Ultrasonics maybe?

[Clyd3]

Hi,
I'm interested in constructing a 3d space tracking setup to track 1 object and maybe more later in an area of 1 meter cubed.
I see that the nintendo powerglove used ultrasonics. I guess if I used ultrasonics i'd have the object as the transmitter and use 3 sensors?

Would ultrasonics be the best sollution or can you suggest something more appropriate?

 

[Clyd3]

that's definitely some positive response
guess i'm on my own then - that sucks

 

[Nigel Goodwin]

You give up too easily! - someone might respond, but you're asking a VERY difficult question!.

 

[dknguyen]

I think 3 IR/ultrasonic beacons distributed around the room would be much easier and more accurate. The target or whatever measures the time difference between an IR and ultrasonic pulse simultaneously transmitted from a beacon. IR light travels fast enough over such a short distance relative to sound that you can assume it to instantly reach the target to simplify calculations. With this 0 propogation time assumption, you can essentially treat the arrival of the IR pulse as the marker for when the ultrasonic pulse was transmitted and then wait to see how long it takes for the ultrasonic pulse to arrive. From that travel time of the sound (really the time difference between the arrival of the IR pulse and ultrasonic pulse) you can calculate how far away the beacon is from each beacon...you might need 4 beacons for 3D tracking rather than just 3 (for 2D tracking). Then you could wirelessly transmit back whatever distance the target-mounted receiver calculated.

Unless you want a single-perspective object tracker, in which case you are into the realm of missiles defence.

 

[hjames]

Some specifications that you might want to narrow down:
resolution, accuracy, update rate, dof tracked (single point or rigid body orientations)

If you haven't already looked around, look at the commercial systems available to get an idea of the technologies available. The other "easy" to build possibilities are magnetic(doesn't like metal), and optical (get some webcams and some software, but likely to have lousy resolution).

Also, how orientation sensitive does the project need to be? I assume ultrasonic transceivers aren't omnidirectional, so you might have some issues there - might require multiple (independent) ultrasonic transmitters on the tracked object.

BTW, entry level commercial systems (magnetic is probably the best bet) start at a couple thousand US$

 

[Clyd3]

i was looking at omnidirectional ultrasonics - the only ones my short googling produced were very large and were for submerging!
So i guess i'll put more than one in / on the moving body, but these generally have a narrow beam (like 9 degrees) so i'll have to look for slightly wider beams, which are available, and see if I could mess with the casing or something.
They do have longer ranges than one metre, so if i were to widen their angles a bit, it shouldn't be a huge problem.

Right now the way i'm planning to go about this is to do it all relatively. I don't need measurements, but rather relative distances. With a bit of math I could obviously turn this into measurements, but that's not what i need.

The speed of the system can be slow / jerky at first, but once it's working, i'm sure it would be easy to modify it to smooth it out.
40khz and 20khz transducers are readily available for the equivalent of about 1 pound. There are others, but they are too large.

I did a search at ti.com and microchip using the keyword ultrasonics, and they both produced very similar pdf docs with application notes for integrating a transducer and two controllers. One, connected to the transducer, sends it a signal, and waits for an echo response, then sends a signal to the mcu, which processes the time taken, and gives the other controller a command to repeat the process.
Their system works on echoing, but if my transmitter object sends a continuous broken signal (beep, beep, beep, beep...), i could make comparisons using that and the time when each of the 3 or 4 (i like 4 - thanks) receivers receive the signal. What u think?

 

[Tarsil]

How about using acceleration sensors? Might simplify the poblem a bit.

