Remote sensing with a switched mode power supply?

Go To Last Post
9 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello,

I figured I might ask on here before trying to reinvent the wheel. I am using a buck converter (4A average output load, LTC3835 controller) and want the voltage regulated at 5V across 6 feet of cable (3 feet on the power line, and 3 feet on the ground line). I don't have the option to use thicker wire, so I was going to run some sense lines out to the load.

I think the best approach is to use a difference amplifier to subtract the power and ground sense lines and feed that voltage to the feedback resistors. I used LTSpice to simulate a DC operating point using 5V and 1V into the difference amplifier and the output comes out to be 4V.

Does anyone have any experience doing this? Am I going to mess up the control loop by adding this external difference amplifier? Suggestions?

Thanks.

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You would need to account for the phase difference of your extended feedback circuit - you don't want your power supply to become an oscillator! :)

There are other LTC parts which might be a better fit. The LTC3833 comes to mind. It has a differential feedback input specifically for applications like this. No doubt digging through the LTC website may turn up some others.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
you don't want your power supply to become an oscillator!

And you need to know the dynamic properties of powered circuit it seems. Never designed such device, but it is a clever thing as you do not need to waste the copper.

Another thing is with a single regulator you need to add 2*I*R_Cu to power supply. Then remote ground will rise I*R above your local ground so make sure you do not short signal lines (if any).

And this single sensing works only when current flowing out is equal to current flowing in. But what if it is not, and a leakage happens?

No RSTDISBL, no fun!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Shall it be feasible, just run in two extra potential Vcc and GND wires and use these for error amplifier inputs.
IIRC this is Lead Resistance Compensation method. No extra parts except wiring in rough approach.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Shall it be feasible, just run in two extra potential Vcc and GND wires and use these for error amplifier inputs.
IIRC this is Lead Resistance Compensation method. No extra parts except wiring in rough approach.

That's a good idea but my load will vary from .5A to 3.5A so that wouldn't work, unless the Vcc / GND wires to the op amp also changed with load.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
load will vary from .5A to 3.5A so that wouldn't work, unless the Vcc / GND wires to the op amp also changed with load.

That will do work especially in such a case. Please, check figure 4 at:
http://www.omega.com/temperature/z/ResistanceElements.html
The OpAmp leads are for measuring Vcc at the load do not carry any current (in theory), thus the current lead losses will be compensated efficiently.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
thus the current lead losses will be compensated efficiently.

Your method is leakage-proof as this will not change output voltage then. But additional sensing wire(s) are needed.
I think my PC power supply has a shunt resistors to measure current through wires and it adds the voltage drop so that U is within range at the end of a known length wire. This does not require additional sensing wires. And that is the main advantage because in case of damaging any of the sensor wires the regulator gets open loop - I feel smoke. With shunt resistors everything is enclosed in regulator so it is less prone for damages..
So many factors need to be considered.

No RSTDISBL, no fun!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Some ATX power supplies do have a sensing wire. I think I have seen two orange wires, one as thick as other orange wires and one much smaller orange wire, that are crimped on the same connector pin. The whole feedback is usually not taken from the sense wire, but it should really keep the wanted voltage at the connector side of the wiring, and thus rise the voltage a bit on the PSU side of the wiring.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dont forget to take some precausions for when the cable gets broken.... and then later re-conncted while power is still applied. This might give high voltage problems.