need help in VR2 Joystick pls help

[dany151]

kindly, i hv wheelchair controlled by VR2 Joystick
http://www.pgdt.com/products/vr2/index.html

the problem is the joystick D50800-3, i wanna use logical control not mechanical cause the disabled hand is cut...pls help to find
another way to control this wheelchair

http://pgdrivestechnology.com/remoteser ... 50800.html

my regards

[Burgerman]

Logical? You mean by switches ON/OFF?

Or by joystick NON PROPOTIONAL switched control?

Or some othe means?

You need to explain more.

Switches and other non analoge input systems usually require an omni input device, and this includes an air switch for blowing or sucking, and an analogue input that can be latched etc.

More explanation of what you are trying to achieve is needed.

[dany151]

thnx for reply
i will build the switching circuit after i understand the work of joystick ... i got the manual of D50800 from below
http://dl.dropbox.com/u/846109/mowstation/D50800-03.pdf
this manual gives the input and output pin connection and the voltage across them ... my problem is
when i turn on the VR2 and start the wheelchair and read the signal it appear as a square wave with negative part
in scope as shown below


is it a type of safety or it is result from the effect of other components in the board ???
what can i do to solve this problem pls?

[LROBBINS]

I suspect that the VR2, as with almost every current-generation wheelchair controller, is using some variant of the CAN bus digital protocol for communicating between modules. In general, it is very, very difficult to reverse-engineer the information "=frames" being sent over CAN bus as it was specifically designed to allow equipment manufacturers to keep their programming "private". There are two signal lines, called CANhi and CANlo, and the difference between the two is read to decide whether a bit is high or low. Over-simplifying: Each frame contains a multi-bit (11 bits for CAN-A, 29 bits for CAN-B) identifier that identifies the sending module and the type of information in the rest of the frame, a bit that is used to decide priority among different incoming information which may arrive simultaneously from multiple modules, a number of bytes of information for the actual message, and an ack bit that a receiving module sends out when it has received an intact message. It is a very robust, noise resistant, error-checking protocol first invented by Bosch for heavy equipment, but now, much faster than in the first version, the basis of most computer communication in automobiles. For example, the OBDIC on board diagnostics information is CAN.

None of this is simple if you want to get into it at the nuts-and-bolts level. If you want a copy of a pdf primer on CAN from a CAN chip manufacturer, send me a PM with your e-mail address and I'll send it to you. I'm only in the CAN programming learning stage myself, so have little to offer beyond what I've written here, or what's in that pdf. In general, for a wheelchair what you need is the EC manufacturer's programming software and cables (and/or dongle), preferably the OEM level software as the dealer or tech level software has many important functions disabled.

Ciao,
Lenny

[Burgerman]

I think he means the analog joystick output, or voltage supply.

It is likely not a smooth DC voltage and is "switched" 5v from the 24v or 12v supply. So will be a fixed frequency and a specific pulsewidth. It works just the same as an analog voltage. Use a voltmeter instead.

[dany151]

dear LROBBINS it isnt a matter of communication, i wanna understand the signal of the joystick
my email :engmoh81@yahoo.com
if u hv an idea pls call me..with regards

[Burgerman]

Joystick output is a simple dc voltage swing. Outrput voltage at mid stick (central) is HALF the supply voltage On 2 axis. The thing you are seeing is the power supply pulses. It doesent supply a smooth 5v DC pulse, but a chopped signal. It derives its 5v from a higher voltage and uses a shorter pulsewidth. You are seeing this on the output too I presume. The voltage swing will be around 1.2v above and below the centre point of the supply.

[dany151]

Right ...i know that dear Burgerman ...and thnx for your help

Know what i do...??
make a driver circuit and use manual switching as a checking circuit...or it waste the time ?
or use microcontroller and DAC ?
how is the figure of signal which i will apply (pure DC or square wave)?

pls advise me

[Burgerman]

DC

[woodygb]

http://www.electro-tech-online.com/elec ... assis.html

[Burgerman]

Woody - this http://dl.dropbox.com/u/846109/mowstation/D50800-03.pdf

shows its a common dc voltage swing joystick. I have several here on my bench. This is the signal he is looking to replicate.

