What DAQ? device can I use to read a Gear-Tooth Speed sensor?

Irving

Joined Jan 30, 2016
3,843
The power feed to J1 and thus to your sensor is from 5v not Vcc. If your sensor requires a pull-up to 5v then this resistor should be on the PCB.

I'd suggest using the 78L05 TO-92-case version for the regulator instead of the TO-220 version as its smaller.

Be careful with the SMD parts - what size have you selected? Also the temperature stability of C1 is important, if you're using a MLCC part you need to specify a C0G or NPO dielectric rather than the more common X5R/X7R.

The line on the connector footprint is the front (this from Kicad)

1657456573526.png
 

Thread Starter

Kulanib

Joined Jul 6, 2022
55
@Irving

1. Thanks for the visual of the terminal block. This help, I can just do the same thing as you did and make a mini PCB to see the CAD model of whatever component I pick - good tip.

2. Wow, I had no idea I can just implement a pull-up resistor right on the PCD/peff-board/breadboard - I thought I would have to go to the Sensor cable and splice it to put a resistor on it lol. Thanks for that one again.

3. I had no idea that I count also power the sensor directly from the board as well, that Vcc was meant for the regulator and LM2917, I was just going to input only the Sensor Output Pin to the board. Now I can basically connect the whole sensor onto the board as well and supply it with 5v, the LM2983 and Regulator with 12 V Vcc(don't see a reason for 15V? since there's a zenner regulator cap at 7.5V). Thanks, I have changed the regulator type to the 78L05 TO92 - I thought the heat would affect it since no heat sink.

4. As for the component type, Let me avoid SMDs (they are end game), I will just revert back to the regular size components. As soon as all works well, then I can attempt the PCB again bearing in mind what you mationed.
 

Thread Starter

Kulanib

Joined Jul 6, 2022
55
The power feed to J1 and thus to your sensor is from 5v not Vcc. If your sensor requires a pull-up to 5v then this resistor should be on the PCB
*Corrected Pin 3 and 4 connection and set final values of R1, C1, C2, Vcc, set Minimum pulse trigger voltage to be 0.5V*
1. Added the power connectors for the sensor with its pull-up resistors (cleaners solutions, really cool).
2. Replaced regulator with the recommended one.
3. I will use a Peff Board for now and avoid the SMD components. But the layout would be similar. Will only move on to PCB later on.
Screenshot.png
4. Got those connectors.
PCB 2.JPG
 

Attachments

Last edited:

Irving

Joined Jan 30, 2016
3,843
VR1 should be a 10k, ideally a multiturn for setting output accurately, in series with a 27k 1% resistor to give 33k for 6v out (6v/180uA). Don't use a single larger value pot, it'll much harder to calibrate. The likely calibration range, based on the tolerances of R1 (1%), C1 (2%), and the charge pump current source (est 1%) is +/-2.5% ie 32.45k- 34.13k = 1.68k. On a 50k pot thats only 3.4% of its range, while using a 10k in series with a 27k uses 28% of its range.
 

Irving

Joined Jan 30, 2016
3,843
Also a 5nF (4.7nF) capacitor will give Fmax (from Eq 7 of datasheet)of 180e-6/(6 * 4.7e-9) = 6.4kHz so your 2kHz will be <1/2 scale. Allowing 10% over 2kHz = 2.2kHz gives C1 = 180e-6/(2200 * 6) = 13.6nF, so use 12nF 2%.
 

Thread Starter

Kulanib

Joined Jul 6, 2022
55
Thank you.

I have built a prototype and been testing it. It works, I was shocked.So that’s 0.997kHz. input according to the multimeter, from an Arduino blink code (5-0 pulse), 50% duty cycle, and that’s about 3 V reading (error from the tolerances of my components and my resistor trim). Doubling the frequency took me to about 6 V.
5F497975-4CBE-4BEB-96C8-C536A78352D6.jpeg
69342275-E611-4BDC-9924-A038C9448E5E.jpeg
7BD1BCF4-E2BD-4AD1-82BB-7F4AB1D7C6CA.jpeg
I am also happy with the response time. (0V is the Arduino resetting)

Notice my little ripple as well on the 3V reading. This is with 4.7nF (found from ceramic tiny ones for now), 12V supply, ~43K resistor. (No regulator yet, still waiting for components. Divided 12V by a 10K and a 1K to get Tach- on pin 11)

What’s the significance of C3 across R3?
 

Irving

Joined Jan 30, 2016
3,843
Excellent! If you get the calibration right I reckon you should be within 10 - 15rpm at full scale (3000rpm),. That's as least as accurate as my laser tacho and outside of the very low speed (high ripple) revs the linearity should be similar or better.

C3 just provides a little bit of noise immunity for the 2.5v bias. If there is ripple on the 5v rail from the varying load of the sensor as it switches on/off this could affect the sample point. Its normal practice to decouple the divider chain. Having said that, the 2917 samples on both rising and falling edges so any variation in sample point due to sensor ripple would likely cancel out anyway.
 

Thread Starter

Kulanib

Joined Jul 6, 2022
55
Excellent! If you get the calibration right I reckon you should be within 10 - 15rpm at full scale (3000rpm),. That's as least as accurate as my laser tacho and outside of the very low speed (high ripple) revs the linearity should be similar or better.

C3 just provides a little bit of noise immunity for the 2.5v bias. If there is ripple on the 5v rail from the varying load of the sensor as it switches on/off this could affect the sample point. Its normal practice to decouple the divider chain. Having said that, the 2917 samples on both rising and falling edges so any variation in sample point due to sensor ripple would likely cancel out anyway.
Thank you so much.

I see the significance now.

I will design a speed pick-up for the hall effect sensor and will try to give it a consistent duty cycle (% even though the value doesn't matter much, long as it's constant and not in the extreme ends) as it rotates by equally spacing out the individual teeth and gaps, as well as giving the gap and the tooth the same arc length. Something like this: (not to scale)

speed pickup.JPG

E.g, So this profile having 8 teeth, producing a frequency of 2000 Hz, would mean my shaft is spinning at (2000)(60)/8=15K RPM max reading.

Doubling the number of teeth would mean (2000)(60)/(16)=7500 RPM max reading and so forth.
 

Irving

Joined Jan 30, 2016
3,843
E.g, So this profile having 8 teeth, producing a frequency of 2000 Hz, would mean my shaft is spinning at (2000)(60)/8=15K RPM max reading.

Doubling the number of teeth would mean (2000)(60)/(16)=7500 RPM max reading and so forth.
Exactly so, but you want your highest required revs to be at or near 6v, so if you require a maximum of typically 3000rpm, which translates to 1800Hz on 36 teeth, but only 400Hz on 8 teeth. More teeth is better as has a wider dynamic range and better ripple for the lower speeds.
 
Top