Digital Speedometer sensors information!

[engr.humair]

Dear:

I am working on Project to design PIC based Digital Speedometer.

Kindly advise me sensor that will perfect to interface with PIC16F876A in order to forward me accurate vehicle speed on LCD.

Waiting for your response.

Thanks in advance!

[solutions]

I assume this is a one-off?  If so, go to a salvage yard and pull a speed sensor off the back of an automotive transmission.  Almost all automotive sensors are 5V    

Barring that, any PM electric motor can be used as a DC generator, with voltage proportionally increasing with speed, and polarity indicating direction.  You can find those here http://www.allelectronics.com/make-a-store/category/400/Motors/1.html

Just watch out for brush noise from the motor. 

[oldvan]

What type of vehicle?  Expected maximum speed?  Desired resolution?

If a modern vehicle, can we grab data from the OBD bus and display it?

[engr.humair]

Thanks for reply.

The module should be able to interface with any vehicle having speed limit of 250kph.

Any PIC based Speedometer schematic, tutorials or speed sensor information that i can interface with PIC?

Thanks

[solutions]

If you say "any", your big problem is mechanical, not electrical....

[king]

Dear Mr.Humair

I have made speedometer and i have tested it at the speed of 5000 RPM.I have simply used Proximity sensor (Metal Detector, NPN , 4mm range, 12V-24V) which gievs me the ON and OFF pulses and by using Pic16f876A I have measure the speed in RPM as well as Length in meters and display it on 4x20 LCD.So as far as vehicle concerned then I think you have a wide area of fixing proximity sensor just fix the sensor at a place where you can easily get pulses.
(Note:If you also want to measure the distance covered by vehicle then before doing coding one thing you should have to keep in mind that how many pulses are equals to '1' meter)

Regards

King

[oldvan]

Use a GPS receiver to determine speed, convert to desired units and send to LCD with PIC.

Q.E.D.

[FriskyFerret]

Here we go again... A noob asks a vague question without supplying details then shoots down every suggestion offered, one at a time.

Lets not drag this out:

1. Use a GPS to obtain speed. Oh forgot, you already shot down the best idea already. Never mind this one.

2. Aim a video camera at the road surface and write the algorithms to determine speed from the changing surface detail. (Has been done.)

3. Measure the air speed of the vehicle with a cup anemometer, hot-wire sensor, or ultrasonic doppler sensor. Should be within 10% of land speed usually.

4. Pick the speed data off the vehicle's On-Board Diagnostic (OBD) network port.

5. Put another variable reluctance sensor reading the same metal rotor teeth that the ABS system uses to compute speed and wheel slip.

6. The old, highly accurate way that the auto builders use: put a "follower wheel" on a pivot arm behind the car. See below.

7. Paint one or more equally spaced white strips radially on the inside-facing side of the tire and use a cheap optical reflectance sensor to compute the speed.

8. Use an accelerometer with DC response to determine speed by integrating the +/- acceleration.

9. Glue a homemade absolute-reading optical rotary encoder disk on the center hub of the speedometer needle and read the speed based on angle of rotation of the needle with a non-contact optical sensor. Duh?

10. Unbalance one or more wheels slightly then record the small vibrations with an AC-coupled accelerometer. Look at the data in the frequency domain. Find the frequency where the power from the wheel shake spikes and multiply by the effective rolling circumference of the tire. That will be the vehicle speed.


Pictured below is how the pros used to do it (now automotive engineers use specialized GPS units at the test track.) A small bicycle wheel strapped on to the rear bumper. It's low cost, low technology, fits any vehicle, and is very accurate. Now tell me how none of these fit your needs.

Do not asked me to supply plans, schematics, etc. That's your job, Mr. Engineer! I haven't had a drink all day and you've really pissed me off with your dumb-ass question.

[solutions]

Is it too late to change your handle?  I'm thinking "CrustyFerret" is more appropo
 

Nice analysis and totally agree and feel the same way - everyone has the ideas, nobody wants to do the work or research, and feels smug in the shoot downs....or they never give us enough information on requirements and constraints to tell us he's building a time machine and needs accuracy at 88MPH only...

Reminds me that you forgot to mention - "measure the voltage across one arm of the flux capacitor", in your comprehensive list.  Everybody should already know they are configured as a bridge topology.

Meanwhile, "your time is worthless compared to mine...."

[rentau]

to engr.humair, 1st you should use GOOGLE, then start with something, then ask for help with what you have and not what you need.

