For Chapulino

[Georges GUAL]

Hello,
I read your answer and I have to confirm that the operation of your tachometer and similar to mine.
Display in step 30 and wait several seconds before updating on Touchdro.
Georges

Bonjour,
J'ai lu votre réponse et je dois confirmer que le fonctionnement de votre tachymètre et similaire au mien.
Affichage au pas de 30 et il faut attendre plusieurs secondes avant l'actualisation sur Touchdro.
Georges

[Yuriy]

Hello,
I read your answer and I have to confirm that the operation of your tachometer and similar to mine.
Display in step 30 and wait several seconds before updating on Touchdro.
Georges

Bonjour,
J'ai lu votre réponse et je dois confirmer que le fonctionnement de votre tachymètre et similaire au mien.
Affichage au pas de 30 et il faut attendre plusieurs secondes avant l'actualisation sur Touchdro.
Georges

Georges,
What exactly are you trying to do with the tachometer?
I designed the firmware so tachometer input is handled on a lower priority thread. My rationale was that a user CAN NOT loose scale input at any time since it will cause defects. Tachometer is not a critical measurement and will be used for informational purposes. 
So, on the lower priority thread it's read 2x per second (i.e. the firmware counts the pulses and twice per second calculates RPM (pulse count * 2 * 60) and sends it to the app. I wrote the code so it's very low-impact, so the board can handle pulses into 10s of KHz. The trade-of is the lower-end. When you start talking about 4-5 pulses per second, the tachometer will be pretty much worthless.
RysiuM added some fancy stuff for the low RPM handling to the Arduino sketch, I'm not doing that to the MSP430 version because it has real tradeoffs for the overall performance. If you really need to measure very low RPM, make a disk with more slits/magnets. Little button magnets are cheap enough that you can add 20 of them if you want to.
Bottom line, the little MSP430 chip has finite amount of resources. In this case the limited resource is the timer. To be able to calculate low RPM I will measure time between pulses. Right now all timers are tied up, so to do that I will need to remove support of iGaiging scale clock  and use that timer for interval measurement.

Regards
Yuriy

[Georges GUAL]

Yuriy,
I just want to measure the speed of my spindle with more precision but I just figured out that the tachometer was not a priority for you as a measure.
Georges 

Yuriy,
Je veux seulement mesurer la vitesse de ma broche avec plus de précisions mais je viens de comprendre que le tachymètre n'était pas une priorité pour vous en tant que mesure.
Georges

[chapulino]

I understand Yuri's compromise solution.

It is appropriate to restrict resources for the tachometer that is never critical.
As for the speed of response therefore in my opinion more than acceptable.

Personalemtne I am happy with these two DRo the milling machine and the lathe.

Thanks Yuri

[Yuriy]

Yuriy,
I just want to measure the speed of my spindle with more precision but I just figured out that the tachometer was not a priority for you as a measure.
Georges 

Yuriy,
Je veux seulement mesurer la vitesse de ma broche avec plus de précisions mais je viens de comprendre que le tachymètre n'était pas une priorité pour vous en tant que mesure.
Georges

Georges,
It wouldn't says it wasn't a priority for me - I have to work with the resources I have in MSP430.
If I compromise the position reading, the DRO will be inaccurate and will result if destroyed parts, so I had to prioritize the position measurement over everything else. Second is responsiveness. Lag of more than 50 milliseconds starts being noticeable. Quadrature and iGaging scales can work pretty fast, so I had to allocate enough resources to make sure I can read the scales fast enough and send the position to the app at a decent rate.
Tachometer is a useful feature but lag is not very critical since the spindle inherently has lag (flywheel effect) and even if the speed is lagging 1 second behind, generally this is not an issue. My mill has a built-in tachometer (SIEG SX4) and it lags more than a few seconds due to the acceleration curves. It took me a few hours tops to develop "muscle memory" to set the dial to the bulk-part position and then fine-tune once the spindle spins-up. Similarly, if you expect to turn at low speeds, use a disk with more slits/magnets. You can easily fit 20 magnets into a 50mm disk and that will give you stable tach output down to 50 RPM or even less (and won't overflow at 1000 RPM either). 
As hobby machinists we have all sorts of compromises to deal with: I can't afford a Mazak 5-axis CNC machining center that can produce 5 Kg of chips in a minute, so I have to live with a "little" SIEG SX4 mill and Jet 1024 lathe. That means that I have to use smaller tooling, more HSS, less carbide and take lighter cuts. Same with TouchDRO: I could use an ARM CPU and have an ultra-precise tachometer but the board would not be DIY anymore and would probably cost 2-3x as much.
Hope this makes sense.
Yuriy

[Georges GUAL]

Yuriy,
I understood very well. You had to make a compromise between precision and RPM. Your choice on precision is excellent.
I made a disc of 20 notches with the use of an opto. It should work well.
Thank you for the time you spent answering me.
But I'm looking forward to the Touchdro user manual. In parenthesis, it is a sacred application. I will do a system for my small milling machine. To see the few explanations and video on the net, it is encouraging.
Georges

Yuriy,
J'ai très bien compris. Vous aviez a faire un compromis entre précision et RPM. Votre choix sur la précision est excellent.
J'ai fait un disque de 20 encoches avec l'utilisation d'un opto. Cela devrait bien fonctionner.
Merci pour le temps que vous avez passé à me répondre.
Mais j'attends avec impatience le manuel d'utilisation de Touchdro. Entre parenthèse, c'est une sacré application. Je vais refaire un système pour ma petite fraiseuse. A voir les quelques explications et vidéo sur le net, c'est encouragent.
Georges