[cnc2]

Hi everybody !

I meant this thread for collecting information about the different encoder models used with the UHU, their price, reliability & the way they are connected to the UHU. (direct connection or through a differential interface ?)

Feel free to post your wiring diagram !

Here's what i found on the net as a DIY alternative :

1) http://users.on.net/%7Emerrifield/opto/index.html it's an optical incremental quadrature encoder i think it has a 5V output, TTL ?

2)http://www.cnczone.com/forums/showthread.php?t=55231 didn't try it yet & have no idea if it can work.


As i'm not an electronician i ain't sure this will work with the UHU so, if you know the answer let us know !


Thanks !
cnc2.

[Al_The_Man]

Those are both crude low resolution encoders, the second link appears to be ~80p/rev, which would come out to ~320p/rev after x4 conversion.
a 1200p/rev (before x4) encoder would require a huge diameter wheel using this technique.
In order to make a compact hi-res encoder, it requires a bit of smoke and mirrors called the Moiré effect grating.
A search in the forums will show an explanation.
A 26LS31 can convert to differential line drive for a $1 or so.
Al.

[vroemm]

A few years ago i have been playing around with making a DIY encoder.
Made a drawing of a encoder disk with qcad.
Had it printed on transparent foil at a copyshop.
Here is part of the result, with pictures.
http://home.wanadoo.nl/vroemm/cnc/encoder/index.html

A other possible way to make encoder wheels:
Take a big white paper, like 1 x 1 meter.
Draw a very big encoder wheel on it.
Take a 6x6 cm photo camera, which has slides or negatives of 6x6 centimeter.
There must be a lot of 6x6 cameras around doing nothing, because of the digital cameras revolution.
Put high resolution (?) slide film in it.
Take a few photos with it of the paper with the encoder wheel.
Develop the film.
Then you have a encoder wheel, but how good will it be ?
A lot depends on the quality of the lens.
Some questions remain open, like:
How expensive is 6x6 slide film ?
How expensive is it to develop the film ?

Other tip.
Put the light sensors and leds on a line in the direction of the center of the encoder wheel.
This reduces the errors in pulse width.

Vroemm.

[Al_The_Man]

Did you happen to test the 50mm 500 line? How successful was it?
The usual problem is the line density is too high for a conventional photo cell arrangement to read.
For example the line spacing on 500 line, 50mm disc will be around .0065" or .165mm.
This is going to make it almost impossible for common photocell transmitter/receiver to distinguish between the lines.
Al.

[contactirfu]

Hi Guys,

I have with me the US digital ones the mini's EP4's I think, at 250ppr they are low res. but still i think will work for me, when I bought them it was like 17 usd now they are 30+,

have to see if they are good for my router or I need to switch over to the 500ppr or more.

and yes I made the Differential line drivers using the 26LS31 based of schematic on the US digital website.

RGDS
Irfan

[vroemm]

Did you happen to test the 50mm 500 line? How successful was it?
The usual problem is the line density is too high for a conventional photo cell arrangement to read.
For example the line spacing on 500 line, 50mm disc will be around .0065" or .165mm.
This is going to make it almost impossible for common photocell transmitter/receiver to distinguish between the lines.
Al.

It is a long time ago.
I used CNY37 led barriers.
And i got a sinus like signal floating somewhere between 0 and 5 volt.
A comparator made it into a square wave.
There was instability in the signal, never found out if it mattered.

Then i stopped experimenting, other priorities :-)
It never was used to actually control a servo motor.

Still interesting puzzle to make a encoder, my hobby hart itches again, but lots of other things to do. :-)

Could a laser from a old cd player be used ?
It has a much narrower beam, i think.

Vroemm.

[cnc2]

Those are both crude low resolution encoders, the second link appears to be ~80p/rev, which would come out to ~320p/rev after x4 conversion.
a 1200p/rev (before x4) encoder would require a huge diameter wheel using this technique.
In order to make a compact hi-res encoder, it requires a bit of smoke and mirrors called the Moiré effect grating.
A search in the forums will show an explanation.
A 26LS31 can convert to differential line drive for a $1 or so.
Al.

Thanks for the reply Al !

For the second link, if you enlarge the picture you will see that the encoder wheel isn't the toothed gear but a 50mm plastic disk stuck (glued) on it, it has so many lines that they appear to be a grey band, you can distinguish the lines if you look at a lamp bulb through the disk, it's 200LPI (i think it stands for Line Per Inch) & 1200CT (count).

I was thinking about using this disk with a mouse sensor, but i ain't sure it can detect such a thin lines ?

About the 26LS31 can you post a pinout or a diagram for connecting it to an encoder ?

Thanks !
cnc2.

[Al_The_Man]

The link for the data sheet is one http://pdfdata.datasheetsite.com/web/58498/26LS31.pdf
Example hookup would be:
Pin1 = Ain
Pin2 = Aout
pin3 = /Aout
pin4 = 5v+
pin5 = /Bout
pin6 = Bout
pin7 = Bin
Pin8 = Gnd
pin9 = Zin
pin10 = Zout
pin11 = /Zout
pin12 = NC or GND
pin13,14,15 spare fourth channel
pin16 +5v
Al.

[cnc2]

Thanks Al !

An encoder like the digikey amt102 has A & B channels, what about the /Aout & /Bout.
from what i uderstand Ain is from the encoder before the diferential interface & Aout is from the diferential interface to the UHU, am i right ? but what about the /Aout (not Aout ?) is it needed for the UHU ?

