Page 3 of 3 FirstFirst 123
Results 21 to 27 of 27

Thread: OT: Low voltage digital circuit and shielded cable

  1. #21
    Join Date
    Oct 2013
    Location
    USA MD 21030
    Posts
    5,555

    Default

    On older vehicles, cars at least, the chassis was used as ground return. This was mostly OK for things like lights and door switches, but not so much for modern electronics. Even with lights, sometimes rust could cause high ground resistance and intermittent operation. In the case of a motorcycle, parts that move relative to each other, like the forks and the frame, need a bonding wire to assure continuity. Such bonding is also required on industrial and commercial equipment, to assure that parts like doors and panels are fully grounded.

    I usually ground shielded wires at one end, but in some cases, both ends must be connected, as in the case of old style "RCA Phono" connectors which use single conductor shielded (coaxial) cable. There is also a separate ground wire that must be connected to avoid AC hum. Oscilloscope probes use BNC connectors and coaxial cable for signals, and the shield at the probe must be grounded to the appropriate part of the circuit to be measured. However, the chassis of the scope must also be connected to earth as well as the circuit being probed. Some scopes are battery powered or floating, but it is still usually required to provide a resistance and capacitance to earth ground to avoid voltage build-up.

  2. #22
    Join Date
    Dec 2016
    Location
    Helsinki, Finland, Europe
    Posts
    3,610

    Default

    Quote Originally Posted by QSIMDO View Post
    STEADY ON, lads!
    Motorcycle, 12 volts DC...I'm using a "Motogadget m-unit" if anyone's interested and a V.2 not the new blue tooth doo-dad.
    I recommend reading these as a starting point
    https://www.amazon.com/Electromagnet...s=books&sr=1-2
    https://www.amazon.com/High-Speed-Di...gateway&sr=8-7
    https://www.amazon.com/Noise-Reducti...5YGSSA5HKDC2D8

    If 1700 pages seems like a bit long read just ignore the shielding and ride to sunset

  3. #23
    Join Date
    Dec 2016
    Location
    Helsinki, Finland, Europe
    Posts
    3,610

    Default

    Quote Originally Posted by Ringo View Post
    Every damn wiring diagram I ever worked with (aviation avionics) the engineer drawing stressed shielding, right up to the cannon plug. And yet when you get into repairs in the field, you may see hack-job repairs, ignored shields, broken shields, and so on. However, at no time have I EVER fixed the problem, and come to the conclusion, that the 'bad shield' was the ultimate cause of the original fault. That is only my 25years of avionics point of view.

    But then, all the engineers will tell you that lower hertz A/C power is the absolute WORST offender of noise, therefore requireing the BEST-UTMOST shielding. This is total horsekrap.
    If this was true then none of us could watch TV, Google, YouTube nor anything...........
    Did you ever look behind your entertainment center, desk, or whatever office place you are all setup?
    What do you see? 110v/60hz (unshielded) power all over the place, with power splitters rampant, Tv cables with dubious connections, internet wires with their own dubious connections, and yet, what????
    We are logging onto internet just fine. Watching TV just fine.
    The engineers opinion of shielding and noise is just that,,,,,,,,,a lot of noise.
    Low frequency AC power noise is hard to keep out if you are interested to measure something in the same frequency range. TV or internet signals are at least 1000 000 times higher frequency and don't care a damn about some 60hz noise.

    Try to make reliable measurements at 2 microvolts 60hz AC and you get in the troublesome department.

  4. #24
    Join Date
    Jan 2005
    Location
    Ct
    Posts
    2,192

    Default

    Quote Originally Posted by MattiJ View Post
    I recommend reading these as a starting point If 1700 pages seems like a bit long read just ignore the shielding and ride to sunset
    You knew that already, didn't you.
    Len

  5. #25
    Join Date
    May 2002
    Location
    SE Texas
    Posts
    12,476

    Default

    This is correct with one caveat. Many shielded cables are grounded at both ends and they work just fine.

    Over 45+ years in TV engineering I have worked with many signal cables and have seen all variations: no shield, shield grounded at one end, and shield grounded at both ends. All can work and all of them can fail: I have seen and fought every variation. It is not just a simple thing. There are three primary factors that are considered when designing a cable to carry a signal from one place to another: the signal level, the impedance of the source and destination, and the frequency of the signal. Of course, the amount of electromagnetic noise in the environment is also a factor because that is what the shield is blocking.

