I'm working on a research project at my school that includes flying a drone with an attached signal source to collect data. The signal source is powered separately from the drone and is reliant on an atomic clock that needs to be reset once it runs out of power. In order to avoid this, I am looking to develop a hot-swap battery system so that we don't have any down time in our data collection. To do this, my initial thoughts are to make a system with a hot-swap controller to control the power switch between the batteries, a capacitor subsystem to provide short term energy, and voltage regulators to stabilize output voltage, as the clock is delicate. Thoughts on this potential system? Any help would go a long way.

 

For just powering signal source separately from two batteries, than a Schottky diode at each battery output would work to isolate the two batteries, but allow the one with the higher voltage to power the source.

Or is it important which battery initially powers the source?

 

For just powering signal source separately from two batteries, than a Schottky diode at each battery output would work to isolate the two batteries, but allow the one with the higher voltage to power the source.

Or is it important which battery initially powers the source?

The initial battery doesn't matter but I only want one of the batteries to be plugged in during flight so as not to add extra weight. The swap would then be done once the drone is on the ground swapping the depleted, lower voltage battery for a charged higher voltage battery. My understanding of what you said above is that with the diodes in place I could plug the higher voltage battery into the system while the lower voltage battery is still plugged in, in which case the diodes will allow power from only the higher voltage battery and I could unplug the depleted battery. Is this correct? Thank you for your help!

 

with the diodes in place I could plug the higher voltage battery into the system while the lower voltage battery is still plugged in, in which case the diodes will allow power from only the higher voltage battery and I could unplug the depleted battery. Is this correct?

That is correct.
I suggested Schottky diodes, since they have a lower forward voltage drop than standard diodes.

 

@cmartinez did something like this for some remote sensors—perhaps he can wisdom-bomb this thread…

 

@cmartinez did something like this for some remote sensors—perhaps he can wisdom-bomb this thread…

Thanks for tagging me Ya'akov ... yes, I think a look at my circuit might be profitable to the TS.

What my circuit does, is first deplete an external battery bank, and then use an internal (to the device) one as a backup while the external battery bank is being replaced. Eventually, of course, the device would have to be shut down so as to change the internal battery bank as well.

For the particular application described here, I'd suggest considering a pFet connected as a diode on each battery bank. Maybe that'd be the simplest solution.

 

I'd suggest considering a pFet connected as a diode on each battery bank.

Just a pfet won't work since the higher voltage battery will conduct to the lower voltage one through the substrate diode.

 

Just a pfet won't work since the higher voltage battery will conduct to the lower voltage one through the substrate diode.

What about back-to-back pFets?

 

What about back-to-back pFets?

They work to block in both directions, but only if there gate-source voltage is actively controlled by additional circuitry to be zero.
They can work alone to block a reverse voltage power to a device, but that is only because the voltage at the device is at zero.

This thread shows a circuit of mine to do something similar.

 

I suggest just having two of the battery pack connectors, with a shottkey diode in series with each, and a switch to bypass the diode for the battery in use. That way there is no diode drop wasting energy during use, but interaction is prevented, and there is a minimum of complexity and weight added. BUT it becomes important for the person changing the battery to know what to do.

 

That is correct.
I suggested Schottky diodes, since they have a lower forward voltage drop than standard diodes.

Would it be smart to include supercapicitors into this system as well? I'm thinking that these could provide extra stability to the power source and minimize fluctuation. I will also include voltage regulators as the piece of equipment I am powering is fairly sensitive and needs a consistent 12V.

 

Would it be smart to include supercapicitors into this system as well? I'm thinking that these could provide extra stability to the power source and minimize fluctuation. I will also include voltage regulators as the piece of equipment I am powering is fairly sensitive and needs a consistent 12V.

No need for supercaps—even if you needed some help with voltage sag ordinary capacitors would be fine—but you probably don’t even need that.

Remember, when a capacitor is discharged and then connected to power it will appear to have a very low resistance until it charges up. This means it will draw a great deal of current just as the power is turned on, called “inrush current”.

This can be a problem with conventional caps, with supercaps is a superproblem! You would need to add a relatively complex soft start circuit in place to limit the current until the caps charge. Fortunately this is moot since you don’t need the supercaps.

 

No need for supercaps—even if you needed some help with voltage sag ordinary capacitors would be fine—but you probably don’t even need that.

Remember, when a capacitor is discharged and then connected to power it will appear to have a very low resistance until it charges up. This means it will draw a great deal of current just as the power is turned on, called “inrush current”.

This can be a problem with conventional caps, with supercaps is a superproblem! You would need to add a relatively complex soft start circuit in place to limit the current until the caps charge. Fortunately this is moot since you don’t need the supercaps.

Ah I see. Thank you for the insight!

 

I suggest just having two of the battery pack connectors, with a shottkey diode in series with each, and a switch to bypass the diode for the battery in use. That way there is no diode drop wasting energy during use, but interaction is prevented, and there is a minimum of complexity and weight added. BUT it becomes important for the person changing the battery to know what to do.

Just to make sure I am understanding correctly, in this system when one battery pack is depleted, another pack is added to the system, both connected to diodes so that the new (higher voltage) pack powers the system. Once the old pack is removed, a switch the bypass the diode is then activated so that there is no longer a voltage drop caused by the diode. Is this correct?

 

Just to make sure I am understanding correctly, in this system when one battery pack is depleted, another pack is added to the system, both connected to diodes so that the new (higher voltage) pack powers the system. Once the old pack is removed, a switch the bypass the diode is then activated so that there is no longer a voltage drop caused by the diode. Is this correct?

That is a good summary of exactly what I was intending to suggest. It also works to save control programs in radio equipment produced with all of the software loaded into a battery backed RAM IC. Those rather expensive transcievers would be great for a few years and then the battery would fail and it was all over.

 

That is correct.
I suggested Schottky diodes, since they have a lower forward voltage drop than standard diodes.

How do they diodes work to only allow the battery with the higher voltage to power the device? Both diodes would be in series with their respective batteries but the batteries themselves would have to be in parallel wouldn't they? I.e, what's to prevent both diodes to be forward biased, increasing the output voltage. I'm assuming in this set up, that a diode is in parallel with each battery, and the batteries are parallel to each other. I'm new to this so all help is appreciated.

 

The batteries would indeed be in parallel during the time when the battery is being changed. The diodes are intended to assure that the battery with the lower voltage does not discharge the battery with the higher voltage. The switches are included to eliminate the diode voltage drop from the battery supplying the load after the other battery is removed.

Do you honestly not understand how a diode functions, or are you seeking the detailed physics explanation?

Batteries in parallel do not "increase the voltage" just because they are both in parallel. And when the fresh battery has a greater voltage, the diode in series with the discharged battery is REVERSE BIASED, and so it does not conduct in the forward direction. THAT is how diodes function. And that is why I suggested putting diodes in series with each battery pack. It was to prevent current from flowing into the discharged battery. The reason for preventing that current flow is to avoid dropping the voltage from the fully charged battery.