Author Topic: Use of twisted pairs with fundamentally single-ended DUT  (Read 918 times)

0 Members and 1 Guest are viewing this topic.

Offline ezalysTopic starter

  • Frequent Contributor
  • **
  • Posts: 329
  • Country: us
Use of twisted pairs with fundamentally single-ended DUT
« on: August 25, 2020, 05:34:49 am »
Let's suppose I'm interested in measuring the voltage response of a DUT to some imposed voltages. The DUT takes V1, V2, V3 as inputs relative to some GND, and outputs V4 relative to the same GND. All signals are below 100 kHz, and there's a 4 m run of wire. Is there any benefit in using four twisted pairs (V1, GND), (V2, GND), (V3, GND), (V4, GND), where for each twisted pair, one of the leads is bonded to ground, and the twisted pair is either stimulated (in the case of an input) or sensed (in the case of an output) by a differential circuit. The DUT can not be extended with differential drivers (it will be deep inside a cryostat at 300 millikelvin.)
 

Offline Berni

  • Super Contributor
  • ***
  • Posts: 5050
  • Country: si
Re: Use of twisted pairs with fundamentally single-ended DUT
« Reply #1 on: August 25, 2020, 06:22:01 am »
It depends.

If the voltages in question are very small then having separate twisted wires for each might be a lot of help. This avoids voltage drops in the ground wire causing offsets, minimizes the chance of accidental thermocouple voltages along the way (They still happen but are close together where the cable is connected as a pair) and reduces the magnetic loop area so that the cable is less likely to pick up noise.

With large clear signals you are probably fine at 100KHz using just normal cables as long as they are sensibly routed. But Using twisted pair never hurts.
 

Online KT88

  • Frequent Contributor
  • **
  • Posts: 334
  • Country: de
Re: Use of twisted pairs with fundamentally single-ended DUT
« Reply #2 on: August 25, 2020, 06:46:41 pm »
Idividual twisted pair connections will very likely improve the signal integrity for two reasons:
1. Having AC signals makes the Impedance of the signal return (aka GND) more relevant. It will be higher than just the resistance. Even the weakest signal will have some influence on this return lead no matter how massive it is. Sharing this return with another signal would superimpose these return signals from different channels.
2. Different impedances of signal- and return path would cause mode conversion: Common mode iterferences will cause different voltage drops for the signal- and return leads.
The remedy for these effects is in fact using individual twisted pairs. It doesn't matter (a lot) if the signals are fully differential or single ended. You only need a (very) good instrumentation amp (or in case of AC only a tranformer) to receive the difference while suppressing common mode noise. Connecting the returns at the receive side is a nono b.t.w.... In case of strong interferences a shielded twisted pair would be a further improvement.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf