Make Battery Last Longer

[SunEagle]

Where did that come from ?

Oh snap. I deleted a few other posts by Oscar but must have missed this one. A member suggest this post and others may be cause by a "bott"

[Mike90250]

They should adjust and counteract one another, including boards includes weight and drag, to counterbalance that you should add push capacity to produce more lift to convey a heavier plane, to do that requires more power from the power plant...

Where did that come from ?

[Mike90250]

OK on the zener voltage. I just wanted to be sure there was not going to be a problem

[jflorey2]

Great, I'm glad it worked. One thing worries me, you have a 12V zener as protection against overvoltage on a 11.1V battery, that seems to be beyond the safety margin by 1.5V or so.

If they are li-ions then they are good up to 4.2VPC, or 12.6 volts total. So a 12V zener should work OK as long as it doesn't remain connected during "normal" charging.

[cagiva]

Mike90250, We used the 12V Zener because this battery is at 12.4V when it's fully charged. Won't a lower 11V Zener attempt to bring the battery down to 11V(10.45-11.55)?

[Mike90250]

Mike90250, Your suggestion was right on. Ditching the charge controller and just adding the schottky diode and a 5w zener has given us the best results. One question, do you see the need to also add a capacitor to the circuit? Thanks!

Great, I'm glad it worked. One thing worries me, you have a 12V zener as protection against overvoltage on a 11.1V battery, that seems to be beyond the safety margin by 1.5V or so.

The large capacitor is not needed, unless the zener generates enough noise/hash that it interferes with the plane's radio, then a much smaller capacitor, 0.1uF / 30V ceramic disc would help.

[cagiva]

I'd add a schottky diode as a reverse bias prevention, and a 5w zener as overvoltage protection.

Mike90250, Your suggestion was right on. Ditching the charge controller and just adding the schottky diode and a 5w zener has given us the best results. One question, do you see the need to also add a capacitor to the circuit? Thanks!

Orbit-Solar (1).png

[bcroe]

That's a big IF One mistake, leaving the PV connected in the dark - will drain the battery. Parked in the sun
between flights, might overcharge, depending on battery SoC.

So, I'd add a schottky diode as a reverse bias prevention, and a 5w zener as overvoltage protection.

That is important. I thought the switch (or connector) connecting the battery to the motor might
have an extra pole to connect the panels. Avoid a zener that size which needs a heat sink. Bruce

[Mike90250]

........ I would even skip the zener, if the cells are never connected when the motor is off. Bruce Roe

That's a big IF One mistake, leaving the PV connected in the dark - will drain the battery. Parked in the sun between flights, might overcharge, depending on battery SoC.

So, I'd add a schottky diode as a reverse bias prevention, and a 5w zener as overvoltage protection.

[bcroe]

i'd be comfortable saying ditch the charge controller and use a zener diode. something
like a 5w zener, at the right voltage for the batteries, would be the ticket, won't need much of a heat sink either.

I will second that, just try to get a good match between the cells and the battery operating voltage. If
that is done you have little to gain from any kind of charge controller. All that other stuff is going to
add weight. With a motor load greater than the cells can deliver, there is no danger of over charging
the battery. I would even skip the zener, if the cells are never connected when the motor is off. Bruce Roe

[Mike90250]

i'd be comfortable saying ditch the charge controller and use a zener diode. something like a 5w zener, at the right voltage for the batteries, would be the ticket, won't need much of a heat sink either.

[jflorey2]

jflorey2, Sunpower cells were our first choice, but our current vehicle wingspan is not big enough to accommodate the number of tiles we need to match our required voltage.

So use a boost MPPT instead of a buck MPPT.

Anyway the MPPT you have posted above looks very big and heavy; those ferrite inductors are not light. There's no way you need those size components to deal with the power you are anticipating (<1 watt per array.) Even if you switch to Sunpower you wouldn't need something even a fraction of that size.

[J.P.M.]

SunEagle We're using a lightweight MPPT (picture below). An additional function of this MPPT is to monitor the current and the voltage of each solar module and make that information available for the control and navigation system through an I2C protocol. This information is communicated to the interface on the ground control station so that the operator is aware of the energy received from the sun in real-time.

