Tachometer

M

miketinte

Guest
Hello all,

I need to design a tachometer for a Formula racing car. We have an FM
output from an ECL going into (most likely) an LM3914 Freq-Voltage
Converter (www.national.com). What most concerns me, however, is the
output of the LM3914. I want it to drive an LED bargraph, but have
been hard-pressed to find any decent schematics of the bargraph with a
good explanation so I can tweak it.

Any help with either a link or the theory of how an LED bargraph works
would be greatly appreciated.

Thanks,
Mike
 
Hi

Try tsm39341 bar graph modules. 10 lights, electronics built in can be
cascaded.

Geoff

"miketinte" <mlt3@lehigh.edu> wrote in message
news:8a641345.0309170332.5ca06e0e@posting.google.com...
Hello all,

I need to design a tachometer for a Formula racing car. We have an FM
output from an ECL going into (most likely) an LM3914 Freq-Voltage
Converter (www.national.com). What most concerns me, however, is the
output of the LM3914. I want it to drive an LED bargraph, but have
been hard-pressed to find any decent schematics of the bargraph with a
good explanation so I can tweak it.

Any help with either a link or the theory of how an LED bargraph works
would be greatly appreciated.

Thanks,
Mike
Click to expand...
 
Thanks Geoff. Who makes the TSM39341. I tried googling it and came up
with one two sites, one was advertising brokerage software, another
was unreachable.

-Mike
 
On 17 Sep 2003 04:32:49 -0700, mlt3@lehigh.edu (miketinte) wrote:

Hello all,

I need to design a tachometer for a Formula racing car. We have an FM
output from an ECL going into (most likely) an LM3914 Freq-Voltage
Converter (www.national.com). What most concerns me, however, is the
output of the LM3914. I want it to drive an LED bargraph, but have
been hard-pressed to find any decent schematics of the bargraph with a
good explanation so I can tweak it.

Any help with either a link or the theory of how an LED bargraph works
would be greatly appreciated.

Thanks,
Mike
I think you meant to say LM2917 for the F/V convertor.
Click to expand...

I built a 20 led dotgraph tach for my motorcycle using LM3914
dot/bargraph drivers. Lots of data on the web on the 3914. The leds
were arranged in a circular pattern and had three auto ranging spans
(the clock could go around a total of four times with an LED to
indicate which range was in use)

I used a series of three comparators sensing the voltage from the freq
to voltage convertor. The resistors used in the comparator references
also provided the voltage to the dotgraph via a summing op amp and
analog switches. The way it is set up - the bargraph "looks" at one
span (say zero to one volt) when that span is over-ranged, it switches
and looks at span two or one volt to two volts, etc.

That way I was able to make 20 LED's do the job of 80, indicating RPM
from 1000 to 5000 (my normal cruising range - no need to indicate zero
since the bike won't idle below 1000 RPM very well)

It works well and is slick looking. I used 2000 mcd leds and drive
them at 6 milliamps and they are plenty bright for daylight use.

If I had to do it again, I'd see about using bi or tricolor LEDs to
indicate the range and just have the display LED's change color
depending on which range I was in.

The national data sheet shows an "exclamation point display" where the
LEDs switch from dot to bar graph and flash to indicate an over range
condition.

My motorcycle only provides one ignition pulse per revolution, so I
ended up using the output of the alternator which is mounted to the
crankshaft and has twelve poles. The optical tach is much faster
indicating than an electro- mechanical meter.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----
 
How about using a PIC to drive the LEDs? I'm looking for about 15 LEDs
to provide a high enough resolution. Can I daisy chain these bad boys
to provide for the higher number of LEDs, or should I just go with a
PIC that has enough outputs?

I'm shooting for an ETA of October 18th with this thing, so any quick
help would be appreciated.

Thanks,
Mike
 
On 4 Oct 2003 09:04:12 -0700, mlt3@lehigh.edu (miketinte) wrote:

How about using a PIC to drive the LEDs? I'm looking for about 15 LEDs
to provide a high enough resolution. Can I daisy chain these bad boys
to provide for the higher number of LEDs, or should I just go with a
PIC that has enough outputs?

I'm shooting for an ETA of October 18th with this thing, so any quick
help would be appreciated.

Thanks,
Mike
Click to expand...

Sure you could use a pic and just calculate the RPM based on interval
between pulses (more accurate for very slow speeds with say one pulse
per rev). I should think that an 8 bit output would be plenty, but
four might be limiting. Then after you decode you would need some
logic if you still want a bar graph and not a dot graph. Or as you
say find something with a 16 bit control output.

It is your project. I can post what I did on the binary group if you
want (and the posting god allows). Right now I can only post jpg or
djvu format. The former occupies a lot of bandwidth, the later, not
everyone has the plug in.

I'm basically an analog designer. I like dealing with analog signals.
For me, that's easy inexpensive, and satisfying.