[sputnik]

The most straightforward way would be to use a hall effect sensor/magnet ,or integrated "gear tooth" sensor type hall sensor that has a magnet in it already, to pick up the speed of any of the rotating parts in the drive train, i.e. drive shaft, wheel etc, then do the math.

There are many projects on the web that describe this very thing, I have a design based on a GLCD that is on several other forums, you will have to look for it to find it, im not going to tell you where.

[FriskyFerret]

Reminds me that you forgot to mention - "measure the voltage across one arm of the flux capacitor", in your comprehensive list.  Everybody should already know they are configured as a bridge topology.

I needed a little humor today...don't forget to insert the phrase "positronic matrix" somewhere.


You kids and your obsessions with Hall effect sensors. A variable reluctance sensor has no semiconductor material and is one of most robust sensors out there. It's been used successfully for decades for sensing gear rotation in the automobile industry. It's one of those sensors that so simple, robust, and reliable that it just can't be improve upon.

See: http://www.instructables.com/id/Magnetic-speed-sensor/


Here's a shot of an axle gear tooth ring with speed sensor. Don't mess with the sensor or the ABS self-diagnostics are gonna have a fit. Install a second one located away from the OEM unit and you'll be good to go. For short term project use or for proof-of-concept purposes attach the sensor with plumber's epoxy putty. It works miracles for prototyping.

Signal conditioning a VR sensor down to logic level is not difficult. National Semiconductor used to make an IC just for the job. In the Application section of the data sheet they gave all the info and circuitry required. I think they also published an App Note on signal conditioning a VR sensor. Can't remember for sure anymore. It was back in 2002 when I did an interface from a VR sensor to an ECU.





The Basics

[rentau]

GOOGLE=QEI decoder
http://ww1.microchip.com/downloads/en/DeviceDoc/70208A.pdf


you be able to determine if you going backward or flying, i mean forward

[Ichan]

Speed or Rpm?

If the engine is gasoline, then you can get the engine rotation (Rpm) by sensing the the ignition spark from the coil cable, just twist a wire several time on the coil cable to make inductive pickup sensor. If the engine is diesel then it is harder, you can try to get the rpm signal from the injection pump or the alternator.

Speed, it is a relative measurement - in this case speed relative to the earth.  Do some calculation, how far the car travel relative to the earth per engine rotation.

Anyway gps is my choice as it is getting cheaper and cheaper.

-Ichan

[oldvan]

I have to use sensor apart from GPS

We all love a receding target.  Keeps us on our toes.

I suppose a GLONASS receiver is out of the question too.

Similar to the "Gong Clock" seen elsewhere on here:  Drive past a police officer
at far over the speed limit.  He will soon tell you how fast you were going. 

Try RADAR or LASER RADAR pointed at stationary object(s).

Drive on rumble strips and measure frequency of the rumble, will be proportional
to vehicle speed.

If you drive the vehicle off a cliff, acceleration will be 1 G until wind
resistance begins to damp acceleration.  Calculating vehicle speed will be easy
based on   S = V0 T + 0.5 A T^2    Not long after speed will quickly become
0. Electronics should be extremely rugged and fireproof.

Eventually this will likely come down to you using one of the technologies
already suggested or the ultimate simplest approach, power down the LCD and
leave it so, use a marker to write "0.000" on the display and know that it is
always perfect when the vehicle is parked.

[solutions]

Speed, it is a relative measurement - in this case speed relative to the earth.  Do some calculation, how far the car travel relative to the earth per engine rotation.

Um - no.  On a stick, you'll be going 10MPH when you push the clutch in while rolling at Interstate speeds. On an automatic, you'll be way optimistic on speed until the torque converter locks up.  If you see a cop, just push the clutch in - your observation of your own speed might hold up in court.

CrotchetyFart:  I HATE variable reluctance wheel speed sensors, not as an engineer for their inelegance, but as the family auto repair technician.  I've changed the axle hubs out twice on my truck (they don't sell the sensor separately) and they still don't read wheelspeed on the ABS (yes, I also changed out the ABS controller...).  The fact they were Chinese made might have something to do with it.  Lived without ABS a decade ago, and doing fine without it now.

Note to our intrepid speed sensor users:  From a system design perspective, and borrowing from the all-knowing auto industry's design practices, make sure you put a "check speed system" light on your system if you use variable reluctance sensors and write the firmware so it comes on randomly after about 30,000 miles.

