Loss of signal on Dataline cable

I will say The 100 cm length will be OK - based on the photo of the cable.

Typically, the workable length of the cable depends on:
1) The amplitude of the data signal. A 15 Volts amplitude signal can traverse longer distances than a 5 Volt signal.
2) The Data Rate. A lower Data Rate signal can travel longer distances without errors than a High Data Rate signal.
3) Normally, a thinner gauge wire will cause more attenuation than a thicker one, but for the distances mentioned, the 22 AWG will should be fine.

If you can tell us the Data Rate and its amplitude, you can get more specific comments.

Hello,
Thank you very much, Ramussons.

I found the shematic of the control board, it corresponds with the photo (dataline plug in P2, top left of the picture):





Not sure of the signal amplitude, maybe 3.3V with both RX and TX signals.

I was told that this dataline cable is used to calculate the battery life, and to manage the power supplied to the motor; Can this give an idea of the data rate?

Is twisting the three wires enough to solve this question posed by the length of the extension cord?

I don't see any issue at all, twisted or not - you're only replacing it with a very short cable anyway.

The twisting is simply to keep the wires together and tidy, nothing else.

Twisted pair common mode rejection :

Noise signals may appear in a cable for various reasons. For example, noise can be capacitive coupled from nearby electric fields, inductively from local magnetic fields, electromagnetically from radio signals or conductively from circuit path leakages. However, when you have a twisted pair line, it intercepts the coupled signal equally, making the incident signals appear only as common mode signals. You have a balanced twisted pair line if there is identical impedance from each line to the local common.

Click to expand...

Regards, Dana.

danadak said:

Twisted pair common mode rejection

Click to expand...

That would apply with balanced signals, or to a lesser extent a signal & ground twisted.

It does not apply with two totally separate signals (in opposite directions, in this case) - the increased crosstalk from close twisting is more likely to be detrimental than beneficial, at a guess?

With this device I'd agree with Nigel, it is a convenience for wiring layout only.

rjenkinsgb said:

That would apply with balanced signals, or to a lesser extent a signal & ground twisted.

It does not apply with two totally separate signals (in opposite directions, in this case) - the increased crosstalk from close twisting is more likely to be detrimental than beneficial, at a guess?

With this device I'd agree with Nigel, it is a convenience for wiring layout only.

Click to expand...

He only seems to interesting in showing off how clever he is, not in actually helping people.

rjenkinsgb said:

That would apply with balanced signals, or to a lesser extent a signal & ground twisted.

It does not apply with two totally separate signals (in opposite directions, in this case) - the increased crosstalk from close twisting is more likely to be detrimental than beneficial, at a guess?

With this device I'd agree with Nigel, it is a convenience for wiring layout only.

Click to expand...

This seems to summarize case for twisted pair in Normal mode transmission.


There was a case study done by Exar, UART king of yore, of twisted pair cabling
in non diff mode transmission, I just cannot seem to find it. I'll keep looking.

Twisted pair also reduces crosstalk - https://en.wikipedia.org/wiki/Twisted_pair

One side comment, not an expert here. I figure at this point I have ~ .001% of the worlds
EE knowledge, and am striving for .002%, even though age is becoming a major factor. I
'welcome correction, fulfills my belief that the duty of an EE is to seek accurate and timely
info, and pass that on. That is a core value of our profession I think, at least thats what I
was taught.


Regards, Dana.

I have no argument with the benefits of twisted pair etc., I use it all the time in both my own gear and industrial gear I work on or design.

However, I can only guess you have not actually looked at the device schematic and have no idea what the signals on that cable are, beyond assumptions?

danadak said:

Twisted pair also reduces crosstalk - https://en.wikipedia.org/wiki/Twisted_pair

Click to expand...

Every mention of crosstalk in that article relates to crosstalk between pairs, in balanced line systems.

Twisting together two or more unrelated signal wires, as they are in this device, cannot improve crosstalk anywhere!

danadak said:

This seems to summarize case for twisted pair in Normal mode transmission.

Click to expand...

There is no ground in the cable, it is three different signal wires. The ground is via the battery negative power cable. It's neither balanced nor "normal mode" transmission.

However, I can only guess you have not actually looked at the device schematic and have no idea what the signals on that cable are, beyond assumptions?

Click to expand...

Schematic and OP references clearly show they are USART Rx and Tx .....up to 4.5 Mb/s

At wiki :

Compared to a single conductor or an untwisted balanced pair, a twisted pair reduces electromagnetic radiation from the pair and crosstalk between neighboring pairs and improves rejection of external electromagnetic interference.

