This is slightly OT for all the groups I'm posting to, but something many of you should have done. I have a need to sample one channel of data at between 200 and 2000 samples per second, 12 bits or so, and store it away. The signal happens to be on a 4-40mA line already, to make life easy for getting off-the-shelf hardware. I need to store hours of guaranteed-uninterrupted data, or failing that, store data that's timestamped at the measuring device (not just in the PC software) so that I can see what's missing and deal with it. (dropouts of even a half second or so would be OK, as long as the timestamps are correct). The mental model I'm carrying around is a laptop PC, connected to some little box via USB, quietly buzzing away in a corner making multi- megabyte files with the information I want. I assume that this is a no-brainer with National Instruments hardware and software -- am I correct? If not, is there a solution that you can recommend? Do you have a favorite other than NI? TIA. -- My liberal friends think I'm a conservative kook. My conservative friends think I'm a liberal kook. Why am I not happy that they have found common ground? Tim Wescott, Communications, Control, Circuits & Software http://www.wescottdesign.com
PC Data Acquisition
Started by ●March 13, 2013
Reply by ●March 13, 20132013-03-13
On 3/13/2013 11:44 AM, Tim Wescott wrote:> I have a need to sample one channel of data at between 200 and 2000 > samples per second, 12 bits or so, and store it away.For one time deal: take your favorite micro eval board from the shelf; connect it to USB to RS232 dongle; put together simple application; record file from terminal program on PC. All preparation should take less then half an hour. Vladimir Vassilevsky DSP and Mixed Signal Designs www.abvolt.com
Reply by ●March 13, 20132013-03-13
On Wed, 13 Mar 2013 12:05:05 -0500, Vladimir Vassilevsky wrote:> On 3/13/2013 11:44 AM, Tim Wescott wrote: > >> I have a need to sample one channel of data at between 200 and 2000 >> samples per second, 12 bits or so, and store it away. > > For one time deal: take your favorite micro eval board from the shelf; > connect it to USB to RS232 dongle; put together simple application; > record file from terminal program on PC. All preparation should take > less then half an hour.It's for a customer over 1000 miles away whose hand I cannot easily hold. And it needs to deal with 4-40mA, for which I have no hardware. But it's a tempting idea. -- My liberal friends think I'm a conservative kook. My conservative friends think I'm a liberal kook. Why am I not happy that they have found common ground? Tim Wescott, Communications, Control, Circuits & Software http://www.wescottdesign.com
Reply by ●March 13, 20132013-03-13
On Wed, 13 Mar 2013 11:44:58 -0500, Tim Wescott <tim@seemywebsite.com> wrote:>This is slightly OT for all the groups I'm posting to, but something many >of you should have done. > >I have a need to sample one channel of data at between 200 and 2000 >samples per second, 12 bits or so, and store it away. > >The signal happens to be on a 4-40mA line already, to make life easy for >getting off-the-shelf hardware. > >I need to store hours of guaranteed-uninterrupted data, or failing that, >store data that's timestamped at the measuring device (not just in the PC >software) so that I can see what's missing and deal with it. (dropouts >of even a half second or so would be OK, as long as the timestamps are >correct). > >The mental model I'm carrying around is a laptop PC, connected to some >little box via USB, quietly buzzing away in a corner making multi- >megabyte files with the information I want. > >I assume that this is a no-brainer with National Instruments hardware and >software -- am I correct? If not, is there a solution that you can >recommend? Do you have a favorite other than NI?I'd look at Measurement Computing http://www.mccdaq.com/index.aspx as well. They've been around since the Dark Ages (when Computer Shopper magazine gave the mailman a hernia). I keep one of their USB acquisition modules in the toolbag for ad-hoc setups. Easy to interface to custom software (DLL and docs are included) but they do have stand-alone data logger software available. A free version is included with most of their modules.
