After a month of dilly-dallying, I'm finally finished the hardware pcb design of my Inertial Navigator.
;======================================================================
Specs:
====
Capability:- 6DOF gyro and force sensing
Dimensions(LxBxH):- 65x45x20mm.
Power:- Lithium Ion Cell Phone battery (replaceable) with temperature sensor and power control.
USB 2.0 capability (still to be tested :bg), with charging
'External' power jack for extra power.
All power points may be connected at the same time.
Endurance:- With 1000mAh battery ~ 50 hours Continuous operation
Longer duration power packs to be investigated.
With USB / External power = unlimited.
Memory:- 1MByte which will be equivalent to 64K waypoints of raw data (uncompressed).
65536 waypoints at 1 second intervals ~ 18 Hours
65536 waypoints at 1 minute intervals ~ 45 days
65536 waypoints at your interval - Device tracks Longitude/Latitude/Altitude all the way up to 'record time'
;======================================================================
My next trick is writing a USB proggy in assembler. I've done a load of USB reading and research but if anyone knows of asm USB example progs for mass storage devices it'll be a great help.
I've done the USB.ORG and some other websites, but and assembler version would be nicer :green2
:U
Main reason for doing this - something to do and having fun :wink
Really... an alternative to GPS, and multi-purpose scalability.
:8)
Interesting.
As I'm not a physicist,...I had to GOOGLE (http://en.wikipedia.org/wiki/Inertial_navigation_system) it,...and, found this YouTube video: Litton LN3-2A Inertial Navigator Platform (http://www.youtube.com/watch?v=ePbr_4yMehs).
...So,...you are also designing your own UFO ???
So are you looking to do PC host side, or device side assembler. If on the device, what processor are you using?
The last mass-storage-device I coded on used an 8051, and most of the SCSISTOR/USBSTOR was in C.
Baltoro:
Nice toy that... a bit big for my pocket, but it'll would be nice to have that follow you around in the mall :bg
Clive:
Both sides, I've worked on the 8051, C and Asm, many yonks ago.
On the device side, it's going to be a mass-storage device (a simplified one :wink) so that the waypoint info is easily accessable from any PC platform.
It would be interesting to see how you've done it. Naturally I will not copy it as I have to put in a lot more funtionality as the device must be able to output single waypoints if connected to a display of some sort. The USB port will be multipurpose - The Host interface/Real world adapter for other types of equipmet will come later.
On the PC side the plan is currently to connect to Google maps, or any other software that can handle waypoints.
Currently 2 modes of operation:
1) Raw data format - The device does not know it's location and just records/sends out sensor data (PC code interprets/corrects this data into Long/Lat...)
2)Longtitude/Latitude/Altitude/Attitude - The device knows it start location (attitude is self determinable) and then stores/sends out Long/Lat/Alt/Att data
I expect when running on battery power it's will consume somewhere between 10-20mA @ 3.3V at low frequency. (datasheet calculations - Physical test will be different))
When USB/external power is connected, the device switches to high frequency and USB 2.0 operation.
I'll post some piccies as progress continues..
Quote from: vanjast
On the device side, it's going to be a mass-storage device (a simplified one :wink) so that the waypoint info is easily accessable from any PC platform.
Well there's going to be a limit on how simple it can get, Windows has some specific expectations about the SCSI commands it expects to be able to send, and the responses it will get. Unless you want to get clever sending SCSI commands directly at the device, the MSD is basically going to a) make Windows happy during initialization, and b) serve up, and write, "sector" data to your storage medium. For this discussion Windows == LINUX, but Microsoft's expectations of how it should behave pretty much ties your hands.
At a guess, you're really going to want to emulate a FAT file system on you flash/eeprom storage space and serve up the sectors to the MSD as requested. When you're logging data you write it too a "file" that you create on this. You can simplify the FAT side of things by say sticking to FAT16, and placing the file(s) in the root directory.
Essentially that's what I've worked out..
I'm not going to implement the whole file system but make it look like there's a file system - 'emulate' if you wish :lol
I've just reduced the battery power consumption to 10mA 'tops', theoretically.This includes leakages, so I'm now hoping for a minimum 100hrs continous operation (~ 4 days)
In reality I think I'll get 72-100 hours depending on battery condition - Lion batteries are fairly consistent over their lifespan
I have a plan to reduce power to approx 5mA, but this requires a lot of tweaking -might just be able to do it.
:U
Here's a of the pcb layout - Made a printout of it.
It's damm smaller than what I thought.. but that's oK
:bg
you should put your name in the silkscreen :U
I try to avoid being sued :P.
