Low current power supply for medical purpose

[alexg-nn]

Hi everyone,

I am currently trying to design a small biomedical/acupuncture device for my father (for personal use, of course). Basically it is a handheld battery driven micro power supply with very low current limit. Patient holds one electrode in the hand and then pokes to some point at the body with the other one. During this operation current will normally rise apparently from zero to some pre-adjusted limit up to 100uA.

Here is my specs for this device:
- powered by single LiPo cell 3,7V;
- charged from USB at 5V;
- digitally controlled;
- some sort of LCD for indication;
- 9-20V output range, digitally adjustable, regulation 1V, accuracy plus/minus 1V will be enough;
- 10-150uA current limit, digitally adjustable, regulation 10uA, accuracy plus/minus 5uA will be enough;
- resistance of human body will change not particularly fast so transition time CV-CC is not so important as in conventional PSU. I can’t tell any number here;
- reasonably efficient and cheap.

Basically it should be some sort of low power CV/CC supply. I found a lot of circuits of PSUs, but I did not see in them the possibility of regulating such low currents and they were too complicated.

I have no problem with digital part, indication and charging. Measurement is more tricky, but I have uCurrent as an example here, so no worries there. My question is how to build a main power supply with specs above.

My first intention was to use some sort of boost/buck-boost converter and then somehow limit current. But efficiency of such converters is poor at low currents. Then I decided to use charge pump (with multiple stages) followed by LDO, maybe even use Dave’s design with LT3080 and his CC mode implementation. But I don’t sure how to maintain low current with proper accuracy and linearity. And what about stability with long inductive leads (say 1m each) and capacitance of human body.

I will appreciate any help and advice.

Thank you,
Alex

[Prehistoricman]

Unless you are simply wasting power, I wouldn't be concerned about efficiency. Your max output power is 3mW. Likely your microcontroller will use more power than the electrodes.
Is efficiency a concern for battery life? What's the intended use time and battery capacity?

[SiliconWizard]

My first intention was to use some sort of boost/buck-boost converter and then somehow limit current. But efficiency of such converters is poor at low currents.

Not for all. There are quite a few of them that are impressively efficient at low output currents.
You can start here for instance:

https://www.analog.com/en/parametricsearch/11446#/sort=5349,asc

[alexg-nn]

Unless you are simply wasting power, I wouldn't be concerned about efficiency. Your max output power is 3mW. Likely your microcontroller will use more power than the electrodes.
Is efficiency a concern for battery life? What's the intended use time and battery capacity?

Perhaps you are right. The controller and LCD will consume much more than this power supply. I mentioned efficiency mainly because the datasheets for converters that I came across did not even show efficiency values ​​for such low currents. As for the battery capacity, I was planning to use something like 1200-2400 mAh.

My first intention was to use some sort of boost/buck-boost converter and then somehow limit current. But efficiency of such converters is poor at low currents.

Not for all. There are quite a few of them that are impressively efficient at low output currents.
You can start here for instance:

https://www.analog.com/en/parametricsearch/11446#/sort=5349,asc

Thank you. I've found some interesting part too, like TPS61096A. But what should I do next?

Let's assume that I have found a suitable 3,7 to 20v boost converter. What should I place next to it to regulate the current? LDO? Or perhaps some over kind of current limiting circuit? Please advise the topology for such a solution.

Here is some math I did. R_L below is the patient's resistance, I took three values ​​as an example. U_OUT is the voltage that the power supply must provide.

For an upper current limit of 150uA plus/minus 5uA (145/155uA)

RL	UOUT
1k	150mV (145/155mV)
10k	1,5V (1,45/1,55V)
100k	15V (14,5/15,5V)

For a lower current limit of 10uA plus/minus 5uA (5/15uA)

RL	UOUT
1k	10mV (5/15mV)
10k	100mV (50mV/150mV)
100k	1V (0,5/1,5V)

My conclusions:
- the output voltage should be regulated practically from 0V;
- U_RIPPLE should be less than 10mV.

Unfortunately, I don't know how to use these numbers for LDO picking. What value of the "load regulation" parameter should appear in the datasheet?

[poorchava]

Lipo has a minimum voltage if about 3V, that's 6.66 boost ratio. Should be doable as far as maximum duty cycle is concerned. You can either use a converter ic or control it from uC directly (through a driver).

[alexg-nn]

I found a suitable power supply circuit from the HP361XA and modified it a bit. At first glance, it works well. However, I don't know how to solve the overshoot problem.




I've attached a simulation in case someone will help me with that.