-1

I have few questions about transformer operation. One of the main points of the transformer is ground isolation.

Let's say if have a transformer in an ethernet transmitter, and my circuit in the primary of the transformer runs at 5 volts.

First quesiton: What is going to be the voltage on the secondary terminals of the transformer?

Second question: Will we be also transmitting a ground reference to the receiving end? Or will it will just be a differential signal?

Third question: Do we always need a reference voltage when we send a differential signal? If not, then how does the receiving end know the reference voltage?

Diego C Nascimento
  • 1,022
  • 1
  • 13
  • 26
Rudy01
  • 481
  • 1
  • 6
  • 11
  • *Transformers* are electrical devices. *Transforms* are mathematical constructs, often to convert values in one coordinate system to another. The fact that you got it right in the title means you know the difference. -1 for sloppiness. – Olin Lathrop Jan 10 '14 at 19:15
  • 1
    The reference for a signal in a differential pair is the other signal. That's why we use them in the first place. – Ignacio Vazquez-Abrams Jan 10 '14 at 19:18
  • Does this mean that we can have unit A have a different ground, and Unit B has a different ground, and we can send a differntial signal from A to B, and the signal can be interpreted correctly by unt B, even though it has a different ground? – Rudy01 Jan 10 '14 at 19:28
  • @Rudy: Yes, that's the point of differential signals. Transformer isolation is a way to create differential signals with very good balance and high common mode range. This is exactly how regular 10 and 100 Mbit/s ethernet using the common RJ-45 jacks works. – Olin Lathrop Jan 10 '14 at 19:38
  • I just think you know it only pass AC signals. Anyway, for example, auto-transformers does not archive ground isolation, this is not the main point of a transformer. Major types of transformer from the electrical generator to your home are auto-transformers. – Diego C Nascimento Jan 10 '14 at 21:10
  • I see, Thanks. But I am still not sure why we use it in the case of ethernet? Because my impression from this discussion is that we are talking about ground isolation, and if coming out of ethenet phy is already differntial, then it is already isolated !! Then why do we want to isoalte an already isoalted signal? – Rudy01 Jan 10 '14 at 21:14
  • Well I think I get your question. You are thinking that because the signal is differential it does not need isolation. Well without galvanic isolation this is not true. Read this http://en.wikipedia.org/wiki/Differential_signaling also RS-485 http://en.wikipedia.org/wiki/RS-485 does not imply the use o galvanic isolation but normally uses a third wire as a ground reference. – Diego C Nascimento Jan 10 '14 at 22:51
  • But, could you read my comment above about having independent ground for two seperate units with a differntial signaling? But I thought the answer to this question is that differntial signaling works with no reference needed! – Rudy01 Jan 10 '14 at 23:02
  • From the RS-485 article that I said you to read: "Connection of a third wire between the source and receiver may be done to limit the common mode voltage that can be impressed on the receiver inputs." That is without galvanic isolation, like a isolated transformer. – Diego C Nascimento Jan 10 '14 at 23:31

1 Answers1

4

What is going to be the voltage on the secondary terminals of the transformer?

If the transformer has a 1:1 winding ratio, then the primary voltage equals the secondary voltage at all times (neglecting leakage and other non-ideal stuff). In the more general case, the voltage ratio between the primary and secondary sides is equal to the the winding ratio.

Will we be also transmitting a ground reference to the receiving end? Or, it will just be a pair of differnetial signals?

If the receiving end is connected to the secondary, and you don't otherwise connect the secondary to ground, then there's no ground reference. That's often the point of using transformer-coupled signaling.

do we always need a reference voltage when we send a differential signal? If not then how does the receive end knows the reference voltage?

The "reference" is the other half of the differential pair.

schematic

simulate this circuit – Schematic created using CircuitLab

Here, VM1 will mirror V1, even if the two sides are connected to different grounds. The ground noise, which might represent natural variations in Earth potential between two distant points, has no effect on the signal. For signal transmission across appreciable distances, this is a really big win.

I've drawn it here with two transformers to make it symmetrical, and so you can clearly see how there's a loop of wire in the middle with no connection to anything else, but which caries the signal anyway. However, even with one transformer, the ground noise is still isolated, although each side is referenced to its own, different idea of "ground".

Phil Frost
  • 56,804
  • 17
  • 141
  • 262
  • +1, but it might be good to point out that *two* transformers are not necessary to achieve isolation between V1 and VM1 in your diagram. – Olin Lathrop Jan 10 '14 at 20:25
  • Thanks. This makes sense. But then if coming out the ethenet chip is already differntial, why can't we just connect it directly to the receive end. Becasue regardless of how the groudning is, since we have a differential signal, we can always get the data, correct? Then what is the main purpose of transformas in this case? – Rudy01 Jan 10 '14 at 20:57
  • @Rudy01 I don't know ethernet specifically well enough to answer that. Also check out this [recent question on balance for audio](http://electronics.stackexchange.com/q/95803/17608). Ethernet has the same problem, and I bet it uses transformers to make the balanced pair, because transformers do it better than most other solutions. – Phil Frost Jan 10 '14 at 21:15
  • @PhilFrost: You pointed out to a nice article, which sorts of explains. But now I got even confused on the other article, because I still don't think that "hot" and "cold" need to be flipped of each other. As long as we can assure they are treated equally, and ending up in having the same impedence, I think it is totally fine :) – Rudy01 Jan 10 '14 at 22:13
  • @Rudy01 that's exactly right. Of course, if you have two transformers like this, then it doesn't matter if you think of one as flipped, or one as zero, because there's no ground. There's only the other half of the pair. – Phil Frost Jan 10 '14 at 22:19
  • @PhilFrost: I posted a question at the comment section in the other post, could you take a look at it please? – Rudy01 Jan 10 '14 at 22:26