Article [13] ST1300 - Gear Position Indicator (Do It Yourself)

[Geditor]

Ok, my final thoughts on a GPI install using the DiGi purchased off eBay for about $25.
I followed the instructions included with the DiGi and combined them with the install instructions found on the last page of this thread.
As I mentioned before, I used a yellow 6' cat 5 network cable that I had laying around, cut the ends off connected them to the DiGi.
for a switch to program it I purchased a $5 switch at Home Depot (GSW-41 Rocker Switch). (any 110/120v switch will work on 12v, just don't try to go the other way)
I soldered the wires together and covered it with a bit of shrink wrap. I am going to let the program switch hang out in the glove box for now.

I suggest getting your Speedometer pulse, power and ground off the left side of the bike under the mid fairing.
Getting them from the right side is a real pain. The wires are easier to identify, but there is so much more work involved. See the picture with my dirty hand pulling back the rubber cover so you can see the yellowish cover where the speedometer pulse, power and ground are available. Its really only accessable after removing the charcoal can and the tip over bar that are in front of it.
The instructions call for a Tachometer Pulse, this is the Ignition pulse found under the seat on the right side of the bike.
The programming was quick and simple, took less than 1 minute with my bike on the center stand.
I currently have it mounted on the dash on the left side, once I am sure where I want it the additional surgery can be done later.
I am thinking it will probably go inside the dash in the bottom right corner of the speedometer in that blank space.

 

[Helitronix]

I wanted to give a status update.

I finished my hardware prototype, and developed two different but related programs that run on some common hardware. One program, let's call it "GPI-Passthrough" is the one that monitors the tach and speed sensor, relays that information to a GPI plugged into it, and when the clutch is pulled in, will output tach and speed sensor data to the GPI from 1.5s before the clutch was pulled in. This maintains the value of the last gear the bike was in, on the GPI. You pull in the clutch while sitting at a light, and the GPI doesn't wander and display the wrong gear.

After developing that first program, I decided to build my own GPI since the hardware was essentially the same with a few minor tweaks. I made one perfboard prototype with the components necessary for both. The second program, the full GPI, monitors the tach and speed sensor, monitors neutral, monitors the clutch input, and functions as a full GPI. It calculates the ratio, and displays the inferred gear to a 7-segment display. And like the GPI-Passthrough, if the clutch is pulled in, the gear that is displayed does not wander because it is not recalculated until the bike is in gear again.

So all of this is well and good, but I'm finding a problem and could use your help. The whole reason for being, for this creation, was to help do a better job of displaying the gear, because it wouldn't wander when the clutch is pulled in. In order for this to work, the clutch switch must work reliably, and ideally, it would engage as soon as you start to pull the clutch in. Instead, I'm finding two issues. 1) The clutch switch doesn't register anything (on my ST at least) until the clutch is pulled ALL THE WAY IN. As in, touching the grips. And 2) The clutch is doing its job, slipping as it should, while the clutch is partially pulled in. When the clutch is slipping, the ratio data that is calculated is completely inaccurate because there is not good correlation between the tach and VSS. Prediction can help a little bit, but it's not sufficient for making it 100% error free, 100% goof proof, 100% perfect. Basically, the clutch switch operates at the "end of stroke" instead of at the "beginning of stroke" where it would be more useful. I just didn't think of that when I got started. I was thinking that shifts would happen more quickly and this wouldn't even be an issue. But, as I was testing it, and shifting with my right foot and right hand, while standing on the left side of the bike (on center stand) I saw just how crazy all of this can be.

At this point, it is clear that monitoring the clutch input as-is, is not quite right. For the whole thing to work as planned, I need to implement a switch that changes state at the beginning of stroke instead. I could certainly build a one of a kind switch with a hall effect sensor, or possibly an optical sensor at the base of the clutch lever, but that might not be all that suitable to anyone else in the community. So I don't know if the project has any value.

So a couple of things then..

1) Do you have a good idea for how to add a clutch switch that would give us reliable beginning of stroke data to use as an input?
2) If you are interested in having this capability on your ST, please speak up so I know if anyone actually cares if this is resolved.

