1st generation AUTO cancel TSCU "banging the bits"

[rustybronco]

Am I not looking at this right or are your Left and Right switch's wired into output lines or is the bus on the 20M2 bidirectional?

The schematic shows a 20M2 pinout. I edited the pin configuration in diptrace using their component editor but I couldn't find a way to incorporate the new part into my diagram. Guess I should read the manual someday. Not to worry, the 20M2 and 20X2 are pin for pin compatible. The only difference is the pin nomenclature.

Here's the pinout for the 20X2... www.picaxe.com/docs/picaxex2.pdf

Almost all of the pins are configurable as either an input, output or (??).

Picaxe 20M2 for comparisons sake. www.picaxe.com/docs/picaxem2.pdf

[bccap]

Means about the same to me.
But I know it will work and work well.

I think its written in Murrican ? I'm so confused I'm going to go out in the sunshine for a ride and contemplate my indignation......

[rustybronco]

And may an electronically challenged Goose lay a nice big solenoid on your clean-dry windscreen while you do.

[bccap]

And may an electronically challenged Goose lay a nice big solenoid on your clean-dry windscreen while you do.

Never mind the TSCU. Figure out how I can put an air conditioner on the bike ? { nice comeback by the way }

[rustybronco]

I had a note cross the local hackerspace mailing list about a guy doing small-batch circuitboards online at circuitboards.com

If enough people are interested in (and can agree upon) a board solution, apparently you can get that guys to make you a few. I havn't checked out his site yet, so proceed with caution.

Thanks for the heads up! I plan on using OSHpark's board service when this project gets that far.

https://oshpark.com/

[rustybronco]

I just found out that seven of the 20X2's pins are Schmitt triggers. That may help with the noise issue on the cancel switch combined-in addition to software debouncing. Moved the output of the reed switch to C.5 of the µC.

http://www.picaxe.com/BASIC-Commands/Digital-InputOutput/inputtype/

https://en.wikipedia.org/wiki/Schmitt_trigger

[posplayr]

I just found out that seven of the 20X2's pins are Schmitt triggers. That may help with the noise issue on the cancel switch combined-in addition to software debouncing. Moved the output of the reed switch to C.5 of the µC.

http://www.picaxe.com/BASIC-Commands/Digital-InputOutput/inputtype/

https://en.wikipedia.org/wiki/Schmitt_trigger

With sufficent sw debounce period the Schmitt trigger should be irrelevant.

[rustybronco]

My thought's were to measure the switches bounce and use that to set the software delay. That or sample the switches state 10ms (?) later to determine if it's sufficiently settled.

I'd rather plan ahead for all possibilities if I can.

Thinking of an RC circuit to the BS170's gate as well. But as Ed likes to say, to measure is to know. Gotta see what's on the lines first.

[Wallowgreen]

I hope that works Dale. Remember though it can only operate within the supply it's connected to. If I remember the characteristics of the noise I saw was full scale (ie 12v or close to it) high Q and high frequency. I had to put a second order low pass filter on each input signal and then I was able to software debounce the output of that.

Fun ain't it

[rustybronco]

Martin, 1.5 great minds think alike. 1.499 yours and .001 mine.

On my way home from work tonight I was thinking about the possibility of having to add a low pass filter to the inputs. You know one other thing I been pondering for the past week or so? Robert Barr used LM7805's to down convert the voltage from the turn signal switch's to the 5 V necessary to feed the input of the 14M2 he was using.

Now why isn't he having issues with false triggering from the coils field collapsing?

http://www.cycleorings.com/piglet/picaxe%20tcu.JPG

[Highway_Glider]

.... having issues with false triggering from the coils field collapsing?

We're in $hit capt'n!

[posplayr]

I hope that works Dale. Remember though it can only operate within the supply it's connected to. If I remember the characteristics of the noise I saw was full scale (ie 12v or close to it) high Q and high frequency. I had to put a second order low pass filter on each input signal and then I was able to software debounce the output of that.

Fun ain't it

Was that power supply noise? I have measured about 250 milli-v rmswith noise synced to the rpm ( mostly from the r/r chopping ac).

Hate to ask but did you have adequate filtering on the regulator? Low ESR?

I know that switches can bounce, and debounce would be effective to a point. The other way is acomplish the same thing as a RC LPF without any parts is treat the inputs as analog and digitally LPF.

Still need a series current limit but that is it. Throw a cap on if nessesarily but just more parts.

[robertbarr]

.

Now why isn't he having issues with false triggering from the coils field collapsing?

http://www.cycleorings.com/piglet/picaxe%20tcu.JPG

Probably because that unit was never installed on a motorcycle. I sort of sold the bike it was intended for. Being 30 feet away from any running machinery doesn't provide much of a test of that aspect.

[rustybronco]

If you throw enough parts at it the problem should go away.

Might as well plan ahead for it. http://sim.okawa-denshi.jp/en/CRCRtool.php

[Wallowgreen]

Was that power supply noise?

No, it's coupled noise on the input lines from the switching at the coils. The wiring from the switch runs right between them. These days the input to whatever you try to detect a switch activation with ends up being low voltage at some point. It's a job to differentiate between what you do want to detect and the noise that's present all the time on the input signals.