# Thread: Glider with GPS guided recovery

1. It probably makes sense to try flying these GPS modules in a conventional rocket before using them in a glider. There's really no telling how any given commercial module will perform in the flight environment without trying it.

2. Originally Posted by Dave92F1
I'm not sure what you mean by "pitching over". It was a windy day so there was some weathercocking off the pad, which led to a non-vertical ascent. There were high Gs until motor burnout of course, but after that the rocket coasted to apogee at 0 Gs, not counting air resistance. (Or you could say it was decelerating, vertically, at 1 G due to gravity.)

After apogee there's a brief period of high G when the elevons go to glide pitch - it's moving fast so it pulls some G force until the glider slows to an equilibrium gliding speed.
For the second flight, I agree that the coast phase of the flight looked like it was pretty much ballistic. But on the first flight, it sure looks to me like throughout the upward portion of the flight, there was some significant lift that was turning the trajectory of the rocket. At those high speeds, even a gradual turn will produce relatively high Gs. Then when the apogee was detected I bet the acceleration spiked up to 30+Gs, and then the spiral had some significant acceleration, too, as it circled.

If the rocket was going in a 50-foot radius circle every 3 or 4 seconds, superimposed on the descent, that's not an insignificant acceleration there, either. Imagine how far a kid would have to lean over to run around a circle that size, that fast.

Originally Posted by Dave92F1
After that the G forces should be low - just whatever is induced by the turn. It's descending at a constant rate, so there is no acceleration there. But my suspicion is that it's not just acceleration that can confuse the GPS - it may also be simple (constant-rate, non-accelerating) vertical motion. If the GPS is designed for use in a car, it may not handle a high rate of descent well, even if the rate isn't changing (and therefore no Gs).
I thought that you set it to mode 5 (airborne low dynamics)? I would expect a mode like that to be o.k. with significant vertical velocity, but it says that it needs < 1 G.

3. Originally Posted by Adrian A
I thought that you set it to mode 5 (airborne low dynamics)? I would expect a mode like that to be o.k. with significant vertical velocity, but it says that it needs < 1 G.
It's impossible to know if that affects the actual PVT solution or just the filtering or both. Assuming that it will simply never lock in that mode is acceleration is > 1 G seems unduly pessimistic to me. But I think the only way to know for sure, short of getting the vendor to actually reveal the internal workings of the chip, is to test.

Note that it didn't lock in the default mode either, so it's just not mode 5 that has the problem.

4. Originally Posted by mikec
...

This is a great project. I've ordered the parts for a 2.6" diameter downscale that I plan to fly as a conventional RCRG, thread to follow.
Thanks. Autonomy 1 glided better than we expected. Autonomy 1 had a 3.1" diameter body tube to accommodate existing electronics.

For Autonomy 2, Dave and I are leaning towards using the same size and geometry for the wing and tail with a smaller diameter body tube to reduce drag. The next on-board electronics will likely be built on a thinner and longer circuit board, so that we can thin down the body to 2.6".

Building an RC rocketplane using the Autonomy 1 design as a starting point, performance would likely be improved by reducing the body diameter as much as on-board electronics and motor diameter will allow.

5. Originally Posted by mikec
I've ordered the parts for a 2.6" diameter downscale...
My parts arrived today. Somehow the thing seems bigger than I was expecting. I'd like to bring it in under a pound dry mass, and the body tube and balsa NC are already about half that (I used BT-80H for robustness.)

I was thinking Readiboard for the wings. Is that a stupid idea? I probably won't fly it on anything larger than an F22J.

6. Originally Posted by mikec
My parts arrived today. Somehow the thing seems bigger than I was expecting. I'd like to bring it in under a pound dry mass, and the body tube and balsa NC are already about half that (I used BT-80H for robustness.)

I was thinking Readiboard for the wings. Is that a stupid idea? I probably won't fly it on anything larger than an F22J.
I hear you as far as building for F to G power, big cost savings in that range over HP.

If your RC receiver will fit in a 29mm tube, I'd use that for body tube.

For wings and tail, could use 40-50% scale of Autonomy 1, building them from 3/16" balsa. If the balsa feels too flexible, cover with extremely thin fiberglass, or glue and paper.

I built the 1/3 scale model of Autonomy with plain 1/8" balsa, and that survived repeated hard flight and glide tests with no damage. The full sized Autonomy needed thin (1.3oz) fiberglass, because the 1/4" balsa was much too flexible across the grain.