    The signal level: this is important because the generally accepted method of measuring noise in the signal is as a ratio; the signal to noise ratio. This is normally expressed in db or deciBells. One db is one tenth of a Bell, named after Alexander Grayham Bell). It is a logarathmitic unit and 10 db represents a ratio of 10 times in terms of power while 20 db represents a ratio of 10 times the Voltage or current (but not both at the same time).

    https://en.wikipedia.org/wiki/Decibel

    So the signal to noise ratio can be improved by either decreasing the noise level or increasing the signal level. Both of these work equally well. With all else in the circuit being equal, the noise that it will pick up from the environment will be the same regardless of the signal level so larger signal levels will produce larger signal to noise ratios. But there are practical limits to to how high the Voltage levels on a signal cable can be and numbers between 0.1V and about 5V are what is normally found.

    The impedance of the source and destination circuits: All electronic circuits that generate or receive signals will have a characteristic impedance as seen looking into their terminals. Designers try to keep these impedances as pure resistances, but high frequencies will bring capacitive and inductive elements into them. For noise purposes, we can look at just the resistive component, at least to a first approximation. The noise generated in a signal cable, a signal wire can be modeled as an induced current while most receiving circuits will be responding to the Voltage it generates. The lower the source and destination impedances, the lower the Voltage that this current will produce and therefore the greater the signal to noise ratio will be. This is due to good old, Ohms Law: E = IR. The lower the R, the lower the E (Voltage). Professional audio circuits are normally 600 Ohms. This is because professional audio started in the days of tube electronics and transformers were used on the inputs and outputs of the circuits. A really low impedance transformer would have very few turns on the external side of the transformer and that would lower the Voltage. So 600 Ohms was chosen as a compromise between low impedance and low Voltages on the transmission line. When professional video came along, things were somewhat better and 75 Ohms was chosen for the standard impedance. This made things easier. For RF or radio frequency signals an even lower, 50 Ohms became standard because this worked will with transmitter circuits and antennae. The 50 and 75 Ohm impedances were also easily compatible with practical coaxial cable construction. The characteristic impedance of the coaxial cable needed to match the source and destination impedances for best power transmission and minimum losses. Losses in a high power RF transmission line can be a real problem and things like 3 and 6 inch coaxial lines have been known to heat to the point where the copper melts. Talk about letting the smoke out; you don't want to be near such an event.

    Note: I was talking about professional audio circuits above. Low impedance audio circuits are more difficult to design and more expensive to construct. Consumer electronics often uses single ended, high impedance circuits, at least on the destination end. Transistor circuits can easily produce a low source impedance, but the receiving end often has a high value (10 kOhms) resistor in series with the input terminal. Among other reasons, this can be cheap protection against spikes in the input line that could destroy a transistor or FET. Many times I had to integrate consumer equipment into a professional system. My answer was to add a 620 Ohm resistor to ground at the high impedance input. This worked well. It lowered the level of the signal by about 1/2 but the consumer equipment was often built for a -10 db signal while the professional system was distributing a 0 or +3 db level so the signal loss was not felt.

    The frequency of the signal: The frequency of the signal is important because much of the energy in electromagnetic noise is in the high frequency components. Simple bypass capacitors can shunt frequencies that are beyond those needed for the circuit to ground. So, it is easier to build circuits that reject noise for low frequencies than for high frequencies. A 0.01 uF capacitor can take care of a lot of high frequency noise in an audio circuit where 20 KH is all that is needed. But that same capacitor may totally destroy a high frequency (high speed) digital signal.

    Oh, one more factor that needs to be considered. There are two general types of transmission lines: single ended and double ended. In simple terms, this means one wire or two. Those audio circuits of old, which used transformers, had two wires on that transformer coil. So, the circuit designer could either ground one and use the other OR use both of them to transmit the signal to the destination. When a single wire is used, it will pick up a certain amount of noise and that noise will be carried to the destination and be received and amplified there. But with a two wire transmission cable, one wire connected to each of the two transformer coil's wires, one wire will be positive while the other is negative and vice-versa. The two wires will have signals that are mirror images of each other. But the noise will be the same on both. When those two wires reach the destination the signals in them will be out of phase and will produce a current in an input transformer while the noise will be close to the same in both wires and when it is applied to the two terminals of that input transformer, it will produce zero current. It CANCELS OUT. This is why professional audio is usually transmitted with a balanced (two conductor) transmission line. Often, audio circuits using twisted pair audio line can be run with no shields. I have seen professional facilities wired this way and it worked well. The older TV antenna lines were also balanced lines with a 300 Ohm impedance.

