Data Logging Without A Computer

[PVfish]

Hello all,

I will be installing a Classic 150 on a floating dock unsuitable for housing a computer to log data. Is there a way to configure the classic so it exports the .csv files containing the minutely data to an external data logger (something more resistant to being outdoors) via RS-232. I have no experience with MODBUS protocol.

Thanks in advance!

[boB]

Just saw this one although it's a very old post now...

This is not a bad idea.  The USB can do this now but you would need
something like a Rasberi Pi which would work very well.  The USB
outputs CSV data at around twice per second.  Voltage in/out
power out and that kind of stuff.

boB

[RossW]

The USB can do this now but you would need
something like a Rasberi Pi which would work very well.  The USB
outputs CSV data at around twice per second.  Voltage in/out
power out and that kind of stuff.

TomW was doing this, but the rPi was running out of legs. The serial data rate seemed to be too high. While it "worked, mostly" it frequently dropped characters, which made the process far less reliable than he was prepared to accept. Ultimately he's ditched the USB and doing it all via ethernet with far better results.

If I'm mistaken in my memories I'm sure he'll chip in and correct me!

[TomW]

TomW was doing this, but the rPi was running out of legs. The serial data rate seemed to be too high. While it "worked, mostly" it frequently dropped characters, which made the process far less reliable than he was prepared to accept. Ultimately he's ditched the USB and doing it all via ethernet with far better results.

If I'm mistaken in my memories I'm sure he'll chip in and correct me!

Absolutely correct.

My theory was that the Pi was being swamped by the data as it uses the same resources for USB and ethernet and logging from USB and storing on a USB stick then accessing the stored data via ethernet was just too much.  Once I started getting it from ethernet, I never looked back.

Tomw

[boB]

Too fast ?  Really ?  There must be some  good reason because for USB this
should be a breeze !  Windows doesn't have any problem keeping up and writing
to disc for the 50 or 60 ASCII characters twice per second.  It might
be the way the rPi does its USB CDC implementation and is some kind
of optimization thing.

Oh well...  Whatever works.

Anyway, you can't communicate with the Classic without using ~some~ kind
of computer or processor unless you want to wire some discrete logic up,
and then you'd still need a computer to read the data unless you print it
out and even then the printer will most likely have a processor in it
unless you have a real old printer with wired logic and I'm not sure if
there was much to be had in the way of a printer without a processor
of some sort in it.

How was THAT for a long run-on sentence ?

boB

[RossW]

Too fast ?  Really ?  There must be some  good reason

Not everyone builds stuff with your care or attention to detail, Bob.
I've seen lots of devices suffer this problem even with surprisingly low datarates.

The usual symptom is that it appears to work fine. But now and then it drops a character or two.
So lines with several numbers  (eg 13 03 15 485, meaning 13:03:15  48.5V) might drop a space and be recorded as  1303 15 485 - which breaks the time and parsing by fields you suddenly have no volts at all, or parsing by character positions you end up with 8.5V (because the 4 has shifted out of position), or you drop the CR/LF and get two lines stuck together... it's horrible.

and then you'd still need a computer to read the data unless you print it
out and even then the printer will most likely have a processor in it
unless you have a real old printer with wired logic and I'm not sure if
there was much to be had in the way of a printer without a processor
of some sort in it.

How about using these: (I've still got some!)

[TomW]

Too fast ?  Really ?  There must be some  good reason because for USB this
should be a breeze

boB;

I think the issue is that the serial data is just dumped and the rPi uses some shared hardware for USB and ethernet so, unlike http or storage device activity there is no error checking / retransmition if errors occur?

Not sure but that is a theory I have.

Tom