7. Senior Member
Join Date
18th January 2009
Location
Lakeland FL
Posts
550
Originally Posted by mikec
My parts arrived today. Somehow the thing seems bigger than I was expecting. I'd like to bring it in under a pound dry mass, and the body tube and balsa NC are already about half that (I used BT-80H for robustness.)

I was thinking Readiboard for the wings. Is that a stupid idea? I probably won't fly it on anything larger than an F22J.
Readiboard will be PERFECT for the wings of this craft if you do not go over the F22J or perhaps the F25.

I have flown many, many gliders built with this stuff on mid F impulse at this size and larger. If you have the 24/40 can, you can use the F24 loads with a 2429 adaptor. Also called 'practice flyin'....

One note on using single plyed Readiboard....warpage. Try to get a FRESH sheet from the BACK of the case. It will still have the tendency to warp along the LONG side regardless. This is not bad.....if you use this on the span side of the wing planform, you get 'free' elliptical dihedral, and that helps especially if yours will be free flight. You may have to splice another small section in front since the sheet will only be 20" chord when cut in this manner.

Please resist the temptation to 'glass' or otherwise 'cover' the wing. OK, if you can iron-on Monokote that is fine, but otherwise, just save the weight. You can do something else though:

Take some 1/8" dowelling and add this to the leading edges of the wings and then tape over this with like 1/2" wide strips of chrome tape or similar. Bingo, leading edge protection and a bit of drag reduction since those edges are now 'roundy'. Also, it stiffens the wing a bit -and that never hurts, plus helps keep warpage down.

Instead of gluing the wings on each side of the body, as Autonomy 1 had, you can just make the entire Readiboard delta planform and crease it down the middle for dihedral and then glue this to the underside of the BT. That also resists warpage. Of course, try to have some dihedral if you do this, bend up the tips at least one half the BT diameter.

You can see Chris Michellsons fantastic blog (hcmbanjo on here) and see a very similar simplified glider of mine called the Astron EconoGlider under the Tampa TTRA launch post. You may just want to build that instead!

8. Senior Member
Join Date
18th January 2009
Location
Lakeland FL
Posts
550
Originally Posted by mikec
I'd like to bring it in under a pound dry mass, and the body tube and balsa NC are already about half that (I used BT-80H for robustness
If you are using a balsa 80K nose, that will weigh a LOT, and may well cause your end up CG to be pretty far forward, since the really light Readiboard will not drag the CG too much aft. This may not be bad at all, if you are doing all up RCRG, so the only deal is to watch where you are putting the bulk of the electronics weight.

If you can get your grubby mitts on an Estes PNC-80K, that might be better if you are weight conscious.

9. Originally Posted by AstronMike
If you can get your grubby mitts on an Estes PNC-80K, that might be better if you are weight conscious.
I think Apogee sells a long, pointy plastic nose cone similar to the PNC-80K.

-- Roger

As Boris said above, I'm laying out a new version of the PCB that will be narrower to fit in the 2.6" body tube.

Another change I'd like to make is to use a single LiIon cell for power (instead of 2 in series now). One cell would be lighter, and much easier to charge (no balancing issues).

I need some advice on this.

Today I use 2 x LiIon cells (7.4v) to drive the servos and also for the ignition current for the parachute ejection. I step-down that voltage to 3.3v for the logic with a Microchip TC105. (Which works great.)

But if I go to 1 x Li+ (3.6v), I'll need to step that up to 6 volts (ideally) for the servos - most of the time the servos don't draw much current, but if they're loaded heavily they can draw up to 2 Amps for short periods (while moving).

So I was thinking of doing this:

3.6v >>> step up to 6v for servos & ignition >>> step down to 3.3v for logic.

Is this a sensible plan? (The Li+ battery could be anywhere from 2.7v drained to 4.2v when charging.)

I was thinking of the following parts for the step-up:

LT1619
LT3579
LTC1871
MAX1771
MAX8625
TPS63020 (only goes to 5.5v, but requires few external parts)

I could use some advice, or a pointer in the right direction. I'm a software guy, not a EE.

Cheers,

--Dave
Last edited by Dave92F1; 30th November 2011 at 11:46 PM.

11. Originally Posted by Dave92F1
I'm a software guy, not a EE.
I'm not an EE either, but I'm not sure I would mess with it because of potential EMI and max current. I considered this for my RCRG and decided I would just use a small 2S cell.

