Hi all.

I bought a bike with messed up wiring, and carbs etc. No farings, modified rear frame...basically the po wanted to have a 'street fighter'.

Well, I wanted a cheap winter bike and don't mind a bit of work...I bought it for $250.

I rewired the digital dash and it now works about 95%...the only issues are speed display and coolant temp display. The coolant temp sensor is the wrong one...I will change it and go from there.

My big question is whether anyone has successfully gotten the stock speed sensor to work with the Acewell?

Currently I have it connecter so that the Pink/Green spped sensor ouput wire feeds the speed input wire on the digital meter. It reads 300kmh (Max) unless yo are pretty much stopped.

My next step is to 'scope the signal, just wondered if anyone has this working?

I rebuilt the carbs and now the thing runs great btw...loads of fun for $250...different than my summer bike ('74 Motoguzzi Eldorado)...but I like it.

Thanks,

Chay

 

Might need to play with wheel-circumference setting. This unit wants ONE pulse per wheel-rotation since it uses magnet on wheel flying past sensor on fork-blades. Try setting wheel-circumference to 1/4th of what's programmed now. Sure 100% way to get speedo working is get magnet and sensor that comes with unit and set it up way they want it.

 

Thanks-I came to the same conclusion yesterday and set the circumference to 100 (Down from 2000). It did make a difference, but the speedo does not track vehicle speed-it sort of tracks rpm which would indicate it's running on noise.

I tried to get info on the pulses per mile, or pulses per rotation, on the stock speed sensor but came up empty. This would help in determining the factor to enter-IE: if there were 5 pulses per rotation, divide the circumference by 5.

This does have implications for accuracy, but anything would be better than what I have now!

I suspect I have a speed sensor issue...

Time to troubleshoot!

Chay

 

I did some more work on the speedo.

I connected a dc voltmeter at the three pin connector on the right hand side of the bike. It's a pretty easy place to troubleshoot. I do have +12, common and the switching 0v, 5V on the pink. I looked at the input connector for the Acewell next, the signal is getting there fine.

What surprised me is the small amount the rear tire has to move to switch the waveform. In other words, the speed sensor is operating on something extremely fine; perhaps splines on a shaft etc.

The Acewell seems to be set up for more like a one pulse per rev, not the many pulses per rev that the stock sensor will give out.

The way the Acewell is acting, with a circumference of 100 entered (Stock = 2000), is that it will count up as the speed increases until the speedo hits 70ish kph (In reality probably only about 30kph). Beyond this it will stay at around 70kph. I suspect the frequency of the signal is too high for the Acewell to resolve at this point, so it just sticks with 70kph. I'm going to guess that there are 50 pulses per rev (May even be higher than this--super fine).


Example: 1 pulse per rev sensor at 100kph, assuming 2000mm circumference of wheel:

100kph = 100 000m/h = 27m/s = 13.88 revolutions of back tire/sec = 13.88Hz

Example: 50 pulses per rev = 694Hz



So I'll probably count the pulses per revolution and figure out a divide circuit (I'm a bit of an electronics guy). Either a TTL divider (Seeing as it's a 5 volt pulse it's a natural) or a PIC device divider-whatever end up being easier.


Chay

 

I did some more work on this....

I built a microprocessor (PIC 16F74) frequency divider that divides the original signal by 48. So for 48 pulses in I get one pulse out.

I put it on the bike and.....drumroll.....nothing. The speedo sits at 0. WTF!

So I checked the incoming 5V signal....and it was staying at 0V.

It turns out that the speed signal generator is a low active device. In other words, when "on" it connect to ground and when "off" it is open circuit. Therefore I needed a pullup resistor in my device build. So, I connected the 5V microprocessor power supply through a 1kohm resistor and sure enough, I get 4.8V on, and about .3V off. This drove the input to the microprocessor just fine. This also means that if you happened to be using a 3.3V microprocessor, just drive it with the 3.3V power supply and you will get a 3.3 volt square wave.

With the divide by 48 in place, the wheel circumference I needed to set for accuracy is about 1250mm. Because this is lower than the actual of about 2000mm, I would guess I'm getting about 1.6 pulses per rotation of the tire.

It works up to about 150k very well, GPS verified. I decided not to go beyond that for safety reasons/road condition etc.


Chay

 

Hey amazing job figuring that out!

I was going to suggest a mechanical method. Such as taking speed-sensor apart to see how it's triggered. If there's some multi-teeth wheel, perhaps remove quite a few of them to get close to 1 trigger per wheel revolution.

 

I toyed with making my own speed sensor with an old abs sensor I have lying around...Mount the sensor and glue a small piece of steel to the wheel which passes in close proximity. It might have actually been easier than what I made lol. But it was a chance to get my geek on.

I'm 99% sure that the original sensor is picking up splines on a shaft, or teeth on a gear-That's the only reason I can think of why the pulses per rotation is so high. Usualy it's something like that. Not ideal candidates for modification.

Chay