Hi all,

I have an Bela>Accelerometer-related enquiry.

I'm using the LilyPad (Sparkfun) ADXL335. It is encased in conformal coating, insulating tape, and many layers of PlastiDip rubber sealant.

It's wired to the Bela via a 2m, 5-core, self-built cable, with 5 individual insulated wires encased in a 2m metal cable sheath, which is in turn wrapped in insulating tape.

None of the wires connected to the Bela are loose or faulty; they're all perfectly conductive and firmly connected.

The accelerometer is wired up exactly as per your example diagram. Up until recently, it has been responding perfectly well to all types of movement.

On my most recent tests, however, I've found two very strange things are happening on Axis 1 (wired into [dac~ 27], when I observe it through the oscilloscope.

First of all, it's oscillating a little more widely than the other axes (the others oscillate pretty imperceptibly, but Axis 1 oscillates continuously within a range of about 0.2).

Secondly - it doesn't respond when the accelerometer is moved, holding the accelerometer in my hand. HOWEVER - and this is the really strange thing - it responds dramatically to the smallest movements of the shoes on my feet, at least a meter away.

My feet/shoes are not making contact with the accelerometer (or anything conductive) in any way. The accelerometer remains totally still in my hand, wrapped in insulation. Its connecting cable is also held totally still; so is the Bela, my computer, and everything else in the room.

I've narrowed it down to the rubber sole in particular - and in particular when I'm actually wearing it - that influences the reading so dramatically, though only when it's at a distance between 1-2m away from the accelerometer.

More mysteriously, Axis 1 only responds to accelerometer movements if it is aggressively 'jolted', but its response to my shoe is dramatic, and perfectly aligned with its smallest movements (across the axis' entire range).

I've contacted Sparkfun for support, but they suggested that the accelerometer is simply broken, which doesn't seem to answer the question at all. Axis 1 is still very responsive - it's just responsive to my shoe, instead of being responsive to the accelerometer's own movement!

I'm at my wits end - I wouldn't have brought this to the Bela forum, but I know it's full of a lot of smart people, and that maybe the Bela has something to do with this. In researching accelerometers, I can't find anything that would suggest they would respond so strongly to a nearby rubber object.

Can you help? Thank you!

More info - the Bela is connected via USB to my MacBook. When I disconnect the power adapter from the laptop, the Axis1 oscillates much less widely (but still more widely than the other axes). The other axes aren't affected at all by the power supply.

The Axis1 sometimes goes back to almost-normal, responding to the orientation of the (heavily insulated) accelerometer.

Also: I have the Bela mounted in a waterproof plastic box, inside of another case (a plastic, waterproof Peli case). When I move the lid of the Peli case, the Axis1 position moves up and down. And when I touch the inner plastic box, the Axis1 position changes slightly, and it oscillates a little more widely.

I'm still struggling to understand, though, why the Axis1 is so responsive to my shoes/feet (which the accelerometer, Bela etc make no contact with) - especially when they are on the floor, but also when they're far off of the floor. This applies whether or not I'm holding the accelerometer, although it's more extreme when I hold it.

I'm sure this comes down to electronics 101, but it's confusing me to no end!

Thanks again 🙂

hard to say what is going on, sounds like vodoo from here 🙂

here is what i would suggest:

-take the bela out of its cases, see if the problem persists.

-how did you measure that the cable is alright? did you just check for continuity from input to output pin or did you also check if there is a short between the different cables?

Thanks for the reply Lokki!

I'm eager not to remove the Bela from either of its cases - it's (more or less) been permanently installed into both of them. I could remove it, but it would cost quite a bit of time/hassle and a little money to put it back in, as I've carefully built the setup to be permanently watertight.

I have two bits of evidence that the wiring is connected properly:

a) all of the intermediate wiring between the 5-core cable and the Bela has been tested as conductive, and has been firmly attached to the board (and the 5-core cable is also firmly connected to this wiring), and

b) all axes of the accelerometer are clearly very responsive, even Axis 1 - except Axis 1 seems to be extremely responsive to my shoes (while worn), and high-velocity jolts, but not orientation (most of the time).

It's responsive to the shoes even when the accelerometer is a great distance from the Bela, so I know it's the accelerometer sensing the shoes, not some energy field around the Bela itself.

It looks like voodoo from this side too! I was finding it hard to believe the readings.

Can anyone offer a possible explanation?

