Any good USB/Network DAQ units out there?

[Gav]

Hi All,

I'm doing a couple of science projects at the moment where acquiring multiple analog channels at once is needed, as well as writing to analog outputs. I'm wanting a PC in the mix, controlling the software, and I'm looking for a good USB/ethernet DAQ board to buy. 

I know I can get a beefy microcontroller, and some DAC and ADC chips and roll-my-own, but I'm just wanting to get something simple off the shelf to save time. 

My rough wish-list of features is: 

• At least 4 analog inputs (12-bit or better resolution, preferrably fast reading)
• At least 2 analog outputs (12-bit or better resolution) 
• Bunch of digital inputs & outputs would be nice. 
• Able to talk to it simply from Python, and also preferably Matlab/Octave, without having to write my own bindings/compile stuff/sacrifice a chicken
• Simultaneous reading of multiple input channels would be nice, but not essential

So far what I've come across are these:

There's the 'Jabjack' range, which looks nice:

http://labjack.com/products

and they even have a cheaper 'embeddable' version:

http://labjack.com/embeddable-daq

They've been around a few years, and are probably my favourite so far. 

There's the beautiful looking 'Red Pitaya':

http://redpitaya.com/

Which looks brilliant from a hardware point of view, but I'm not crazy about as it's an embedded device you put an image on then webbrowse to. (Might be just me, but my experience with the Raspberry Pi leads me to believe the more removed a device is from my PC, the less likely I am to fix various problems and the more it just sits in a box on my shelf) They also don't appear to have finished or documented the bindings for labview/Matlab yet. 

There's the 'OpenDAQ', which I'm quite tempted by:

http://www.open-daq.com/

And they also have cheaper 'embeddable' boards. Sadly they only have a single analog output pin. 

Has anyone used something similar in their travels? Any good stuff to get, or traps to watch out for? 

Cheers,

Gavin 

[tALSit de CoD]

That's a pretty good list.

There's always this, but it's not cheap:

https://www.youtube.com/watch?v=hNksWHJ2rZE

// talsit.org

// +81 90 6736 1127

--
You received this message because you are subscribed to the Google Groups "Robots & Dinosaurs" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sydney-hackspa...@googlegroups.com.
To post to this group, send email to sydney-h...@googlegroups.com.
Visit this group at http://groups.google.com/group/sydney-hackspace.
For more options, visit https://groups.google.com/d/optout.

[tALSit de CoD]

On that related note, I need a +/- 10V analogue output for a DAQ board - preferably 4+ channels, as well as some ADC input channels. Anyone know of such a device?

// talsit.org

// +81 90 6736 1127

On 6 March 2015 at 07:17, Gav <the.mechat...@gmail.com> wrote:

--

[Rodney Berry]

Hi Gavin,

I used this one a lot for EEG/EMR stuff a few years ago. It was quite affordable and no-nonsense with passive and active scalp electrodes. I think we also fed a doppler heart monitor to it as well.

http://www.dataq.com/products/di-149/ they might have a newer one by now.

You can configure it via the USB port. here's some video of sending it instructions etc. from the TouchDesigner environemnt.

https://vimeo.com/49832767

Rodney

--

You received this message because you are subscribed to the Google Groups "Robots & Dinosaurs" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sydney-hackspa...@googlegroups.com.
To post to this group, send email to sydney-h...@googlegroups.com.
Visit this group at http://groups.google.com/group/sydney-hackspace.
For more options, visit https://groups.google.com/d/optout.

--

Rodney Berry rodberry.net

[Tristan Steele]

Just buy a Labjack and be done with it. :) 

For the price, the provide a pretty good unit, plenty of software and in my experience the Python code works really well. I've got about a dozen of these deployed around the place, and none of them have ever given me any trouble. They work pretty much out of the box on a RPi too, if having ARM drivers is an issue.

Tristan

On 6 March 2015 at 09:17, Gav <the.mechat...@gmail.com> wrote:

--

[Luke Emrose]

Wow, Labjack looks really good.

I've been on the lookout for a data logger with analog input and over 16bit resolution but I'd had no luck.

I was going to use them to assist me in calibrating exponential VCOs for analog synth calibration.

The idea would be to measure VCO frequency and then tell the user which direction to turn a trim pot to finish a calibration step.

OpenDAQ also looks pretty tempting since I'm more interested in ADC than DAC for my purposes (though with both I could auto-calibrate which is also appealing as an idea).

thanks so much, very informative thread.

