Help with LM2917 F to V conversion!

Thread Starter

funkychicken9000

Joined Jun 21, 2007
2
Hi guys, I'm new here!

Basically, I'm wanting to convert a frequency into a voltage. This is for a tachometer for a racing car.

The car's ECU outputs a logic level pulse every engine revolution. I was going to use a LM2917 to convert this into a voltage signal varying between 0 and 4v. Since we're using a superbike engine, the frequency of the signal from the ECU will vary anywhere from 0 Hz up to around 15kHz.

I thought I could do this relatively simply with an LM2917, but I'm really struggling on component values etc. I also was told by someone that my 8-pin LM2917 may not work, since it requires positive and negative-going AC, whereas I'm using pulsed DC.

Can anyone do me a massive favour and suggest a circuit diagram for this unit? Time is short and my efforts so far have proved relatively futile!

Thanks a lot!

[edit] I should probably mention, this is going to need to run off an 8V supply...
 

Papabravo

Joined Feb 24, 2006
21,159
The very first thing that you should do is download a copy of the datasheet. Then you can verify the stories that your little friends tell you. Otherwise you're at sea!
 

Thread Starter

funkychicken9000

Joined Jun 21, 2007
2
I've been reading the datasheets, and it seems that the 8-pin LM2917 does indeed require a signal varying around 0v. Since I have no way of getting the 14-pin version in time (which has a user-settable reference rather than ground), is there anyway I can create a signal varying around 0v from my digital pulse? Someone on another forum said "You simply feed it through a capacitor, and a resistor to chassis (0V), it will then be centered around 0V". Am I right in thinking that this means putting the input through a capacitor and resistor in series to ground, and then connecting the LM2917 input to the middle of the two components? Or should they be in parallel?

Also, I'm not entirely sure what values I should use for these two components. I thought maybe the RC time constant should be small so at high frequencies the signal will still fully charge and discharge the capacitance - does this sound right?

[edit] thinking about it a bit more, I'm guessing the R and C should be in series. This would make a high-pass filter, which presumably would differentiate the signal. With an RC time constant a lot smaller than the minimum period (say 1/20th?) then with a 15,000rpm engine speed giving a 900kHz digital pulse, I'd be looking at R and C values of 1.8k and 100pF respectively. Sound about right?
 
Top