Discussion in 'Electronic Basics' started by electro, Apr 6, 2006.

1. ### electroGuest

im trying to explore the operation of the chip in zener regulated F>V
circuit with 250mVpp @ 1khz,12V Vcc as prescirbed in the datasheet.my
threshold levels are rated +-50mVpp so im quite sure that proper output
Voh(8.3) and Vol(2.3) shoudl be produced at pin 2 or 3...
my R1 is 100k C1 is 20nF and C2 is 1uF and a 10k transistor
output(pin5)...

my problem is:

1.my reading at pin 2 is 580 mVdc max
2.my reading at pin 3 is 520mVdc max
-> shouldnt it be equal to ic x R1 = 185uA x 100k = 18.5V where my ic
or I2=I3 is rated to be about 185uA @ 25degree celsius as based from
the datasheet graph
3.my reading at pin 5 is 420mVdc max
-> the output voltage which says Vout = 66Hz/volt is nowhere to be
attained. changing the input frequency from my function gen. doesnt
alter the Vout in anyway...why?
4.my application is in metal detedector. i have a lc tuned oscillator
and i want to use the frequency change as my indicator of metal
presence as opposed to "mixing"...or dou you have any other way to do
this? meaning a simple switch in which when a certain threshold
frequency is attained, it outputs a certain voltge level?

thanks!

datasheet is here:http://www.national.com/pf/LM/LM2917.html

electro, Apr 6, 2006

2. ### electroGuest

sorry forgot to connect pin 12 to ground..however my values for the
said pins are now equal

pin 5,4,3 an2 = 80mV

electro, Apr 6, 2006

3. ### Noway2Guest

electro wrote:
> sorry forgot to connect pin 12 to ground..however my values for the
> said pins are now equal
>
> pin 5,4,3 an2 = 80mV
>

Are you sure that the chip is functional? Did you possibly damage
something by accident when you didn't have the ground connected?

When faced with unexpected results, I try to back up and see if I can
get things working in a known environment. What I am trying to say
here, is perhaps you are biting off too much unknown in one pass. Try
implementing the F-V converter on the bench, using one of the sample
circuits in the data sheet and see if you can get "good" results that
way first.

The LM2917N has been around a long time and should be a reliable
device. I was planning on using one in the project I am working on,
but the requirements were changed on me and I no longer need the
functionality of it, otherwise I would offer you better specifics on
the circuit implementation.

Noway2, Apr 6, 2006
4. ### ChrisGuest

electro wrote:
> im trying to explore the operation of the chip in zener regulated F>V
> circuit with 250mVpp @ 1khz,12V Vcc as prescirbed in the datasheet.my
> threshold levels are rated +-50mVpp so im quite sure that proper output
> Voh(8.3) and Vol(2.3) shoudl be produced at pin 2 or 3...
> my R1 is 100k C1 is 20nF and C2 is 1uF and a 10k transistor
> output(pin5)...
>
> my problem is:
>
> 1.my reading at pin 2 is 580 mVdc max
> 2.my reading at pin 3 is 520mVdc max
> -> shouldnt it be equal to ic x R1 = 185uA x 100k = 18.5V where my ic
> or I2=I3 is rated to be about 185uA @ 25degree celsius as based from
> the datasheet graph
> 3.my reading at pin 5 is 420mVdc max
> -> the output voltage which says Vout = 66Hz/volt is nowhere to be
> attained. changing the input frequency from my function gen. doesnt
> alter the Vout in anyway...why?
> 4.my application is in metal detedector. i have a lc tuned oscillator
> and i want to use the frequency change as my indicator of metal
> presence as opposed to "mixing"...or dou you have any other way to do
> this? meaning a simple switch in which when a certain threshold
> frequency is attained, it outputs a certain voltge level?
>
> thanks!

> sorry forgot to connect pin 12 to ground..however my values for the
> said pins are now equal
>
> pin 5,4,3 an2 = 80mV
>

Hi, Electro. It looks like you're doing the 14-pin version of the app
circuit, "Zener Regulated Frequency To Voltage Converter" on the top of
p.9 of the data sheet.

It _does_ work. That happens to be the first one I tried the first
time I got the IC.

