-2
\$\begingroup\$

I seek a device that can be programmed to output physical voltages based on a set of datapoints. For example:

[Data](2, 5, 10, 6)mV -> [Device] -> [Voltmeter](2.05, 4.90, 9.95, 6.05)mV

The device should also support multi-channel outputs (and inputs), and preferably be able to operate for a large # of channels (by combining several devices, etc). The minimum desired resolution is 1mV.

Any device capable of accomplishing this?

UPDATE: Application:

Clinical EEG data is fed into the device, which then transmits real voltages into a single-board computer (SBC). Specs:

  • Channels: 22+
  • Sampling rate: 200+ Hz
  • Resolution: 12+ bit
  • Output range: \$ 2 \cdot 10^3 \$ (1 \$ \mu \$V to 2 \$ m \$V)
| improve this question | |
\$\endgroup\$
  • 2
    \$\begingroup\$ You need to provide more information about your specific application and required sample rate. Your list of requirements could be solved by many different existing devices and instruments. DACs, programmable signal generators, synthesizers, DAQ cards, sound cards, etc. \$\endgroup\$ – Edgar Brown Dec 25 '18 at 6:22
  • 1
    \$\begingroup\$ No. You must figure out things work and how to specify everything that makes the difference in "capability" , so think about what you really want to accomplish, how soon?? for how much?? the "must haves = specs"?? and nice to haves (language, load, speed, max channels, interface, performance ( rate, BW, range, current?) . User input?? Spreadsheet,? MP3 file? Commodore Basic program? Output ? a Voltmeter? Speakers? a Flight Simulator?? a CNC Stepper Servo-controlled 3 axis Gantry?? Get the idea ? \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Dec 25 '18 at 16:03
  • \$\begingroup\$ Updated with relevant information. @EdgarBrown \$\endgroup\$ – OverLordGoldDragon Dec 26 '18 at 17:33
  • \$\begingroup\$ Updated with relevant information. @TonyEErocketscientist \$\endgroup\$ – OverLordGoldDragon Dec 26 '18 at 17:36
4
\$\begingroup\$

I seek a device that can be programmed to output physical voltages based on a set of datapoints.

This kind of device is called a digital-to-analog converter or DAC.

The device should also support multi-channel outputs (and inputs),

Up to 8 output channels per chip is pretty common. You can probably find 16 channels if you look, but you might have to compromise on other specs. Normally there'd be one input interface per chip, not one for each output channel.

The minimum desired resolution is 1mV.

The real limitation is dynamic range. 1 mV resolution is easy if the maximum output voltage is 1 V or so. But pretty challenging if you want a maximum of more than 5 V. (A 16-bit DAC could be amplified to give you 64 V output with 1 mV resolution, but designing an amplifier to do that is not a beginner's project)

Before you choose an actual DAC device you'll also want to work out what are your requirements for update speed, absolute accuracy, output current, etc.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Updated with relevant information; one or two specific devices that qualify should suffice for now. \$\endgroup\$ – OverLordGoldDragon Dec 26 '18 at 17:36
  • 3
    \$\begingroup\$ @OverLordGoldDragon, part selection questions are specifically off topic, because the list of parts available changes over time, and might be different depending where you live. What we can do is give you advice on how to select a part, like I've done in the answer above. \$\endgroup\$ – The Photon Dec 26 '18 at 17:42
0
\$\begingroup\$

Based on your updated requirements, a custom design with far more specifications including CMRR, Gain, Nyquist filter, electrode probes, motion restraints , ESD protection and cabling specs are needed to pickup the uV level signals with > 40dB SNR and amplify to TBD output levels for a suitable MUX and ADC.

So rather supplying a commercial solution or suggesting this much broader question than the scope of this forum to be deleted, I suggest you examine commercial solutions that have solved all these requirements and ask yourself, how did they do it?

enter image description here

BTW a range of 2â‹…10^3 needs 2^11 range or >=11 bits

There are many commercial solutions to this problem, so your problem is a good one, yet disliked by many moderators of this forum. It is a fizzy question that is solved by experience using a low cost ( yet ugly) portable EEG scalp monitoring system that can combine many hardware channels with very high gain, high impedance and very high common mode noise rejection ratio with customizable analog filters, digital filters and a wireless 32 channel mobile computer with a host that performs the fuzzy logic learning. Once the algorithms have been optimized for one person with a trans-cranial electrical stimulation people with this affliction can be treated in future. I would imagine a self-contained system under a wig or a tight fitting toque with the hardware and interface to a smart phone.

More Examples enter image description hereenter image description here

  • wired scalp with wireless sender to host CPU/GPU
| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks for your input; a DAC was more appropriate to my ends, but ultimately decided against real-voltage prototype testing altogether - shall test locally instead. \$\endgroup\$ – OverLordGoldDragon Jan 5 '19 at 0:25
  • \$\begingroup\$ No need for real voltage if you have the data to process the results. But I thought you might want to create your own data too \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jan 5 '19 at 1:45

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