Just fooling around on that silkscreen :green2
(http://www.masm32.com/board/index.php?action=dlattach;topic=17035.0;id=9469)
Front side is a CPU, GPS, accelerometer, and on the backside is a Quad Band GSM modem, and two SIM-on-chips.
what do you use for an accelerometer ?
Quote from: dedndave on July 15, 2011, 10:01:31 PM
what do you use for an accelerometer ?
Freescale MMA7455
http://www.digikey.com/us/en/ph/Freescale/MMA7455.html?WT.z_Tab_Cat=Featured%2520Products
Funny how all the accelerometers look similar - there must be industrial espionage :eek.
I'm using that ST- LIS302DLH.
One thing that's not explicit with the VTI, and your Freescale one, is the sensisity..
- They say Count's-per-G, but do not explain Counts (Maybe you can clarify this for me) whereas the ST version just says 1mG (12bit at +-2G) sensistivity - which is far better for me.
:8)
Does anybody know of a suppler/source for surface mount tooling and accessories (paste, glue, etc), as I'll need this to assemble the cards.
I must just slap my reflow (pizza) oven together quickly - had it stacked away in my garage (in pieces) for a year or more. :bg
and then I'm free to go.
you can almost use a box with a light bulb
(http://www.almostgotit.com/wp-content/uploads/2009/05/easy-bake-1.png)
thing is, it wants to follow a rather strict schedule
the closer you get to it, the better the results
The customer was more interested in high G (blunt force trauma) events, than orientation/motion. I did look briefly at free-fall detection with the Freescale demo board.
I think primary customers for the accelerometers are the cell phone companies, and perhaps Nintendo, which might explain the footprints.
vanjast,
My favorite, and most helpful video:
http://www.youtube.com/watch?v=3NN7UGWYmBY
For those who can't solder, and for prototyping:
http://www.schmartboard.com/index.asp?page=products_schmart
hth,
farrier
a little additional info...
the schedule or "profile" for curing solder paste goes something like this...
(http://www.maxim-ic.com/images/appnotes/3377/3377Fig06.gif)
what makes it fun is that, for different runs, you may have a different total mass in the oven
of course, it takes longer for larger masses to achieve temperature (hot or cold)
in your case, that is working to your advantage, as you have low mass, and probably a single unit (on the first run)
you want to be sure to leave any melting plastic parts off until after reflow
the balance is:
don't overheat the parts, or keep them at peak temp too long
you have to be hot enough, long enough for the paste to flow
the pre-heat and soak times help get the paste ready to flow without damaging components
the temperature of the board itself is typically monitored - not ambient
Quote from: dedndave on July 16, 2011, 05:06:02 PM
..the temperature of the board itself is typically monitored - not ambient
Yeah! I saw that... nice to catch these little titbits.
I'll get hold of one those tagged sensors that I can attach to the board.
A laser sensor cold be very useful here.
:U
edit: Just figured out that I'll use a PT100 with a CrocClip to clamp it to the pcb - simple
Been reading/remembering what I studied yonks ago.
I'm using the ST gyro chip with the accelerometer one. After a bit of 'catching up' I figured out that it's possible to implement a 'rate gyro' with only the accelerometer + some maths hashing. The micro sucks about 3mA at 1Mhz, the gyro sucks 6mA standard, and the accello=0.5mA. i can increase the micro frequency to handle the maths (+2mA) and shut down the Gyro in code (~uA)... I actually think the rate gyro would be more accurate :bg I'll keep the normal gyro, as it would be more usefull with regard to aeronautics.
I can save about 30-40% suction...
Ahhh .. another glass of red for me
:U
i have had occassion where we were running on the cheap - lol
the company i was contracting to did not have the oven, or even a thermocouple
i added a 1N4148 to the board layout (very inexpensive switching diode - upgrade of the old 1N914)
this part would hardly ever be stuffed, of course
but, for making temp measurements, a typical diode junction is very linear
you forward bias it with a few mA, then measure the Vf (forward voltage drop)
i used an op-amp and a couple pots on a little board outside the oven (one pot for offset, one for gain)
once the pots are adjusted, the diode makes a very nice thermometer
i scaled the output so that the DC voltage could be read with a DVM
to make things simple, i set it up so 1.00 VDC from the op-amp was 100 degrees C, 0VDC was 0 degrees C
it gives a good representation of the board temperature, as it is hand-soldered into it
of course, if you leave it at high temp for too long, the diode solder could flow - lol
but, that isn't likely to happen because you don't leave it hot that long and paste flows before regular solder
other companies i have worked for had all the nice equipment, of course
the commercial ovens are way over-priced, if you ask me
but, they will store several profiles and settings
once you have made the first run and stored the profile, anyone can operate it
Did a similar thing with airflow measurements.
For consistent reliabilty, it probably better to stick with 'standard' sensors.
:U