3
\$\begingroup\$

I'm actually an experimental physicist who does software-development for a living. So my knowledge basically stops at writing micro-controller programs and putting them on a breadboard, and hooking the wires the simple way.

enter image description here

Back then in my student days, I also managed to get a PCB with holes, and solder electronic parts on it to create an RCD programmer for a PIC16F84 microcontroller, and it worked (that was like 10 years ago).

enter image description here

Now, I would like to make a high-resolution ADC with Arduino, but it has to be reliable and eventually I'd like to put it in a case/box. The ADC chip I have my eye on is ADS1256.

I'm targeting 24-bit with about 10 kHz sample rate, and this chip seems to be perfect. However, I have absolutely zero knowledge on how to get such a relatively small chip on a board, so that I could interface with it.

I'm wondering whether there's a simple way to do this myself (e.g., with a PCB with holes, like the one in the picture), or whether there are services that would do this for money without having to make 10k pieces. One challenging part when thinking about this as DIY is that if I screw this up, everything will act like an antenna and I'll have lots of noise in my digitized signals.

What are my options? Please advise.

PS: If it's relevant, I live in Germany.

\$\endgroup\$

closed as off-topic by Chris Stratton, Wesley Lee, The Photon, PeterJ, m.Alin Aug 22 '17 at 6:35

  • This question does not appear to be about electronics design within the scope defined in the help center.
If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    \$\begingroup\$ It's not clear whether you wish to experiment with PCB layout or not. If yes, the fabrication part is quite viable, there are many fab services where you can make few PCBs - proper layout and mitigating noise will be another issue though. If not, then you should probably look into some sort of freelance portal. \$\endgroup\$ – Wesley Lee Aug 21 '17 at 18:35
  • 4
    \$\begingroup\$ I'm voting to close this question as off-topic because "what are my options" questions are not a fit for the stack exchange model of specific answers to specific questions. Additionally, the title of the question bears no relation to he body. \$\endgroup\$ – Chris Stratton Aug 21 '17 at 18:48
  • 5
    \$\begingroup\$ To get the benefit of a high precision ADC you need a carefully, professional designed board for it. See if the manufacturer offers an evaluation board, or look for an alternate which is available on a board intended to practically demonstrate its theoretical capability. \$\endgroup\$ – Chris Stratton Aug 21 '17 at 19:33
  • 1
    \$\begingroup\$ @TheQuantumPhysicist - eurocircuits.eu, pcbway.com, oshpark.com, etc \$\endgroup\$ – Wesley Lee Aug 21 '17 at 19:40
  • 2
    \$\begingroup\$ @marcelm 20 bits and I'll be thankful. There's no way to know without trying. \$\endgroup\$ – The Quantum Physicist Aug 21 '17 at 21:05
7
\$\begingroup\$

What's your budget? You can always breadboard using breakout boards as mentioned in another answer, but you'd maybe get half of those 24 ADC bits as useful information that way. Getting even 16 useful bits out of an ADC requires a PCB with good power and signal integrity practices.

You could try your hand at laying out your PCB, but it might quickly get overwhelming.

Alternatively, TI offers an evaluation kit for the part which will have been designed to squeeze as many bits as their Application Engineers could manage. See here:

ADS1256 Performance Demonstration Kit

From reading the User's Guide on that page, it looks like the kit comes in two parts, which gives you some options. The smaller board is the ADC and its supporting circuitry. I've only skimmed the guide, but it looks like the analog signals are broken out on J1 and the digital signals on J2. Thus, J2 can be breadboarded to your Arduino (it's a 100mil pitch, which matches your typical breadboard, so you might just have to turn the board upside down and plug in), and J1 can potentially be wired directly to whatever you're measuring or preferably broken out on a board to a proper cable connector. The digital control is just I2C so your Arduino will have no trouble.

It looks like you used to be able to buy just the smaller board, but it's been obsoleted. You can get both boards together for $200 from TI. The second board is a DSP-powered interface to your computer. You'll have to read up on what that offers, but you might even find that's all you need for your purposes, assuming it offers an SDK and you're not trying to do any remote monitoring. If you're cost-sensitive and just want the smaller board, try contacting TI directly. They are often accommodating with demo kits, especially if you're associated with a big-name institution or company.

ADS1256 EVM and DSP board

Lastly, since I work for a TI competitor that rhymes with "shmanalog shmevices" (nowhere near their chip business though), I'll just add that when looking at evaluation boards it's worth checking out as many devices as you can. You might find some eval boards more useable than others for your applications, depending on how the Apps Engineer wanted the demo platform to be used.

