Olderguy and other Electrical Wizards
#1
Elite Member
Thread Starter
iTrader: (6)
Join Date: Sep 2005
Location: Chicago
Posts: 4,143
Total Cats: -5
Olderguy and other Electrical Wizards
Alright guys lets put our thinking caps on. I bought a new bike, but because of the new sprocket and tires I have put on the bike my speedo is off by 7.76%.
I want to be able to put something inline to correct this error. The speed sensor on the tranny output, sends a frequency to the speedo cluster and that displays the MPH. Based on two data points given in the service manual, 181 Hz = 60mph and 88Hz = 60kph, there is approximately ~ 2.35 Hz/mph.
How do I intercept the frequency and and change it?
I want to be able to put something inline to correct this error. The speed sensor on the tranny output, sends a frequency to the speedo cluster and that displays the MPH. Based on two data points given in the service manual, 181 Hz = 60mph and 88Hz = 60kph, there is approximately ~ 2.35 Hz/mph.
How do I intercept the frequency and and change it?
#2
Did you get the 7.76% from comparing indicated speed to a GPS?
I wonder what the accepted error of the unmolested setup is? On my completely stock Tribute the speedo is out by nearly 5km/h @ 110km/h. What's that nearly 5%? I've heard +-5% is typical. My Miata is even worse.
Who knows, maybe the changes you are making will improve the accuracy Check with a GPS or a stop watch and mile markers.
As to your actual question, I don't have a clue since I'm just a uselss ******* Canadian.
I wonder what the accepted error of the unmolested setup is? On my completely stock Tribute the speedo is out by nearly 5km/h @ 110km/h. What's that nearly 5%? I've heard +-5% is typical. My Miata is even worse.
Who knows, maybe the changes you are making will improve the accuracy Check with a GPS or a stop watch and mile markers.
As to your actual question, I don't have a clue since I'm just a uselss ******* Canadian.
#6
You still have access to whatever equipment your engineering dept. has, right? Or are you gone for good now? If you had access to a decent scope, I'd hook it up to the signal line and look at the shape of the wave. That'll shed some light onto what you'd have to do to change the frequency. In general though, wave signals aren't the easiest to change, unless it's like a pulse-counting system, which might be easier.
#9
Elite Member
Thread Starter
iTrader: (6)
Join Date: Sep 2005
Location: Chicago
Posts: 4,143
Total Cats: -5
You still have access to whatever equipment your engineering dept. has, right? Or are you gone for good now? If you had access to a decent scope, I'd hook it up to the signal line and look at the shape of the wave. That'll shed some light onto what you'd have to do to change the frequency. In general though, wave signals aren't the easiest to change, unless it's like a pulse-counting system, which might be easier.
Im pretty sure its a counting pulses its a square wave. ... from the service manual
#10
Supporting Vendor
iTrader: (33)
Join Date: Jul 2006
Location: atlanta-ish
Posts: 12,659
Total Cats: 134
In your miata, you can look at the tach and see 4k rpm, then know you're doing 80 w/o looking at the speedo right? You can do the same thing on a bike. You'll very quickly learn how fast you're going by feel anyway. I almost never look down at my speedo when riding; I don't have to. And if you're going so fast that you can't tell, you shouldn't be looking at the speedo anyway.
Anyway, if it's PWM based, you can build a comparator circuit inexpensively.
Ben
#14
Boost Pope
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,022
Total Cats: 6,590
One option would be a frequency-to-voltage converter, an op-amp, and then a VCO.
You didn't say whetehr it reads high or low. If it's reading high then maybe you could just build a simple divide-by-n circuit, something as simple as a shift register and a big multi-input AND gate to drop every nth pulse?
You didn't say whetehr it reads high or low. If it's reading high then maybe you could just build a simple divide-by-n circuit, something as simple as a shift register and a big multi-input AND gate to drop every nth pulse?
#18
get a simple PIC microcontroller (16F series for example) and write a simple program to take counts in (on an interrupt) and use a variable to drop as many counts (or as little) as you want. Make sure you scope the signal beforehand and do some signal conditioning circuitry before the PIC input as it is a 0-5volt and current must also be watched.
The whole thing won't cost you more than a few dollars if you already have a PIC programmer and know some basic C (you need a compiler for that) or assembler.
The whole thing won't cost you more than a few dollars if you already have a PIC programmer and know some basic C (you need a compiler for that) or assembler.
#20
Elite Member
Thread Starter
iTrader: (6)
Join Date: Sep 2005
Location: Chicago
Posts: 4,143
Total Cats: -5
get a simple PIC microcontroller (16F series for example) and write a simple program to take counts in (on an interrupt) and use a variable to drop as many counts (or as little) as you want. Make sure you scope the signal beforehand and do some signal conditioning circuitry before the PIC input as it is a 0-5volt and current must also be watched.
The whole thing won't cost you more than a few dollars if you already have a PIC programmer and know some basic C (you need a compiler for that) or assembler.
The whole thing won't cost you more than a few dollars if you already have a PIC programmer and know some basic C (you need a compiler for that) or assembler.