Reply by ●March 13, 20132013-03-13
Tim Wescott wrote:> This is slightly OT for all the groups I'm posting to, but something many > of you should have done. > > I have a need to sample one channel of data at between 200 and 2000 > samples per second, 12 bits or so, and store it away. > > The signal happens to be on a 4-40mA line already, to make life easy for > getting off-the-shelf hardware. > > I need to store hours of guaranteed-uninterrupted data, or failing that, > store data that's timestamped at the measuring device (not just in the PC > software) so that I can see what's missing and deal with it. (dropouts > of even a half second or so would be OK, as long as the timestamps are > correct). > > The mental model I'm carrying around is a laptop PC, connected to some > little box via USB, quietly buzzing away in a corner making multi- > megabyte files with the information I want. > > I assume that this is a no-brainer with National Instruments hardware and > software -- am I correct? If not, is there a solution that you can > recommend? Do you have a favorite other than NI? >I usually use this for such jobs: http://labjack.com/u3 Plus a few resistors. It also comes in a version for LAN connection in case that's more practical. It comes with a light version of Azeotech SCADA software which is really handy if you don't want to mess with computer programming or Excel-VBA too much. If you get stuck both Labjack and Azeotech have rather responsive forums where their respective support staff chimes in. And if someone screws up a PID you can become the expert and help them out :-) -- Regards, Joerg http://www.analogconsultants.com/
Reply by ●March 13, 20132013-03-13
Tim Wescott <tim@seemywebsite.com> writes:> On Wed, 13 Mar 2013 12:05:05 -0500, Vladimir Vassilevsky wrote: > >> On 3/13/2013 11:44 AM, Tim Wescott wrote: >> >>> I have a need to sample one channel of data at between 200 and 2000 >>> samples per second, 12 bits or so, and store it away. >> >> For one time deal: take your favorite micro eval board from the shelf; >> connect it to USB to RS232 dongle; put together simple application; >> record file from terminal program on PC. All preparation should take >> less then half an hour. > > It's for a customer over 1000 miles away whose hand I cannot easily hold. > > And it needs to deal with 4-40mA, for which I have no hardware.Like a resistor? :) -- John Devereux
Reply by ●March 13, 20132013-03-13
On Wed, 13 Mar 2013 10:54:17 -0700 Joerg <invalid@invalid.invalid> wrote:> Tim Wescott wrote: > > This is slightly OT for all the groups I'm posting to, but something many > > of you should have done. > > > > I have a need to sample one channel of data at between 200 and 2000 > > samples per second, 12 bits or so, and store it away. > > > > The signal happens to be on a 4-40mA line already, to make life easy for > > getting off-the-shelf hardware. > > > > I need to store hours of guaranteed-uninterrupted data, or failing that, > > store data that's timestamped at the measuring device (not just in the PC > > software) so that I can see what's missing and deal with it. (dropouts > > of even a half second or so would be OK, as long as the timestamps are > > correct). > > > > The mental model I'm carrying around is a laptop PC, connected to some > > little box via USB, quietly buzzing away in a corner making multi- > > megabyte files with the information I want. > > > > I assume that this is a no-brainer with National Instruments hardware and > > software -- am I correct? If not, is there a solution that you can > > recommend? Do you have a favorite other than NI? > > > > I usually use this for such jobs: > > http://labjack.com/u3 > > Plus a few resistors. It also comes in a version for LAN connection in > case that's more practical. > > It comes with a light version of Azeotech SCADA software which is really > handy if you don't want to mess with computer programming or Excel-VBA > too much. If you get stuck both Labjack and Azeotech have rather > responsive forums where their respective support staff chimes in. And if > someone screws up a PID you can become the expert and help them out :-) > > -- > Regards, Joerg > > http://www.analogconsultants.com/Wow. I was actually just wondering about something like this the other day. And the fact that it comes complete with cross-platform drivers and Python bindings makes it pretty seriously nifty. -- Rob Gaddi, Highland Technology -- www.highlandtechnology.com Email address domain is currently out of order. See above to fix.
Reply by ●March 13, 20132013-03-13
Followups set to sci.electronics.design . In sci.electronics.design Tim Wescott <tim@seemywebsite.com> wrote:> I have a need to sample one channel of data at between 200 and 2000 > samples per second, 12 bits or so, and store it away.You already own A/D hardware and software that can sample two 16-bit channels at up to 44 kHz with an amplitude of roughly one volt. These days it's integrated but the board used to say "SoundBlaster" on it. Just bypass the DC blocking caps on the input.> I need to store hours of guaranteed-uninterrupted data, or failing > that, store data that's timestamped at the measuring device (not just > in the PC software) so that I can see what's missing and deal with it.With some soundcards, it's possible to record from the left and right channels separately. You could feed the same signal to both channels and then stagger the file gaps so that one of the channels is always recording. I'd do something like "new file every 5 to 10 minutes"; that way you can go with the file timestamp and the sample count to establish an exact time of the event of interest. Use something like Audacity to look at the files, or use SoX (Sound eXchange) to convert the .WAV files you will probably get into raw binary files, which you can then concatenate back together, or scan through in your favorite programming language. If you can, it might be worthwhile to run the application on the target PC for several hours and see how bad the PC's internal clock drifts. Some are good and some can be out several seconds over the course of a day.