\$\endgroup\$
  • \$\begingroup\$ Actually the evaluation board is very expensive for my budget. I'm looking for something like 30€. Interestingly, googling Ads1256 and looking at the images showed very interesting, small boards from sources like aliexpress. \$\endgroup\$ – The Quantum Physicist Aug 21 '17 at 19:55
  • 1
    \$\begingroup\$ Check this out: eckstein-shop.de/High-Precision-AD-DA-Board \$\endgroup\$ – The Quantum Physicist Aug 21 '17 at 19:57
  • \$\begingroup\$ In that case you should consider contacting TI to see if you can get just the ADS1256EVM (the smaller board); a significant portion of that $200 is the larger DSP board and the technical support they offer on the kit. And like I said they're very accommodating (and they're also known to give away demo kits for free to students and researchers). I'd be skeptical of the signal integrity of boards on aliexpress, if you go that route you might want to scale back the number of bits you need. \$\endgroup\$ – jalalipop Aug 21 '17 at 19:59
  • 1
    \$\begingroup\$ @TheQuantumPhysicist that link looks like a reasonable option. I would also like to point out that the users guide above includes a schematic, bom and images of the layout for the smaller breakout board, so it should be fairly straight forward to design your own if you so desired. \$\endgroup\$ – Redja Aug 21 '17 at 20:07
  • 1
    \$\begingroup\$ @TheQuantumPhysicist Beware when buying from sources like AliExpress; overall quality varies wildly, and getting (products with) fake chips is a very real concern. \$\endgroup\$ – marcelm Aug 21 '17 at 21:00
4
\$\begingroup\$

You can do surface mount soldering yourself without any new equipment if you buy a breakout board for that specific package (form factor) of chip. I didn't read the entire data sheet on the chip you specified, so I'm not 100% sure of the package spec, but this might be what you're looking for. Then you can do the surface mount soldering at home with flux and a regular soldering iron.

enter image description here

Eventually, if you're going to make a few of these, you can just draw this type of breakout pattern in the PCB layout software and get it printed. If you don't need them urgently, you can send off the design to some services out of China and get 20 or so boards printed for about $1-2 each in about 2 weeks.

\$\endgroup\$
  • \$\begingroup\$ Wow... This is great! Thanks a lot! Unless I get a better suggestion, I guess your solution fits very well! \$\endgroup\$ – The Quantum Physicist Aug 21 '17 at 18:40
  • 13
    \$\begingroup\$ You will not get anywhere near the advertised capability of an ADC chip if you mount it like this and put it on a breadboard. These breakout boards can be very useful for other parts, and for making sure you understand the digital communication between an MCU and an ADC, but to leverage the full capability of the chip you need a carefully designed purpose made circuit board where extreme attention has been payed to avoiding noise issues. Even then, the low order bits will only be "real" in a statistical sense. \$\endgroup\$ – Chris Stratton Aug 21 '17 at 18:49
0
\$\begingroup\$

I evaluated a 22bit ADC once, running perhaps 50Hertz word rate, on a 2-sided PCB. There were NO digital lines anywhere near the input nor near the ADC, except for the SPI (3 wire) serial interface that was quiescent between ADC conversions. And I had 1 Kohm resistors in the clk-data-enable 3-wires, to "encourage" the remote MCU trash to remain remote.

How much magnetic fields can you tolerate? Consider 60Hz, 1amp, 1 meter distance, with surges to 10 amps with 1uS risetimes as the rectifier diodes abruptly turn on. 1uS edges are poorly rejected by the copper foil (just one or 2 dB attenuaton by skin-effect) so we'll assume no magnetic shielding. Assume your vulnerable loop-area is the area enclosed by Vin+ and Vin- of the ADC, modeled as 100milliMeters (4") by 1mm. Here is the math (I don't know the answer; I'll be as surprised as you).

$$Vinduce = [MU0 * MUr * Area / (2 * pi * Distance)] / dI/dT$$

Knowing MU0 is 4 * Pi * 10^-7 Henry/meter, we rewrite the Vinduce as

$$Vinduce = 2e-7 * Area/Distance * dI/dT$$

and we have

Vinduce = 2e-7 * (100mm * 1mm) / 1meter * 10amps/1uS

Vinduce = 2e-7 H/m * 100e-6 m^2 / 1m * 10^+7 amp/sec

Vinduce = 2e-7 * 100e-6 * 1e+7 = 200 * e-6 = 200 microVolts.

Do you feel lucky? (Clint Eastwood)

What is the topology of this interferer?

A single infinite-length wire carrying 10amps with 1uS Trise (occurring 120 times a second, which we don't need for our computation), the return path located at infinity. In reality, power cords have the hot/rtn leads just a few millimeters apart, so most of the magnetic field is self-cancelling. Most. How lucky do you feel?

And the vulnerable loop is the 100mm long input signal to the ADC, with the Vin+ located 1mm from the Vin-; distance over planes does not get modeled.

Another error? electric fields from spikes on the 60Hz, and 80,000Hertz from electronically controlled fluorescent lights.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
  • \$\begingroup\$ The question asks for a cheap sub $200,- 24bit ADC. This is likely not the answer he was looking for. This answer does show it might be hard to create such a device. \$\endgroup\$ – JWRM22 Aug 22 '17 at 6:37
  • \$\begingroup\$ That's why we do differential measurements :) \$\endgroup\$ – The Quantum Physicist Aug 22 '17 at 10:29
  • \$\begingroup\$ "nard to create such a device"? Yeap---that is purpose of all the details I provided. \$\endgroup\$ – analogsystemsrf Aug 22 '17 at 17:24
  • \$\begingroup\$ The 200 uV was induced into the loop formed by Vin+ and Vin- of the differential input traces on the PCB. \$\endgroup\$ – analogsystemsrf Aug 23 '17 at 18:00

Not the answer you're looking for? Browse other questions tagged or ask your own question.