Click to expand...

Maybe I misread the above ?

There is no ground in the cable, it is three different signal wires. The ground is via the battery negative power cable. It's neither balanced nor "normal mode" transmission.

Click to expand...

So the qualifier for normal mode has to mean the ground is in the cable ?


Regards, Dana.

For any of that to apply, the signal must have no overall signal current in the bundle; either overall balanced or signal & ground in each pair, so all signal current is within the same pair (or each triple/quad etc).

Balanced current means minimal or zero net field, signal radiation or signal coupling at a distance from or to the pair or bundle, and the twist means any stray coupling that does occur it likely to be shared by both (all) conductors, so mostly nulled out.

That is the essence of why twisted pairs (or triples/quads as appropriate) are so good at minimising stray emissions or noise pickup.


None of that applies with this piece of equipment; it's three totally different and separate signals with ground returns via a completely separate path.

Each has an external current return via the main ground, NOT part of the twisted bundle, so no magnetic field / interference cancelling.

Maybe part of my difficulty here is not having conveyed what I keep thinking, why should
he not do 3 twisted pairs, eg. not just twist the 3 signal wires together as shown as current
solution. The OP asked is it OK to extend current solution. Clearly none of us know what his
actual data rate is, neither apparently does he as he stated, but if its chip max certainly one
would entertain that in light of the need to extend cable length.

He states :

I was told that this dataline cable is used to calculate the battery life, and to manage the power supplied to the motor; Can this give an idea of the data rate?

Click to expand...

When I hear motor currents in the vicinity I immediately think coupling, of course qualified
by motor size. "Managing power supplied to motor", that could or could not mean high
speed links....

And were the original designers out in left field twisting the 3 signal wires together, eg. no
valid reason, other than mechanical "cleanness" of installation ? Was that a poor decision ?
Or was it rational, concerns over coupling.

Would be great if we could do a sim of the 3 signal layer twisting solution currently in
effect, not sure if there is a tool out there to do it, free of charge of course.



Regards, Dana.

Thank you all for your nice expert answers, I can move forward with my project now.

Just one last question:

The device is an electric vehicle (note: only legal modifications).

The dataline extension will come out of the vehicle, and will be very close to another extension, that of the power supply with a maximum current of 20 A.

Will it be necessary to add a shielded sheath around the dataline cable (see photos), and connect the metal shield to the aluminum body of the vehicle?

flolan said:

Will it be necessary to add a shielded sheath around the dataline cable (see photos), and connect the metal shield to the aluminum body of the vehicle?

Click to expand...

I don't know if screening will help; it may reduce interference pickup, but also increase capacitive loading, which could affect the signal - that all depends on the data rate.

My main concern is voltage drop on the battery ground. If that becomes significant due to a longer cable, the "ground" offset between the battery monitor and main circuit board could get large enough to affect the data wire voltage levels and disrupt the data, or even cause damage.

(You are in effect adding a resistor between the battery negative and circuit board negative, so under load the circuit "ground" will have some level of positive voltage offset relative to battery negative).

Make sure the battery negative cable is adequately sized so the voltage drop on it does not increase too much, or bond it through the vehicle metalwork if the system is grounded to that anyway?

Why not experiment, try a wrap while looking at data signals with a scope ?

Just a thought.

Dont forget you might have load dump considerations in a vehicle :






Regards, Dana.

rjenkinsgb said:

My main concern is voltage drop on the battery ground

Click to expand...

rjenkinsgb said:

Make sure the battery negative cable is adequately sized so the voltage drop on it does not increase too much

Click to expand...

I've tried several "voltage drop calculator" to check the power supply (battery) extender.
Results: between -1% and -1,30% (= -0,47V) in the worst scenario (20 amps)

With these parameters:

Battery: 36 V nominal
Wire size: 14 AWG 200° silicon
Distance (one way, same for positive and negative): 1.15 meters (45 inches)
Load current: 20 A (there is a 20 A fuse on the control board)

I don't know if these -0,47V are ok or not.

flolan said:

Results: between -1% and -1,30% (= -0,47V) in the worst scenario (20 amps)

Click to expand...

The data levels are only between 0V and 3.3V, so 0.47V is quite a big chunk off the noise margin and may be enough to invalidate the logic levels, depending on the receive thresholds at either end.

Thank you all for your help, without you I probably would have damaged my device.
I can move on to another solution now.