Maker Pro
Maker Pro

Sensor cable design

W

whit3rd

Jan 1, 1970
0
"whit3rd"
** Simply reducing source Z to 100 ohms did the trick for your case - but
it did not eliminate cable C.

Oops; omitted the other part of the solution. I also put a 100 ohm load on the
distal end of the cable, thus applying the full transmission-line treatment.
That was the easiest way to deal with cable C_stray (driving and driven ends both
terminated into the characteristic impedance).
 
R

RobertMacy

Jan 1, 1970
0
Oops; omitted the other part of the solution. I also put a 100 ohm
load on the
distal end of the cable, thus applying the full transmission-line
treatment.
That was the easiest way to deal with cable C_stray (driving and driven
ends both
terminated into the characteristic impedance).

From memory, the characteristic impedance goes through some wild
transitions determined by frequency. Something like 90 up to 120, or 120
down to 90, can't remember which.

Transferring energy either is no biggie, but transferring V(t) signal can
become important
 
R

RobertMacy

Jan 1, 1970
0
---
Generally, as long as the OD of the center conductor, the ID of the
shield, the concentricity between the center conductor and the shield
and the dielectric constant of the insulation between the center
conductor and the shield don't change, the impedance of the cable will
remain constant and will suffer only a frequency dependent attenuation
which increases as length and frequency increase.

I was referring to the twisted pair's characteristic impedance. However,
since you described a coax, your comment that the characteristic impedance
is constant is not true, unless you apply 'engineering' tolerances to
it...The characteristic impedance of a coax changes with frequency.

For a 'word argument': Characteristic impedance is approximately the
square root of the ratio of inductance per length over capacitance per
length. At low frequencies, the whole center conductor's cross section is
involved in carrying current and has a value of inductance per length. At
high frequencies, skin effect rears its ugly head and you see the
increased loss that you mentioned, but also the carriers are traveling
along the outside of the center conductor, which changes the center
conductor's inductance per length, thus the characteristic impedance of
the coax changes with frequency.
 
I was referring to the twisted pair's characteristic impedance. However,
since you described a coax, your comment that the characteristic impedance
is constant is not true, unless you apply 'engineering' tolerances to
it...The characteristic impedance of a coax changes with frequency.

At audio frequencies, the coaxial impedance can be several hundred
ohms. At LF to UHF it is quite constant close to nominal impedance.

At extremely high frequencies the cable no longer act as a coaxial
cable, but rather like a waveguide with different propagation modes
depending on frequency.
 
R

RobertMacy

Jan 1, 1970
0
At audio frequencies, the coaxial impedance can be several hundred
ohms. At LF to UHF it is quite constant close to nominal impedance.

At extremely high frequencies the cable no longer act as a coaxial
cable, but rather like a waveguide with different propagation modes
depending on frequency.

At 10kHz, an RG-58 cable is what?

At 10kHz, an RG-59 cable is what?
 
J

josephkk

Jan 1, 1970
0
I'm having difficulty finding a suitable cable for a wired sensor application, and I was wondering if anyone could check my design approach. Basically, my control circuit powers a wired sensor located about 2 meters away. The sensor circuit uses a single-ended op-amp to amplify a 5 MHz transducer signal, and then sends it back to the control circuit where it is digitized. So, I need 3 wires between the control circuit and the sensor: Vcc, ground, and signal.

Since the cables from multiple sensors will be running next to each other, I thought I should use twisted-triad wire to reduce cross-talk. External EMI shouldn't be a problem, so I don't think shielded cables arenecessary. However, I haven't been able to find twisted-triad wire without a jacket or shield. What kind of wire should I use for this? IfI used two twisted-pair wires for each sensor, what would I connect the extra wire to?

The sensor has a Vcc bypass capacitor, so I think most of the higher frequency content will be between signal and ground. Since I'm bringing back the signal and ground wires together, would this be considered a balanced circuit?

Thanks,
Doug

Cannonically, the two twisted pairs would be one pair for sensor power and
one pair for sensor output. Bond return/ground wires of the pairs at the
receiving point, and only maybe at the sensor.

?-)
 
Cannonically, the two twisted pairs would be one pair for sensor power and

one pair for sensor output. Bond return/ground wires of the pairs at the

receiving point, and only maybe at the sensor.

Was waiting for someone to bring that up about the GND bonding location...
 
F

Fred Abse

Jan 1, 1970
0
In order to measure the impedance at 10 KHz, you need a chunk roughly
30 kilometers long.

No you don't.

Open- and short-circuit impedance measurements on an electrically short
sample are all you need.

I've done it using my HP4274A, on samples a couple of feet long.

In earlier life, I used to use a Marconi TF1275 Q-meter.

Zo = sqrt(R+jX/G-jB) = sqrt(R+jwL/G+jwC) = sqrt(Zsc/Yoc) = sqrt(Zsc.Zoc)

G is so small, you usually can't measure it, unless you have the most evil
dielectric imaginable. It can be neglected.
 
F

Fred Abse

Jan 1, 1970
0
[quoted text muted]

At 10kHz, an RG-58 cable is what?