If you do use a 1S, consider a COTS solution like http://www.dimensionengineering.com/lvboost.htm

Some people run 5V servos off 1S, but this seems like an even worse idea.

12. Originally Posted by mikec
I'm not sure I would mess with it because of potential EMI and max current.
That's kind of where I am. At first I thought it'd be nice to use 1S, but it started to look more complicated than I expected.

I'm hoping somebody out there will say "Oh, that's easy....".

I guess an alternative is to find a way to charge both cells off the 5v USB input. But that looks kind of complicated, too.

13. Originally Posted by Dave92F1
I guess an alternative is to find a way to charge both cells off the 5v USB input. But that looks kind of complicated, too.
A cheap 2S balancing charger is only \$10 or so. Buy an RC plane or heli that uses 2S and you get one for free

Actually, I'm surprised you don't fry your servos running straight from 2S. Many aren't rated to work above 6V, some not even above 5 or so.

14. Originally Posted by mikec
A cheap 2S balancing charger is only \$10 or so. Buy an RC plane or heli that uses 2S and you get one for free

Actually, I'm surprised you don't fry your servos running straight from 2S. Many aren't rated to work above 6V, some not even above 5 or so.
What do RC folks usually do to drop voltage from a 2s LiPo for a 4.8-6.0V servo?

Does the RC Rx electronics drop the voltage?
Like for example: http://www.servocity.com/html/optima_6_2_4ghz_rx.html

15. A cheap 2S balancing charger is only \$10 or so. Buy an RC plane or heli that uses 2S and you get one for free [/QUOTE]

I know how to charge it... http://nerdfever.com/?p=896

But charging 2 batteries is twice as much work as charging one. (And I'm a perfectionist, which is the real reason.)

Originally Posted by mikec
Actually, I'm surprised you don't fry your servos running straight from 2S. Many aren't rated to work above 6V, some not even above 5 or so.

Most every spec has a healthy fudge factor built in; unless you're unlucky enough to get the 1 part in 1000 that is the worst-case unit, things will usually work well past the spec numbers.

16. Originally Posted by delta22
What do RC folks usually do to drop voltage from a 2s LiPo for a 4.8-6.0V servo?
Usually in a powered model the motor speed control has a "battery eliminator circuit" (BEC) that steps the battery voltage down to 5V or so. In gliders people either use that or NIMH receiver packs that run at 4.8V. Or you can use a separate BEC. Most receivers just loop whatever the input voltage is out to the servos.

17. Since you're making your own board, it's simple to make your own BEC by just using a linear regulator to step it down to 5.5V or 6V for the servos. Get one in a big package like TO-25 so that it can handle the current. It shouldn't cost more than a buck or two and take about a square inch of board area.

18. Originally Posted by Dave92F1
But charging 2 batteries is twice as much work as charging one.
Only if your charger can't charge more than one cell at a time, but if you don't want to buy a charger...

All I know about servo overvoltage is that some fast servos used for tail rotor pitch control on helis are rated to 5V or less, and those who use 6V servos elsewhere use a step-down diode so the tail rotor servo doesn't exceed spec. People claim to have problems otherwise, but who knows for sure? You're right about margins, and you're using high-quality servos.

19. Originally Posted by delta22

Does the RC Rx electronics drop the voltage?
Like for example: http://www.servocity.com/html/optima_6_2_4ghz_rx.html
No. All of the 3-pin servo connections are bussed together. You can power the receiver from any of them.

20. ## Charging & balancing 2 LiIon cells

OK, I've given up on the idea of running the system off a single Li+ cell.

I found a part that would supply enough current (the MAX1709), but it needs lots of low-ESR capacitors and the solution ends up taking up almost half my board space. So I conclude that's not a good solution.

So - instead - I'd like to find a simple way to charge both cells in the 2S pack off the +5v USB input while balancing them at the same time.

I see there are lots of chips that will charge a single LiIon cell, or 2 cells in series (without balancing them).

But I haven't been able to figure out how to charge both cells independently (so they'll be balanced) from a single supply (+5v USB supply).

Right now I charge them on the workbench using 2 separate power supplies, but I'd like to be able to charge them just by plugging in a USB cable.

21. Originally Posted by Dave92F1
OK, I've given up on the idea of running the system off a single Li+ cell.

I found a part that would supply enough current (the MAX1709), but it needs lots of low-ESR capacitors and the solution ends up taking up almost half my board space. So I conclude that's not a good solution.