--

Luke Emrose
aka evolutionary theory
www.evolutionarytheory.com
www.soundcloud.com/evolutionarytheory
www.reverbnation.com/evolutionarytheory
http://www.facebook.com/pages/evolutionary-theory/110717958952413

[Mark Pearson]

A lot of research organisations use Powerlab from ADInstruments. They are not cheap, but very reliable.
http://www.adinstruments.com/products/powerlab

[Ada Lim]

On 06/03/2015 12:11 PM, "Luke Emrose" <evolution...@gmail.com> wrote:
>
> Wow, Labjack looks really good.
>
> I've been on the lookout for a data logger with analog input and over 16bit resolution but I'd had no luck.

Isn't that called a minidisc recorder?

> I was going to use them to assist me in calibrating exponential VCOs for analog synth calibration.
> The idea would be to measure VCO frequency and then tell the user which direction to turn a trim pot to finish a calibration step.

You don't really need 16 bits for frequency analysis, do you?

-A

[Luke Emrose]

On 6 March 2015 at 15:22, Ada Lim <a...@panda2.net> wrote:

Isn't that called a minidisc recorder?

A minidisc recorder isn't computer controllable, and definitely not via Python, at least not without a reasonable amount of work.

You don't really need 16 bits for frequency analysis, do you?

Not the frequency analysis itself, but I do for recording the exact value of a control voltage going to an analog oscillator.

The control voltage is between 0-10V (but generated by it's own DAC, whose voltage I won't know exactly without measuring it, and what I'd be recording is the control voltage vs the resulting frequency. This plot I can then analyse to see if it's giving me an exponential response to match the usual V/Oct standard used for modular synthesizers.

The frequency is going to just be measured by a timer and interrupt from any microcontroller, but the monitoring of the CV would need to be done by an accurate data logger that I can control.

[Ada Lim]

On 06/03/2015 3:39 PM, "Luke Emrose" <evolution...@gmail.com> wrote:
>
> On 6 March 2015 at 15:22, Ada Lim <a...@panda2.net> wrote:
>>
>> Isn't that called a minidisc recorder?
>
> A minidisc recorder isn't computer controllable, and definitely not via Python, at least not without a reasonable amount of work.

You said logging, which is usually offline.  My point still stands - audio cards are usually the cheapest way to get >16 bits of data acquisition, depending on how much signal conditioning you need.

> Not the frequency analysis itself, but I do for recording the exact value of a control voltage going to an analog oscillator.
> The control voltage is between 0-10V (but generated by it's own DAC, whose voltage I won't know exactly without measuring it, and what I'd be recording is the control voltage vs the resulting frequency. This plot I can then analyse to see if it's giving me an exponential response to match the usual V/Oct standard used for modular synthesizers.
> The frequency is going to just be measured by a timer and interrupt from any microcontroller, but the monitoring of the CV would need to be done by an accurate data logger that I can control.

Is the rest of the analog pathway going to give you 100dB of SNR?    more pertinently - how would you ever tell?

(I've got some fancy 6 digit chinese multimeter that will do USB data acquisition.  what I don't have is the fancy calibration gear that lets me trust past the third digit.)

-A

[Ada Lim]

On 06/03/2015 9:18 AM, "Gav" <the.mechat...@gmail.com> wrote:

> There's the 'Jabjack' range, which looks nice:

I didn't realise this was a typo at first, but JabJack sounds like the sort of port you would use for medical data acquisition (hooked up to an invasive blood pressure monitoring transducer, for example).

It would be a great name for a continuous sugar reading device for diabetics, I think.

-A

[Luke Emrose]

You said logging, which is usually offline.

OK 

> Not the frequency analysis itself, but I do for recording the exact value of a control voltage going to an analog oscillator.
> The control voltage is between 0-10V (but generated by it's own DAC, whose voltage I won't know exactly without measuring it, and what I'd be recording is the control voltage vs the resulting frequency. This plot I can then analyse to see if it's giving me an exponential response to match the usual V/Oct standard used for modular synthesizers.
> The frequency is going to just be measured by a timer and interrupt from any microcontroller, but the monitoring of the CV would need to be done by an accurate data logger that I can control.

Is the rest of the analog pathway going to give you 100dB of SNR?    more pertinently - how would you ever tell?

As discussed the control voltage, which controls frequency, is what I'm measuring, not audio.

In my case the DAC output goes to an exponential current generator then to the oscillator.

Yes, 100dB of audio fidelity would not be audibly different from something lower, but I wasn't talking about that, which might not have been clear.

To support the full MIDI note range, which is C-1 to G9 is approx 10 octaves.

Assuming we have an exponential converter for the VCO let's map out 0-10V DAC range to 1 volt per octave to control it.

There are 12 notes per octave = 10 octaves * 12 total notes = 120 notes

Get get 1 cent accuracy is 100 cents per note implies we need 120*100 steps = 12000, which could be handled by a 14bit DA converter.

The two extra bits for measuring are useful for getting enough latitude to measure the 14bit signal above to a reasonable accuracy.