If you are asking if the LM3914 can display 15 LEDs, two of them can.
Ten can light 100 LEDs (but you'd be hard pressed to get 100 on a
display unless you use surface mount LEDs. I settled for 20 (or 20 x
4 RANGES = 80) as plenty of resolution. With a mere 15, your
resolution is going to be in 250 RPM steps? - is that enough?

It wouldn't pay to use my idea for only 250 rpm resolution. It takes
me 7 IC's (counting regulators) to do it.

Interval and digital smoothing is probably the best way to do it.
Using interval you aren't concerned with the number of pulses per
revolution, one is all you need.

If I use one per rev, in an analog circuit, the integration capacitor
has to be large - and then I only know what the rpm was three seconds
ago.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----
 
"miketinte" <mlt3@lehigh.edu> wrote in message
news:8a641345.0310040804.abf6234@posting.google.com...
How about using a PIC to drive the LEDs? I'm looking for about 15 LEDs
to provide a high enough resolution. Can I daisy chain these bad boys
to provide for the higher number of LEDs, or should I just go with a
PIC that has enough outputs?

I'm shooting for an ETA of October 18th with this thing, so any quick
help would be appreciated.

Thanks,
Mike

Mike,
I don't know you're exact requirements, but to measure rpm, try this:

Use BCD ripple counters with bcd-seven segment drivers and seven segment led
displays.

Use a timer chip circuit to make a pulse whose width, (time) is adjustable.
If something is putting out pulses and you want that pulse rate to indicate
a certain value, then adjust the time window pulse width for that time and
the count will be what you want. If the shaft slows to one half that speed,
then you will count exactly one half (+-1) as many pulses in that same fixed
time window. Latch the count into display drivers and reset the counters at
the end of the time window.

Example: Engine goes 2400rpm which is 40 revolutions per second. If you
have two magnets on opposite sides of the shaft, you will get 80 counts per
second. If you count for one second, the counter will read 80. But if you
count for 24/80 of a second, (0.3sec), the counter will read 24. You can
read in 100s of rpm. Similarly, if the engine goes 1200rpm = 20 rps then
the counter will read 12. A zener regulated voltage operating a 555 timer
with good quality resistors and capacitors and a calibration pot can provide
the adjustable time width pulse good enough for normal use. Crystal
oscillators and dividers can give more accurate pulse widths. Here the
count updates or changes 3 times a second. A simple and accurate
"speedometer". Put 2 cabinet door magnets clamped with 2 radiator clamps
onto your rear-wheel car's drive line, mount a pickup coil and you have a
good digital speedometer.

The example used "two teeth" per revolution. A gear with many teeth can
count much faster. If you had 20 teeth and counted for 3 seconds you could
actually get a count of 2400 with a resolution or 1rpm. (10 times as many
teeth and 10 times as long to count). This is somewhat a waste of effort
because the time window oscillator isn't accurate enough to really show +-
1rpm out of 2400, (0.04+%). But use 20 teeth and 0.3sec count window to
readout in 10s of rpm and just show the units digit always 0. You get fast
update and the counter resolves 2400, 2390, 2380, etc, every 10 rpm.
Same story with 10 teeth and 0.6 sec window etc.
Ghost
--
"The difference between fiction and reality
is that fiction must make sense."
Tom Clancey
 
(snip)

In the initial post Mike is looking for a bargraph display. Digital
tachs are the best choice for accuracy and resolution, and easier to
implement, but it takes too long to read them when one is "busy."

Analog LED displays shine in that app. Concentrate on the driving and
let the peripheral vision watch the tach. Using LEDs buys a faster
response time than a damped mechanical movement.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----
 
Good points.
Ghost
--
"The difference between fiction and reality
is that fiction must make sense."
Tom Clancey


"default" <R75/5@defaulter.net> wrote in message
news:g380ovkdlk8b4ch961l37fpiv6d87egdvf@4ax.com...
(snip)

In the initial post Mike is looking for a bargraph display. Digital
tachs are the best choice for accuracy and resolution, and easier to
implement, but it takes too long to read them when one is "busy."

Analog LED displays shine in that app. Concentrate on the driving and
let the peripheral vision watch the tach. Using LEDs buys a faster
response time than a damped mechanical movement.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----
Click to expand...
 
Yes. The bargraph is going to be arranged in an arc shaped patter so
as to mimic the look and feel of an analog tach. My school is known
for their design every year, so we're doing our best to make this
original and eye-catching.

Just to clarify...so I want to put a frequency modulated signal into a
PIC, and then have that control a voltage at one of the outputs? Why
not just use a F-V converter directly driving the bargraph driver? Why
bother with the PIC?

Thanks,
Mike

default <R75/5@defaulter.net> wrote in message news:<g380ovkdlk8b4ch961l37fpiv6d87egdvf@4ax.com>...
(snip)

In the initial post Mike is looking for a bargraph display. Digital
tachs are the best choice for accuracy and resolution, and easier to
implement, but it takes too long to read them when one is "busy."

Analog LED displays shine in that app. Concentrate on the driving and
let the peripheral vision watch the tach. Using LEDs buys a faster
response time than a damped mechanical movement.


-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----
Click to expand...
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top