120,000rpm tachometer help

[decoder]

Hi Smurf, i believe you ought to use the black and white paint on the moving part of your engine and use a REFLECTIVE OPTO SWITCH which are very cheap, about £2.00 in the uk and are available from R.S.COMPONENTS and they are on the webb (put it in google) but i'm sure you would have similar electronic stockists in your country.
The reflective opto switch is about the size of your thumbnail and you fix it about 5mm away from the painted area. It comprises an led which you put a series resistor on, and a battery, and the led shines on the black/white surface and its beam is reflected off the paint and into the switches detector.
When you go onto the r.s. site a data sheet is available showing how it is used and how to connect it. In its very simplest form you could use a multimeter as the display, but you would have to calibrate it to relate volts that the meter is reading, into rpm that the engine is doing.If you used the LM2907 integrated circuit (also available from R.S with data sheet) then much better, more linear results can be obtained, and still use a multimeter for the display. Another option for the display is, you could probably get a rev counter from a car scrap yard very cheaply, as the modern ones are electronic and you could connect the reflected opto switch to the input of the rev counter and write new numbers on the rev counter dial.
I wish you luck and i hope i havent bored you.
Clive
(Bedford UK)

[Smurf208]

Thank you decoder that was helpful useful info, thanks for the good post!!

[decoder]

Hi smurf
If you cant find the info i refered to in my previous communication, please email me direct.
dec0der@ntlworld.com (the o in decoder is a zero)
Regards
Clive
(Bedford UK )

[lerman]

For a frequency counter kit (around $55), see:

http://www.apogeekits.com/frequency_counter.htm

Connect this to your optical, magnetic, or other sensor. Multiply the reading that you see by 60 to convert to RPM.

If you prefer to not buy a kit, you can get the schematic and source code from that site.

http://www.invent-electronics.com/freq_Counter.htm has a $50 kit

If you google: frequency counter kit, you will find lot's of others. Or, if you prefer, go to ebay and search for frequency counter. You'll find lots of them; new and used.


Ken

[NC Cams]

Smurf: You obviously can't be impressed or convinced or swayed with engineering logic. Moreover and unfortunately, you are apparently electronically and mathematically challenged.

For this reason, allow me to see if I can stimulate your interest in the use of the LM2907 via a technology forum that you don't need a college degree to understand - namely ab out of pocket cost analysis.

The LM2907 can be turned into a tachometer via the interconnection of 3 subcomponents:

1. a motion sensor (at this point undefined and unpriced)
2. a device to calculate RPM (the 2907 tach circuit)
3. a Volt/Ohmeter (probably digital that you probably already own) to display the speed output calcuated by item 2 above.

Now that you have had some alternate tachometer devices to look at (see post #28 for $$$ comparisons), the charm of the 2907 should become readily apparent even to a mathematically challenged person when you consider the following:

A minimal parts count tachometer can be built using the cirtuit diagram found at the bottom of page 5 of the following link:

http://eshop.engineering.uiowa.edu/N...9/DS007942.pdf

via the use of:

1 @ optical sensor (TBD)
1 @ LM2907M-8 for about $1.90 (Digi-Key)
2 @ capacitors (values TBD) (Digi-Key, $1 or less)
2 @ resistors (values TBD) (Digi-KeY, $0.025 or so for 2)
1 @ prototyping board (Radio Shack or Digi-Key, Cost TBD)
1 @ VOM (that you should already have).

Thus, for the cost of the motion sensor and probably $5 more in TOTAL parts, you can have the "simple" tach that you want/need. Is that "simple" and cheap enough for your budget??????

Or you can buy the units as outlined in post #28 and still have to mess with math and other F-to-V stuff that you admittedly don't want to or can't deal with.

Unfortunately, until you or someone defines EXACTLY how many counts per rev you intend to generate, the resistor and capacitor calcs can't be performed for the 2907 circuit.

Until/unless someon identifies exactly what type of signal is being output from the optical signal generator that is yet to be specified, NOBODY can interface it with the rest of the 2907 tach circuit.

Note: sometimes you need to do some signal conditioning to make the optical sensor work with the tachometer circuit. Hence, a chip or two may need to be added to the mix. Even so, we're still talking relative peanuts in cost and complexity compared to the alternatives.

Sadly (for you), just about any non-commercial, DIY tach circuit WILL have to have some mathematical calculations performed so as to take basic building block parts and make them do specific things - high tech electromechanical devices (ala tachometers) actually do have to be engineered by SOMEBODY as the Good Lord grow them on trees like apples, pears or peaches that get ripe at certain times of the year for the picking....

EDIT: To the potential tach builders who are following this thread and are interesed in the engineering of the circuit, a note or two FYI.