One goody that came out of all this playing around is that I built a GPI device that I can eventually use to provide telemetry to the heads up display on my Skully AR-1 (And no, it hasn't arrived yet. They are thinking it will be here in May, umm, 2016). Now that I have the GPI data from my device (clutch or no clutch), I can send that over Bluetooth to a smartphone and use the Skully software API to relay that information to the HUD. This will allow me to avoid mounting a 7-segment display to the bike itself and put the data in my field of view where I want it. I'm fairly excited about that! So, if nothing else comes of this experiment, I'll have some useful telemetry for my helmet. And that, I couldnt' do with an existing GPI.

So 3, if anyone is interested in the design for the purposes of the telemetry, for whatever purpose, let me know too. I'll still be releasing full project details and source. I'm just wondering if anyone else would care to do the same thing.

That's all for now.

Joel

 

[T_C]

Joel,
At this point would it be easier to just put a hall effect sensor set-up on your shift lever and track the gear changes there? Eliminate the clutch issue all together? A little more work figuring out how and where to mount the two sensors but it eliminates a lot of other work of correlation of VSS and IPG signals.

 

[MaxPete]

Hi All: I think that T_C has put identified something worthwhile here. MY '79 GS850G had a GPI and I am pretty sure that it only interfaced with the transmission (I suppose the shift drum) and had nothing whatever to do with the clutch or the VSS (if that old girl even had one of those....). It was pre-electronics days for goodness sakes - that bike had a kickstarter and points ignition (as well as the GPI, a self-cancelling turn signal and a decent fuel gauge).

Anyhow, as I recall it, the GPI didn't seem to care what was going on with the clutch or the bike's over-the-road speed - it simply indicated what was going on inside the transmission (which is of course the information we're after).

Obviously, the difficulty with emulating the same approach on an ST is that we do not have a graceful way of interfacing with the mechanical bits of the transmission itself - but is there a way of sensing motions of the gearshift lever and then keeping track of "ups" and "downs" to determine what gear the bike is in - regardless of speed? All the system would need to know is when is the bike in neutral, and then what happens with the gear lever after that - and I think that the neutral signal is readily available.

I may have missed something here but its just a thought in the vein of keeping it simple.....

Pete

 

[T_C]

A pair of hall effect sensor mounted on the bike and a magnet on the shift lever.

Not the easiest thing to mount the sensors on the bike side but neither is mounting a second clutch switch.

 

[Helitronix]

If it were possible to mount the sensors on the shift lever so that if the clutch lever is ever past a certain rotational position, then it guarantees that a change in gear has occurred, it would work. However, my foot, which isn't a perfectly reliable indicator, tells me that the engagement point is slightly different from 4th to 5th, and, if you preload the clutch lever and then pull in the clutch, the engagement point seems closer than if you pull the clutch in and then manipulate the lever. I don't know if the second point is true. It just feels that way to me. Maybe someone has some data. Unless the shift lever position was a perfectly accurate method to determine # of up shifts and # of down shifts, I think it would still be problematic. That said, if someone wants to mount a sensor and try it to see if it's reliable, I'd like to see the data. An optical sensor with a encoder wheel disk, or a digital potentiometer would help a lot. I'm just not sure what it would be like to mount one to the lever. I haven't given that a good look.

Joel

 

[MaxPete]

What about two simple limit switches each with a plunger that is tripped when the (or just before) the gear level reaches the stop....

 

[Helitronix]

I'll take a look at some options to see what I can reasonably make and mount.

Joel

 

[Geditor]

After riding a few hundred miles with my new GPI I have found 3 things.
1) I don't know how I went without it for so long.
2) It is very accurate
3) China quality still isn't great. The top left vertical segment died.

So Heliotronix, if you want to put out a manual or a kit for your design, I will jump in, and get a better quality segment display.

 

[MaxPete]

Good to hear Geditor - I can't wait to get mine. In fact, yesterday on only my second ride of the season I found myself howling along at 4000 rpm in 3rd gear when I should likely have been in 4th or 5th. Oh well.

