I have been asked to build a digital to analog circuit.
The flow sensor:
has a digital pulse output proportional to the flow.
I would like to create an analog detector, so when it has reached a minimum flow rate, a logic level is triggered.
I am a firmware guy and know how to do this with a PYFMC[1].
I would like to get some suggestions on how to do this in analog.
thanks
hamilton
You could use my Double-Tach circuit followed by an analog comparator.
ftp://jjlarkin.lmi.net/DoubleTach.jpg
Or maybe just a couple of cascaded retriggerable one-shots, if all you need is the single underspeed indication. That's simple.
Nice looking sensor. What does it cost?
John
The solution depends quite much on the required accuracy of the detection.
An analog approach would be a monostable for constant-width pulses, smoothing the pulse train to a DC level, and finally comparing it against a threshold voltage. The achievable precision and stability are modest.
A digital approach would include a monostable set for the threshold interval, and a latch to capture the previous monostable output at next triggering instant.
PYFMC controlled by a crystal may be the most cost-effective solution for good accuracy.
On a sunny day (Sun, 09 May 2010 13:28:51 -0600) it happened hamilton wrote in :
If it is just a pulse, one shot -> low pass -> analog comparator. The one shot makes a fixed pulse width, the more pusles the higher the voltage after the lowpass, the comparator flips on a specific voltage level, and then triggers the lever that releases the apple that falls on the detonator that triggers the nuke.
YAW
Thanks,
Yes, underspeed is the goal.
Just an output logic level going to another micro that controls the rest of the system.
AAck $100 each !!!
The boss thinks he need the accuracy.
With the other comments, I will use a PYFMC [2] and just count pulses.
The data sheet states the range is 37 to 917 Hz. (.1 to 2.5 LPM).
hamilton
[2] This time I'll use an AVR.
Here is a frequency to Voltage converter .. Look below and see many examples..
Its old technology but it still works, especially well for what you're doing..
for a logic level set point, you need to pass that into a schmitt trigger buffer, or use a comparator like the LM311, LM393 with a little (+) feed back to get a hysteresis effect.. Or! if that is over the hump too much, you can do what I do, and that is use the trigger and threshold inputs tied together of a 555 timer. The timer has a 1/3 ladder voltage of the Vcc, it works nicely for a RS flip flop and a window comparator. The timer output will give you the logic level.. There will be no components added to this chip :)
have a good day..
On May 9, 2:28=A0pm, hamilton wrote: (ker-snippity...)
OK - just off the top of my head, and I'm sure others will cruely and casually punch holes in my suggestions...
1) how about a low-pass filter feeding into a voltage comparater.2) frequency-to-voltage converter (LM2907/2917) - might have to put a one-shot on the front end.
3) missing pulse detector - easily made w/a 555 (again, may need a one- shot)4) how big does the max flow deviate from the minimum you want to detect? Would a PLL work? The output would be the deviation, and again, into a voltage comparator
5) the way we do it - hire a grad (or undergrad) student to watch the flow (or indicator)punch away...
A retriggerable monostable will provide the required functionality.
A TLC555 connected as a monostable and a comparator (with hysteresis?), plus a low-pass filter with an RC or two. Probably more than a dozen parts.
I think I'd use a micro with 1% internal calibrated oscillator (or maybe a 0.5% ceramic resonator) and that's about it. Writing the code might take a bit longer than designing the analog circuit, but the parts count reduction would probably be worth it.
Best regards, Spehro Pefhany
"hamilton"
** Small but important point - why did you call the pulse ouput "digital" ?There is nothing digital about such a varying pulse train - it's pure analogue.
.... Phil
A re-triggerable one-shot will have an output that stays high until the minimum pulse rate occurs, then it will be toggling. ...Jim Thompson
What if it were capacitivly coupled at the input?
Output toggles when input rate is less than one-shot period.
Grant.
OK, got it, Misunderstood parent post.
Sheets/FT210...
I presume this is a fixed-width pulse with well-conditioned peak and trough values? Feed it to an R-C series pair, clamp the other end so it alternately sources current and returns to GND. Then integrate (resistor, op amp, capacitor in negative feedback) and limit the integrator gain with a resistor across the feedback capacitor. It's your basic capacitor current source and ammeter, with that clamp diode feeding only the positive half of the current to the meter.
=20
Next question. How fast does it have to respond? Third question. How touchy is the system at detection threshold?
Have something to add? Share your thoughts — no account required.
Ask the community — no account required