Noise on comms wires?

[Faringdon]

Hi,
Just today seen a big test jig with unshielded ribbon cable comms wires and it was working intermittently. I suspect noisy comms. There was an approx 15way ribbon cable (approx 30cm long) between 2 boards. Can you confirm that the comms signals in this need to be differnetial type?, and not the unipolar (TTL) type that you commonly see between components that are near to each other on the same board? So, eg, the ribbon cable should contain eg CAN bus signals, or RS485?, rather than TTL?

The test jig was for a switching circuit for three phase mains.
..............................-----------------...............................

Also, presumably on a test jig, which will not be made in volume, you should do super-low-bit-error-rate comms?.....eg, the most noise immune comms protocol that exists? With the highest magnitude signals (bits), so that SNR is maximised?
Also, common mode chokes should be used at each end of the cable to reduce bit error rate?
Also, shielded cable will reduce bit error rate?
And if its ribbon cable then it must have GND interleaved with all signals? (to reduce BER)?

BTW, what is the comms protocol with the lowest BER?

Presumably slower comms will always be less noisy than faster comms?.......and if possible, comms should be slowed down so much that the comms lines can be RC filtered?

[Gyro]

 

[floobydust]

You've already asked this question here, what was wrong with the answers ?

We have no idea what the data rate is, voltages, what data is being sent etc. How about a picture of the test jig?

It's most likely burst EMI from the mains switching- which is what you'd address- instead of blaming the ribbon cable. It can even be poor grounding of the jig and two boards. Many possible causes.
Physical layer and data link layer are different and it's good to learn the differences.
A robust comm protocol has error checking such as CRC etc. to verify integrity, along with a packet resend/retry, at the data link layer. You'd properly fix the lowest (physical) layer first.

[Faringdon]

It's most likely burst EMI from the mains switching

Thanks, thats i think a very likely cause of bit errors, which is bad because the test result data  couLD be corrupted by it and we wouldnt know....eg we could read an acceptable value of leakage current when it wasnt really.
In order to gaurd against this, would you think we should do a really noise resilient comms method?...eg   coaxial cable diff pair........ie, diff pair with the ground of the diff pair signals on the sheath?

Do you agree that unshielded ribbOn cable is the dirtiest roughest type of comms, with highest BER?...after all, youve got crosstalk between adjacent signal lines, and no shielding, and if no interleaving  grounds then its even worse?......ribbon cable = highest BER?

[floobydust]

I hear you want to make a blanket statement that "ribbon cable is the best antenna in the world, proven by having the highest BER". The circuit's impedances, bit rate, signalling method, logic levels TTL etc don't know, you won't tell us. If you'd like to demonize the ribbon cable, okay go ahead but would you switch the Test Fixture over to something like Ethernet just to leap 12"? Reading high leakage current due to noise picked up by a ribbon cable seems like a crappy design all around. Where is the A/D?
Just put a foil shield on the ribbon cable, to check your hypothesis and see if it's the cable at all.

EMI is best dealt with at its source instead of at the receiver. The 3-phase mains wiring and arcing contacts are easier to look after with bigger payoff.
Most test fixtures just have bad connections, bad pogo pins, crappy software etc. as the reason for their moodiness.

[Nusa]

Hi,
Just today seen a big test jig with unshielded ribbon cable comms wires and it was working intermittently. I suspect noisy comms.

The rest of what you're saying is irrelevant if you haven't identified your problem. Prove that's the problem before you go off trying to solve it! Because if you're wrong, you're literally spinning your wheels and getting nowhere.

[Siwastaja]

TTL, really?

Or maybe you mean CMOS?

For unidirectional signals, series-terminated single-ended CMOS (i.e., put a 33R in series right at the driving source) is often quite acceptable in a ribbon cable especially if you can dedicate every other wire as ground.

But if the environment the ribbon cable lives in is very noisy, or if you have any risk of ground loops causing common mode voltage differences in excess of a few tens of mV, then use RS422/485 or CAN instead.

A cheaper option might be slight RC filtering the CMOS signals and addition of true schmitt trigger buffers at the input, but usually if you go this far you can just add RS422 transceivers instead.

[radiolistener]

Also, common mode chokes should be used at each end of the cable to reduce bit error rate?
Also, shielded cable will reduce bit error rate?
And if its ribbon cable then it must have GND interleaved with all signals? (to reduce BER)?

