MSP430FR5989: ESI module doesnt detect rotation on water meter

Hello Bálint ,

I am looking into this and will get back to you.

Hello Balint,

The magnetic distortion caused by rotation of the wheel of the meter is different to that one caused by your test setup. You would need to re-calibrate the setup with the meter. For that the meter would need to run... the wheel in its final mounting configuration needs to rotate...

To do that you need an water circulation with a speed controlled pump; alternatively you could use a valve to throttle the flow... of what I've seen from your pictures your setup is able to do that.

Johann

Dear Johann,

I calibrate the setup every time I'm doing a test. 

It makes no difference if I test the device on the water meter or with the screw. Also this doesnt make any difference if set_dac() completes without any rotation.

I can provide more pictures of my setup if needed.

Bálint

EDIT:

I have an exact same rotation wheel that is currently removed from the water meter, I can use that for calibration too.

Also, if calibration doesnt complete automatically, then I use for example the water meter to finish the calibration. After that, I rotate the water meter but I doesnt get any ESI interrupt, besides ESIIFG1, the ESISTOP flag.

I think I might get wrong DAC values, but I couldnt prove it (and I dont know how to fix it), its just my theory.

Bálint

EDIT:

In TSM_auto_cal funtion I struggle to meet this criteria: if (mathTemp > 12)

If I manage to make past TSM_auto_cal, I noticed that if the noise_level is greater than 8, I cannot complete the calibration.

If I complete the calibration, the difference between the upper and lower DAC values is between 2-6

Is that the expected values?

Also, the code that I use was generated with FLOWESI GUI, and that used SMCLK=2MHz, since I use SMCLK=8MHz, I changed a few things:

in ESI_TSM_INITPARAMS_DEFAULT structure, I changed ESI_TSM_SMCLK_DIV_1 to ESI_TSM_SMCLK_DIV_4

I added 2 extra lines in esiConfig() function

Hope this helps.

Bálint

UPDATE:

We figured out that we had a faulty capacitor on our custom board. It was responsible for some of the problems.

The device can complete the calibration, detect the rotation and it counts the impulses correctly. However tests showed us, that if the device is constantly counting, after about  an hour it stops counting. The program doesnt freeze, if I move the coils a little bit closer it starts counting again. 

I couldnt figure out what causes this, but I am thinking that the lack of recalibration is responsible for  this. Can someone confirm/deny this?

• set_dac() still completes if noise level is less than 5 (is this an issue or is this intended?)
• SEPARATION_FACTOR is 8
• The difference between DAC1 upper and lower register is between 6 - 10

Any help is greatly appreciated.

Bálint

Can somebody provide me with some information/help?

Hi Bálint,

coulf you request friendship... this would allow to close this case ofline with higher efficiency; the results will be postet here again later

Johann