A review of my R/C motor dyno circuit notes showed that a simple photo cell that sensed 2 pulses per rev at 60,000 rpm (1 pulse/rev at 120,000 rpm) required a special trigger in order to function properly with regard to RPM sensing. We had to feed the optical sensor into a 555 based "one shot" that fed into a Schmidt trigger CMOS IC then to the 2907. T

his gave us a fast occurring but narrow, well defined square wave pulse that the 2907 could better integrate into a voltage proportional to/with speed. I seem to recall that we also added a small capacitor to the RPM voltage signal pin on the 2907 to do some voltage smoothing/integrating as well. Although the min parts count circuit doesn't show this, we empirically found that a small cap on the output pin made the readings a bit more stable and easier to read.

NOTE: If you just used the raw signal from the photo cells (as we did from a slotted window in our flywheel), the variable duration of the photo cell output(it got shorter as speed increased) resulted in non-linear operation of the 2907's charge pump. This led to bizarre "that can't be" high rpm readings that took a while to figure out.

By generating a true fixed duration pulse from the optics with a one-shot, a pure "frequency" as opposed to a "psuedo PWM'd frequency" was ultimately generated. THis sort of thing (one shot trigger) might also be needed for this circuit due to the speeds involved, come to think of it. Once we did this, the thing performed amazingly well for a $5 circuit.

END EDIT

[JRoque]

Hello Ken, great links thanks for that. A frequency counter seems like a good idea for this application. My DMM has a built-in counter and others might do as well.

Hi Gerry, please consider leaving some of the good information on the thread posted and, if needed, remove just the "spam".

JR

[pastera]

Smurf,

I am trying to be the better person and help you even though you have insulted me.
The LM2917 datasheet gives the following formula for output voltage
Vo = Fin * Vcc* R1 *C1

The only set number is Fin (input signal frequency) = 2000Hz (120kRPM/60)
Choose an easy output voltage - 1.2 volts will work nicely with your digital meter.
Next we choose a voltage to run the circuit from; A 9V battery is cheap and portable: Vcc = 9v
Next we have to choose either R1 or C1, there are restrictions on the size of R1 so lets choose the resistor.
Vo (output voltage)/R1 must be less than 140uA therefore R1 must be greater than 1.2v/140uA = 8571ohms
Lets choose 12000 for R1
C1 = Vo/(Fin*Vcc*R1) = 5.55e-9 ~ 5600pF

Here is a circuit with the calcs done. Use a 1% resistor for the 12K. The 5600pF capacitor should be accurate as well. The circuit is so simple, it can be put together on a simple perf board. More accuracy can be gained by replacing the 12k resistor with a 10K in series with a 3k variable to allow for calibration.

If you use the lm2917, you need to add a 1000pF cap in parallel with the 5600pF cap (or use two 3300pF)

You will need to feed this circuit with a signal that switches from <0.7v to >0.7v once each revolution. This signal would be generated by your optical sensor.

Aaron

[Smurf208]

Pastera, i don't remember saying anything bad to you, if i did it i'm sorry i don't remember haveing anything against u. Now thats helpfull information, thats the kind of post that people should be doing, this isn't a site where people expect something out of the other in return for there help, this is a site where people share knowledge, thats a post that thats helps greatly. Why can't things just be simple and simple answers and questions.

[Smurf208]

O and i just need to double things cause i will need 2 magnets to balance things, u don't need to redo the calcs, i can double the numbers and see what i can do, thanks tho

[Smurf208]

Or i could find a optical from ebay or something that will do, which would leave the 1 pulse and everything is good

[JDC]

Hello ,THIS is JDC as my experiments with this type system you should
look at laser feedback sys. with quad support converstion
GoodLuck,SIR

[NC Cams]

Note: Unbiased, unemotional, well intended advise follows:

Smurf: you can't simply double the values if you double the number of counting pulses that are being input into the circuit - been there, done that, doesn't work. You (or someone) will have to recalculate the RC timing relationship.

Reason: the tach won't give the correct voltage output for the actual revs being turned with doubled values, especially at ultra high RPM where the "averaging" of the charge pump circuit is affected by the much shorter times involved (long technical reason that doesn't need to be explained, however it just does).

Note: at high speeds (120K or 240 pulses/min for 120K rpm), you may find that the DURATION of the input pulse width can start to affect the linearity of the tach circuit. We did and we fought that issue for a week on our R/C car race motor tach.

Result: the tach starts to "lag" behind in its ability to give true RPM as you reach your terminal speed if the time constant of your trigger pulse changes drastically from low to high speed. Thus, even though the engine may truly be running faster, the tach won't show it or may show a bogus speed. The "one-shot" mentioned in a prior post that is triggered by your sensor will solve that dilemma.

The duration of the one shot pulse had to be empirically determined in our high speed circuit - it took a bit of experimentation to get RC constants of the one-shot and tach worked out so that the tach would work properly but it can be done - the 2907/2917 circuits are quite adaptable.

The need to test and verify/callibrate can not be overemphasized.