You can't expect an output voltage greater than the supply voltage in
(Vz - 2V * (1V/66Hz) ). Anything over that will show about 5.5V.

Here's what to do. First, turn the power off. Then, copy the circuit
shown on p.9 on a sheet of paper, and transpose the pin numbers for the
14 pin IC. Make a Xerox of your artwork. Now get a cup of coffee and
think about something else for 10 minutes. Once your mind is fresh,
re-trace your wiring point-to-point, using a colored marker to track
out each node of the circuit on a piece of paper. Once every line is
colored in, you know the circuit is wired correctly. It's important to
draw out the entire circuit before you wire it up.

You have already indicated that you had at least one wiring problem.
Sometimes running an IC without the GND pin (12 on this IC) connected
is fatal, sometimes not. You'll definitely have some interesting
current paths.

I don't know if you remembered that the 8-pin version has the charge
pump connected to the input of the output op amp, where you have to
make that connection externally with the 14-pin version. If you
haven't made this connection (pin 3 to pin 4 on the 14-pin IC), that
might be the source of your problem.

If you triple-check the wiring, and it still doesn't work, there's a
possibility you smoked the IC. Try a new one and see if it works.

You haven't given enough information to determine whether this is the
way to go here. Metal detectors typically operate at RF (radio
frequency). This is too high a frequency for the LM2917 to detect.
Are you mixing your sensing signal with a known oscillator frequency?
I know it's sometimes easier to read a meter than listening, but your
ears are very sensitive to small changes in tone, actually much better
than a voltmeter. Not only that, but your ear can normalize for slow
drift in the reference frequency, where you'll have to be constantly
nulling the voltmeter. There's a reason why most of these circuits use
an output tone.

Do you have a link, or can you provide a schematic for your proposed
circuit?

Good luck
Chris

Chris, Apr 6, 2006
5. ### electroGuest

thank you for your kind thoughts...
yes im referring to the zener F>V converter circuit

initially, i made a circuit that operates in mixing mode.i have a fixed
oscillator of about 480khz and then another one
with almost the same frequency and is made variable by the loop
inductance..my problem however was that i cant get
my signal mix or extract the difference signal. i tried single ended
bjt mixer as well as diode but cant get a good output..
also, i would like to indicate metal presence not by an audio tone but
by a pic mcu.my initial setup was to rectify and extract
the input frequency and input it to my pic mcu which is then compared
to a reference voltage which will be pre-calibrated.
but having found of chips that can do frequecny to voltage conversion,
i opted for this chip now...
my application is very simple.i will just calibrate my pic reference
voltage and then from there any change in frequency will correspond to
a certain voltage...
these are my problems:
1. the maximum input frequency based on datasheet is fmax =
I2/(C1xVcc)...from datasheet, for a supply voltage of 12V,my I2=I3 is
about 185uA..also, C1 should be at least 500pf due to timing
considerations...so that will give me about 30.8khz maximum..my plan is
to further increase my loop inductance to b e in the range of
20-30khz.now my question is how sensitive will this be with the
presence of metal?
2.what waveform will i expect from pin 2 and 3?
I2 is mirrored to I3 right? isnt it i should be seing a waveform in
which its voltage changes from about 8.3 then discharges to about 2.3
which indicated the Voh and Vol for the pin3?
3.my Vout should be varying when i vary my input frequency right?
it seems im not gettng the sensible outputs.....
_i connected pin3 and 4

lastly,

any idea on a simpler way to realize my application?