[Ichan]

Um - no.  On a stick, you'll be going 10MPH when you push the clutch in while rolling at Interstate speeds. On an automatic, you'll be way optimistic on speed until the torque converter locks up.  If you see a cop, just push the clutch in - your observation of your own speed might hold up in court.

Yup, you are right but i am not wrong too.

I was grin when i type "Do some calculation", it did not meant to be a simple calculation - slips and open clutch need to be considered too.

-ichan

[FriskyFerret]

CrotchetyFart

Damn, I like that!

Lived without ABS a decade ago, and doing fine without it now.

Oh, you're one of those guys. I suppose you're also one of those that bemoans the passing of the carburetor. Tell you what, ABS has saved my butt more than once in the last 10 years. But I suspect we're in agreement about how evil full drive-by-wire systems are.

From a system design perspective, and borrowing from the all-knowing auto industry's design practices, make sure you put a "check speed system" light on your system if you use variable reluctance sensors and write the firmware so it comes on randomly after about 30,000 miles.

You're full of it. A VR sensor has a permanent magnet and a coil of wire. Two parts. What exactly failed in all those sensors you changed out over the years? Both the magnet and the coil are usually potted in industrial, high-temperature epoxy.

[solutions]

You're full of it.

Yes I was.... by intent - and anyone frustrated by a "check engine" light will know that. A smiley would have been too obvious.

The reluctor gap due to bearing wear was the original problem on one wheel.  Replacement TWICE of the hub and sensor assembly (same vendor who charged half what factory hubs cost) didn't fix the ABS - like you, I realize it doesn't get much simpler, which is why I am still scratching my head.  Plugging in one of the old hubs (only one side had fragged bearings, but I "wisely" decided to change both) removes the fault on that wheel without the wheel rotating at all (smells like wrong impedance....).  But, I needed functional hubs on the mountain for the 4x4 this winter a lot more than ABS.  In 4WD, the ABS is disabled in the system, anyway...

I also think drive by wire is a good thing as long as the lawyers don't tell the engineers how to do the system design (my drive by wire SUV doesn't even list it ANYWHERE as a feature, and got slammed in the auto magazines for "not having it these days" - I found it out when it refused to do doughnuts in the snow, backing off the throttle and applying diagonal braking WHEN I DIDN'T WANT IT...no way to disable it - at least without a pair of dykes).  BRAKE means STOP.  Ignition key means OFF.  No exceptions.  And traction controls should have a "performance mode", but I guess I can understand not having it, since Toyota drivers have demonstrated that drivers don't have the sense (just the mere act of buying the car shows they have little sense in their herd mentality) to turn the key off, the media has not seen fit to educate the public with "turn the key off, stupid", with one dead woman's car apparently having had its handbrake applied.

All my vehicles, including the motorcycle are thankfully fuel injected, though I am glad the bike is not drive by wire like its newer model is (that's my only luddite opinion).  For the record, I do hate carbs, both on my intake and as my intake.  

[Ichan]

I dig my project archive, here is attached my Universal Tachometer.

In the proteus design the photocoupler used for simulating photosensor input triggered by ac signal generator, no amplifier for this as the photosensor is feed to the comparator input of the ATMega8. On the real hardware the alternator input is protected by resistor - zener combination.

Remember this is a Tachometer not a Speedometer!

Hope this give some help to the OP.

-ichan

[Roka]

I would use an accelerometer and an MCU. Need only one sensor, you can put it into your pocket.
If the acceleration (and time) is known the velocity could be calculated.
The sensor:
http://www.analog.com/en/mems/low-g-accelerometers/adxl345/products/product.html
The theory:
http://www.freescale.com/files/sensors/doc/app_note/AN3397.pdf
I used this older analog output sensor:
http://www.analog.com/en/sensors/inertial-sensors/adxl330/products/product.html
Very sensitive, fantastic device. It has "null" position you can measure any little tilt of any coordinte (x, y, z).
As my idiot boss always said at the end of the weekly meeting "After all, That's only question of software"
Roka

[FriskyFerret]

I already came up with that, nimrod. Read the thread first.

[solutions]

Is a "nimrod" in metric or English units and how many are in a thread?  Sound like it might easily convert to knots....

Speaking of nimrods in a thread, casting a chip log (not a euphemism for going to the toilet) is one speed measurement technique you did miss, FriskyFarad.  See "Origin" here: http://en.wikipedia.org/wiki/Knot_%28unit%29