    In professional TV stations and other AV facilities audio is run with 600 Ohm, balanced lines that are shielded. That shield on the audio cable is normally connected on only the source end. But analog video is run in 75 Ohm, coaxial cable and the shield is connected at both ends. What is the difference, other than tradition? Well, the two conductors of the balanced audio cable produce a complete, back and forth circuit path. So, the shield is not part of the signal circuit. It is only a shield. With the video, there is only a single wire to transmit the signal so the ground path is needed to complete the circuit. If the shield were only connected at one end, then the return path would have to pass through the equipment's metal enclosures, the racks that they are mounted in, the power ground that interconnect those racks, and heaven knows what else. This ragged path is not what is called a constant impedance path. Different frequencies in the video signal would react differently to it and the signal would be distorted. The use of the shield for the return path is essential. Much the same is the case for a high frequency digital signal or an RF signal. So the signals that have higher frequency components and that have to be run for a significant distance, usually have the shield connected to ground at both ends. But that ground may not be the same as the power ground. It will be a SIGNAL ground and may be separate from power ground in the source or destination circuit or both.

    My experience has shown me that the grounding of a shield is not a simple subject. There can be problems with all kinds of signal cables and often they must be considered separately. I have seen problems that took much time and experimental effort to resolve. Theory be damned, we just had to find what actually worked. The OP did not say much about what his digital circuit was, what frequency it used, etc. I would say that if the manufacturer did not specify a shielded cable, then it is probably not needed. Run the shielded cable. Either connect the shield at one end or not at all and see how it works. If there is a problem, then change something. Repeat until things work OK. That is what I offer with 45+ years of tackling these problems.

    As for the aircraft practice, I am sure it is what works for most circuits on aircraft. But I would bet that there are exceptions where something else must be done.



    Quote Originally Posted by Ringo View Post
    yes you can ignore it
    yes you can terminate 1 end of the shield; this is standard practice in aviation avionics
    no, you cannot ground both ends, this makes a loop antenna and gathers noise as per standard avionics

    it sounds like running wires inside the metal frame, the tubing frame itself becomes a crude shield in and of itself
    Paul A.

    Make it fit.
    You can't win and there is a penalty for trying!

  6. #26
    Join Date
    Dec 2012
    Posts
    508

    Default

    Paul, I appreciate your post; I agree with all of it except that 99.9% of professional audio circuitry is no longer 600 ohm-that went out in the 60s.

    Since then, low-source impedance outputs (on the order of 10-100 ohms) are used to drive cables(not transmission lines) which connect to high-impedance inputs (usually between 7.5K and 15K ohms).

    I have put in a few pieces of equipment that truly needed to see 600 ohms on input and output-but even that was back in the 90s...

  7. #27
    Join Date
    May 2002
    Location
    SE Texas
    Posts
    12,476

    Default

    Yes, I am totally aware of that and was tempted to include it. But, it was long enough already. I can be very verbose and sometimes I just have to stop.

    In truth is is still loosely based on the old 600 Ohm system. ICs, transistors, and FETs have made very low impedance outputs with a frequency response that is flat to well under 1 db from 20 Hz to 20 KHz and ultra low distortion very easy to make. A common device in professional audio is a distribution amplifier. It is made to allow an audio signal to be sent to many destinations. Modern designs will have a solid state amplifier with an output impedance of just a few Ohms. Then a bunch of 600 Ohm (or 300 Ohms for balanced outputs) resistors will each lead from that low impedance point to the individual output terminals. I have seen distribution amplifiers with this design that had ten or more outputs, all of them just separated by those resistors. Those resistors bring the output impedance of each output back to that old 600 Ohm number. Because the amplifier has such a low output impedance, loading is not a problem and if one of the outputs, past that 600 or 300 Ohm resistor is shorted to ground, it is barely felt at the others which continue at very close to the design level. It is a very robust design. And you can still follow the old design rules for the external lines: there is almost complete compatibility. Any input that they go to can be almost any impedance from a few Ohms up to 10 KOhms or more and again it will not have any appreciable effect on the others.

    I worked in TV and other professional facilities well past the 60s. I only retired in the present decade. I always used the 600 Ohm model when working with audio. It was and still is the basis for professional audio distribution. In that time I have selected, purchased, installed, and trouble shot more than a few pieces of audio equipment.

    PS: a cable IS a transmission line. Anything conduction path that carries a signal or even AC or DC power is a transmission line. Digital design engineers will look at the foil traces on their circuit boards as transmission lines when high frequencies are involved. Even the power companies call their power lines "transmission lines". That is not a misuse of the term. And seeing it that way is a very good way to avoid problems, even with low frequencies like audio.



    Quote Originally Posted by andywander View Post
    Paul, I appreciate your post; I agree with all of it except that 99.9% of professional audio circuitry is no longer 600 ohm-that went out in the 60s.

    Since then, low-source impedance outputs (on the order of 10-100 ohms) are used to drive cables(not transmission lines) which connect to high-impedance inputs (usually between 7.5K and 15K ohms).

    I have put in a few pieces of equipment that truly needed to see 600 ohms on input and output-but even that was back in the 90s...
    Paul A.

    Make it fit.
    You can't win and there is a penalty for trying!

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •