Take the 2-minute tour ×
Electrical Engineering Stack Exchange is a question and answer site for electronics and electrical engineering professionals, students, and enthusiasts. It's 100% free, no registration required.

I am looking for a simple way to encode some data over a dc power line. (Probably using a very low cost IC) In my design, I have a port that provides 5V, 200mA for connecting some device, but now I like to extend the capability of my port to send and receive some data over that port. The idea I have is to encode data over 5V line. The distance the signal needs to travel is in the order of 10-15cm. Never more than this.

I have no data rate requirement but If I could achieve a 1-2Mbps it would be fantastic. If this is not possible, I am OK with a lower transmission rate. This would be a closed loop system, i.e. I control both ends. Therefore, I don't need to worry about standards etc.

share|improve this question
    
How far from ideal your DC supply is tolerant? Because any communication will require that your DC is not DC anymore. For instance: adding modulation would be seen as ripple in your "DC" (which is not DC anymore because it changes over time). –  Blup1980 Aug 4 at 10:51
    
What is your lowest speed data rate? –  EM Fields Aug 4 at 10:57
    
Maybe you can use something like this application note? ti.com/analog/docs/… –  rve Aug 4 at 18:34
    
@rve this is a great paper and leverages the idea that Andy Aka proposes. –  Ktc Aug 5 at 14:03

1 Answer 1

up vote 11 down vote accepted

You could think along these lines: -

schematic

simulate this circuit – Schematic created using CircuitLab

The data you send needs to be DC balanced i.e. something like Manchester encoded or scrambled. The amplitude of the data needs to be significantly smaller than 5V - maybe 1Vp-p.

You'll need a comparator/data slicer on the data out signal to recover the amplitude to logic levels.

L2 and C1 form a 1600 Hz low pass filter to smooth the DC recovered - this means your data rate is best as high as you can feasibly go and with Manchester encoded data there will be no content at DC up to the data clock rate - whatever frequency that is it ought to be at least 30 times higher than 1600 Hz i.e. approximately greater than 95kbits per second. If you can send M-encoded data at 1Mbps then that would be a lot better.

If your data is intermittent then send a preamble of 1s and 0s to get the data slicer aligned correctly with the data.

share|improve this answer
    
This idea is really neat. –  Dean Aug 4 at 14:35
    
One nice thing about this approach that might be worth mentioning is that it can work with half-duplex communication among any number of devices. Instictively, though, I'd be worried about the possibility of resonance if a load's demand varies periodically. Is that a danger, and if so how should it be mitigated? –  supercat Aug 4 at 15:44
    
Power Over Ethernet works in a similar way except that it's used to transmit 48V DC over data cables. –  slebetman Aug 4 at 16:08
    
It's worth mentioning that you might have voltage reference issues if you pass a lot of current (or if you have a long and very resistive cable). Everything is referenced back to your common ground. If you have a significant current flowing through the ground pin on that connector it can mean that your reference is a few volts higher at one end or the other. It is easy to work around and mitigate if you're aware of it up front, but can hurt you later on if your load current changes at some point while the system is operating and you don't expect it. –  Nick Aug 4 at 16:12
    
@supercat - it's exactly the same scenario in a buck converter but worse because, on a buck the supply (before the LC low pass) is switching between 0V and Vcc*2. Same rules apply - keep away from resonance both with the data and load effects. Luckily, as more load is applied, damping becomes greater so that this is not really a big problem. Another similar example is on an AC inverter - you choose output L and C to be significantly above 50/60Hz but not too high or you barely get rid of harmonic distortion - choose a natural frequency that is too close to 50/60Hz and there's a lot of smoke! –  Andy aka Aug 4 at 16:30

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.