Hi all,

Just to sum up, here is the issue:

While connected to the Bela, one axis on an ADXL335 accelerometer is responding dramatically to movements around the room that should have no influence on it (slow movements, not causing sound or vibration, making no contact with the accelerometer, the Bela, or anything conductive).

The axis responds especially to movement near the Bela itself, or near the accelerometer itself. However, it also responds to (quiet, non-vibrating, non-conductive) movements up to 2m away.

The accelerometer's axes 2/3 respond normally.

Axis 1 always responds to (intense) dynamic acceleration (jolts), but rarely responds to static acceleration (orientation), although sometimes it inexplicably becomes more responsive to orientation.

The accelerometer is heavily insulated and securely connected to the Bela, which is fixed within 2 waterproof plastic cases inside each other. The Bela is connected to a MacBook via USB, and I'm taking readings off of the IDE's oscilloscope.

THE SHOES:

Axis 1 is very sensitive to the movements of my shoes, which make no contact with the sensor, the Bela, the computer, or anything conductive.

This happens when the shoes are within a range of 0.5 - 2m from the accelerometer (and not making any contact with either the accelerometer or the Bela), and provokes a dramatic response from Axis 1.

Here is some of what I observed:

  • axis 1 responds to even small movements of my shoes, especially the lifting of my toes from the floor

  • smaller pieces of rubber (or any other objects, incl large metal objects) have no influence on the axis whatsoever, although other (less thickly-soled) pairs of shoes provoke a less dramatic response

  • response to shoes is much more dramatic when macbook power cable is disconnected

  • distance of shoes under accelerometer makes no difference to the readings - it responds equally dramatically even if shoes are 1m or 2m beneath it

  • distance of shoes horizontally from the accelerometer weakens the response of the axis

  • axis also responds to movement of feet when they are far off the floor (1 metre)

  • response is much more dramatic when i am holding the accelerometer in my hand (although it is held perfectly still, is heavily insulated, and makes no contact with my feet)

  • response is much more dramatic when i am actually wearing the shoes (instead of moving them around independently)

  • still responds to my feet/shoes when neither of them are touching the ground

OTHER MOVEMENTS THAT INFLUENCE THE READING:

  • inserting power cable into mac (makes an enormous difference to the reading: the Axis1 oscillates maybe 4 or 5 times more widely)

  • removing power cable from mac (doing this, the axis at one point stopped oscillating almost completely, but, when power is not plugged in, the axis becomes much more responsive to the movements of my shoes)

  • moving the flight case lid on its hinges (the axis reading moves up and down, but it doesn't affect the amount of oscillation)

  • bringing hand near bela box (makes axis oscillate more widely, and move up a little)

  • putting hand on the bela box (makes axis oscillate especially widely)

  • removing hand from bela box (makes the axis reading drop suddenly)

  • placing laptop on top of bela box (axis oscillates less)

  • touching either end of the audio cable going into the Bela (makes the axis oscillate much faster and less widely)

  • when moving my body, objects in the room, or the accelerometer, the axis value sometimes makes sudden jumps to a new position

Hopefully this information helps reveal what is happening! Thank you all again

Hi @papa_sangre

It sounds to me like you either have a loose connection between Bela and the accelerometer, probably near or at the accelerometer, or the accelerometer itself is broken leaving that connection floating. Bela's analog inputs have extremely high impedance, which is generally useful in a lot of ways, but when unconnected they react quite dramatically to charges in their vicinity - try scoping the unconnected analog inputs whilst poking your fingers near the input pins, both with and without your hand grounded on a laptop or similar.

If the accelerometer itself or your connection to it is broken, your cable will be floating, and if it is two metres long it will make a fantastic antenna, picking up electromagnetic fluctuations in the air. Do the oscillations on that channel happen to be around 50/60Hz (or double that at 100/120Hz)? You can check using the FFT mode of the Bela scope. If so it is probably picking up signals from the mains wiring in your room. The reaction to your rubber shoes is probably something to do with static build up.

So I would double-check the wiring around the accelerometer, and if you can't see anything wrong try replacing it, or at least looking at the outputs on a standalone oscilloscope to check it's working.

Thanks so much Liam, that sounds about right.

Apologies for the length of the posts - I have a performance with the instrument upcoming, and I couldn't make heads or tails of all of the seemingly disconnected things I was observing.

I'm dreading taking things apart to locate the loose wire, but I'll post back here if I find anything of interest. Thanks again!