Multi-channel data acquisition/logging on the cheap

[LaurentR]

Hi all,
I have a few simple applications where I need to log 5-15 channels of data (essentially voltage, current through a sense resistor, temperature), all this at rates going from somewhat fast (a few times a second) to slow (every few seconds/minutes).

I already have bench DMMs (34461A/34401A) and script them using Matlab (for now - maybe Python/C later), so I am looking for an option to do the channel switching that can also be scripted. Without breaking the bank if possible.

I assume the common instrument-based solution would be to grab an Agilent 34970A and one of the multiplexer modules off EBay ($500 in all?). If I understand well how these work, I could either use a standard module and use the internal DMM or get just a matrix switching module, use the 34970A just for switching and use my own DMM?
Is a 34970A with the multifunction module something nice to have around anyway?

Another cheaper route is to get RS232 or USB (VCP) relay boards. EBay is full of those like the KMTronic modules (e.g. USB 8-relay board for $58). These all seem to use a very simple serial protocol for realy control, so the scripting should be very straightforward. Any recommendation for one of these?

Any other option?

Laurent

[dom0]

3488A and new(!) modules can be had cheaply. It's essentially a modular, royal version of USB relay cards.

[Keylectric]

If you need the precision/accuracy of your DMM ( I doubt that?) I guess your plan makes sense. Otherwise you could check something like Picotech and their wide range of data loggers for less money.

[Fsck]

3488A and new(!) modules can be had cheaply. It's essentially a modular, royal version of USB relay cards.

agreed. since you already have the DMMs, this is the cheapest and most effective solution.

[Mark]

http://labjack.com/
I book-marked the above site when it was mentioned on here, never got around to buying anything.  They may have something of interest to you.   

[LaurentR]

3488A and new(!) modules can be had cheaply. It's essentially a modular, royal version of USB relay cards.

agreed. since you already have the DMMs, this is the cheapest and most effective solution.

Thanks both, I didn't know about that unit. It is indeed cheap. I am not that familiar with HPIB. Can I use any modern USB GPIB interface to drive it (like the NI or the Agilent) or do I need some more ancient RS232>HPIB or something like that? Need to drive it from a modern PC (USB-only). Have FTDI RS232 cables if needed.

[LaurentR]

If you need the precision/accuracy of your DMM ( I doubt that?) I guess your plan makes sense. Otherwise you could check something like Picotech and their wide range of data loggers for less money.

http://labjack.com/
I book-marked the above site when it was mentioned on here, never got around to buying anything.  They may have something of interest to you.   

Thanks. Those are both interestingly cost-effective data acquisition options. In my particular case though, some of the experiments involve battery discharge, so, while absolute accuracy is not necessary, high resolution and some amount of time stability is good.

[dom0]

3488A and new(!) modules can be had cheaply. It's essentially a modular, royal version of USB relay cards.

agreed. since you already have the DMMs, this is the cheapest and most effective solution.

Thanks both, I didn't know about that unit. It is indeed cheap. I am not that familiar with HPIB. Can I use any modern USB GPIB interface to drive it (like the NI or the Agilent) or do I need some more ancient RS232>HPIB or something like that? Need to drive it from a modern PC (USB-only). Have FTDI RS232 cables if needed.

The point with GPIB is that it's universal. Works with pretty much any computer and OS of the last 50 years (and everything in the next few decades, at least).

The nice thing is that it's a bus, so you can connect quite a slew of devices with just one connection to your computer or network. And the cables are very cheap (3-5 € / each) and widely available, so it's really quite a nice and useful thing.

[Keylectric]

Thanks. Those are both interestingly cost-effective data acquisition options. In my particular case though, some of the experiments involve battery discharge, so, while absolute accuracy is not necessary, high resolution and some amount of time stability is good.

Did you look at the ADC-20 and ADC-24 devices?

[LaurentR]

Did you look at the ADC-20 and ADC-24 devices?

Just did 
They do look good, and surprisingly cost effective. I am going to have to look at the software aspect of those.

[PedroDaGr8]

Don't forget the Keithley 199. It has 8 channel with the switcher card (make sure it is included in the auction), they can usually be found for pretty cheap. I have a couple I am going to clear out in the next two weeks, as I need the money, which is why I know about them.

[macboy]

Another cheaper route is to get RS232 or USB (VCP) relay boards. EBay is full of those like the KMTronic modules (e.g. USB 8-relay board for $58). These all seem to use a very simple serial protocol for realy control, so the scripting should be very straightforward. Any recommendation for one of these?

Any other option?

Laurent

A relay board is not a bad idea until you realize that most or all will have power relays, not signal relays. There is a big difference. If you try to put small signals through a power relay, you will be in for a world of trouble. Their contacts are designed in such a way that they require a minimum amount of current to flow in order to be reliable. Otherwise, you can get high resistance contacts, or even apparently open contacts that should be closed. You must use small signal relays (a.k.a. telecom relays) for switching low current signals.

The nice thing is that it's a bus, so you can connect quite a slew of devices with just one connection to your computer or network. And the cables are very cheap (3-5 € / each) and widely available, so it's really quite a nice and useful thing.

I like GPIB, but I'd really like to know where to find these mythical $5 cables!?

[dom0]

Not that uncommon... keep an eye out for bundles of cables, preferably auctions. Bought seven cables a while back for about 35 €. I also bought a few by means of "buy-now offer" (eBay), they were pricier (+2 € shipping ea), so total maybe 10 € / cable for those. Single cables not made by HP or NI sometimes go for 1 € + shipping. So for example Meilhaus or R&S cables.

Built a little test jig (2× 2 € for 24 pin sockets plus some perfboard) to check for broken contacts ; only one cable out of the ten I bought had a bent contact, which was fixable.

New and unused GPIB cables are an entirely different story, price wise.

[Keylectric]

Did you look at the ADC-20 and ADC-24 devices?

Just did 
They do look good, and surprisingly cost effective. I am going to have to look at the software aspect of those.

I only tried Picolog (logging software) briefly but it seems nice. Picoscope (scope software) I have used for several years with some 3000 and 4000 series scopes and I really like it.

A nice thing with Pico is that the software is common for all devices and free to try out in demo mode. Once you've learned your way around the software and need a new type of scope for a new task you feel at home at once. I have the 16bit 4262 Picoscope as a complement to an Agilent DSOX-3032A.

[mtdoc]

Multichannel PCI based DAQ boards from NI can be found on eBay for low prices. You'd likely want a copy of Labview to either program your own solution or used one of the included DAQ examples. The student version, which is fully functional except can't create executable stand alone VIs, can be bought from Sparkfun for $50 - no student verification needed.

Occasionally on eBay you'll see entire NI DAQ systems for cheap with DAQ board and PC with LabView already installed with a VI ready to go. Usually these come from university labs upgrading or closing down.