[woodygb]

Perhaps it is ... never quite sure with the language barrier ...I'm merely supplying possibly pertinent info.

[LROBBINS]

I think that I did misunderstand what you were asking about. Are you wanting to substitute just the joystick pot and feed directly into the electronics in the joystick pod? If so, yes that can be done, and indeed on my daughter's first chair that's exactly how her head and foot switch control was implemented 20+ years ago. I don't seem to have a copy of the old schematic on my computer (it was probably just drawn on a piece of paper), but basically the switches just selected among 4066 digital-analog switches to engage voltage divider taps for forward, reverse, left and right. The outside of the voltage divider was attached across the supply leads to the joystick pot, and the taps fed to the signal lines from the pot to the chair's single board analog computer. On that chair, the joystick pot was turned 45o to the chair axis, so the two outputs were directly left motor and right motor. On modern chairs, the joystick is usually in line with the chair axis so that one output is forward-aft and the other is left-right with the mixing for the two motors done in the computer. You could also directly connect voltage dividers with switches, but by using the 4066 as an intermediate a single SPST switch can select the dividers for both channels; it's actually cheaper, simpler and more reliable. (It also means that the analog signal is only traveling between the box with the 4066s and the pod, not up to your head or feet or whatever, and what's on those "long" wires is just a nice solid on or off; see below where I mention EMF concerns.) These days it would probably be even cheaper to use a microcomputer to do the selection.

It actually doesn't much matter what the voltage to the pot is on your controller, nor whether it is real DC or pulsed DC - you're going to be dividing that with some resistors to get the two outputs. 50% will surely be "neutral" and you could even figure out the swing needed using some regular potentiometers - it's usually less than a 0 to supply voltage swing so that 0 or supply are sensed as errors that mean something is disconnected.

Basically simple, but there is a serious thing to consider about doing things this way. These are analog signals going into a (probably high impedance CMOS) circuit. There is a risk of electrical noise sensitivity. Within the chair pod those wires are short, and perhaps just twisted to reduce noise. If you are going to have a remotely mounted voltage divider, it would be best to keep the lines from it as short as possible, shielded and with ferrites, and then do some testing around strong EMI sources before trusting your bottom to the chair. On Rachi's old chair this wasn't as much a concern as the analog computer was not CMOS and had some fairly good built in noise rejection because it was receiving it's own analog joystick input on a long cable too. So, it can be done fairly simply, and Burgerman has done this for RC control, but some caution is in order.

The other approach is to go from your switches to something that puts out the appropriate CAN signals - and for P&G that's the OMNI, but it is a pretty expensive option and you'd still need the P&G programming stuff. On Rachi's current chair with a Dynamic DX controller, the switches connect to a DX-4SW box that sends CAN to a DX-SCR so there are no voltage dividers and no analog signals. It still has an adapter that contains 1 4049 and five 4066, but that's so that the switch outputs can be multiplexed to control her voice output computer and other stuff as well as the chair rather than for selecting voltage dividers.

I'll send you that pdf about CAN basics.

Ciao,
Lenny

[dany151]

thnx woodygb ...it is very helpfull..
im thinking of remove the cover of D50800-03 then find the contact points of the 4 directions (Forward,backward,right&left)
then use 4 relays instead of it? is it good idea ...??
pls advice me cause i didnt hv spare parts ...

[dany151]

dear LROBBINS ....thnx alot i got the datasheet of 4066 as shown below
http://www.fairchildsemi.com/ds/CD/CD4066BC.pdf
it is good....can i use it with microcontroller instead of D50800 joystick? is it good idea?

