AVR Freaks

Thank you, everyone!

That 500 mA current limit was the number I was looking for, thank you!

With the LED Driver putting out 12 V x 150 mA = 1.8 W.

If one assumes 80% efficiency, (hopefully a little bit better), then the power input to the LED driver = 2.25 W.

So a 3.7 V LiPo (1xcell), would have to run at about 608 mA during the time the LEDs are on.

So that, unfortunately, is a bit over the 500 mA number, albeit used for a 10% duty cycle per Second.

The LED Driver I was looking at is single channel, so the LEDs have to be stacked.

I could easily use two of the chips, but I was hoping to keep the PCB space low, (it’s a small drone).

I thought about rolling my own constant current drivers with an NFet, low side driver, and a sense resistor, and either a small signal transistor or an op-amp for the Gate driver, but that starts to be a lot of little parts to put on the board, and the NFet is run in the linear mode, so I‘d have to do some thermal calcs on it.

Heat sinks obviously add space and weight, weight is the enemy on this project.

Then there is the issue that I’ve never been good at thermal calcs…, and I’d rather my learning curve be limited to LiPos and driving high current LEDs.  

I selected the battery above in large part due to its size, and from my initial reading of the data sheets, it looked like charging a single cell from a USB source was pretty easy.  I’ll have to look at it again, as perhaps a 2 cell charger is also easy, once I find the right chip.

I’ll look at the MC parts, the reason I selected the shunt charger was that I don’t care about rapid charging, and it was a trivial solution.  Some of the chargers required extensive support parts, which will take up space, especially since I like 1206 parts for my hand & frying pan soldering.

I’ll have a look at the other suggested Tinies. 

I envision the code as being pretty trivial, also, (famous last words), as I am not anticipating adding mission critical, failure detection, failure correction / mitigation, etc. to the project.  It is truly envisioned as a simple, small, LED flasher, (although flashing LEDs at 150 mA, not 20 mA, and incorporating a Lithium battery recharger, (What could possibly go wrong? <smirk>).

In any event, I hope to write the code, burn it, and be done with it!  As the device is portable, and I usually put an ISP header on the PCB, it wouldn’t be difficult to update the software at my work station, and I don’t envision changing any code in the middle of the device being used.  I do appreciate the reminder to think about a Bootloader, however.

LiPo C ratings is now added to my search vocabulary, thank you.  I’ll have a look for some small footprint 2 C battery packs, and see what their capacity and weight are.  That does, of course, mean that I’d now how to squeeze a small voltage regulator on the PCB, also…  The world is full of trade-offs!

I’ll just tell my wife I need a bigger drone, I’ve already outgrown this dinky one!

Thanks, everyone!

JC

All good points, and I appreciate the input.

The drone system is full of rechargeable batteries, (The drone itself, the controller, and the tablet/display), so one needs a process for recharging everything post usage.  Adding one more device to plug in seemed easier that buy batteries forever...  Plus, its about time I learned a little bit more about LiPo's.  The kicker is that if one wants a secondary mode with a forward looking "headlight" that is always on, then the power consumption really goes up, (no longer a 10% duty cycle), so going rechargeable seems to make sense.

There are certainly many Lithium based battery charger chips, and LED driver chips, available.  I've never delved into switching power supplies, and it seems that both of these are descendants of those, and they come in all of the various topologies.  It is taking me a while to gain even a basic understanding of the field, but that's what learning curves are all about.

I have done a bit of reading, mostly data sheets, but I'll have a look at some of the projects that are available to see what hurdles those designers encountered and overcame.

Your mention of a Fuse is a good one.  Truly I wanted this to be small, really small, and light weight.  My original plan was to use a polarized connector on the LiPo battery and call it a day.  No other reverse polarity protection, or overcurrent protection, (poor decision, especially since we're talking about Lipos!).  You are guilting me into doing the right thing.  Sigh.  Reminds me of the other current Thread discussing grams of potatoes, only I'm counting grams of components!

