Cousin of OSW (Open Sim Wheel)

ok i will run the tests again

Hi,
I would like to ask for some help with setting up the rotary encoder for the wheel.
I'm using an OMRON E6B2-CWZ6C Rotary Encoder 2000P/R.
So far I managed to get the A B channels working with the Ai-wave wheelconfig software and Arduino Leonardo, by using 1k pull-up resistors to 5V.
I can't get the encoder's Z channel to work, with our without pull-up resistor. When it's connected, the rotation of the wheel is not measured anymore.
If I simply leave out the Z channel, or use a switch instead, I get a calibration error message from wheelconfig. It never shows the Ready message.

I also tried the MMOS software and the STM32 discovery board, but there I couldn't even get the wheel rotation counter to work at all.

If anyone is using a similar encoder, I'd greatly appreciate any help as I'm stuck at this point. My friend, Google was no help either.
Thank you.

I haven't used an omron type encoder before but it comes to mind that you should be atleast have the Z(index) pulled down to ground instead of tied high. It should only see a high pulse when it passes the index mark not constant +5v with the index pulse of low signal.

What type of pullup resistor are you using ? I run a pulldown resistor on the z channel

Post up your settings for the advanced screen and the main screen

Are you running the 'dev' mode of the software ?

I'm using a 1 kOhm pullup resistor on A and B.
Sadly, on my encoder, the Z index is always at 5V with a low voltage signal when index is reached. Using a pull-up has no use as it should be working inverted. Is there a trick for it?
I fiddled some more with the switch type index connected to GND and switched when calibrating. It worked once somehow, but it's dead again.
Also the wheel center button does not do anything. I don't know if it should center the wheel.
What is 'dev' mode?

Leave the z channel disconnected and fit a resistor to gnd on channel 6

Turn everything off and on again then connect the channel 6 to ground when you select calibrate.

Look in your wheelconfig folder you should see a icon called dev - it displays the counts of the encoder

I had a resistor between 5v and ch6 and a wire shorted ch6 to gnd. I momentarily removed the wire from ch6 gnd after pressing calibrate. Do you mean the resistor should be between ch6 and gnd, and short to gnd when calibrating?
I wonder why it worked once randomly.

I can't find a dev icon in the wheelconfig folder. Can you provide a link to a version that has it?

create a new shortcut - it should be "wheelconfig.exe -dev"

you only need to connect ch6 to ground - if you fit a switch you can break the connection to get the wheel calibrated

does the wheel icon rotate when you turn the encoder ?

are you using a leonaro to control the wheel ? what type of motor driver are you using ?

Yes, the triagle icon does rotate.
The only problem is the calibration error message. It pops up as I open wheelconfig and can't make it go away even with the ch6 gnd switch.
I use a leonardo.
Earlier I flashed a pro micro with the EDtracker firmware, since then every leonardo I plug in has the EDtracker name in game controllers.

It's early to plug in a motor driver I suppose. I will be using an IBT-2.

if the triangle rotates than all is good

get the dev version working and you can count the pulses

watch the videos carefully at the start of this thread make sure you have covered each step

set the advanced settings to pwm +/- if you are using a ibt 2

you can run a light bulb instead of a motor if you want to check it is working

I gave up on the pull-up resistor mess. Strangely when I connected Z index directly to ch6 on Leonardo, wheelconfig reported "Ready" state with no calib error, even the numbers were counting below on the dev panel. Only the triangle is dead, except when I reach index, then it moves a bit then snaps back to zero. That is because index is inverted.
When I disconnect the index lead, the triangle starts working, but as soon as I hit calibrate it responds with error until I reconnect the index. Grounding ch6 or using switch does not help.
I wish there was a way to modify the FFBWheel.0.16.leonardo.hex firmware to make the input pins pinMode(6, INPUT_PULLUP); and have an option to invert the index signal in software.

Therefore I pulled out an Arduino UNO and wrote a simple sketch to pass thru each encoder signal from an input pin to an output pin and invert Z index in the process.
Now the encoder is only connected to the UNO, receives 5V from it, and the outputs are connected to the Leonardo's 0,1,6 inputs. No pull-up resistors are needed since the pull-up is set inside the UNO.

As a result, surprisingly it works without any errors and calibrates nicely.

wow thats an innovative approach !

why not try and track down the software writer and see if there is any software and firmware updates ?

i have found you need to break and make the connection on pin 6 when calibrating it seems to work ok

After some more testing, this Arduino UNO passthru idea has some flaws. At fast rotational speeds, it can't keep up and loses steps.
Need to look for another approach.

my guess is your pull up resistors need to be larger in the original board

perhaps you need to use an interrupt so you don't loose pulses

Maybe I will use a hybrid method. I found an Arduino Pro mini (smaller than teh UNO), I will use that on the Z index and connect A,B channels directly to Leonardo as it was working originally.
Or I make a logic inverter circuit with a transistor for the Z and toss the extra Arduino.

if you were to pulse the z channel once every second would it affect the a,b channels ?

I guess it won't affect them much, that is, if you give the pulse where it needs to be. Every pulse of the Z channel centers the wheel.
Anyhow, I made some progress with the logic inverter. At this point, I can say it's all working, including the Z index.


R1=10k
R2=1k
Q1: BC546 NPN transistor
X is the input from encoder Z index, not shown on the picture there is the original 1k pull-up resistor to 5v
Z is the output to Leo pin6

Just for the record, this is the final result for OMRON E6B2-CWZ6C rotary encoder.
Tested and working. Any other simple NPN transistor should work.