J.P.M. Our Matlab model estimates that 3 % to 5 % percent of the solar energy will be lost and converted into heat, especially in the diodes, transistors, and inductances. Considering the very small surface of the MPPT, this can make its temperature increase up to 110◦C at noon. In order to monitor this effect, two temperature sensors are placed on the printed circuit board and connected to the microcontroller. It is thus possible to react from the ground if necessary by stopping the MPPT or limit the current to a certain threshold.

I'll post results once we have completed new test flights with the panels.

Screen Shot 2018-08-07 at 9.52.52 AM.png

We're talking about two different animals. That would be 3% to 5% of the electricity that's generated. Actually, depending on the type of PV device you'll wind up with something like ~ 85 - 90+ % or more of all the incident solar radiation that will hit the wing/PV assembly will be rejected as heat via convective and radiation heat transfer mechanisms with the remainder of the incident solar energy being turned into electricity. You may well lose about 3 - 5% of that generated electricity through other parts/components of the system, but compared to what's incoming, that's a relatively small amount. Like ~ 0.05 * 0.15 = ~ 0.008 of the incident solar radiation.

Given all the likely heat transfer going on via convection and the probably thin cross sections and lack of thermal insulation, I seriously doubt that's enough thermal energy to raise the temperatures by anything that can be measured much less of any impact.

[cagiva]

SunEagle We're using a lightweight MPPT (picture below). An additional function of this MPPT is to monitor the current and the voltage of each solar module and make that information available for the control and navigation system through an I2C protocol. This information is communicated to the interface on the ground control station so that the operator is aware of the energy received from the sun in real-time.

J.P.M. Our Matlab model estimates that 3 % to 5 % percent of the solar energy will be lost and converted into heat, especially in the diodes, transistors, and inductances. Considering the very small surface of the MPPT, this can make its temperature increase up to 110◦C at noon. In order to monitor this effect, two temperature sensors are placed on the printed circuit board and connected to the microcontroller. It is thus possible to react from the ground if necessary by stopping the MPPT or limit the current to a certain threshold.

I'll post results once we have completed new test flights with the panels.

Screen Shot 2018-08-07 at 9.52.52 AM.png

[J.P.M.]

[QUOTE=Garrett;n381898]
I'm curious to know your results on wind velocity over the wing and how it effects the panels delta-T. /QUOTE]

FWIW, from convective heat transfer considerations, and doing an energy/heat balance on the wing/panel assembly, and using a stated air velocity over the wing of ~ 8.3 m/sec, that the wing/panel assembly temp. will probably be fairly close to the ambient air temp., particularly given the likely thin wing section and light weight.

While convective heat transfer theory is quite well developed, empirical wind convective coefficients over (mostly) flat plates, either one or 2 sided, are all over the board, and quite dependent on the application. For a couple of examples only, for one sided heat transfer, McAdams suggests a value of h ~ = 5.7 + 3.8 V for one sided surfaces (with the other side insulated), where h is the convective heat transfer coeff. in W/(m^2*deg. C) and V is the wind vector in m/sec. But, as Duffie & Beckman note, that correlation may have radiation effects buried in it. A guy by the name of Watmuff (some names you can't make up !) suggests using h = 2.8 + 3.0 V, with the radiation heat transfer effects removed but without stating the mechanism of removal. Stuff I've done with my array suggest a gross two sided heat transfer coeff. of 21.7 + 6.32 V W/(m^2*deg. C) per m^2 of panel area. My number has thermal radiation effects in it, but preliminary stuff I've done looking at that suggests that lowering the result by about a third seems to produce a number for convective heat transfer alone that makes some sense with respect to what the radiation heat transfer might be after accounting for panel, effective sky and the roof temp. under the array.

I too would be interested in what investigations by the OP may turn up. There are about as many empirical heat transfer correlations in the literature as there are investigators to look for them. Fertile ground for the white collar welfare of graduate theses in mechanical or chemical engineering. Theory is a good guide, but for convective heat transfer, the number of variables that can and do influence the actual heat transfer means that theory will never be as accurate as actual and careful experiment and measurement for a particular application, probably more so than for some other disciplines.