electro, Apr 6, 2006
6. ### ChrisGuest

electro wrote:
> thank you for your kind thoughts...
> yes im referring to the zener F>V converter circuit
>
> initially, i made a circuit that operates in mixing mode.i have a fixed
> oscillator of about 480khz and then another one
> with almost the same frequency and is made variable by the loop
> inductance..my problem however was that i cant get
> my signal mix or extract the difference signal. i tried single ended
> bjt mixer as well as diode but cant get a good output..
> also, i would like to indicate metal presence not by an audio tone but
> by a pic mcu.my initial setup was to rectify and extract
> the input frequency and input it to my pic mcu which is then compared
> to a reference voltage which will be pre-calibrated.
> but having found of chips that can do frequecny to voltage conversion,
> i opted for this chip now...
> my application is very simple.i will just calibrate my pic reference
> voltage and then from there any change in frequency will correspond to
> a certain voltage...
> these are my problems:
> 1. the maximum input frequency based on datasheet is fmax =
> I2/(C1xVcc)...from datasheet, for a supply voltage of 12V,my I2=I3 is
> about 185uA..also, C1 should be at least 500pf due to timing
> considerations...so that will give me about 30.8khz maximum..my plan is
> to further increase my loop inductance to b e in the range of
> 20-30khz.now my question is how sensitive will this be with the
> presence of metal?
> 2.what waveform will i expect from pin 2 and 3?
> I2 is mirrored to I3 right? isnt it i should be seing a waveform in
> which its voltage changes from about 8.3 then discharges to about 2.3
> which indicated the Voh and Vol for the pin3?
> 3.my Vout should be varying when i vary my input frequency right?
> it seems im not gettng the sensible outputs.....
> _i connected pin3 and 4
>
> lastly,
>
> any idea on a simpler way to realize my application?

If you want to operate a metal detector at low frequencies, you'd
better bring a plastic wheelbarrow to carry the thing. You'll need a
lot of windings to get the inductance you'll need for a large air coil
to oscillate at lower frequencies.

Small pieces of metal will make small changes in the oscillating
frequency -- possibly tens or hundreds of Hz out of hundreds of KHz. A
tenth of a percent or less. Dividing it down to a frequency accessible
by a PIC will mean you lose the information you need.

Metal detection is inherently an RF enterprise -- hate to say it, but
get used to it. It's an analog world. And the LM2917 is made for
tachometer-speed applications (RPMs rather than KHz).

If you really want a voltage output, get the beat frequency oscillator
working, and then work on changing the audio beat frequency into a
voltage. You can do that easily by using a comparator to change the
audio into a digital output, and then use a 555 to output a brief pulse
for every beat oscillation. You can then use an R-C filter to get a DC
voltage you can read with a meter or your PIC. You also might want to
try just counting the beats with a PIC, but that will have limitations,
too. You have to watch out for maximum frequency.

If you are allergic to the whole mixer concept, you might just want to
get a handheld frequency counter, and pick off the RF oscillator
output. Inexpensive handheld frequency counters can give you the
resolution you'll need, but it gets tiring looking at the display. The
idea is to wander blithely down the beach waving the finder back and
forth, and having an excuse to scope out the scenery while you're
nominally being industrious. ;-)

Good luck
Chris

Chris, Apr 6, 2006
7. ### electroGuest

i have a working lc tuned oscillator now.my loop is about 5 inches wide
with 20 windings,im getting a 37.9uH
from that setup...can i simply port that to the ADC unit of my
PIC876A?how will the digital value relate to my frequency?
also, since resonance is dependent on capacitor and inductor, why cant
i select capacitor values and still have a tolerable amount of
inductance air coil winding?

thanks!

electro, Apr 6, 2006
8. ### BanGuest

electro wrote:
> i have a working lc tuned oscillator now.my loop is about 5 inches
> wide with 20 windings,im getting a 37.9uH
> from that setup...can i simply port that to the ADC unit of my
> PIC876A?how will the digital value relate to my frequency?
> also, since resonance is dependent on capacitor and inductor, why cant
> i select capacitor values and still have a tolerable amount of
> inductance air coil winding?
>
> thanks!

maybe you first read up a bit how such a thing works. when you have 400kHz
oscillator, it will maybe change its frequency 100Hz with a big chunk of
metal, lets say a treasure buried a few inches below ground, that is 0.025%.
http://home.clara.net/saxons/bfo.htm that is why another oscillator is used.
It is very easy to mix the two outputs, just with resistors.
Since both oscillators are similar the frequency drift gets cancelled and
you can listen to the more stable beat frequency. A PIC will not help, when
you do not have the front end right.
--
ciao Ban
Apricale, Italy

Ban, Apr 6, 2006
9. ### electroGuest

or cant i use the built-in comparator of my pic...sine i have a sine
wave signal at zero dc offset,
i will count the interval between successive zero crossings that is i
will set my zero volt level as my reference...
will that be possible?

electro, Apr 7, 2006