For those following my work, thank you and sorry for the long wait for an update! I am currently writing up my doctorate so as you can imagine I have been a social hermit recently!
Nevertheless, I have now had a journal paper published which may be of interest to those fellow electronics geeks out there and as a bonus it is open access so no charge to view it!
http://www.sciencedirect.com/science/article/pii/S0360319915308065
The paper focuses around balancing a hydrogen fuel cell with a battery with ideal diodes. These give a tremendous efficiency advantage over traditional diodes. The powertrain exploits the ability that a fuel cell has to operate at different voltages in order to create a potential long endurance system which still has all the power of a battery at an instant when needed.
Once again, thank you for your interest and I will do my best to answer any questions you have! The paper is a creative commons license so feel free to use it and build upon it, giving credit for the original work where due!
You need to be a member of diydrones to add comments!
Comments
@John, thank you for the schematis and source code.
I think I googled for supercapacitors, but didn't find any suitable. Could you point me in the right direction?
@Martin - here's the URLto schematics and firmware
I must also apologize Simon for hijacking his original thread.
Your problems seems rather obvious - your battery can't handle high currents or has relatively high internal resistance (likely both).
Have you considered adding a supercapacitor in parralel to the battery? It'll come to you at around 150g (100F, 3 cells voltage, 70 euros). Or, of course, you can do it the same way I did - put 1000mAh 90C battery for takeof/maneuvring and have another-one for idle/hover flight to prolongue flight-time (my board weights ~14g, 1000mAh - high-C - is another 150g).
This is a very interesting discussion. Some time a go we wanted to add a lipo parallel to li-ion cells to use during the launch phase because the voltage drop was too great and the resulting power output too low. We didn't get quite the result we wanted, also the usable voltage range for a li-ion goes down to 2.5 V vs 3.0 V for a lipo.
@John Rambo, could you share the schematics and source code of your design?
Wow, that is awesome. I need to get back into the fun hobby side rather than the academic work side ;-)
@Simon,
LTC4416/17 Family is an interesting thing indeed. Wonder how did I miss it. I've rejected LTC4370 for the same very reason as you did - it only enables 50/50 load-sharing. Plus, package is to small for diy PCBs.
@Christian,
I've implemented proportional current load-sharing by employing two mirrored (phase-correct) PWM channels on ATTiny85. Was quite a challenge, but Its performing rock-solid. If one FET goes 25% duty, another is going 75% duty inverted. That prevents currents from flowing from source-A to source-B and vice-versa, thus eliminating a need for back-to-back FETs.
LTC1155 is Gate Driver (driving 2x N-FETs on High-side bus thus giving the lowest voltage drop possible - P-FETs can beat this). I've got CSD18536KCS (N-FET) samples from TI - impressive-ones! At 60Amp they're a bit warm even without a heatsink.
ATTiny takes PPM input from receiver. Then you've got an absolute control on behavior of the two power supplies. You can merge both 50/50 or 20/80, you can switch smoothly from one to another based on throttle, you can set a transient window by using trim knobs on receiver, you can use a throttle-cut (or failsafe pin) to i.e. switch to reserve battery.
Hi Christian,
Very interesting idea, I think it probably hasn't been done much because of the limited payload. However, if rather than 2x 4Ah batteries linked in parallel to double your endurance you linked them through a LTC4416 in parallel you would get the same increase i endurance but the security that if one pack fails then it won't knock out the whole power supply. Bear in mind that with the LTC4416 both batteries would be used simultaneously in normal operation.
The LTC4416 has been the most robust part of my whole setup if I am honest. It has worked exactly as expected from day 1. I am using Infineon Power MOSFETs which are good up to 80-100A ish in my setup with passive only cooling, no dedicated heatsink except the PCB (specially designed).
Thanks for your comments gents.
Rambo, that looks very interesting congrats. I am actually using the LTC4416, unfortunately the 4370 is designed for load sharing equality which my system isn't. The fuel cell is not electrically similar to a LiPo so wouldn't be able to share 50:50 like two batteries would, making the 4370 somewhat over-engineered for this purpose (although it would of course still work). This project started a couple of months after the 4416 was released. (yes academia is that slow!) The follow on project is to look at controlling the chemistry within the fuel cell to anticipate a load change (eg feed forward throttle for an upcoming manoeuvre) and therefore be able to sustain itself in order to reduce the battery capacity (therefore size/weight) with no drop in performance. Basically an energy optimisation study.
Birkeland, yes I am well aware that 5C is achievable with LiPos. Interestingly though if you look at battery electric cars they tend to be around the 0.3C mark for safety and battery life. The peak charge limitation is to do with the fuel cell voltages. As the voltage climbs, its current output reduces. So in a series system the overall load sharing ratio between fuel cell and battery would be very low in the normal operating window because the fuel cell output cannot go above the charged battery voltage. In a parallel/combined system you can have a fuel cell with a higher nominal voltage than the battery allowing for recharge and/or a better load sharing ratio under normal operating conditions.
Hope this makes sense, if not ask again and I will try and reclarify or revisit anything! I am grateful for the interest.
I shall also point my successor (when appointed) in the direction of this so he/she can keep the community posted.
Great stuff. I skimmed your article, and the peak charge limitation that is used as the reasoning for not doing the more traditional serial system, seems a bit dated. You can easily get LiPo batteries today that will accept a 5C charge, and some even claim to handle 10C charging.
I've been working on something similar, but for powering a machine with different capacity/voltage LiPos (i.e. 3S and 4S in parallel or 3S 4Ah+3S2.2Ah).
Industry has already came with a solution for your problem (see LTC4370 for example).
I, however, had to come with a more intelligent prototype - it enables me to manually balance/switch the current drawn by i.e. adjusting trim knob on transmitter, or by reacting to the throttle curve (hover from battery-A and maneuver or failsafe from battery-B automatically).
-
1
-
2
of 2 Next