Help with cleaning up noisy 10 bps communication

What kind of transmission signals you are using?

- RS-232
- RS-485
- Something else?

RS-232 is notoriously bad choice for longer than 10 meter cables, noisy environments or environments that could create ground loops between the device communicating. RS-485 uses a twisted pair cable providing a differential balanced line, which will improve noise immunity and is more tolerable to ground loop problems.

Slow signaling rate (10bps) doesn't mean that the cable would not be a transmission line: It is the slew rate of the signal which determines whether the cable should be considered as a transmission line or not. If the transmitter is fast, for example signal goes from -5V to +5V in 100ns, you may have a transmission line. If you do have a transmission line, you can terminate the cable either in the sending end of the cable or at the receiving end of the cable. For example, if the cable's characteristic impedance is 100 ohm, just place a 100 ohm resistor in series with the transmitter and the cable. Or, you can terminate the cable in the receiving end by placing a 100 ohm resistor across the cable signal line and the ground, if the receiver and transmitter can handle this extra 100 ohm DC-load. You may want to experiment a bit to get desired results.

If you are experiencing ground loop problems, you may want to use some kind of [optical] isolation which will break the ground loop and save your day.

If you just cannot get a clean signal, you could add redundancy to the data, which would provide sufficient forward-error-correction. :)

Br,
Kalvin

>> signaling rate (10bps) doesn't mean that the cable would not be a
transmission line: It is the slew rate of the signal which determines
whether the cable should be considered as a transmission line or not.

No. 
The problem is not the transmission line but that you're assuming a
_receiver_  bandwidth of 100 MHz for a 10 Hz signal. You're trying for a
perfectly square eye diagram at the detector.

If the receiver uses adequate filtering, it will simply reject any high
frequency components, regardless how many times they were reflected. 


	 

_____________________________		
Posted through www.DSPRelated.com

On Thu, 14 Aug 2014 08:39:36 -0700, kalvin.news wrote:

> What kind of transmission signals you are using?
> 
> - RS-232 - RS-485 - Something else?
> 
> RS-232 is notoriously bad choice for longer than 10 meter cables, noisy
> environments or environments that could create ground loops between the
> device communicating. RS-485 uses a twisted pair cable providing a
> differential balanced line, which will improve noise immunity and is
> more tolerable to ground loop problems.
> 
> Slow signaling rate (10bps) doesn't mean that the cable would not be a
> transmission line: It is the slew rate of the signal which determines
> whether the cable should be considered as a transmission line or not. If
> the transmitter is fast, for example signal goes from -5V to +5V in
> 100ns, you may have a transmission line. If you do have a transmission
> line, you can terminate the cable either in the sending end of the cable
> or at the receiving end of the cable. For example, if the cable's
> characteristic impedance is 100 ohm, just place a 100 ohm resistor in
> series with the transmitter and the cable. Or, you can terminate the
> cable in the receiving end by placing a 100 ohm resistor across the
> cable signal line and the ground, if the receiver and transmitter can
> handle this extra 100 ohm DC-load. You may want to experiment a bit to
> get desired results.
> 
> If you are experiencing ground loop problems, you may want to use some
> kind of [optical] isolation which will break the ground loop and save
> your day.
> 
> If you just cannot get a clean signal, you could add redundancy to the
> data, which would provide sufficient forward-error-correction. :)
> 
> Br,
> Kalvin

You need to read the whole thread.  Tucked up about a dozen responses deep 
is a comment by the OP that it's powerline communication of some sort -- 
the only wires to the motor are the wires for three-phase power.

-- 

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

torstai, 14. elokuuta 2014 20.20.32 UTC+3 Tim Wescott kirjoitti:
> On Thu, 14 Aug 2014 08:39:36 -0700, kalvin.news wrote:
> 
> 
> 
> > What kind of transmission signals you are using?
> 
> > 
> 
> > - RS-232 - RS-485 - Something else?
> 
> > 
> 
> > RS-232 is notoriously bad choice for longer than 10 meter cables, noisy
> 
> > environments or environments that could create ground loops between the
> 
> > device communicating. RS-485 uses a twisted pair cable providing a
> 
> > differential balanced line, which will improve noise immunity and is
> 
> > more tolerable to ground loop problems.
> 
> > 
> 
> > Slow signaling rate (10bps) doesn't mean that the cable would not be a
> 
> > transmission line: It is the slew rate of the signal which determines
> 
> > whether the cable should be considered as a transmission line or not. If
> 
> > the transmitter is fast, for example signal goes from -5V to +5V in
> 
> > 100ns, you may have a transmission line. If you do have a transmission
> 
> > line, you can terminate the cable either in the sending end of the cable
> 
> > or at the receiving end of the cable. For example, if the cable's
> 
> > characteristic impedance is 100 ohm, just place a 100 ohm resistor in
> 
> > series with the transmitter and the cable. Or, you can terminate the
> 
> > cable in the receiving end by placing a 100 ohm resistor across the
> 
> > cable signal line and the ground, if the receiver and transmitter can
> 
> > handle this extra 100 ohm DC-load. You may want to experiment a bit to
> 
> > get desired results.
> 
> > 
> 
> > If you are experiencing ground loop problems, you may want to use some
> 
> > kind of [optical] isolation which will break the ground loop and save
> 
> > your day.
> 
> > 
> 
> > If you just cannot get a clean signal, you could add redundancy to the
> 
> > data, which would provide sufficient forward-error-correction. :)
> 
> > 
> 
> > Br,
> 
> > Kalvin
> 
> 
> 
> You need to read the whole thread.  Tucked up about a dozen responses deep 
> 
> is a comment by the OP that it's powerline communication of some sort -- 
> 
> the only wires to the motor are the wires for three-phase power.
> 
> 
> 
> -- 
> 
> 
> 
> Tim Wescott
> 
> Wescott Design Services
> 
> http://www.wescottdesign.com

