explain that if signal frequency results in a wave length which is low enough so that coaxial cable length is less than 1/4 of signal wave length. In other words cable length < 0.25*signal wave length. In this case do i need to worry about characteristics impedance of tranmission line ?

or should i ask this question in a seperate thread ?


Cheers

mmuj

 

[vimalkhanna]

For audio tx the winding self capacitance has no role to play.However , when we use this for MODEM (HF applications)this cap kills/attenuates the entire signal and virtually nothing passes to the coax. which is used to convey signals .It shall be better to have pri:sec isolation barrier as well as have intermediate barriers after every few turns of the primary winding .Better to use PTFE sheathed multicored wire .The isolation/self capacitance is lower between the windings .As such ,the HF response improves ..

 

[FvM]

meanwhile, if someone can explain that if signal frequency results in a wave length which is low enough so that coaxial cable length is less than 1/4 of signal wave length. In other words cable length < 0.25*signal wave length. In this case do i need to worry about characteristics impedance of tranmission line ?

Yes and no. You don't need to analyze cable in transmission line terms, but you need to consider the cable capacitance. I made a calculation in one of your previous threads: https://www.edaboard.com/threads/226391/#post965970

For audio tx the winding self capacitance has no role to play.

At least not in the present impedance range. There are audio transformers with kohm impedance in use, there capcitance definitely plays a role. For the present problem, we rather need to worry about leakage inductance. Depending on the windings layout, it may be a problem.

If you imagine a toroid core with one winding on the left side and the other on the right...

 

[mmuj]

modem frequency is 1.2kHz/2.4kHz.

is it applicable to this frequency as well ?

Thanks

 

[FvM]

If my calculation of a 700 Hz low pass is correct, the capacitance matters. The effective impedance is however 600 ohm/2 due to both side termination, so you get about 50 % level reduction at 2400 Hz. This most likely will still work.

The suggestion to connect the 600 ohm system as is to the 50 ohm cable has been already made in your previous thread, if I remember right.

 

[vimalkhanna]

My calculation is showing 800Hz lo pass which provides for 6dB attenuation.
The mod and demod signals at 2.4Khz shall get rounded off pass as a clipped (chopped)sine wave ..
Hopefully ,it functions .
One part is correct ...the self winding capacitance as well as leakage inductance shall play a major role in the line transmission .
The parity check for signal identification and trigger may /maynot be recd to sync the Modem data .

 

[mmuj]

Just to summaries what is been suggested so far by gurus (especially FvM)

use cmx 868 interface as it is i.e. 600 complex impedance termination or cmx side. Refer datasheet figure 4a (r13 is setting source impedance)
use standard 600:600 voice/data transformer with high inductance to minimize leakage current and leakage inductance. (one of the part i found to do this is z1656-4 from walters OEP)
coaxial cable of 5-8 miles can be considered as RC circuit. where R is coper losses, for given cable it is 25 ohm/mile, and C is 1.7nF/mile.

look the link for very simple rlc circuit to simulate frequency response, -3db point range is from 50 Hz to 13kHz.

If i have missed anything or understood wrongly please let me know

Regards,

mmuj