It depends on the signal frequency. Higher frequency leads to higher dependency on impedance matching and impedance of transmission line.

GND interleaved wires helps to keep stable impedance across transmission line and reduce cross talk, so it helps to keep working condition at MHz frequency. But that's just a part of story. You're also needs to match transmission line with the load in order to avoid ringing in the transmission line due to signal reflections. You can use resistors to absorb reflected waves in the cable to reduce standing wave effects.

Presumably slower comms will always be less noisy than faster comms?

It depends on transmission line geometry, isolator and coductor materials and impedance matching. Properly designed and shielded transmission line can keep high enough communication speed. Usually about 40 Gigabit per second or even more for copper wire.

The main issue with high speed transmission is transmission line length. Due to skin effect high frequency current flows at very thin surface of wire so it leads to a high loss on resistance of the wire conductor. Also the wave propagates in the insulator between wires, so it leads to a high loss due to dielectric losses in the insulator between wires. As result, higher frequency leads to higher signal loss in the transmission line.

For a ribbon cable you can expect about 150-200 MHz limit for about 20-30 cm length. Also it is very critical to keep exactly the same length of wires for all data lines, otherwise they will have different phase delay at receiver.

[BradC]

BTW, what is the comms protocol with the lowest BER?

In my experience it's Morse by Aldis lamp unless one of the operators has been on the rum.

[perieanuo]

Hi,
Just today seen a big test jig with unshielded ribbon cable comms wires and it was working intermittently. I suspect noisy comms. There was an approx 15way ribbon cable (approx 30cm long) between 2 boards. Can you confirm that the comms signals in this need to be differnetial type?, and not the unipolar (TTL) type that you commonly see between components that are near to each other on the same board? So, eg, the ribbon cable should contain eg CAN bus signals, or RS485?, rather than TTL?

The test jig was for a switching circuit for three phase mains.
..............................-----------------...............................

Also, presumably on a test jig, which will not be made in volume, you should do super-low-bit-error-rate comms?.....eg, the most noise immune comms protocol that exists? With the highest magnitude signals (bits), so that SNR is maximised?
Also, common mode chokes should be used at each end of the cable to reduce bit error rate?
Also, shielded cable will reduce bit error rate?
And if its ribbon cable then it must have GND interleaved with all signals? (to reduce BER)?

BTW, what is the comms protocol with the lowest BER?

Presumably slower comms will always be less noisy than faster comms?.......and if possible, comms should be slowed down so much that the comms lines can be RC filtered?

in some industrial medical equipment we used rs485 on 1 meter long 10-pin ribbon cable (connecting 3 'intelligent'  step motors communicating with rs485, each motor with his different address, but that was software stuff ), 10 centimetres away we had 2.4ghz custom wifi comm module. all approved by french official labs regarding emi and all working 100% reliable. rs485 wasn't 'terminated' by termination resistor at any end. the comm was 128kbps if i remember correct, but 1M tests were reliable also.
of course we avoided ttl for that length.
can is also a good choice, in fact the choice is specific to what you need/know best to implement as differential comm.
if you have only 30cm, non-differential will work with ber=0 if some precautions are taken when implementing (for example, i2c works at that bus length, depends also of your desired bus speed)
i think you have noisy environment, also ttl should work

[Just_another_Dave]

Hi,
Just today seen a big test jig with unshielded ribbon cable comms wires and it was working intermittently. I suspect noisy comms. There was an approx 15way ribbon cable (approx 30cm long) between 2 boards. Can you confirm that the comms signals in this need to be differnetial type?, and not the unipolar (TTL) type that you commonly see between components that are near to each other on the same board? So, eg, the ribbon cable should contain eg CAN bus signals, or RS485?, rather than TTL?

The test jig was for a switching circuit for three phase mains.
..............................-----------------...............................

Also, presumably on a test jig, which will not be made in volume, you should do super-low-bit-error-rate comms?.....eg, the most noise immune comms protocol that exists? With the highest magnitude signals (bits), so that SNR is maximised?
Also, common mode chokes should be used at each end of the cable to reduce bit error rate?
Also, shielded cable will reduce bit error rate?
And if its ribbon cable then it must have GND interleaved with all signals? (to reduce BER)?

BTW, what is the comms protocol with the lowest BER?