I still remember banks of 2200 uF or 4700 uF caps on 7805 based power supplies...  I'm not sure what duration of "flash" is adequate for spotting the aircraft against a light blue sky.  The contrast differential is what helps one's eye to spot the drone (led) while scanning the sky.  During the day I suspect, (totally unproven), that a decent duration flash will be helpful.  Once the project is up and running, flying, it will be easy to test 10, 20, 50, & 100 mSec pulses. 

An alternative that I have seen is a single, bright, LED on the top of the drone.  But then one ideally has some sort of diffuser to take some of the vertically directed light and aim it outwards, sideways, so that a ground based observer can see it.  

Plan B, using a zenon flash tube, is beginning to look a whole lot better!

JC  

Edit: Typo

Xenon vs Argon

I think classic “flash tubes”, such as those used in classic cameras, and emergency vehicles, use / used xenon, and have a wide spectrum of emission, and therefore appear white.

Argon, IIRC, glows blueish when electrified.

The old Neon bulbs, I think, had neon gas with a little bit of argon mixed in…

I guess the possibilities are endless, which is why bar / pub / tavern lights come in so many colors!

I promise not to harass any paraglider pilots! 

I bet you don’t like lasers pointed at you either!  (Sorry, just being a smart a…)

The goal is really to be able to spot the drone in the distance, when it tends to fade into the sky and is hard to be seen.  I can tell from the panel which direction it is pointing, and in which direction it is moving, (which aren’t the same, because like a copter is can go in any direction…), so I don’t need to gain that info visually from the lights.  I’m hoping just a bright, brief, 1 Hz flash, will suffice.

I found it a little bit odd that the TinyCircuits LiPo didn’t tell the battery’s “C” rating, as after doing some reading that seems to be a pretty important spec.  I could go with a different battery, with a “real” data sheet, but I don’t think I will need that level of documentation and detailed info for this project. (Hopefully).

I presume the 500 mA “max” current sourcing would be for “in spec” designs, where one is concerned about maximizing the battery’s life, recharging cycles, and safety from physical expansion and fire, etc.  I am not surprised that the drone / RC model industry pushes the batteries to the limit, as flight performance is probably a higher priority that battery integrity for that user group.

It is still, from a design perspective, nice to know how many mAs, (or Amps!), one can pull from the battery without its internal resistance limiting the current and dropping the supplied voltage “too much”, (another poorly defined term!).

The tentatively selected battery does have its own internal safety monitoring circuit, (primarily to prevent over discharge), and the tentatively selected LiPo charger chip also has circuitry to prevent overcharging.  I looked though 20-30 Analog Devices LiPo charger chips and the nice thing about the one selected is that it needs a by-pass cap and a resistor, that’s it!  K.I.S. to the max!  One can add an NTC to throttle back the charge current if the battery gets too warm, (2 more parts). 

I don’t think a long charging time will be an issue, as my flight time is ½ hr / drone battery, and I only have two batteries at the moment.  The others that I fly with also have pretty short flight times.  As the entire project is only a couple of dollars / board, I can have several of them, if needed, and Velcro them to the airframe.

Some of the charger chips have many options and features, but can require 20+ external components, and that isn’t what this project needs.

I can’t imagine flying my drone at 8500 feet above ground level!  The USA - FAA rules are a maximum altitude of 400 feet AGL.  If one is inspecting a building or a tower then one can fly up to 400 feet above that object, within a 400 foot (IIRC) radius of the building / tower.  I don’t think my little drone has the battery capacity to fly up to 8500 feet, and back down, before the battery would be drained! 

It is already a little bit nerve racking deciding how soon one has to start flying back to the starting point without running out of power before making it. 

I appreciate everyone’s input and guidance!

JC

you're not the only user of the air.

Well understood! 

Normal recreational drone pilots, and "Part 107" commercial drone pilots, are not permitted to fly at night.

When I fly for fun it is during the day.

When flying at night the drone Pilot In Command is required to have an FAA waiver, and the drone is required to have the strobe/beacon.

Even when flying at night the aircraft's altitude is maintained < 400 AGL.

Normal aircraft should be well above that.

When we are flying in the copter at night there are so many lights around it truly won't be noticed.

