I need to send serial data ( approx. 60kbs continuous ) over a non contact interface (basically a rotating pulley). I can transmit power to the device with a coil rotating inside another coil (inside diam 65mm) OK.

I have a PIC processing lots of maths to calculate the data I need to transmit.

I have researched lots of wireless charging and data transfer and have built circuits but have limited success in the back scatter data transfer. Not fast enough.

There are many RF modules to use but I will have many units in close vicinity so cross talk could be a problem.

In my design there will be 2 PCBs almost touching each other, but rotating. I keep coming back to a simple low power FM data transmission with a PCB circular aerial close to another with limited angle of same diameter.

No bi-directional data transmission is required. Any suggestions.

  • 1
    \$\begingroup\$ I think it would help if you included a sketch of your design if you have any. \$\endgroup\$
    – 0x6d64
    Feb 17, 2016 at 11:36
  • \$\begingroup\$ Can you explain more about how many units you have in close vicinity (presumably each transmitting 60 kbps?)? Do you have access to the centre of the "shaft"? What does " back scatter data transfer" mean? Do all "units" share the same power rails from the rotating transformer? \$\endgroup\$
    – Andy aka
    Feb 17, 2016 at 11:36
  • 3
    \$\begingroup\$ Your question doesn't really tell us about your requirements. What I managed to interpret is that you have a sensor in some rotating machinery, you need to get data back wirelessly from said sensor at 60 kbps, you do not need to transmit data to the sensor but you want to transmit power to the sensor wirelessly trough inductive coupling. Are my assumptions correct? If they are, have you considered transmitting the data with LEDs and receiving it with photodiodes? It would be electrically and mechanically simple but could susceptible to ambient light. \$\endgroup\$
    – jms
    Feb 17, 2016 at 11:40
  • 1
    \$\begingroup\$ I agree with @jms ... been there, done that. We transferred some tens of watts of power to a rotating device using a rotary transformer (two cup-shaped ferrite cores facing each other, each containing a coil) operating at ~25 kHz. We also transferred about 1 Mbps of data up the center of the hollow shaft using an IR LED and photodetector. Worked great! \$\endgroup\$
    – Dave Tweed
    Feb 17, 2016 at 12:50
  • \$\begingroup\$ @Andyaka I'm guessing his reference to 'back scatter data tranfer' was to using some method of powersupply modulation 'back' through the transformer (probably by using a 'dump circuit' to modulate the xfmr load similar to USB-OTG host-negotiation power-pin signalling). \$\endgroup\$ Feb 17, 2016 at 16:52

1 Answer 1

  1. For your power transmission, there are many effective ways to make a usable rotating transformer; my personal favorit for your described use would likely be to use 2 concentric coils over a single core (possibly even using the 'axle' as a core), but other methods exist & are effective.

  2. You stated that you have 2 PCBs 'almost touching each other,' with one rotating & the other stationary.

    • With this close proximity, I think that rf transmission adds a large amount of 'unneeded complexity' to your design (and as a HAM operator, I like using rf).
    • If you use a circular trace on at least 1 board, and at least a portion (arc) of the same circle on the other, you can simply transfer your data (at up to many mbit/sec) by capacitive coupling between the 2 traces.
    • In most applications we consider capacitive coupling to be 'parasitic,' but here it would allow you to transmit your 60Kbaud data using simple OOK (on-off keying) high/low voltage outputs using the circular traces as a 'pass-through capacitor' (or 'DC blocking capacitor') for your signal. No rf or optical elements necessary, and it would be incredibly difficult for other units to cause interferance.

NOTE: If you cannot use 'full circle' traces, your traces won't effectively ground any 'accidental inductive coupling' to your inductive power transfer coils. Keep your signal traces as physically distant from the power coils as possible, and consider filtering out the power coil frequency from any trace that cannot be made 'full circle' in your capacitive coupling.

  • \$\begingroup\$ I can make the 2 coils full circle if need be. The device will need to comply with noise emissions so the dV/dT need to be keep in check. I do not understand..... high/low voltage outputs using the circular traces as a 'pass-through capacitor' (or 'DC blocking capacitor') for your signal..... What type of "PCB track" drive circuit? \$\endgroup\$
    – Joe Ben
    Feb 18, 2016 at 11:30
  • \$\begingroup\$ @JoeBen (1 of 2) Your 2 circular traces would be forming the 'plates' of a parallel plate capacitor. To calculate their capacitance (and thus impedance vs. signal frequency), you can use this calculator: daycounter.com/Calculators/Plate-Capacitor-Calculator.phtml - as an example, using a 2" diameter & 1/8" trace width with your pcbs 0.04" apart would make a ~ 4.43pF capacitor. Then, using this calculator: qsl.net/pa2ohh/jslcimp.htm we find that this capacitance would have an impedance of ~600kOhm@60khz (or ~60kOhm@600khz) to your signal. \$\endgroup\$ Feb 18, 2016 at 15:08
  • \$\begingroup\$ (2 of 3) 600Kohm is a high impedance, so I'd recommend using wider traces if possible (3/4" wide traces with circle inside diamater of 2" and outside diameter of 3.5" would drop impedance to ~100Kohm @ 60KHz), or use a >60KHz 'carrier signal' (like an rf application; so adds some circuit complexity). To 'drive' this capacitor from the 'transmit' side would simply require outputting the data through an amplifier that has been designed around the signal impedance of the cap (imagine there's a same-impedance resistor in place of the capacitive traces), so for 600K if using the 2" by 1/8" traces. \$\endgroup\$ Feb 18, 2016 at 15:19
  • \$\begingroup\$ (3 of 3) Then, on the receive side, you'll want to use a 'voltage buffer,' or 'voltage follower' (en.wikipedia.org/wiki/Buffer_amplifier#Voltage_buffer), or a high input-impedance comparator/schmitt trigger to give a useful output current/signal for further processing. Here, luckily, since we can use a 'square-wave' signal (PWM on a 60Khz square-wave 'carrier' would be easiest to process), distortion in the amplifier stages is mostly a non-issue, so we don't have to worry about keeping the amps 'linear' (very important in rf tx/rx applications), which makes things much easier. \$\endgroup\$ Feb 18, 2016 at 15:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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