Power supply for wearables using EEG

[kampa]

Hi all,

I’m starting out with hardware development for wearables. My first design uses EEG, and I need to buy a bench power supply that won’t interfere with the EEG signals.

I researched some options:
- rigol dp832(A)
- rigol dp711
- Siglent SPD3303X-E
- R&S® NGE103B

But I'm not sure what should I get because buying a power supply is quite an investment and  I’m worried about noise/leakage currents corrupting the EEG readings.

- Requirements
Output voltages: 1.8 V, 3.3 V, 5 V (and rarely 12 V)
Current: up to 1 A
Quality and long-term reliability
USB control for automated testing (Python + SCPI)
Nice to have: small footprint (limited desk space)
Nice to have: battery simulation (CR2032 or flat Li-ion style)

Any help would be appreciated , thanks!

[Kean]

I don't do medical stuff, but isolation is key for safety as well as the issue of leakage currents.

Do you really need a fully programmable PSU?  I'd consider if you can used fixed voltages or battery power for all use cases where a human is involved.

Can I suggest that anything being tested while connected to actual humans should probably be using a medical grade PSU, plus some isolated DC/DC converters.  Even if you use one of the above listed bench PSUs, you should still add isolated DC/DC converters if there isn't some other isolation already in the design.  You may have to add filtering for any switching noise.

Or stick to battery power and battery powered DMMs for measurement.  Battery power will have very little noise.

If you must have something programmable, then I'd look at using a USB isolator or wireless data link.  One low power programmable PSU is the siro,box which you could pair with a common USB 2 isolator based on an ADuM3160 or similar.  There is also the Qoitech Otii Arc if you want fancy battery simulation, but that is quite expensive.

https://sirubox.de/en/
https://www.eevblog.com/forum/crowd-funded-projects/live-on-kickstarter-tiny-usb-2-5w-vi-controllable-power-supply/

[kampa]

I don't do medical stuff, but isolation is key for safety as well as the issue of leakage currents.

Do you really need a fully programmable PSU?  I'd consider if you can used fixed voltages or battery power for all use cases where a human is involved.

Can I suggest that anything being tested while connected to actual humans should probably be using a medical grade PSU, plus some isolated DC/DC converters.  Even if you use one of the above listed bench PSUs, you should still add isolated DC/DC converters if there isn't some other isolation already in the design.  You may have to add filtering for any switching noise.

Or stick to battery power and battery powered DMMs for measurement.  Battery power will have very little noise.

If you must have something programmable, then I'd look at using a USB isolator or wireless data link.  One low power programmable PSU is the siro,box which you could pair with a common USB 2 isolator based on an ADuM3160 or similar.  There is also the Qoitech Otii Arc if you want fancy battery simulation, but that is quite expensive.

https://sirubox.de/en/
https://www.eevblog.com/forum/crowd-funded-projects/live-on-kickstarter-tiny-usb-2-5w-vi-controllable-power-supply/

I'm looking for a power supply during the hardware development until the hardware is stable and safe and then I was planning to use, as you suggested, a battery power when a human is involved.

[Doctorandus_P]

The ubiquitous 2mA - 2A 0-30V "Hiland" is a quite usable kit, and worth it's EUR7 or so. You do have to add a box, transformer and (panel) meter(s).

It does have a few issues, but there are long threads on this forum which discuss it's limits and suggest some modifications.

There are also a few variants of a 16V 2A linear power supply you can buy for around EUR 50. It's not a great power supply, but it's plenty for lots of projects.

"30V 3A" has been a sort of standard for lab power supplies for 40+ years, but in that time, many projects have become much smaller and less power hungry. I've never bought such a big / expensive power supply. I don't really have a need for it, and they are too big to fit on my desk.

It's also very easy to cobble something together with an LM317. It's a bit of a shame the L200 has become obsolete, as it has a built in adjustable current limit.

But it's also not so difficult to add current limiting to an LM317. You can start by adding a series shunt resistor on the input of the LM317, Use an opamp in a differential configuration to measure the voltage drop over that resistor, and pull the adjust pin of the LM317 to GND when the set current is exceeded.

I'd say that not only building from a kit, but designing your own power supply is a good exercise for beginners in electronics.

[thephil]

All of the candidates you found should be fine – in fact, any linear power supply should be suitable for your task. Personally I love programmable, remote-controllable supplies with a numeric keypad best (I have a Uni-T UDP3305S) but you should be able to go a lot cheaper than that without problems, depending on how many bells and whistles you want.