Good PRECISION low frequency data acquistion ADC -> Ethernet or USB

Hi, all:-

Any suggestions for a really good low frequency (up to a few hundred Hz would be great) 24+ bit ADC box with very low low frequency noise and drift? Preferably something like 20-30ppb RMS noise from 10uHz to

1Hz, with at least two-four channels input (that's roughly a true 20 bits at 1kHz).

The usual NI boards and their clones have really high low frequency noise and others have low drift but a bit too high a noise floor.

I don't really want to build something here (that will come later). Any suggestions would be most welcome. A few K$ would be fine.

Best regards, Spehro Pefhany

Reply to
Spehro Pefhany
Loading thread data ...

How about two to four good DVMs?

--
John Larkin         Highland Technology, Inc 

jlarkin att highlandtechnology dott com 
 Click to see the full signature
Reply to
John Larkin

Yes, I've suggested that (and lent an Agilent/Keyshite LXI 34401A to the cause). I was hoping to get a bit better, for less than the ~$7K that would cost.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward" 
speff@interlog.com             Info for manufacturers: http://www.trexon.com 
 Click to see the full signature
Reply to
Spehro Pefhany

I've got a labjack that sorta meets those specs. (~$3-400) I've not looked closely. The low end (10uHz.) sounds hard, I mean what's a day? 3600x24 =~10^5 seconds

George H.

Reply to
George Herold

Specsmanship is a problem. For example, the PSD on one product looks almost good enough until you start actually measuring something that isn't almost zero, then the reference 1/f noise dominates.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward" 
speff@interlog.com             Info for manufacturers: http://www.trexon.com 
 Click to see the full signature
Reply to
Spehro Pefhany

How about use a good soundcard? Or, a 24 bit soundcard like Creative Labs EMU1212 system. setup as 'professional' balanced to drive a cheap MC1496 analog multiplier-like for each channel and recive balanced, too, then knowing the frequency of the drive [sine wave], synchronously demod to recreate the DC to ?? spectrum, easily get two channels DC to 8kHz, and run at 44.1 kS/s which allows for 8 channel operation, could result in getting 8 channels DC to 4 kHz, and probably still get better than 20 bits. Haven't looked at resulting specs but awfully easy to try out. From memory the input noise is around 12nV/rtHz, which is pretty equal to the 10k input Z.

Reply to
RobertMacy

Sound cards are stunning and I am using a USB sound card for lab measurements here, a Behringer UCA202. But these will not go down to DC without a serious hack or an "inverse sampler" a.k.a. modulator up front.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

HP3457A? I've been pining for one all day and would be willing to buy one when you're done.

ChesterW

Reply to
ChesterW

Can you tell us a bit about the application - a lot of my time now is spent designing high end data acquisition stuff (current emphasis on massively multichannel and quite high speed) but spreading out to to other things. It won't do you any short term good but we might eventually look at very low noise if we could see enough demand. As you are probably aware audio stuff is poor to dreadful at low frequency noise/stability (one of the reasonable trade-offs is to loosen up on stuff that can't be heard). From my (not exhaustive) experience I think that the DMMs (suggested earlier) might be the simplest off the shelf route to truly low noise at ultra low frequencies.

Michael Kellett

Reply to
MK

I think that might have been my suggestion, a year or two back. I've not used mine for anything but audio yet. Did you make any modifications to it? I must have a more extensive play with mine...

Clifford Heath.

Reply to
Clifford Heath

I'm afraid it won't do you much good. It's going to have to be crammed into a tiny satellite form factor at some point (and work in a hard vacuum deep space environment), you can consider this as a lab experiment.

Best regards, Spehro Pefhany

Reply to
Spehro Pefhany

I left it as is since I didn't need to go down to DC so far. I found that higher priced sound modules were often not so good and returned them.

Just avoid anything with 3.5mm phono jack. That whole phono jack concept is junk, always has been. RCA is the way to go.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

On a sunny day (Wed, 20 Aug 2014 11:48:57 -0700) it happened Joerg wrote in :

RCA is unreliable too, I have lots of problems with bad contacts in RCA connectors for video. Some are gold plated.

PL295 is better... for RF and LF. I just unsoldered a huge PL259 (for 2 cm thick coax), was outside, moisture got in, cable oxidized for a long length (central conductor copper black and green, screen black). Connector after cleaning still perfect.

formatting link
formatting link

I had a power audio amp that used PL259 for speaker connectors.

DIN 3 and 5 pole is OK for audio and data. XLR works too.

BNC is OK, but too easy to pull the wires out.... D connectors, Amphenol has real springs... are OK, some others not.

USB and micro USB connectors are terrible. what else do we have...

Actually 3.5 mm phono jacks are OK, my Tecsun SSB receiver uses it for antenna... And USUALLY those headphones work.

Reply to
Jan Panteltje

I suspect that a large part of the problem with sound cards (and USB audio devices) is ground isolation and noise. A good USB audio device that has a well-designed common-mode transformer in the USB cable should give the best isolation.

The TI chip in the UCA202 is (was?) available, but I didn't get any; I wanted four channels with four perfectly synchronised ADCs at the time, to do directional microphony based on phase correlation.

Reply to
Clifford Heath

What is your sensor source impedance ? What is the optimal source impedance of the ADC for best noise performance ? Perhaps a suitable amplifier in front of the ADC might help.

Have you looked for ADCs with CAN bus interface. After all, in some Amateur radio satellites (AMSAT) the CAN bus is used to communicate between different units.

In CAN bus with 64 bit payload, you could transfer two 24 bit or three

21 bit ADC channel samples. With a 1 Mbit/s CAN bus, more than 6000 Hz sampling rate could be used.
Reply to
upsidedown

The "UHF" connector is a nice anode voltage connector to at least a few kV :-).

TNC and SMA/SMC have a bit more realistic size for DC/audio applications.

With low voltage (near) DC measurements, there is always the issue of electrochemical between dissimilar metal connection surfaces. If you are going to measure microvolts at DC, one should study e.g. thermocouple wiring practices.

Reply to
upsidedown

I've already designed, built and tested signal conditioning circuitry that's better than needed- the signal is optimally scaled to fit within +/-10V range or less, as required, (so most ADC boxes will actually divide it down internally). Still short by about an order of magnitude (a few bits).

The bus isn't particularly important as long as there is some way to get the data off. Data rate is low and any practical design is probably going to have the input channels isolated. CAN would add something of a layer of complexity, I think, since you can't just plug it into a PC directly. Ethernet or USB are fine and relatively simple to deal with.

Thanks for your suggestions.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward" 
speff@interlog.com             Info for manufacturers: http://www.trexon.com 
 Click to see the full signature
Reply to
Spehro Pefhany

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.