 

[Clyd3]

what are they - guess i'll check them out - report back later

 

[Clyd3]

OK NOW I FEEL LIKE A COMPLETE CHIMP!!!
I knew what accelerometers were - never thought of using one -
Nice no interference
i'M Now just looking for a cheap asian source or a sample or something...
need a nice 3 axis one i guess - looking more into it

https://www.dimensionengineering.com/DE-ACCM3D.htm
nice - but a bit expensive - certainly what i'm looking for... very tempting, but i'll keep looking

 

[hjames]

Incidentally the two bigger manufacturers that I know of are:
Analog Devices (Their part is probably the one on that board you posted)
FreeScale

 

[Clyd3]

On that note, how in this world does one deal with the LFCSP package??? Do I need lots of fancy big equipment $ $ $ or skill
**broken link removed**

 

[hjames]

On that note, how in this world does one deal with the LFCSP package??? Do I need lots of fancy big equipment $ $ $ or skill
**broken link removed**

1) A big hole in the bottom of the board
2) just reflow them using a toaster oven or paint stripper. (Skill required though)

 

[Clyd3]

the risk of destroying it sounds huge - i'm gonna look for a socket...
reflow? looking that up
ok, that looks like i need an accurate temperature probe, among other things - hard as hard - but it's always nice to be able to do that

 

[hjames]

Parts are really robust - they are built so that they can withstand a good fraction of a minute floating on molten solder.

1) It's the right temperature when the solder melts, so if you've got a pair of tweezers, poke the part until it moves
2) The first thing to go is probably the pcboard - you'll see it discolor, warp, and start smoking if you get it too hot
3) They make specific temperature sensitive "crayons" which melt at the proper temperature
4) Or you can get a (relatively expensive) IR thermometer and not bother with a physical sensor.

It's a skill that's relatively easy to gain - get some useless pc boards (for example - useless/broken computer or cellphone parts) and practice on them.

 

[Tarsil]

OK NOW I FEEL LIKE A COMPLETE CHIMP!!!
I knew what accelerometers were - never thought of using one -
Nice no interference
i'M Now just looking for a cheap asian source or a sample or something...
need a nice 3 axis one i guess - looking more into it

https://www.dimensionengineering.com/DE-ACCM3D.htm
nice - but a bit expensive - certainly what i'm looking for... very tempting, but i'll keep looking

I've scavanged one from a helicopter model. I've bought that from a flea market for 10$. The accelmeter was encased in a hard epoxi? confused: ) capsule with the pins stiking out. It was v simple to desolder and reuse. This was a model made prob in the 80's. Newer ones ...i don't know.
In the model u also have RF modules, servos rtc.

 

[Clyd3]

thanks - i'm gonna have to, aren't i
but i found a source - right here in Durban, South Africa - HOT!!!

 

[Clyd3]

i wouldn't mind a little helicopter hehe

i have - just got - this heat gun - not for electronics - i wanna try it with heat shrink though!!!
It has a huge nozzle, and it says its 300 - 560 degrees celcius
i safely heated up my foot on the lower settings from abt a meter away, so it should be alright to use, if not exposing the board to it for too long, what u think?

I'm gonna turn up a smaller nozzle for it, but i don't want it to have too much pressure or feedback and burn itself out or something - twisted
How big a nozzle should i aim for???

Or should i return to the oven method???
i tried a stove out - but i wasnt in control - i destroyed an isa graphics card with the stove
at least with the heat gun i could choose what quater of the board (next isa card) to heat - more than 10 seconds and it fries the board itself- i use the lower setting - i'm guessing around 350 right close up

i'm also considering that if i make a nice small nozzle - too small causing feedbank - i'll just have a vent to relieve some of the pressure - i just need some accuracy i think

 

[hjames]

Stove? Keep this stuff away from anything that you might ever use for food... Most of the time people use a dedicated toaster-oven - something cheap and mostly disposable. Lead isn't exactly healthy.

When I solder stuff with a heat gun(random 1200W handheld one, ~3cm nozzle) and solder paste, I usually have it on the low setting and keep it ~10-15cm away from the board. It usually takes a couple dozen seconds for the solder to reflow. I usually throw a piece of solder on the board so I can see roughly when the solder should melt. When I need more accuracy I switch to another piece of equipment, so I haven't tried making any sort of nozzles.

 

[Clyd3]

thanks - that was some very handy info
i'm going to have to try things - its the only way