The copter is at several thousand feet AGL unless we are on take off or final approach.

My copter pilots obviously don't like drones around the airport or their landing zone, but otherwise its "not a factor", and they move on with their scan or other tasks.

BTW, (off topic), when I fly the drone at night I'm working, not doing it for fun.

Many drones have small LEDs on them to help make them visible, and to help with discerning the orientation of the drone.

Many of them have a small "beacon" LED to help make them visible, even  during the day time.

Especially the small ones are easy to lose track of when one looks down at the controller and then looks back up at the sky.

The controller for one of the larger drones actually includes an ADSB (transponder) receiver, so the controller software pops up a box telling you there is a C-172 X miles SW @ Y altitude.

Very cool what one can accomplish with current technology.

That would not have been fesible even 10 years ago!

All that said, Yes, I hear your concerns!

My goal is to fly safely, whether I'm in the aircraft or on the ground!

JC 

Thank you, the kind words are certainly appreciated!

can lead to some volume 

Truely I am better at design and prototyoing than at production of more than a handful of anything!

Needless to say I learned a lot on this project, and have a few improvements in mind if I elect to built another, Ver II, model.

Production would best be farmed out to another Freak!

I haven't built any PCBs this past calendar year, what with Covid keeping me pretty busy.

Perhaps I was just rusty, or perhaps I'm just getting older, but it took 2+ hrs to paste (no stencil), place, and frying pan solder the board!

Clearly I need a Pick-and-Place machine for Christmas if there should ever be any interest in a production run!

The PCB's were very inexpensive, and the key non-micro parts were small SMDs, hence I went directly from a paper design to the PCB.

I usually like to breadboard prototypes, but that didn't seem feasible for this project.

I used PCBWay.  The boards are 40 x 50 mm. It was 10 PCBs for $4 USD, plus DHL Shipping & Handling ($18 USD), and PayPal ($2 service charge), for a total of $24 USD / 10 boards.

That is 2-sided, through vias, top silk screen, color of your choice. etc.

If I could convince my wife to move to CHina I wouldn't have to pay DHL shipping and handling charges!

Weight is a key parameter for this project!

I ended up using a 500 mAh battery that is about 2 cm x 3 cm x 0.75 cm in size, including its own little LiPo safety circuit board, (that prevents over charging and over discharging the cell).

There were several sites that had 500 mAh cells listed, but they were out of stock.

Mouser had the TinyCircuits battery in stock.  Decision made.

I could easily go with a smaller battery, but I still have limited experience in how much current the smaller cells can source for the LED flashes.

The 3W LEDs can handle a continuous current of 700 mA, but that is with a big heat sink.

Needless to say, heat sinks are heavy, hence I derated the current to 150 mA, which is still incredibly bright.

Coupled with a low duty cycle on time and no heat sink is required, and the battery life per charge is markedly extended.

My test software is flashing the micro's small LED for 50 mSec, then the Driver #1 pair of LEDs for 50 mSec, pause 50 mSec, and then the Driver #2 pair for 50 mSec, wait (cool down) for 300 mSec, and then repeat.

That gives two flashes / Sec.

Ignoring the little LED on the PCB thats a 200 mSec / Sec on time, 20% duty cycle.

On its first charge it has been flashing for 6 hours so far, and the LiPo voltage has dropped from about 4.2 V to 3.8V.

My drone is a Mavic Mini, which weighs 248 grams on my scale, (0.546 lbs).

A Small UAV that weights < 0.55 lbs doesn't have to be registered, (in the USA).

The Team's drones are larger, and registered.  

The added weight for them is essentially immaterial regarding flight time.

If I want to fly it at night, or beyound visual sight, (both under an FAA waiver), it does have to be registered.

The PCB and battery weight ~ 28 grams.  

So putting it on the drone automatically puts the drone over the minimum weight limit registration exemption and also makes registration required just to add it to my drone.

(248 + 28 = 276 grams, = 0.61 lbs, which is over the 0.55 lbs limit)

It was quite a fun little "weekend" (1 month!) project!

JC

Edit: Typo