Recommendation for micropower boost converter?

[davegsm82]

Hi Everyone, I'm looking for recommendations for a micropower boost converter, output up to 40V, input suitable for 3V coin cell (down to 2.5V~).

Output current is virtually nil, way less than a milliamp, just provides the bias current for a SiPM for radiation detection.

Preferably driverless, something self contained, something like the MAX630 but as low power as possible and without the quite surprising pricetag.

Cheers.

[amyk]

Have you looked at charge pumps?

Also worth considering: http://gyraf.dk/schematics/Voltage_multipliers_with_CMOS_gates.pdf

[Marco]

Maxim has an example to combine a comparator with reference with some logic to make a low power boost converter. You don't have to use the Maxim comparator, there are lots more with similar operating currents. You can combine that Maxim boost converter with a diode multiplier to get less extreme duty cycles for high ratios if necessary, the Theremino geiger power supply uses that for instance.


For slightly higher supply current you could use something like AP3015A or XC9120 with a diode multiplier too.

[David Hess]

How low of a current do you require?  A micropower boost part like the TI TPS61041 requires a quiescent current of 28 microamps from the supply, however the 4 microamps drawn by a 10 megohm feedback divider on the output will require 53 microamps from the input.  Is 81 microamps what you had in mind or do  you require even lower input current?

[SiliconWizard]

Have you looked at charge pumps?

Charge pumps for going from 2.5V to 40V? Yeah.

[SiliconWizard]

You'll find a number of them from LT (now Analog Devices). Although with 2.5V min Vin and 40V max Vout, you're pushing it a bit. You options are limited.

The now "venerable" LT1615 would be a reasonable fit, unfortunately its max Vout is "only" 34V if I'm not mistaken. It's in stock at Mouser. It's not the cheapest.

[Marco]

Anything meant to be used with an external diode, which is going to be the older stuff, can be made work with a multiplier even if it has an internal switch which can't reach 40V.

[davegsm82]

Hi everyone.

Ok, so, charge pumps are out really. I'm looking for absolute minimum component count as this is basically a watch and space is already tight.

The AP3015A looks like a really good bet, its definitely the sort of device I'm looking for, covers all the bases although I wouldn't know what the current consumption would be without building a prototype first. The datasheet seems to suggest the quiescent current is about 17uA when the device isn't switching. Believe it or not this is actually more than the present circuit draws, which is based on a Max9117, comparator and driver transistor etc.

The most quiescent current that the SiPM detector would ever draw from the 30-32v bias voltage is in the region of tens to hundreds of nano-amps, radiation detection events trigger pulses in the region of 50uA for very short periods of time, maybe 10-20uS.

The current SiPM has a bias voltage of around 31V, others may be in the region of 40V but I'm not thinking of using those at the minute, I would be happy with a 32V supply.

Thank you everyone for your suggestions.

[NiHaoMike]

Most likely you'll have to resort to switching the feedback network to get the quiescent current way down. An alternative would be to run it open loop with a zener switched in every few cycles to regulate the voltage. (Or perhaps there are extremely low current zeners for applications like that?)

[David Hess]

The now "venerable" LT1615 would be a reasonable fit, unfortunately its max Vout is "only" 34V if I'm not mistaken. It's in stock at Mouser. It's not the cheapest.

A boost converter can directly drive a capacitive voltage multiplier lowering the voltage requirements of the output switch.

Most likely you'll have to resort to switching the feedback network to get the quiescent current way down. An alternative would be to run it open loop with a zener switched in every few cycles to regulate the voltage. (Or perhaps there are extremely low current zeners for applications like that?)

Inexpensive small (0603) thick film high value surface mount resistors are available.  A divider resistance of 100 megohms reduces the divider current down to 400 nanoamps and the required additional supply current down to 5.3 microamps, which is a good match for the lowest supply current converters.  Total supply current could be under 15 microamps this way.