Presumably slower comms will always be less noisy than faster comms?.......and if possible, comms should be slowed down so much that the comms lines can be RC filtered?

Shielded cables are less susceptible to noisy environments and, consequently, you can expect them to reduce the error rate if that’s the cause. Twisted pairs are not as good as coaxial cables, but better than doing nothing. Common mode chocked, on the other hand, reduce common mode noise, so they can improve the signal of that’s the cause of the noise

Another possibility for improving noise immunity when long cables are required is to sent data encoded in current. Although I haven’t seen it used in digital communications, 4-20mA is sometimes used in analog sensors when measurements need to be sent through noisy environments.

In any case, knowing a bit more about the device could make providing a solution easier

[Nusa]

Pattern detected. Treez is back, isn't he?

[Gyro]

Yes, old news now... https://www.eevblog.com/forum/supporters-lounge/treez-how-many-more-accounts-are-acceptable/
Quite why this thread is allowed in the beginners section in the first place eludes me, he allegedly has a 1st class honours degree in electronics and 20+ years in the industry.

[Terry Bites]

No one thing or technique is bad or good for this or that, its just that you know how and when to apply them that matters.
It seems like you havent had the experiece of debugging simple stuff and now you've jumped in the deep end unprepared.
Did you have examine the EMC issues at the design stage? Even if you are a lone wolf, you need to effectively project manage from day one.

What kind of noise, what frequqency or amplitude, bursts or continuous? How did you measure it?
I know you are working on a secret project that will make you richer Jeff Bezos. Still, you might share a system and wiring diagram instead of shooting randomly in the dark. Some one may actually be able to help you with your problems.  Its time for tough love baby!


You might attend a course: https://learnemc.com/emc-resources.

[Simon]

Yes, old news now... https://www.eevblog.com/forum/supporters-lounge/treez-how-many-more-accounts-are-acceptable/
Quite why this thread is allowed in the beginners section in the first place eludes me, he allegedly has a 1st class honours degree in electronics and 20+ years in the industry.

The section used is not meant to be chosen based on your qualifications or experience. I post in any topic I deem fit for the nature of the question including the beginner section if it's a simpler problem. EMC related stuff is beyond the beginner.

I am no transmission line expert or overly experienced in EMC. Ribbon cable was used in computers to interface to hard drives at up to at least 33MHz (ATA133 - 32 bit bus wasn't it? so 4 bytes at a time at 1/4 the frequency). But then recently I was involved with an issue of corrupt comms on an RS485 bus that was caused by a metal punching machine on the factory floor we were working on.

[floobydust]

We've asked OP for pictures, voltage, frequency, etc. any details whatsoever- and nothing is provided, ever.
There's no spirit of learning or finding a remedy, contrary to beginners. Cross-post had a few tidbits more information about the psychology of ribbon cables.

[Simon]

What happens on other forums is non our our concern here. When I used more than one forum I would post the same topic on different forums. No rule against it on any forum. I know many people use multiple forums so will see it elsewhere but no forum is an exact overlap of other forums in membership and just the nature of the forum can lead to a different approach to the question.

It does appear though that the OP cannot provide enough information and is not informed enough to work with the subject of the topic. In the past I would often say to my boss that I thought some thing was so and so but I would ask on the forum and get some other opinions. This is what I am doing on my latest topic. I have a fair idea of what is going on and what is wrong but I want other peoples opinion as it will on the one hand tell me I am right and on the other hand stretch my knowledge.

[JustMeHere]

Are you using multiple grounds?  It helps to have at least one ground wire next to each data line.  The data wires will couple with a ground wire.  Anything in between the signal and ground will pick up noise.

https://electronics.stackexchange.com/questions/180811/why-do-cables-have-multiple-grounds

[floobydust]

We will never know any details about the ribbon cable because OP is not involved beyond speculation.

Crossposting is bad etiquette, Eng-Tips does not allow it and a few other sites.

Difference between asking for information here to learn, compared to working beyond one's means and appearing to be intelligent, or acting above your level in the workplace.
It stops with safety and other forums will lock threads if it's obvious OP is working with something hazardous beyond their skill level. So innocent posting does have limits.

I consider it a colossal waste to start a non-sequitur thread, withhold information, continue asking questions, then abandon it. It just litters this site with trash threads.
All threads discussing this frustration which many members have expressed, are locked, so the only thing you can do is click on ignore the bonehead.