So - instead - I'd like to find a simple way to charge both cells in the 2S pack off the +5v USB input while balancing them at the same time.

I see there are lots of chips that will charge a single LiIon cell, or 2 cells in series (without balancing them).

But I haven't been able to figure out how to charge both cells independently (so they'll be balanced) from a single supply (+5v USB supply).

Right now I charge them on the workbench using 2 separate power supplies, but I'd like to be able to charge them just by plugging in a USB cable.

It could be possible with a bunch of on-board switching to put the two cells in parallel to charge them and then put them in series to run them. But I don't think it's worth the trouble. Just get a 2-series lipo with a normal 3-pin connector, and plug it into a dedicated charger (like the hobbyzone one for the Radian sailplane) when you want to charge it. USB charging works great for single-cell lipos, but not when you need 2 cells in series.

If you expand the different RC flying you do, it will eventually pay off to get a full-featured battery charger that can deal with packs that have up to 5 or 6 cells in series and a variety of chemistries.

22. Originally Posted by Dave92F1
Good question. I looked at a bunch of Maxim datasheets and appnotes but didn't find anything that would balance multiple cells in parallel. Maybe it's not critical for low power-draw applications.

I opened up one of my cheap RC chargers but couldn't immediately identify the chips. It also has a bunch of large magnetics and a DC-DC converter, I think, so it is not that simple.

If it were me I would just charge the battery externally.

(I did run across TI SLUA463, "Cell Balancing With the bq77PL900", but didn't read it in detail.)
Last edited by mikec; 1st December 2011 at 06:26 PM.

23. Originally Posted by Adrian A
Just get a 2-series lipo with a normal 3-pin connector, and plug it into a dedicated charger (like the hobbyzone one for the Radian sailplane) when you want to charge it.
That's exactly how I've been doing it so far, except I use bench power supplies instead of a dedicated charger.

I used to have a dedicated RC charger and never used it - it was just easier to use the power supplies I had on the workbench. I eventually sold it on eBay in a fit of anti-clutter.

24. Senior Member
Join Date
18th January 2009
Location
Austria
Posts
733
Some servos can be directly run from 2s lipos. Besides the gained simplicity, these servos usually perform better than standard servos. On the downside, they are more expensive and not as widely available.

If there is an PWM capable IO pin left on the micro controller, it can be used to drive a simple step up converter to charge the 2s battery. An example for this approach can be seen here:
http://www.atmel.com/dyn/resources/p...ts/doc1659.pdf
The whole charging strategy is controlled via software. I guess the current shunt can be omitted, if the charging current is limited by design to safe values.
To balance the cells, a precise voltage divider and an operational amplifier, capable of driving "high" currents (10s of mA maybe?) should be sufficient. Some way to disconnect the balancer is likely necessary too, otherwise it will drain one of the cells all the time, due to tolerances, noise etc. involved. An alternative approach with some kind of hysteresis involved would be more elegant (eg. monitoring and selective discharge, controlled via software).

Reinhard

25. You guys could always just use a "sacrificial lawn dart" - I don't know, a \$10 Estes kit - to test your hypothesis about the GPS and it's supposed inability to read fast descents because it's made for automotive use.

Put a non-ejecting, 1st stage motor into a cheap throwaway LPR (a C6-0 perhaps?) to see if the speed descent is causing your GPS challenges.

Would this be in violation of your club policy?
Last edited by rifleshooter; 19th December 2011 at 10:26 PM.

26. Originally Posted by delta22
Analysis of the telemetry showed that the GPS had lost lock again...
How did this manifest? Did you just get GGA messages with a fix status of 0?

I recently flew an MT3329-based GPS and seemingly had it lock up sending the same incorrect coordinates until it was rebooted, but apparently with fix status of 2D at least. See the last few messages in http://www.rocketryplanet.com/forums...d.php?p=193340

27. Originally Posted by mikec
How did this manifest? Did you just get GGA messages with a fix status of 0?

I recently flew an MT3329-based GPS and seemingly had it lock up sending the same incorrect coordinates until it was rebooted, but apparently with fix status of 2D at least. See the last few messages in http://www.rocketryplanet.com/forums...d.php?p=193340
All I know is that the "good fix" flag in the \$GPRMC message was false.

However the GPS never required rebooting - once the glider landed, the fixes were good.

I've got a new model of the PA6 GPS (PA6C) that I put on a board this winter - it is _much_ more sensitive than the older one (gets a fix indoors right away, every time).

I'm hoping it will be better.

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•