L

[Ada Lim]

On Fri, Mar 6, 2015 at 5:18 PM, Luke Emrose
<evolution...@gmail.com> wrote:
> As discussed the control voltage, which controls frequency, is what I'm
> measuring, not audio.
> In my case the DAC output goes to an exponential current generator then to
> the oscillator.

> To support the full MIDI note range, which is C-1 to G9 is approx 10
> octaves.
> Assuming we have an exponential converter for the VCO let's map out 0-10V
> DAC range to 1 volt per octave to control it.
> There are 12 notes per octave = 10 octaves * 12 total notes = 120 notes
> Get get 1 cent accuracy is 100 cents per note implies we need 120*100 steps
> = 12000, which could be handled by a 14bit DA converter.
> The two extra bits for measuring are useful for getting enough latitude to
> measure the 14bit signal above to a reasonable accuracy.

I'm not saying you don't need it - i'm saying that the rest of your
analog signal path is going to inject noise so you will have to be
very careful with how much you trust your measurements. I don't think
that any affordable off the shelf device will do what you want it to
do reliably.

consider that a typical high precision voltage reference is only
accurate to 0.01% (15 bits) and will have its own temperature
coefficient etc.

great news, everyone! if you didn't buy the second edition of H&H in
anticipation of the third edition coming out, your wait is soon to be
over! and you too can start designing ultra accurate measurement
devices.

(my copy of H&H is boxed up, but you did post once that you ordered
it, so none of this stuff should be news to you.)

[Nick Johnson]

How about the Analog Discovery? http://www.farnell.com/datasheets/1639278.pdf

Alternately, a PSoC 5 development board, though you'd have to write some code.

-Nick

--

[Dr Hugh]

A small trap is that the Windows usbser driver is unreliable, and likely will get into a 1-way communication state when abused.
My previous company changed all our USB boards to HID (using usbhid driver).
So suggest looking for HID rather than serial devices.

[tALSit de CoD]

Dr.Hugh, for which DAQ is this for?

--

[spanner888]

If you want cheap - look at the Teensy 3.1

[Gav]

Cheers for the suggestions everyone. I think the labjack will be the best fit for my application, I'll try and get one this week. 

" I've got about a dozen of these deployed around the place, and none of them have ever given me any trouble."

@Tristan - That's exactly the sort of recommendation I was after, thanks. 

By the way, how's the OpenMCA going? 

@Nick, that analogDiscovery looks pretty nice. I'm tempted to get one just to have on the shelf :-)

On 8 March 2015 at 09:40, spanner888 <spann...@usabledevices.com> wrote:

If you want cheap - look at the Teensy 3.1

--

[Andrew Larkin]

Just received in a email from Digikey, this may be of interest:

http://www.digikey.com.au/product-highlights/en/max11300-integrates-a-pixisupsup/53135

 

MAX11300

·     Up to 20 12-bit ADC inputs, single-ended or differential ·     ADC absolute input voltage ranges: 0 to +10 V, -5 to +5 V, -10 V to 0 V, and 0 to +2.5 V (direct ADC connection) ·     Programmable sampling averaging per ADC port ·     Up to 20 12-bit DAC outputs with 25 mA current capability ·     DAC output voltage ranges: 0 to +10 V, -5 to +5 V, and -10 V to 0 V ·     Up to 20 general-purpose digital I/O ·     +5 V analog supply ·     20 MHz SPI/QSPI-compatible ·     Small 6 mm x 6 mm 40-pin TQFN and 9 mm x 9 mm 48-pin TQFP packages

·     0 to +5 V input range for GPI ·     0 to +2.5 V programmable threshold range for GPI ·     0 to +10 V programmable output range for GPO ·     Internal or external voltage reference for DAC and ADC ·     Individually selectable voltage reference for each ADC-configured PIXI port ·     Internal and external temperature sensors, ±2°C accuracy ·     50 Ω analog switching capability between adjacent PIXI ports ·     1.8 V to 5.0 V compatible serial interface

Do a breakout board and plug into you favourite SPI host...

[tALSit de CoD]

 Do a breakout board and plug into you favourite SPI host...

This is where the sacrificing of chickens that Gav was trying to avoid comes in.... 

[Andrew Larkin]

Some chickens deserve to be sacrificed.

 

I'll have a leg thanks.

 

[tALSit de CoD]

Oh man oh man. So, as per your suggestion, I got myself a LabJack. And man, did I regret buying one!! (*) Within 3 minutes of downloading the software, I had an example running. And 45 minutes later, I had a python program up and running, replacing my original "solution". That meant that I no longer had a LabJack to play with. So I ordered another one!!

AAA+++, would buy again, and, well, I did.

The only thing I'm finding less than ideal us their quadrature speed, which is capped to 70/20kHz, depending on your mode. That's because it's interrupt based, not native clock-based.

By "one", I mean that literally, as in, "just one".