72.78 -j52.89, calculated from Belden 8262 published L, C and R figures.
At 10kHz, an RG-59 cable is what?

79.05 -j25.29 (Belden 8263)

Any transmission line that has resistive loss must have a complex
characteristic impedance, which is frequency dependent.
 
F

Fred Abse

Jan 1, 1970
0
[quoted text muted]

---
Generally, as long as the OD of the center conductor, the ID of the
shield, the concentricity between the center conductor and the shield
and the dielectric constant of the insulation between the center
conductor and the shield don't change, the impedance of the cable will
remain constant and will suffer only a frequency dependent attenuation
which increases as length and frequency increase.

Only true in the case of lossless lines. Resistive losses mean a complex,
frequency-dependent, characteristic impedance.
 
F

Fred Abse

Jan 1, 1970
0
Well, you measured the L, R, and C, in separate open/short measurements,
and calculated the impedance. My point was that impedance is a function of
frequency, and it's not meaningful, as a transmission line impedance, if
the line is a tiny fraction of a wavelength. There's no reason the OP
should worry about line impedance or termination. Capacitive loading of
the opamp could matter.

If I put a 10 foot coax into one black box, and the equivalent R, L, and C
into another black box, as lumped elements, you couldn't tell the
difference at 10 KHZ.

I have no disagreement with any of that. What I was commenting on was your
statement that:

"In order to measure the impedance at 10 KHz, you need a chunk roughly
30 kilometers long."

Which is patently untrue.
 
R

RobertMacy

Jan 1, 1970
0
[quoted text muted]

At 10kHz, an RG-58 cable is what?

72.78 -j52.89, calculated from Belden 8262 published L, C and R figures.
At 10kHz, an RG-59 cable is what?

79.05 -j25.29 (Belden 8263)

Any transmission line that has resistive loss must have a complex
characteristic impedance, which is frequency dependent.


that was my point with the question. I believe complex and more in the
range of 72 ohms stuff, not that kiloohms touted earlier.
 
R

RobertMacy

Jan 1, 1970
0
[quoted text muted]

---
Generally, as long as the OD of the center conductor, the ID of the
shield, the concentricity between the center conductor and the shield
and the dielectric constant of the insulation between the center
conductor and the shield don't change, the impedance of the cable will
remain constant and will suffer only a frequency dependent attenuation
which increases as length and frequency increase.

Only true in the case of lossless lines. Resistive losses mean a complex,
frequency-dependent, characteristic impedance.

Your method of quoting implies I said that statement. I did not. I refuted
it with a 'verbal' description of the complex relationshi between loss, Z,
and frequency.
 
R

RobertMacy

Jan 1, 1970
0
...snip...
If I put a 10 foot coax into one black box, and the equivalent R, L, and
C into
another black box, as lumped elements, you couldn't tell the difference
at 10
KHZ.

Actually you can, that's around 100 ppm effect shows up gangbusters in my
work. Actually a bit more since the velocity down a cable is less than
speed in air.
 
F

Fred Abse

Jan 1, 1970
0
You're playing with words. You didn't measure the impedance, you
calculated it.

Describe a practical method of determining the characteristic impedance
of a lossy transmission line, at a specific frequency, without calculation
from indirect measurements.

The use of specialized instruments, such as vector network analyzers,
(which invariably do the calculations for you), isn't allowed.
 
F

Fred Abse

Jan 1, 1970
0
Your method of quoting implies I said that statement.

No, it doesn't, the very first line says:

"On Wed, 09 Oct 2013 20:45:34 -0500, John Fields wrote:"

The insets and ">" quote characters say the rest.

I was replying to John, and his reply to me shows that he knew that.

Looks clear to me.

Maybe your newsreader doesn't honor the usual quote characters.
 
F

Fred Abse

Jan 1, 1970
0
According to your own words: "Well, you measured the L, R, and C, in
separate open/short measurements, and calculated the impedance."

How is that less valid than measuring the voltage across a conductor, then
the current through it, then calculating its resistance?

Or sampling voltage and current, multiplying individual samples, and
integrating to derive power?
 
F

Fred Abse

Jan 1, 1970
0
Actually you can, that's around 100 ppm effect shows up gangbusters in my
work. Actually a bit more since the velocity down a cable is less than
speed in air.

Quick and dirty simulation with a 50 ohm load on both network and line
shows a difference in amplitude of 0.053dB and phase difference of 0.113
degrees at 10kHz. Using Belden 8262 figures for cable constants, the same
values for the lumped network.

Difficult to measure. Yes, there *is* a difference, and it depends on load
resistance. Above about 1k, there's little difference.

John's assertion is close enough, the way he described it, for readily
measurable quantities.
 
R

RobertMacy

Jan 1, 1970
0
No, it doesn't, the very first line says:

"On Wed, 09 Oct 2013 20:45:34 -0500, John Fields wrote:"

The insets and ">" quote characters say the rest.

I was replying to John, and his reply to me shows that he knew that.

Looks clear to me.

Maybe your newsreader doesn't honor the usual quote characters.


Evidently not! Glad you were responding to someone else.

I just didn't want to leave the impression that I did not understand the
effect of 'skin effect' on characteristic Z, etc.
 
Top