 

[Helitronix]

I looked at the gear select pedal and the clutch lever. I think it will be easier to mount a hall effect sensor and magnet to the clutch lever than it will be to reliably detect and up/down motion on the gear pedal. Either could be done, but I think the clutch lever solution will be easier to implement. In the first photo, I've circled where the Hall Effect sensor could be mounted, and in the second photo, I've circled where a magnet could be mounted, with an appropriate mount that puts it at the right distance to the sensor. That second photo shows what the lever looks like when it is extended, just to show how the magnet will pass over the sensor. In the "clutch out" position, the magnet would be aligned to the sensor, and once the clutch lever starts to move a few degrees, it would no longer be in range of the sensor and would give the output I want to signify beginning of stroke.

I'll mock it up with a hall effect sensor and magnet that I temporarily fix in place to see how it works. If I like it, I can make something more permanent. I'll update here as I figure it out.

Once I have something together that makes sense, I'll share the details for how to build the GPI and/or Passthrough for those interested.

Joel

 

[Helitronix]

I found this on Amazon, a nicely packaged, weatherproof hall effect sensor, in a form factor that would work. http://www.amazon.com/Polaris-Sportsman-Effect-4040081-4010298/dp/B0100LMD3U/ref=sr_1_19?ie=UTF8&qid=1459958008&sr=8-19&keywords=hall+effect

But before I buy that, I am purchasing this instead, just to make sure it works how I like: http://www.amazon.com/gp/product/B00979RAL2/ref=crt_ewc_title_dp_2?ie=UTF8&psc=1&smid=ATVPDKIKX0DER and this for good measure: http://www.amazon.com/gp/product/B01A09LAV8/ref=crt_ewc_title_dp_1?ie=UTF8&psc=1&smid=A2TOL213PF2CAG

 

[MaxPete]

Yup - sounds like a good plan indeed. Its amazing how cheap some of these little electronic devices can be.

Thanks for the note Helitronix.

Pete

 

[ST Gui]

I found myself howling along at 4000 rpm in 3rd gear when I should likely have been in 4th or 5th. Oh well.

I've been in the same situation a few times. The 4000rpm always makes me want to shift up a gear. Going in to "6th" is a bigger problem for me. My 750 Duc was geared a little tall and I never felt like the RPM were a little too high.

3) China quality still isn't great. The top left vertical segment died.

Hah that affects the most useful bit of the GPI: telling me that I'm in 5th. It still works but if only one gear were properly displayed I'd want it to be 5th. LOL

Maybe you got the only bad one and everyone else will be ok. The $25 price tag is still pretty tempting.

 

[MaxPete]

True - but it was a bit too greasy for that sort of horsing around I'm afraid.

 

[Helitronix]

The hall effect sensor arrived and it works well on the bench. Now I have to give some thought to how to mount it to the ST. I have some ideas and will make something soon. I'm thinking of machining a bracket to hold the magnet. The bracket would use screws to hold it onto the clutch handle. The hall effect sensor itself would get a bracket, that maybe, I can design in such a way that it can be held captive by the acorn nut. However, there might not be enough thread exposed to do that, and I might want more than one nut to hold it in place so that it doesn't move over time. In the worst case, I'll be drilling and tapping two holes on the underside of the clutch handle bracket to hold the sensor. I'll sketch it out and see.

Joel

 

[MaxPete]

Sounds like progress! Looking fwd to seeing how it works.
Pete

 

[Geditor]

Hah that affects the most useful bit of the GPI: telling me that I'm in 5th. It still works but if only one gear were properly displayed I'd want it to be 5th. LOL

Maybe you got the only bad one and everyone else will be ok. The $25 price tag is still pretty tempting.

Oddly that segment started working again, died then came back and has been steady for the past 200 miles. But I am not ready to cut a hole into the dash to mount it quite yet.

 

[thunderpipes]

Sorry,i just do not see why bikes need gear position indicators. Its simple... If the bike sounds like its pulling heavy then you need to tread down a gear or if it sounds like its revving high,then you need to change up a gear. If you intend to stop,i go down through all the gears,when i get to walking pace,to guarantee that i have the only responsive gear for pulling away, i "tap down" twice with the gear lever before i stop. If i was in 2nd the first push puts you in neutral and the second push down will guarantee 1st the only gear required for pulling away safely. This procedure enables you to keep your eyes looking up in the right place,not looking down at something you do not need.

 

[ST Gui]

Sorry,i just do not see why bikes need gear position indicators.

Uh ok.