I need to multiply a low voltage signal of 0-100 Hz by
a range of about 4 to 6.
One solution I'm considering is using a frequency-voltage IC
followed by a voltage-frequency IC with the appropriate
resistors and capacitors to get the right calibration. Specifically,
I've identify the Nat. Semi. LM2917N and the Analog Devices AD537
for this application. This combination seems to minimize the number
of circuit components. I was able to get a few samples and
have already built the first half before moving on to the second half.
However, I don't have any much experience in electronics and
I'm wondering if I might have overlooked a simpler approach.

Project Background:
I'm trying to build a replacement speed sensor for my car. The
gear in the transmission has stopped and would require a trans. rebuild
to fix. The mechanic thinks the transmission is otherwise in good shape
and doesn't recommend that solution at this time.
I am building a system based on a magnetic sensor and magnets on
the drive shaft.
The speedometer seems to measure the frequency of an
oscillating signal coming from the speed transducer. From a few
tests, it seems to expect a frequency about 24 times the rotation
frequency of the wheel. I can mount up to 6 magnets around the shaft
but would still need to multiply the signal by 4. It would also be nice to
be able to calibrate the system to display the correct speed.

 

Another solution would be to use a CD4046 phase lock loop with a CD4017 counter. One input to the 4046 phase detector is your sensor, the other input is the output of the 4017 counter which is fed by the VCO of the 4046. The counter is set to divide by 4 or 6. The VCO also feeds
your pickup circuit.

__________________
The great thing about electronics is unlimited ways to do the job. The only limit is one\'s imagination. I generally think my way is best.
Show me a different way. I have an open mind.

 

k7elp60.....

 

Bogdanfirst, perhaps you don't understand phaselocked loops (PLLs). A PLL connected as k7elp60 describes will, in fact, multiply the input signal frequency as described. It won't, however, work down to 0 Hz. Neither will the F-V-F scheme proposed by solmantx.

Solmantx, you need to pick a reasonable minimum rotational speed, below which you really don't care about functionality. Even then you'll have a challenge. Any circuit you use will have lag. The lower the cutoff frequency, the longer the lag.

 

instead of mounting magnets on the shaft, have 24 vanes welded (or mounted on a hose clamp) on the shaft then put one magnet behind the hall effect sensor.

__________________
see my website: www.geocities.com/russlk

 

What wheel are you referring to? What is the likelihood that it rotates at the same rate as the driveshaft? The tires certainly don't rotate at the same speed as the driveshaft.

 

A couple of questions -

Is the "meter" part of the speedometer a voltmeter? On my car it is a voltmeter that is driven by the electronic speed sensor systems. If is is in fact a meter then you might develop a more direct D/A conversion and drive the meter directly.

Resolution at low speeds could be/will be a problem unless you get more pulses per rev. If you have an ABS system you might steal the pulses from that system then divide down to get what you want. The ABS sensors on my car must have 100 teeth which I would presume generate 100 pulses per turn of the wheel. While you'd only see the rotational speed of one wheel it's at least something. Driveshaft RPM doesn't tell the whole story either in some situations.

__________________
stevez

 

Steve wrote:It tells no more or less than the speedometer output from the transmission, or, for that matter, any other sensor that you can put on the drive train. Steve, I'm not disputing what you said - just trying to offer some clarification.

 

Sorry to just be jumping in right here but I have a similar problem. I am doing a conversion of interiors on a vehicle and need to simply double my incomming frequency to supply the speedometer circuit.

I have a 4,000 Pulse per Mile AC SineWave Signal that I'll be needing to change to a 8,000 Pulse per mile Signal. The out going signal can be either Sine or Square. It needs to be able to handle up to 4Volts incoming and supply close to 4Volts out. The voltages are unimportant as long as they don't get too high.

Input frequencies to handle will be As close to 10hz incomming as possible if not lower and up to 240hz.

Output Freqs. will obviously need to be 20hz - 480hz.

This will give the speedometer an operating range of 10MPH - 216MPH.

Any information will be great. Thanks.

 

Do you know how the existing speedometer measures the pulses? Are they counted or is there some form of frequency-to-voltage convertor?

 

Earlier on in this thread, it was suggested that a 4046 PLL with a divider in the feedback should be used - it's a classic frequency multiplier.

This was 'rejected' as it won't work down to 0MPH, well neither do mechanical speedo's.

The 4046 solution could be made to work quite easily, as I remember, the 4046 has an 'out of lock' output - simply use this to mute the outgoing oscillator signal. As the car stops the input pulses stop, the 4046 loses lock and the output is muted - the 4046, as with all PLL's, will only lock over a certain range, but it doesn't need to lock at very slow speeds (0-5MPH), mechanical speedo's don't read that low.

By using a comparator on the VCO control input, you could actually preset a point where the output gets muted.

 

WellI was pointed by a friend to try the 4011 since it's supposed to DOUBLE the frequency. Well that idea has failed me.

 

You can configure a 4011 to double a frequency, but not over a wide range, and you need to feed it a square wave.

 

Does the car in question have ABS or traction control? If the car has either of these, it will have at least two wheel speed sensors that are independent of the main speedometer. I've had some experience with a GM model, and it puts out a pretty clean sine wave at around 1.7V peak- to peak. The one I've worked with has 47 little teeth cut into a wheel that is right next to the brakes. Figuring the wheel circumference and the Frequency based on 47 pulses per rotation, I've been able to get the approximate vehicle speed.

 

This is ridiculous! I am a transmisson repairman, and if the governor is faulty, replace it! If it is the governor driving gear that is at fault, than replace that one. They are not expensive. The transmission dont have to be rebuilt for this repair.

//Staigen