|
Message
Tracy,
Just a couple of questions/comments...
How will the data transfer be triggered? The most
obvious would be a command sent by the PC, but you could have a physical push
button send on the EM2 if your end is not bi-directional.
I assume when you are talking about markers like $F9, you
are not talking about the ASCII characters but Hex... but then word length is 16
bits. I may be missing something there. I guess it doesn't really
matter since I would read binary, and just keep binary format for each marker...
or look for fixed length returns for each channel if the individual values can
get high enough to encroach on the markers.
Flow control isn't a problem since even the most basic PC
has much more memory available to buffer the transfer than most Integrated
Control chip memory allocation. The latest real time logging I have
been working on can file append 10,000 lines (80 ASCII characters
a line) a minute clocking less than 10% CPU time. An application
should be able to write as fast as you can send, or just save in memory to be
written after transfer.
I'm not a huge fan of Excel. I would probably
graph out each channel in the application display with some form
of movable correlation line marking the identical time position in each graph
with text fields showing the text data for a given time position. I
already do something similar to that with factory Schedule/Demand
monitors. If users want to view/manipulate the data in their own
way, a dump to a .CSV file that any spreadsheet program can import is fairly
simple.
I work for a manufacturing company, so this is the sort of
programming I do. Computer integrated manufacturing with single point
control / communication / data display with PLC's, scales, scanners,
mainframe computers, etc.
If Joe decides against working on this, let me know and
I'll take a shot at it.
-Tim
----- Original Message -----
Sent: Wednesday, April 19, 2006 7:36
PM
Subject: [FlyRotary] Re: Possible
data-logging solution?
Thanks Tim,
Joe Hull is a pretty good programmer and said he would take a look at it
but don't have any commitments at this point. Any and all help would be
appreciated on either part of this project. I'm simply out of time to do
it all myself. Any contributor will get full credit for their efforts
and if this stuff ever makes any profit, I'll count you in for
royalties! (hint, it won't be like winning the Power Ball Lotto :
)
For anyone interested, Boring technical details of the task follow
below: All others may delete message now.
Data logger protocol
This protocol is still being worked out so if you have any suggestions to
make it easier of more reliable, fire away.
The first part needed is a PC Program to input a stream
of data words on a serial port from the data logger and
store them as a disk file. The words will represent multiple blocks of
28 14 bit values. Each 14 bit word in the block is one
of the 28 data channels in the EM2 engine monitor. There is one 28
word block for each 2 second sample period. Definition and order of data
words still being developed.
The serial data file stream is preceded by a
file start marker byte of $F9. The data words are then sent as two
bytes having the msb (bit 7) cleared and the lower 7 bits holding the
data. The least significant 7 bits of the 14 bit word are sent
first. At the end of each 28 word data block there will be an end of
block marker of $FA.
The end of the data file will be signified by an end of file
marker byte of $FF.
The async serial data will come from the data logger at a baud-rate of
19,200. Currently there are no flow control lines (CTS / RTS) on
this serial interface but these could be added if necessary.
The second half will be an EXCEL spreadsheet that displays the data file
as 28 channels in chart form with time being the X axis (2 seconds per data
block) and the value of each channel plotted on the Y axis. Don't
know if it is practical to display all 28 channels on one chart (would be
nice, in order to see unexpected correlations) or to have separate
charts for various groups.
Each channel will be associated with a different formula to convert the
raw sensor data word to human readable units (degrees F, PSI,
etc.) These formulas are already developed.
I have only a little EXCEL experience so I don't know if
I'm expecting too much from it or not. Your input welcome
here.
Tracy
Tracy,
If you don't have any volunteers by now, get me the
details and I'll take a look at it.
-Tim
----- Original Message -----
Sent: Monday, April 17, 2006 10:14
AM
Subject: [FlyRotary] Re: Possible
data-logging solution?
I'll check this out but in the mean time I added a simple data
logger to the fly-by-wire board since it only required the addition of 1
inexpensive part. Data logging time is limited to 74 minutes
at a sample rate of once every 2 seconds. This should be
plenty for analyzing cooling & performance issues.
It ain't over yet since this only stores the raw sensor
data. As Rusty points out, I (or a volunteer with some
programming/EXCEL chops?) need to write the program to download the data
and an EXCEL spreadsheet to display it in a nice
chart.
BTW, Congrats Buly! Now, give us more details on
that needed cooling tune-up : )
Tracy (did I mention I was looking for volunteer
programmers?)
Subject: [FlyRotary] Re: Possible data-logging solution?
Interesting
Phil. I just bought a serial data logger for my Dynon DS180 (EFIS
and EMS). I just received the Dynon, and data logger, so I
haven't had time to try anything with them, but this setup has
promise for Tracy's stuff as
well.
At the moment, the
unit only logs one of the two serial channels, but in the next
version of software, it will log both channels. It uses SD or
MMC memory cards, so you can take them home to read them. The OEM
version has this card slot available on the bottom of the
board, where the boxed version is enclosed. I'm told you can use
any branded memory, so you don't have to buy theirs. I got the 128
mb version, since they seem to screw you on the memory cost,
and I can buy larger cards much cheaper from other
vendors.
They don't use the
FAT file system, so you need their "reader" program to read the memory
card on a PC, or you can use a laptop and download the data via
X-modem. Once I get the data, the next problem will be figuring
out what to do with it. I guess I'll be learning more about Excel
than I cared to know.
Cheers,
Rusty (one week
until my "sentence" in Cleveland starts)
|