> The mental model I'm carrying around is a laptop PC, connected to some > little box via USB, quietly buzzing away in a corner making multi- > megabyte files with the information I want.If you use Windows, try to have the machine not be on a network, and shut off all the Windows updates, anti-virus updates, etc. These can randomly munch large amounts of your disk and CPU and maybe affect the data acquisition.> I assume that this is a no-brainer with National Instruments hardware > and software -- am I correct?They probably sell a card or USB dongle that can do it, in conjunction with Labview. Be sure you are sitting down when you ask about the price, and plan on having a fairly stout PC to run Labview on. Measurement Computing probably sells something equivalent for slightly less money. If you don't already own Labview, and you have the ability to write your own code, you can probably reduce the cost somewhat; despite NI's advertising, a first-time Labview user is *not* going to get something to work in 5 minutes. Matt Roberds
Reply by ●March 13, 20132013-03-13
On Wed, 13 Mar 2013 11:10:30 -0700, Rob Gaddi <rgaddi@technologyhighland.invalid> wrote:>On Wed, 13 Mar 2013 10:54:17 -0700 >Joerg <invalid@invalid.invalid> wrote: > >> Tim Wescott wrote: >> > This is slightly OT for all the groups I'm posting to, but something many >> > of you should have done. >> > >> > I have a need to sample one channel of data at between 200 and 2000 >> > samples per second, 12 bits or so, and store it away. >> > >> > The signal happens to be on a 4-40mA line already, to make life easy for >> > getting off-the-shelf hardware. >> > >> > I need to store hours of guaranteed-uninterrupted data, or failing that, >> > store data that's timestamped at the measuring device (not just in the PC >> > software) so that I can see what's missing and deal with it. (dropouts >> > of even a half second or so would be OK, as long as the timestamps are >> > correct). >> > >> > The mental model I'm carrying around is a laptop PC, connected to some >> > little box via USB, quietly buzzing away in a corner making multi- >> > megabyte files with the information I want. >> > >> > I assume that this is a no-brainer with National Instruments hardware and >> > software -- am I correct? If not, is there a solution that you can >> > recommend? Do you have a favorite other than NI? >> > >> >> I usually use this for such jobs: >> >> http://labjack.com/u3 >> >> Plus a few resistors. It also comes in a version for LAN connection in >> case that's more practical. >> >> It comes with a light version of Azeotech SCADA software which is really >> handy if you don't want to mess with computer programming or Excel-VBA >> too much. If you get stuck both Labjack and Azeotech have rather >> responsive forums where their respective support staff chimes in. And if >> someone screws up a PID you can become the expert and help them out :-) >> >> -- >> Regards, Joerg >> >> http://www.analogconsultants.com/ > >Wow. I was actually just wondering about something like this the other day. And the fact that it comes complete with cross-platform drivers and Python bindings makes it pretty seriously nifty.It should just enumerate as an FTDI serial port, so there would be no need for drivers and DLLs. -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
Reply by ●March 13, 20132013-03-13
John Larkin wrote:> On Wed, 13 Mar 2013 11:10:30 -0700, Rob Gaddi > <rgaddi@technologyhighland.invalid> wrote: > >> On Wed, 13 Mar 2013 10:54:17 -0700 >> Joerg <invalid@invalid.invalid> wrote: >> >>> Tim Wescott wrote: >>>> This is slightly OT for all the groups I'm posting to, but something many >>>> of you should have done. >>>> >>>> I have a need to sample one channel of data at between 200 and 2000 >>>> samples per second, 12 bits or so, and store it away. >>>> >>>> The signal happens to be on a 4-40mA line already, to make life easy for >>>> getting off-the-shelf hardware. >>>> >>>> I need to store hours of guaranteed-uninterrupted data, or failing that, >>>> store data that's timestamped at the measuring device (not just in the PC >>>> software) so that I can see what's missing and deal with it. (dropouts >>>> of even a half second or so would be OK, as long as the timestamps are >>>> correct). >>>> >>>> The mental model I'm carrying around is a laptop PC, connected to some >>>> little box via USB, quietly buzzing away in a corner making multi- >>>> megabyte files with the information I want. >>>> >>>> I assume that this is a no-brainer with National Instruments hardware and >>>> software -- am I correct? If not, is there a solution that you can >>>> recommend? Do you have a favorite other than NI? >>>> >>> I usually use this for such jobs: >>> >>> http://labjack.com/u3 >>> >>> Plus a few resistors. It also comes in a version for LAN connection in >>> case that's more practical. >>> >>> It comes with a light version of Azeotech SCADA software which is really >>> handy if you don't want to mess with computer programming or Excel-VBA >>> too much. If you get stuck both Labjack and Azeotech have rather >>> responsive forums where their respective support staff chimes in. And if >>> someone screws up a PID you can become the expert and help them out :-) >>> >>> -- >>> Regards, Joerg >>> >>> http://www.analogconsultants.com/ >> Wow. I was actually just wondering about something like this the other day. And the fact that it comes complete with cross-platform drivers and Python bindings makes it pretty seriously nifty. > > It should just enumerate as an FTDI serial port, so there would be no > need for drivers and DLLs. >Not quite. It has a whole slew of settings that are handled by the driver, AFAIK. It even has a simple set panel where you could goose a DAC, set a port or read an ADC value (they all show up in the window). Then there's things like streaming mode and so on. It also needs to know whether it has foreign device connected to it and whether they should be addressed with RS232, "Her Majesty" or SPI. -- Regards, Joerg http://www.analogconsultants.com/
