[openmulticopter] OMC mini - a few notes

[Heinrich Warmers]

Hello
i think the  temperature drift is not real a problem, since the
accelerator has only a small.
If we use the filter from Antonie  or our Kalman filter the gyro bias
and the drift are corrected. Only the yaw is a problem but this will
corrected by the magnet sensor like the HMC5843 Pleas look for the
master work from the ETH with the IDX500 and the ADX610. The price of
the IDX is low. They had verry good results for flying helicopters.
http://www.sky-sailor.ethz.ch/docs/Development_of_a_Planar_Low_Cost_Inertial_Measurement_Unit_for_UAVs_and_MAVs_(Fux2008).pdf
Reagads
Heinirch

2009/11/29 Crossfire <crossfire at gmx.ch>:
> I was quite busy lately and didn't have the time to contribute to the project. Still, I follow your discussions and the development progress in the wiki.
>
> Having said this, I would like to write down a few notes which may be useful to you.
>
>
>
> Let's start with the gyroscopes. When we were selecting the components for the OMC mini there were several candidates for the gyroscopes. Ideally, they would be operated from 3.3V and there would be different types of the same model which solely differed in the axis being measured (roll, pitch, yaw). The latter feature would allow mounting the gyroscopes entirely horizontal. This makes the IMU smaller and saves costs for the breakout boards.
>
> There are two models from different manufactures that offer those features. On the one hand it's the new gyroscope product line from ST, on the other hand there are the IDG-* gyros from Invensense.
>
> ST's first gyroscope LISY300AL turned out to be unusable in other projects due to its low excitation frequency in the MEMS structures which caused intense resonance effects combined with the BLDC motors and controllers used in multicopter projects.
> When I suggested looking into ST's line of products I always expressed my concerns that the new gyroscopes could suffer from the same problems as their predecessor. For this reason we wrote an email to ST asking them about the MEMS frequency of their new gyroscopes. It took them quite a few weeks to write back but we eventually got our not so encouraging answer: 4.1kHz. This means the ST gyroscopes are disqualified for the use on a multicopter. They might be used in a head-tracker for FPV flights (or some other robot projects), though.
>
> What about the Invensense sensors?
> The IDG-300 is famous for its incredible drift over temperature which is even worse than the drift generally found in piezo gyros. Those of you who had a X3D or 1. generation Mikrokopter might remember the effects when flying on hot summer or cold winter days. Although some crazy compensation routines allowed using those piezos no-one ever really found the IDG-300 to be useful in their projects.
> This leaves the concern that the new IDG-500 and similar models share the same design with the IDG-300 and features like the "zero-point calibration" don't awake much confidence. It more or less seems that they took the same design and put a few half-hearted countermeasures in place. Unfortunately I failed to extract any useful information from the datasheet. There's some piece of information about zero-drift but they do not specify it in °/s/°C. Please write a mail to the mailing list if you have some answers to this question.
>
> After all, I think that we should stick to the gyroscopes that are known to be reliable and are commonly used in other projects as well, since better MEMS gyroscopes fail to exist at this time. This leaves us with the well known ADXRS610, which are hard to solder and quite expensive but really accurate, and the MLX90609, which provide the best mix of price, quality and ease of use. The downsides of this decision would be the necessity to use break-out boards to mount the roll and pitch gyros and the higher complexity in the analog part due to the 5V operating voltage.
>
>
>
> The second thing I would like to write a few lines about is the switching regulator. Someone suggested that we should use a switching regulator with a high frequency in the range of 500kHz to 1MHz in order to reduce the impact on the analog part of the IMU. The idea behind this is that a very high switching frequency will fail to have a negative impact on the sensors and ADC when it is just too high for those parts to follow the noise. In order to prevent an alias effect on the signal after the analog-to-digital conversion there would of course be some analog low-pass filter in the analog circuitry before the ADC.
> While all of this might work just as expected (though I never tried it and would have to rely on the experience and knowledge of those who proposed this idea) I fear that using such a switching regulator might cause severe problems in the RC part of our multicopters. I heard some stories about a line of motor controllers for RC planes and helicopters which caused interference with the generally used 35MHz RC equipment and resulted in pilots losing control over their planes.
> Although new 2.4 GHz receivers might be too far off the regulator's frequency and therefore not show any of those problems we still have to consider that many people are using 35MHz RC equipment.
> If you have a good answer to this question, please write a follow-up to the mailing list.
>
>
>
> The last thing I would like to suggest is to separate the circuit into two PCBs if it turns out to be too difficult to put everything on a single one. One PCB could contain the STM32 microcontroller, the voltage regulator, connectors, SD card slot and so on. This PCB should be designed really flat which means that the second PCB could be mounted only 5mm above it (shortest available distance bolts). The second PCB would consequently contain all the sensors and analog parts.
>
>
>
> Please let me know if you have any questions, objections or counter- proposals.
>
>
> --
> Patrick
> _______________________________________________
> openmulticopter mailing list
> openmulticopter at openmulticopter.org
> https://openmulticopter.org/cgi-bin/mailman/listinfo/openmulticopter
>