[dany151]

dear woodygb the pdf which u posted...is perfect .... thank you very much,
and i wanna ask if there is a way to read the serial data from the joystick?
i tried to use scope to read the data between (white & black)wires
but it give a signal with 12khz with different shapes i will post it asap..
any other way to read this data by scope?
or by data acquisition card with matlab ?
is it help idea or there is a protection prevent to read the correct data?

[woodygb]

The data being sent and it's translation is beyond me...Good Luck!

[dany151]

okay..dear
thnx for your help...

[woodygb]

If it helps I believe the protocol is J1939 or CANopen.

http://www8.cs.umu.se/education/examina ... Astrom.pdf

[dany151]

thnx, but i expect that i will find an easy way to get the signals of joystick ...
i read your posts
viewtopic.php?t=1965

viewtopic.php?f=2&t=1947&st=0&sk=t&sd=a&start=40

they are helpfull.....to control the wheelchair.....

but i need jest replace the joystick...

[LROBBINS]

CAN open specifies the format of the frames, not the content, and I'm afraid one needs to know what P&G's content is if you want to communicate with a P&G power module. Once one manages to figure out a few of the P&G messages, it's probably not all that hard to crack the rest, but figuring out the first one, when lots of messages are simultaneously circulating on the bus seems like a pretty rough task if the manufacturer doesn't want you to know. If you sniff all of the binary information in a frame, how do you know if this was the left pot output, the right pot output, some other signal related to fault checking or motor current, or battery level or .....

4066 can definitely be triggered by a microcontroller. A few months ago I added an accelerometer-based shock and tilt sensor to Rachi's chair (there's a thread here somewhere about that, and I think it includes a schematic). I started out with comparators and discrete components, but in the end I used an Arduino Uno to read the analog-output accelerometer and control two gates of one 4066 chip. If the chair gets in a situation in which driving is "uncomfortable" one gate opens leaving a resistance on the inhibit pins of the Dynamic power module putting the chair in slow. If the chair gets in a dangerous state, the inhibit pins see an open circuit and the chair stops. The Arduino also monitors the difference between CANhi and CANlo to detect when the chair is turned on so that, if the chair has been off for an extended time, such as at night, the accelerometer is re-calibrated when the chair is turned on. In any case, the control pin on a 4066 gate needs almost no current, just a very small fraction of what the microcontroller pin can handle, so you can even simultaneously control quite a few 4066 gates from just one digital output pin if you wanted to.

A 4066 is much better than a relay for switching small AC or DC currents with non-inductive loads. Just stick to what the data sheet says and you won't get in trouble with these.

Ciao,
Lenny

[dany151]

Dear LROBBINS ...is it this post

viewtopic.php?f=2&p=22226

[LROBBINS]

That was the first attempt. More information here:
http://www.wheelchairdriver.com/board/viewtopic.php?f=2&t=2001&p=25821&hilit=+accelerometer#p25821
Ciao,
Lenny

[dany151]

i found this Semi Automated wheelchair pdf
http://www.eecs.ucf.edu/seniordesign/fa ... 2Final.pdf

page No.99 - 100
5.5 Joystick Interface
and there is a circuit of interfacing....is it ok? pls check it

[Burgerman]

Yes, its a simple analog voltage swing joystick. As found in almost all mobility controllers.

Read my RC page, or look at the joystick PDF on my roboteq/BM3 page for more specific info.

[woodygb]

It also shows ..as referenced in one of my links .. a second output for each axis ( two hall effects per axis ) ..this is probably a safety double check feature ...thus both outputs are being checked against each other...any major deviation between the two would probably throw up an error.

[Burgerman]

Those are not evident in all mobility products. Including the pilot plus on my desk. So likely not required.

[dany151]

ok i will read them...

[dany151]

loooook at this work ... it is very nice

http://www.instructables.com/id/Psycho- ... /?ALLSTEPS

[Burgerman]

All my wheelchairs are radio controlled. Theres a page somewhere onm site that shows how. I used to drive my ex gf around like that. But its just useful for loads of stuff. Like when swapping chairs, taking one to my garage, loading another in the van etc. Or walking the dog remotely up and down the street when its raining...