Choosing motorencoders

We have an installation where the encoder cables are 50 metres long. The cables have a section of 0.35 mm2. But we have spare wires so we can use double wire for the power and zero.
Question : won't the voltage drop be to high for 5V TTL encoders? On the other hand 24V HTL maybe capture too muche noise. What about capacative weakening of the signal, will it be a problem at that length for the 5V TTL, would it be less for 24V HTL, will it be less if I use 1024 incr encoder in spite of 2500 incr encoder? The encoders serve as speed feedback for closed loop speed control.
What would you advise?
Thanks

Here is one option.

http://www.encoderproducts.com/literature/datasheet-rxtxrepeater.pdf

I'll add a thought about voltage drop. As you would expect with Ohm's law if there is little current flow it's nothing to worry about. As a test I use a 24vdc icecube relay and many 500-foot rolls of 18-gauge wire. I started with one roll and progressed to six rolls. With six 500-foot rolls of 18-g wire (total 3000 feet or 914 meters) I saw 22 volts at the coil of the relay. Of course not the same as an encoder application.

dYoast's link is very promising. I did not know it existed and appreciate the information.

But if it were me, I'd go for it if I could.

You should not be using 5V TTL encoders for that distance. You should use 12-24v differential output encoders with sufficient size power wires.

Choosing motor encoders

theDave2 said:

You should not be using 5V TTL encoders for that distance. You should use 12-24v differential output encoders with sufficient size power wires.

Click to expand...

What do you consider as a maximum length for 5V encoders, considering a crossection of 0.35 mm2?
And how about resolvers? Are longer cables possible? For encoders you have to take care of maximum pulse frequency because of rounding of the pulses by capacitive effects of the wires, is that less a problem for resolver signals?

Plc_User said:

What do you consider as a maximum length for 5V encoders, considering a crossection of 0.35 mm2?
And how about resolvers? Are longer cables possible? For encoders you have to take care of maximum pulse frequency because of rounding of the pulses by capacitive effects of the wires, is that less a problem for resolver signals?

Click to expand...

Actually, I have no idea what size "0.35 mm2" is, being in the USA. But that doesn't matter. For power, what matters is how much current draw there is for power and the voltage drop. Standard TTL power is +/- 5% (4.75 to 5.25vdc). You would need to see the specs on encoder power needed.

Next, you are of course correct that signal cable capacitive does matter. It will alter the rise time of the signal. TTL signals have a very fast (<5 nano-seconds) rise time and are meant for short distance (<10 feet) runs.

Resolvers are no problem at long distances, due to the waveforms used. But that's an expensive solution for speed control.

Check out the spec sheet of:
http://www.electro-sensors.com/product_info.php/cPath/8/products_id/69

http://www.electro-sensors.com/documents/1101-931 ES104 RevE.pdf

While it may be overkill in terms of Haz. location usage, it is very typically the type of speed sensor /pulse generator that is widely used. You would most likely use it with a ring kit: http://www.electro-sensors.com/index.php/cPath/7



If you do want to use an 'encoder' , get one that operates on 12-24vdc, has either Line Drivers that operate at least 12 vdc or something like 12-28vdc open collector output.

theDave2 said:

Actually, I have no idea what size "0.35 mm2" is, being in the USA.

Click to expand...

Here are some guidelines to follow from the Dynapar FAQ section:

http://www.dynapar.com/content.aspx?id=46#Q:_How_far_away_can_I_place_my_encoder_from_my_system

0.35mm^2 is about 22AWG, pretty small..

Choosing motor encoders

theDave2 said:

You should not be using 5V TTL encoders for that distance. You should use 12-24v differential output encoders with sufficient size power wires.

Click to expand...

Why are 5V TTL encoders used so commonly, as they are limited in max. distance?
Are differential outputs always pushpull and single ended outputs always open collector?

OK, going with my understanding, which I can't promise is perfect.

Volt-drop - as others have said is not too much of an issue because of the very low current.

Signal type - Single ended signals will have a lower maximum length than Differential. See here for a description of Differential.

Cable Type - Use twisted pair cable with a Differential signal, see the RS422 description above.

Cable Screening - Always use a screened cable with the screen grounded at one end only.

High Voltage versus Low Voltage - Though high voltage will tend to be less susceptible to electrical noise it also means having to change the voltage of the cable through a much greater range. It takes time to change the voltage which will limit the maximum pulse rate on the cable. Also due to the capacitance there will be a larger current needed at signal change and the output circuit has to be able to cope with that current.

Cable Diameter - It feels like a bigger cable will be better than a small one and with low speed pulses it will be, providing that the output circuit can switch the larger current due to increased capacitance. For higher speed signals it is actually better to go with a thinner cable. If you don't have a supply voltage available at the encoder end then use multiple cores for the Supply and Zero volt. Get the cable resistance specs and current needed to calculate supply voltage drop.

Cable placement - always run signal cables away from sources of electrical noise, cables carrying mains voltages, VSDs etc.

So the answer to your question is that there is no simple answer. It is a balance of maximum speed with maximum distance with minimum cost, working with the measuring device inputs that you have available. The best way is to get the encoder manufacturers specs for maximum distance and speed. Then if it all goes tits up you can blame the supplier. Your distance of 50 metres isn't that far but if you don't use screened twisted pair cable, or if you want a very high speed rate of pulses, or if you run the cable in the same tray as mains cables, you will have problems. To give you some perspective RS422 works typically in the 5V range and at 1200 metres has a maximum speed of 100kbps, that is 100000 pulses per second. That would be a 100ppr encoder spinning at 60,000rpm.

If ever distance is a major issue then look at optic fibre rather than electrical cable transmission.

Oh, and finally, why so many 5V encoders out there, History, when encoders were first made the kit they had available was 5V so they made them 5V.

Bryan

Hi, Bryan,

Thank you for your information.
'To give you some perspective RS422 works typically in the 5V range and at 1200 metres has a maximum speed of 100kbps, that is 100000 pulses per second. That would be a 100ppr encoder spinning at 60,000rpm'.
This was very interesting. Do you links to articles where some more data is given : f.i. for 10-30V encoders and other cable lengths.

I don't have any links for other encoders/distances, the RS422 data is from the RS422 standard, shown in tables such as this:
http://www.lammertbies.nl/comm/info/RS-485.html#char

You can get Encoders with an RS422 output and then use a converter like this to go to htl.
http://www.motrona.com/level_converters.html

When I was researching encoders for a job a couple of years ago I did find some manufacturers who give maximum distances on their data sheets but I can't remember who it was. Please don't ask me to go back through two years of emails!!

I was looking a job where I was working in collaboration with a company who do a lot of under sea ROVs (remote operated vehicles). The electronic type boffin they had, ignored all the recommendations and tested things for himself. He got a 100 Metre reel of 8 twisted pairs and then sent the signal back and forth through the cable using a standard 24V encoder. The system would still work after going through the cable 4 times, a journey of 400 Metre. I don't know how much electrical interference there was where he did the test, but again, it put things in perspective.

Bryan

Note about encoder voltages...
Simply going to a higher voltage is not a solution. 24V output encoders would be much more likely to develop waveform problems then 5V encoders over the same wire length.

The very best output for long-haul encoder signals would be line-driver differential.