Thanks, Tim.

10k feet (3km) cable as a transmission media polluted with motor induced noise, and fairly slow 10bps transmission rate might just need a plain old voiceband modem or a simple FSK modulator/demodulator - implemented either in analog hardware or DSP. For better reliability, some sort of forward-error-correcting coding might be useful.

Br,
Kalvin

Artificial Lift? 

Unfortunately, in most cases I am familiar with, one doesn't have control over the downhole transmitter as often it is determined by the manufacturer of the pump. The best you can do is then mitigate the surface receiver side. Modulations over cable are usually very simplistic: think FSK, BiPhase, UART-ish. 

Ron, are you allowed to tell us which modulation is used? We can then guestimate the best approach for you. 

For FSK it is most likely asynchronous FSK, so you can easily match signal bandwidth with a receiver filter. But you mention ringing, so I am guessing UART-ish or BiPhase-ish. In that case you need to learn how to do waveform captures and make "eye diagrams". Then play around with a filter while evaluating filter bandwidth vs. impact on "eye diagram". This is how I supervise my software guys when we don't want to make too many changes and just want to improve SNR with a simple solution. 

IIRC most AL motors are directly driven from surface, so current/voltage noise fundamental frequency will be changing a bit. 

Cheers. 
Julius

On Tuesday, August 12, 2014 1:11:27 PM UTC-4, mako...@yahoo.com wrote:

> 
> It seems clear Ron's task is in  a professional setting, i.e.  his empoyer is paying him to figure this out.  I'd like to hear a discussion about "giving away" this kind of information given the job market we are in.  If Ron were not able to come here and get this kind of help for free, Ron's employer might have to hire one of us. 
> 

Plenty of jobs related to signal proc and comms in oilfield domain, but most smaller companies think they can't afford to hire one guy just to do comms. So they tend to look for "software" hires thinking that it's a "software" issue. When it's solved, then they think they can re-use the software person to do other software things. 

On the other hand, many signal proc / comms candidates write their resume in a way that is hard for oilfield tech people to understand. Anyway, comms is viewed as an enabling technology, but not yet a differentiator, in many aspects of oilfield.

On 08/14/2014 08:58 AM, Tim Wescott wrote:
> On Wed, 13 Aug 2014 19:23:50 -0500, Richard Owlett wrote:
>
>> BasicPoke wrote:
>>> [*SNIP*]
>>>>
>>>>
>>> The 3-phase motor windings are part of the circuit, in series with 10k
>>> feet of the motor's power cable.  The signal passes over the power
>>> cable.  There are other reactive components.  ...
>>
>> You *HAVE* problems!
>> You need someone competent dealing with low level signals in presence of
>> interference orders of magnitude larger.
>>
>> I have experience as an engineering tech.
>> I once was fired for refusing to pretend to be an engineer. [On a
>> subsequent job I got a positive review for same stance ;]
>> I'm past 3 score and 10, so I'm safe ;)
>>
>> See thread on "good questions".
>
> Line carrier communications have worked since the 1940's or earlier
> (apparently it's what some frustrated amateur radio operators did during
> WW-II -- there's a chapter on it in one of my wartime ARRL Handbooks).
>
> AFAIK it's mostly done by choosing a frequency that's high enough so that
> the line frequency stuff is easy to filter out, and low enough that it's
> not horribly attenuated by the wires.  And then use a really high-power
> amp on the transmit side!
For the old school power line communication "really high power amp" 
means an injection transformer pumping bursts of signal onto the line so 
the mains voltage ripples at a few hundred Hertz (a frequency chosen to 
lie between the harmonic lines of the mains). This is frequently called 
voltage ripple signalling. They send short messages with on off keying 
of the ripple. They inject 10's of kW with these things. It is widely 
used for street light control, off peak tariff control, load shedding, 
and so on. The receivers used to be resonant electromechanical devices, 
but electronics is more common now.

Regards,
Steve