Solar to SuperCap Charger

Take a look at the bq24650.  It can be modified to charge a SuperCap instead of a battery.

I looked over the part, what are the modifications done to charge a supercap instead of a battery?  Specifically, how is the precharge mode and the 30 minute fault timer bypassed on the VFB pin?  All I can see is to add external circuitry to monitor the super cap voltage and use the VFB pin as a charge shutoff signal.

We do not have a released IC to support solar panel and charge a supercap directly. You have to add a MPPT (input voltage regulation) circuit in front of bq24640 to support this application.

What is an annual usage for this application?

Our projected usage is about 300 per month.  I have looked at the bq24640 and thought about doing just that.

Thanks.

Jim

In theory, you should be able to add a resistor divider and diode that clamps the FB pin voltage above the VLOWV threshold.  We are in the process of testing this circuit.

My thinking was not quite at such a simplistic level (simple is good), I was still thinking at using an opamp to buffer the two signals--a clamp and the charge limit.  It would be outstanding if your idea works and the tolerances allow operation over the entire temperature range.

Thank you

It's possible to use the bq24650 and bypass the precharge problem with supercaps by the addition of a resistor between the 3.3V output and the VFB input.

The resistors in the feedback divider along with this added resistor allow the VFB pin to be at ~1.6V at startup with discharged supercaps and at 2.1V when the output is at the desired regulated voltage.  It's simply a set of two linear equations for the two conditions solved to yield ratios of two of the resistors in the network to a third.  It isn't possible to solve for all three but a good choice for the upper divider resistor might be 1Meg as the deadload on the supercap(s) would then be (Vout-2.1) in uA.

The equations would be:

(0-1.6)/R2 + (3.3-1.6)/R3 = 1.6/R1

And

(Vo-2.1)/R2 + (3.3-2.1)/R3 = 2.1/R1

Where R2 is connected between the output (Vo) and VFB, R1 is connected from VFB to ground, and R3 is connected between 3.3V and VFB.

As an example with Vo=5V, R1 = 0.241*R2 and R3 = 0.206*R2.  Since paralleling resistors effective decreases the tolerance stackup by the square root of the number in parallel, with R2 arbitrarily chosen to be 1Meg, R1 can then be a 487K||475K+412 and R3 can be 412K||412K+249, to be excessively exact since the significant digits in the given data don't support this many digits in the derived values.

I'm sure if I'm missed something obvious and there's a problem with this suggested solution someone will correct me here, TIA.

Hope this helps,

Pete

Hi Pete,

I built a disable circuit as you described in the attachment. I forced Vfb to 1.8V.  I was able to disable the pre-charge region till about Vbat=2V. Anything above Vbat>2V the charger works in fast charge mode. But if Vbat<2V, the charger works in the pre-charge mode as shown in figure 04 in the attached document. The reason for not being able to disable completely the pre-charge region is because of the SRP and the SRN pins. They force the device into the pre-charge region and ignoring the Vfb as Bbat<2V.  

Thanks!

Tahar

Tahar,

Excellent!  Thanks a million for the quick followup.  I didn't have an EVM to test the theory and the datasheet block diagram shows the precharge threshold comparing the 1.5V reference (as it's annotated) with the voltage coming from the VFB.  If instead the SRP and SRN pins bypass or complicate the circuit then naturally a simple change to the internal regulation reference via the added resistor won't work across the entire range.  Good to see there's a work around with a diode switched overdrive that keeps the pin close to or above the threshold, very clever.

Thanks again,

Pete