Zin, Zout & /Zout can be used for an index channel, am i right ? so, if i don't have an index channel Z won't be used ? like the remaining fourth channel.

Is there any redy made layout for such an interface or a bread board will do the job ?

Thanks !
cnc2.

[Al_The_Man]

The advantage to differential driver comes when both A & /A etc are used for noise elimination on long conductor runs, if using only one output, you may as well use the single ended output direct as shown in the first link.
If you really wanted to use differential but only had single ended input at the receiver, a 26LS32 receiver could be used.
Al.

[cnc2]

Thanks Al !

After verification the AMT102 from digikey has only A,B,Index,5V & GND. And i ain't sure if UHU uses (needs) /Aout /Bout.

is there any way of implementing the /Aout /Bout from Aout & Bout on the differential board before the 26LS31 ?

Else, how to make the differential interface assuming we only have A & B, 5V & GND (index isn't used by UHU) can i expect any noise elimination this way ?

Thanks !
cnc2.

[kreutz]

Thanks Al !

After verification the AMT102 from digikey has only A,B,Index,5V & GND. And i ain't sure if UHU uses (needs) /Aout /Bout.

is there any way of implementing the /Aout /Bout from Aout & Bout on the differential board before the 26LS31 ?

Else, how to make the differential interface assuming we only have A & B, 5V & GND (index isn't used by UHU) can i expect any noise elimination this way ?

Thanks !
cnc2.

The original UHU does not use differential signals, so your encoder seems to be OK.

In order to use a differential interface you need then a differential transmitter chip on the encoder side and a differential receiver chip on the UHU side. The interface cable should be shielded twisted pair (with 4 pairs minimum) for best results.

Regards,

Kreutz.

[cnc2]

The original UHU does not use differential signals, so your encoder seems to be OK.

In order to use a differential interface you need then a differential transmitter chip on the encoder side and a differential receiver chip on the UHU side. The interface cable should be shielded twisted pair (with 4 pairs minimum) for best results.

Regards,

Kreutz.

Thanks for the reply Kreutz !

So, if i understend well to eliminate the noise over a long encoder cable i need to build a differential interface with a transmitter chip & a receiver chip, so, is it possible to do it with 26LS31 & 26LS32 ? and how ? (any pinout or diagram)

Is it possible to use an FTP cable as an interface cable ? it has 4 pairs of twisted wires enclosed in a sheath of aluminum foil, is this the needed shielding or do the pairs have to be shielded from one another ?

Do you have an idea about the bandwidth used by the encoder data ? (how many Mbit/s)

(sorry for my english that is reaching its limits, i hope i wasn't talking chinese)

Thanks !
cnc2.

[contactirfu]

Hi CNC2, we are all from earth, where exactly are you from!

just wanting to know better

here are the circuits for the driver and reciever on this thread.
http://www.cnczone.com/forums/showthread.php?t=56547

using the network cable may work, but since those cables are of single strand they are not rated for continuous movement. you may consider multistrand twisted pairs.

RGDS
Irfan

[contactirfu]

Vroemm / henrik / Josef,

can any one of you build simple schematics for a the differential driver based on the above chips and driver schematics( meanig PCB design). , probably a D9 output and soldered inputs of A B + and -

RGDS
Irfan

[H.O]

Hello,
I'm not Kreutz but I think I can help answering this one....
The HP-UHU already has the differential receiver interface on board so if you have a single ended encoder you need to add an differential transmitter such as the 26LS31.

(Please note that it is possible to bypass the differential interface all together and just use the single ended encoder as is but it's not as noice tollerant as the differential interface)

The 26LS31 takes the A and B signals from the encoder and produces complementary signals (A, /A & B, /B) and is capable of driving rather long cables. As an alternative to the 26LS31 you may look at the DS9638, I've used that a couple of times with good results.

On the HP-UHU board there already is a differential receiver which converts the signals back to plain old A & B that the servo-processor reads.

As far as the cable goes, shielded, twisted pair cable is the best. You run each complementary signal-pair down one pair. TP-cable will work just fine from an alectrical point of view but is usually built up from solid wire and will fail in a rather short period of time if used on a machine where it moves all the time. There is TP-cable with stranded wire availble and that should be alot better.

HTH
/Henrik.

[H.O]

Hello,
Attached is a schematic of a possible circuit using the 26LS31. Please note that this has not been tested by me but it 'should work'. With that being said if anybody sees any mistakes, please let me know.

HTH
/Henrik.

[contactirfu]

Here is what is there from US digital web site, if any one can design a pcb around these then it would certainly help!

presently I have done that on a general purpose board, its working well but a little messy.

RGDS
Irfan

[vroemm]

I can not find 26LS31 and/or 26LS32 in the eagle library http://www.cadsoftusa.com.
The search function in eagle does not always work well.
Does anyone know in what lib folder to look ?

And which encoder connector is there on the HP-UHU ?
Do you have the eagle name for it ?
Can the same connector be used on the 26LS3x receiver for the original UHU.

Which pins belong together as part of the twisted pair ?

Vroemm.

[contactirfu]

Vroemm, use standard 16 pin dip footprints

the encoder connector can be a D9 similar to the one used on the HP UHU (rs232) we can have two pins shorted which is required for HP UHU to be active.

A and -A one twisted pair, B and -B on another twisted pair, +5V and GND on one and another pair to short - to keep HP UHU active.

there is no requirement for a receiver for HP UHU, but will help for regular UHU users, we can just have wires coming out from the receiver for the A B and 5v and GND lines.

RGDS
Irfan