Maker Pro
Maker Pro

Frequency to voltage conversion

S

Simon Morden

Jan 1, 1970
0
I'm trying to remember my limited 20-year old electronics knowledge and
failing miserably...

Some background: the school I work at is involved in a science
competition run by Rolls Royce. Our project is to make a wind tunnel and
test turbine designs. I am in charge of the instrumentation.

The problem: I need to be able to record the rpm of the test pieces.

What I've done so far: I have availed myself of the National
Semiconductors' LM2917 (14-pin), as detailed here:
http://www.national.com/mpf/LM/LM2917.html
It's a dedicated frequency to voltage IC, and I'm running it off a 9v
battery. I've set it up so that it works - I can input an ac voltage and
get a proportional dc output.

What I can't do: I can't make my rpm detector talk to the chip. The
detector set up is a bike computer sensor and a 3v battery. Output is
pulsed dc which does not fall below ground - and the chip requires that
pin 1 goes at least -30mV on each cycle.

What I can't do (part2): I know I have to use something like capacitor
coupling to strip out the dc and keep the ac. I know I have to keep RC<T
because I'm trying to just get the voltage spikes +/-. What I really
can't do is make what on the face of it should be a really simple
circuit to convert the pulsed dc train (at 10-100 Hz) into a pseudo ac
voltage with a large enough variation to fool the chip.

Any help gratefully received!

Simon
 
J

Jon Slaughter

Jan 1, 1970
0
Simon Morden said:
I'm trying to remember my limited 20-year old electronics knowledge and
failing miserably...

Some background: the school I work at is involved in a science competition
run by Rolls Royce. Our project is to make a wind tunnel and test turbine
designs. I am in charge of the instrumentation.

The problem: I need to be able to record the rpm of the test pieces.

What I've done so far: I have availed myself of the National
Semiconductors' LM2917 (14-pin), as detailed here:
http://www.national.com/mpf/LM/LM2917.html
It's a dedicated frequency to voltage IC, and I'm running it off a 9v
battery. I've set it up so that it works - I can input an ac voltage and
get a proportional dc output.

What I can't do: I can't make my rpm detector talk to the chip. The
detector set up is a bike computer sensor and a 3v battery. Output is
pulsed dc which does not fall below ground - and the chip requires that
pin 1 goes at least -30mV on each cycle.

What I can't do (part2): I know I have to use something like capacitor
coupling to strip out the dc and keep the ac. I know I have to keep RC<T
because I'm trying to just get the voltage spikes +/-. What I really can't
do is make what on the face of it should be a really simple circuit to
convert the pulsed dc train (at 10-100 Hz) into a pseudo ac voltage with a
large enough variation to fool the chip.

Read the 2nd paragraph on page 7. You need to set up the chip to use dc.

It has a comparator that can be setup to be ground(or actually a bit above
ground) and it says something about hysteresis in their too.

Optionally you could shift the voltage of the pulsed dc down by using a
negative supply or possibily shifting the ic chip up. e.g., put a diode on
the ground pin or use an active resistor(if the ic doesn't pull enough
current for the diode to turn "on"). The output signals will all be shifted
up of course but you can unshift them. At least this should work in theory
;)
 

neon

Oct 21, 2006
1,325
Joined
Oct 21, 2006
Messages
1,325
provide an input bias to pin1 then apply the signal trough a cap it should work.
 
S

Simon Morden

Jan 1, 1970
0
Jon said:
Read the 2nd paragraph on page 7. You need to set up the chip to use dc.

It has a comparator that can be setup to be ground(or actually a bit above
ground) and it says something about hysteresis in their too.

You star. My only excuse (and I'd read that document *and* the AN-162
applications pdf) is that I'm a geologist by trade, and only picked up
the physics I needed...

I now have the sensor and the chip talking to each other, with only one
more problem to solve - it only works when I have a voltmeter between
pin 1 and ground (I was measuring V(in) so I could balance the V(bias)).
I have a feeling a M-ohm resistor will do the trick, but not sure why.

But I've stopped the night - I can't see the numbers on my capacitors
anymore.

Simon
 
J

Jon Slaughter

Jan 1, 1970
0
Simon Morden said:
You star. My only excuse (and I'd read that document *and* the AN-162
applications pdf) is that I'm a geologist by trade, and only picked up the
physics I needed...

I now have the sensor and the chip talking to each other, with only one
more problem to solve - it only works when I have a voltmeter between pin
1 and ground (I was measuring V(in) so I could balance the V(bias)). I
have a feeling a M-ohm resistor will do the trick, but not sure why.

But I've stopped the night - I can't see the numbers on my capacitors
anymore.

Well, try it and see ;) If it works then it works ;) I'm not sure what is
going on though. Maybe adding just enough capacitance to the circuit or
something else? It could be acting as a pulldown to ground which is needed
if the input ends up floating for some reason.
 
J

Jamie

Jan 1, 1970
0
Simon said:
You star. My only excuse (and I'd read that document *and* the AN-162
applications pdf) is that I'm a geologist by trade, and only picked up
the physics I needed...

I now have the sensor and the chip talking to each other, with only one
more problem to solve - it only works when I have a voltmeter between
pin 1 and ground (I was measuring V(in) so I could balance the V(bias)).
I have a feeling a M-ohm resistor will do the trick, but not sure why.

But I've stopped the night - I can't see the numbers on my capacitors
anymore.

Simon
A pull up resistor is needed most likely.


http://webpages.charter.net/jamie_5"
 
S

Simon Morden

Jan 1, 1970
0
Jamie said:
A pull up resistor is needed most likely.
I will experiment. My sensor (and thus V(in) is working at +3v, my
f-to-v IC at +9v. Would a pull-down on V(in) be better? Or would I be
better off trying to run the sensor at +9v too?

Simon
 
S

Simon Morden

Jan 1, 1970
0
Simon said:
Jamie wrote:


I will experiment. My sensor (and thus V(in) is working at +3v, my
f-to-v IC at +9v. Would a pull-down on V(in) be better? Or would I be
better off trying to run the sensor at +9v too?

Simon

Just to reply to myself. Everything is now working perfectly (I used a
2k pulldown - it's probably too large, but I need to go to the shops
again anyway).

The sensor is still on the +3v, with a bias (controlled by a variable
pot) of half that. I actually feel like I almost understand what's going
on in the circuit, and can make adjustments accordingly.

Thanks again for everyone's help.

Simon
 
Top