Fitting an encoder on a Tormach 1100

I have updated my 1100 to version 2.1 and would like to connect an encoder to the 5i25 Mesa card P2 from what I have read on another post Peter (PCW) suggested kindly that he could make a bit file to suite. I understand that on a future update the bit file would be overwritten but I think that can be solved by stopping it following this very good tutorial by Sarah (cncbasher) forum.linuxcnc.org/pathpilot...tutorial-part1

Ok questions.
a) What could I use to connect to P2 on the 5i25 , I am thinking that a PP card would be to slow for a 1000 or 2000 P/R A,B,Z encoder. I also have a 7i76 and 7i73 but if possible would prefer not to use these on this project.

b) Could I get help in modifying the Hal & ini file to achieve this.

c) Would the new bit file interfere with the general operation from P1 on the 5i25 (ie does PathPilot have the same P1 side bitfile as the general stock 5i25 board.

a.) If you have a full size case you can get a bracket/cable adapter from Mesa ( or elsewhere, this is basically the second parallel port hardware from dual parallel port cards ) A parallel port would limit you to a few hundred RPM if you use a 1000/2000 PPR encoder

b.) Sure, encoder setup is not very difficult

c.) Maybe... Tormach mill bit files use a custom encoder component the does hardware position extrapolation for use with RapidTurns 1 wire index/encoder QA. This encoder uses the same register addresses and register map as the standard 3 wire encoder so the two encoder types cannot be used simultaneously (and least without hacking the firmware build system). So theres no issue making a Tormach compatible bitfile with a standard encoder interface on P2 but you would lose the ability to run Rapidturn (at least simultaneously)

I should note that there is no "standard" 5I25 bitfile, they are all basically custom made for a particular BOB/combination of BOBs

Many thanks Peter for the rapid response
.

A parallel port would limit you to a few hundred RPM if you use a 1000/2000 PPR encoder

Is this because of the slow opto's? would a 600PPR be better? Do you have a rough idea how fast I could run the spindle?

If I used my 7i76 Would I have enough encoder inputs to have 3 mpg's plus the spindle encoder (if I put 2 mpg's on TB5 pins 1-4)?
Is there a better card to do this?
.

Tormach mill bit files use a custom encoder component the does hardware position extrapolation for use with RapidTurns 1 wire index/encoder QA. This encoder uses the same register addresses and register map as the standard 3 wire encoder so the two encoder types cannot be used simultaneously (and least without hacking the firmware build system). So theres no issue making a Tormach compatible bitfile with a standard encoder interface on P2 but you would lose the ability to run Rapidturn (at least simultaneously)

I don't intend to use Rapidturn.

Its not a problem with OPTOs, the issue with a parallel port counting encoder pulses is that the counting has to be done by the host computer in software, typically the maximum rate that software can sample the inputs is about 50 KHz and 25 KHz is more likely. A 2000 PPR encoder has 8000 edges per turn. For reliability you need to sample around 2 times for every edge so a 25 KHz sample rate would only get you to about 1.5 RPS or 90 RPM before you started to lose counts

If you use the 5i25s encoder counter, you do not have this speed limitation ( the only limitation will the encoder itself )

If you are just adding an encoder to the 1100, I would chose one with differential outputs and I would use the second 5I25 connector and add a differential to single ended converter ( either one from AMT/CUI or a Mesa 422ENC or ENCY )
to convert from to differential to the single ended signals the 5I25 expects (and protect the 5I25 from noise/wiring mistakes from the encoder leads)

If you are just adding an encoder to the 1100, I would chose one with differential outputs and I would use the second 5I25 connector and add a differential to single ended converter ( either one from AMT/CUI or a Mesa 422ENC or ENCY )
to convert from to differential to the single ended signals the 5I25 expects (and protect the 5I25 from noise/wiring mistakes from the encoder leads)

.
Unfortunately All the encoder I have are single ended and I have already mounted the encoder on the mill.
.
So If I used my 7i76 Would I have enough encoder inputs to have 3 mpg's plus the spindle encoder (if I put 2 mpg's on TB5 pins 1-4)?
or is there a better card to do this?

keep your leads as short as you can and use twisted pair wires for the A/B phases. a feritte core might help too.
depends on how electrically noisy your environment is.

I believe the default behavior is for 1 encoder per channel on the 5i25 so that's 2 total.

smgvbest wrote:

keep your leads as short as you can and use twisted pair wires for the A/B phases. a feritte core might help too.
depends on how electrically noisy your environment is.
I believe the default behavior is for 1 encoder per channel on the 5i25 so that's 2 total.

Tnx Sandra. Yes I understand the reasons for the twisted pairs and ferittes etc. I have two projects on the go, one with the 1100 and the other with my Warco converted mill that I already have running with Linuxcnc with a 5i25/7i76 combo with mpg's on pins 1-4 TB5.
.
What I wanting to achieve first is the 1100 and the best way to use the second port on the 5i25 (I don't want to use Rapidturn) I believe that Peter (PCW) would be willing to make a bit file to use the second port.

Ideally I would like three Mpg's and the spindle encoder (that I have physically fitted) what are my options re cards to come off the 5i25.

thanks for posting.

Thank you Andy for the advise I will look back into it in the near future.

andypugh wrote:

And the second connector on the 5i25 remains a good bet.

Ok I am going to have another crack at this. If I understand correctly I can connect a simple bob to the second port on the 5i25 and use the fast encoder A,B Z on the 5i25 P2.

I don't need rapid turn. So will I need a bit file made for the P2 port for it to be used? if so will this affect the the general outputs driving the Tormachs pcnc 1100 BOB connected to the 5i25 DB25 port.

If the above is correct would Peter (PCW) kindly make me a bit file to suit when he has the time.Thanks