Jump to content
robmitch

DIY HT Noise/measurement issues

Recommended Posts

I have been working on my own gyro/acc/compass hybrid HT. I am using a 2-Axis accelerometer and single axis gyro, and a 2-axis compass sensor. These are all loaded on my custom pcb. I am supplying 12v power, gnd, PPM in and PPM out to the board via a 4' multicore shielded cable from my JR transmitter module. At the moment, however, I am using a Zilog mcu on a Zilog development board for processing, and it is connected to my custom pcb using a 8" ribbon cable (12 wires to the jumpers on the development board). I have been able to monitor the incoming PPM signal, modify the last two channels with either tilt or roll from the accelerometer output, and either pan using integrated yaw rate or compass heading. The aim at the end of the day is to use a kalman filter to combine yaw rate and compass heading to a pan angle. However, there are two issues with the yaw rate output that I need to resolve:

1) At startup, the zero rate reading starts at a high value and over a period of about a minute, it asymptotically settles to its zero rate value, which remains relatively stable from then on (If I supply power to the board via 12v wall-wort). Could this be due to a temperature transient at startup (the gyro should be fairly independent of temperature - in datasheet), or perhaps the voltage regulator 3.3v LDO linear on startup (the gyro output is ratiometric with input voltage)???

2) If I supply the power to the setup from my tx (36MHz), I get the additional problem of small random variation in gyro output with tx proximity (even my hand coming within a couple of inches of the sensor pcb changes the output). The tx will be within a couple of feet of the unit at the end of the day, so I need to sort that out. What sort of power filtering should I have to the circuit - LC filter? Do I shield the circuit with a metal housing? Do any of the commercial HT's have any of that... I saw Melih's circuit board for his - it doesnt seem to have any filtering components at all?!?

I have attached the STMicro gyro datasheet. I am using the suggested circuit on page 8 (4x rate output pin goes straight to ADC of mcu. From the gyro datasheet, this output is a high impendence buffered output). There is a filter cap on the input pin of the mcu development board too.

Any issues to be seen so far?

Cheers,

Rob

ly530alh small.pdf

Share this post


Link to post
Share on other sites

My best guesses:

(1)

The Gyro's High Pass Filter Reset (HP) feature should be used after power is applied.

(2)

To attenuate any conducted RF, I would start with a common mode Toroid filter on the interconnect cable, placed near the HT board.

If radiated noise is causing grief then your PCB layout may need some attention. If your prototype is a rat's nest of wires then going to a clean board layout may be the holy grail.

Share this post


Link to post
Share on other sites

My best guesses:

(1)

The Gyro's High Pass Filter Reset (HP) feature should be used after power is applied.

(2)

To attenuate any conducted RF, I would start with a common mode Toroid filter on the interconnect cable, placed near the HT board.

If radiated noise is causing grief then your PCB layout may need some attention. If your prototype is a rat's nest of wires then going to a clean board layout may be the holy grail.

Thomas,

1) I thought simply grounding the 3 control pins of the gyro would save me time :lol: I will have to scrape some track off the HP line and try that. Good idea.

2) (Conducted noise) Toroid like the ones we use on servo cables? Should I also have any additional caps where power is supplied to the board (I already have the recommended caps on input and outputs to the regulators)?

Cheers,

Rob

Share this post


Link to post
Share on other sites
2) (Conducted noise) Toroid like the ones we use on servo cables?

Yes. All the wires that interface with the R/C Tx would be wrapped through a Toroid donut. As many tight wraps as possible.

Should I also have any additional caps where power is supplied to the board

I suggest LC filters on all of the wires that connect to the R/C Tx. If you need help on this then open up a Futaba Tx and review how they LC'd their trainer port. Do the same sort of thing on your Gyro Interface's design.

Share this post


Link to post
Share on other sites

One other thing..........wall warts can give you problems here. Try using a battery for power supply to see if it helps, much smoother!

Share this post


Link to post
Share on other sites

OK,

used HP filter reset at startup and it improved the transient from about 30s to 2s. However, this was using a third option for power supply - the regulated supply from the mcu development board... It has a larger 3.3V regulator and doesnt get very warm at all. The regulator on my custom pcb gets quite hot, supplying both the sensors and the development board (I measured 120mA, and my little regulator is rated for 800mA). The startup transient is still better using the HP reset (large zero rate values have gone) with my regulated supply, but the startup transient still takes some time to settle to a steady value (I am sure it is the reg output voltage is changing ever so slightly due to temperature changes on startup). The development board has a few peripherals that might be eating all that current, but maybe I am sinking (sourcing?) current from the sensor outputs into the adc of the mcu as well? Do I need current limiting resistors between the sensors and adc? Values?

WRT LC's on lines, I dont have a Futaba radio :( What sort of LC values should I aim at?

Share this post


Link to post
Share on other sites

You probably should speak with the gyro IC app engineer to determine what is acceptable for calibration times. Honestly, if it is ready to use within 3-5 seconds then I would say that it is on par with a lot of hobby gyros.

The start up drift can be from an endless number of things. You'll just have to spend the time checking off the list and hopefully you find ways to improve it.

For example, if you are using the internal ADC Vreg feature built into your microcontroller, then it might be wise to reconfigure it to use external Vref from a accurate (1%) source. Keep in mind that the Gyro IC is an analog component and it should not be sharing the Gnd and Power buss from your digital stuff. Typical mixed analog/digital PCB layout practices are required.

Do I need current limiting resistors between the sensors and adc?

Honestly, what you have sitting in front of you is just a mysterious black box to us out here. So, specific answers to these sort of things are best answered by your data books and host board manuals. But as a general comment, if your ADC circuitry is like the usual MCU configuration, current limiting is not used. If anything, you want low'ish impedance connections to the analog inputs to ensure the signal can adequately drive the ADC's input (due to its input capacitance). But, your design may have special needs.

I dont have a Futaba radio :( What sort of LC values should I aim at?

I've got a Futaba Tx in a box somewhere, but I don't have any free time to tear it apart and see what's in there. So, my gut feeling is that the LC's should be around 4.7uH and 1000pF. These are common RFI filtering values for VHF RF apps.

Share this post


Link to post
Share on other sites

Thanks,

By the sounds of it, I am lucky anything is working at all. I will do some reading on mixed pcb layout etc. Thanks for the rough LC values. If that Futaba radio you have sitting around is a 9C and you want to sell it cheap - look no further ;)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×