Interesting discovery with OpenXVario

[aduncan]

I have put together several varios, but apart from one, all showed unsteady reading for vspd and alt. Alt especially would creep upwards all the time.
Then I put one in an Elf dlg, between the trailing edge and the boom. But it was a tight fit, and the reset button kept pressing itself. So I had an uncommon brainwave. I made the MS5611 module separate from the Arduino board by using about 60mm of wire. The MS5611 was fitted between the servos and the front wing screw, and the Arduino board stayed under the trailing edge.
After this everything worked perfectly. With glider on the ground alt and vspd were steady. After a flight I put the model on the same spot on the ground, and the alt showed zero.
I then modified another vario which was not steady, and got the same results.
Could there be some electrical interference between the 2 boards? Before and after the vario was either covered with black tape or in black heatshrink, so that should have eliminated any effect of light on the sensor.

Edit. Just discovered another interesting thing. I built my varios with guidance from the original webpage. This page showed the output from the receiver going to the Raw pin on the Arduino, and from raw to the bottom left pin on the MS5611 board. When writing this post I forgot the name of the barometer board, so went to the website for a look. I noticed that the new site has the power to the MS5611 going from the VCC pin on the Arduino to the bottom left pin on the MS5611. One of the varios I built has always been unusable, it gave an altitude reading of about 28000 metres, and the vspd varied so much as to be useless. After reading the new webpage I altered the wiring so that the MS5611 was fed from the VCC pin instead of the Raw pin.
Result - the vario now works perfectly!

[mstrens]

Personally I never had issue with unsteady reading for vspd and alt. In my case the ms5611 (gy63 board) is always put directly on top of Arduino pro mini (5 volt) and I use Arduino Vcc pin (and not Raw) to power the MS5611.
An issue that I have is a slow alt drift. If I let Arduino run on a table, it is not uncommon that after 15 minutes, alt is about 5 or 10 m (instead of zero).
I do not expect that this is the result of fast changes in the barometric pressure.
I suppose that this drift is related to an increase of the internal MS5611 temperature when it is read (50 times/sec) by the Arduino.
In recent version of openXsensor I added a parameter in order to better compensate this effect. Still the results I got are not so good.

For unsteady vspd & alt, I could imagine 2 reasons:
- noise on the power supply
- noise on the I2C bus.

When you put MS5611 60 mm apart from Arduino, it could be that the MS5611 voltage regulator provide a more stable voltage (because the condensator at the IN pin of the voltage regulator can filter a little more).
When you use Vcc arduino pin in fact you get an already stable (filtered) voltage to power up the MS5611.

It would be interesting to know the kind of power supply you normally use (type of battery, Voltage of the battery, ESC or DC/DC converter) and if the issues you had (before recabling) occured with different types of power supply.
Normally if you use a 4.8 Volt Nimh battery, I expect that the voltage on the Raw would already be very stable to power the MS5611 and that there should be no issue.

The other reason for unsteady vspd and alt could be that the signals on the I2C bus between Arduino and MS5611 would be to noisy (adding resistors to Vcc could help). Still it is strange that you get a better result when the cables become longer. I would just expect an opposite result.

So I presume that the reason is more related to an "unfiltered" Raw voltage.

[jhsa]

This is interesting. My Alt also drifts with time.
Maybe it is not a matter of cable length but the sensor not being on top of the arduino. Could it be that the arduino regulator warms up and changes the temperature around the barometric sensor? I have a current sensor on one of my planes and the raw connects directly to the lipo (need to change it to connect only to 2 cells instead of 3 or solder a 8V regulator before the arduino ). But I have checked that precise regulator's datasheet and it can handle the 12V. I have noticed though that it gets a bit warm to the touch. With much less voltage on the reg it wouldn't get that warm but it could still change the temperature around the sensor. If connected directly to the VCC pin then the regulator has no effect, but I wonder if the chip itself or any other component get a bit warm and vary the temperature. Could it be that small changes could affect the reading by a few meters?
Some testing is in order as I need another vario but it might take little while as I'm still working on another projects..

João

[mstrens]

Temperature of MS5611 has a significant impact on the values reported by MS5611.
That the reason why this temperature is sent by MS5611 to arduino and is used in combination with a raw pressure in order to calculate the effective pressure (the one being used to calculate the altitude).
MS5611 calibration provides the parameters to be used for this calculation and so in theory, temperature impact should be compensated.
Still I expect that compensation provided by ms5611 calibration is not 100% correct.

I do not think that, in my case, the temperature of the arduino voltage regulator has an impact on the internal temperature of the MS5611 because the 2 components are just on opposite faces (button face - without component - of Arduino is closed to button face of MS5611 and so both component are outside).

It is clear that internal temperature of MS5611 increases after startup - more than 5 degree (values reported to arduino show it).
I think that increase of temperature is just the result of the internal electronic of MS5611 when he has to make and transmit many conversions per sec.

[aduncan]

I think you are correct about the unfiltered voltage. I have now changed 2 of my varios to feed power to the MS5611 board from VCC. In one case it has made the vario much more stable. The second case was a vario which I was never able to use as it fluctuated so wildly. Now it is almost usable, but I suspect the MS5611 board is faulty.

[kalle123]

Which arduinos are you using?

br

[aduncan]

Arduino Pro Mini 5v 16 Mhz

[kalle123]

I think you are correct about the unfiltered voltage. I have now changed 2 of my varios to feed power to the MS5611 board from VCC. In one case it has made the vario much more stable. The second case was a vario which I was never able to use as it fluctuated so wildly. Now it is almost usable, but I suspect the MS5611 board is faulty.

You know this diagram. There is this 5V to 3.3V LDO on the MS5611 board.