PDA

View Full Version : A little propeller board



Graham Stabler
08-15-2011, 09:27 AM
I've been meaning to post a pic of this for a while, something I developed as part of my research. Propeller, eeprom, two h-bridges and buck-boost converter (to run from lipoly 3-4.2v in 3.3v out) and the somewhat more application specific magnetic encoder on homemade flex PCB.

Double sided PCB with wire for vias but components on the top only. Soldering the dual mosfets was a real pain but got there in the end. I think they are rated for something like 7A with suitable PCB but I will run sub 1amp.

Graham

84054

Cluso99
08-15-2011, 09:39 AM
Nice and tiny Graham. What is the size?

Graham Stabler
08-15-2011, 09:41 AM
There is a scale on the picture, about 30mm X 30mm, ~1.2"X1.2"

Franklin
08-15-2011, 03:57 PM
Which chip did you use for the voltage converter?

Graham Stabler
08-15-2011, 04:01 PM
It is a TPS63001 by TI, another pain to solder :)

Sapieha
08-15-2011, 04:03 PM
Hi Graham.

Nice board.

What type of driver IC's -- that ones 6 pin's You use to drive H-Bridges!

Graham Stabler
08-15-2011, 04:08 PM
Sapieha, I don't think it is a proper solution for high switching frequencies, I took inspiration from the output drives of small brushless speed controllers as my main hope was a light weight development system. So they are just dual transistors, the pull gate down to ground otherwise it is pulled up by a resistor. The part is MUN5211DW1T1G

This thing is supposed to go in flying machines if I work out the kinks.

Graham

Sapieha
08-15-2011, 04:21 PM
Hi Graham.

Thanks

I don't need big frequencies to.

Zap-o
08-15-2011, 04:40 PM
Nice board ! Can I ask what the part number for the xtal is? I have been looking for a xtal smaller than what I am using and that looks like the one.

Graham Stabler
08-15-2011, 04:44 PM
Zap, I can't remember if I nabbed the part number from Phil's (PhiPi) propeller backpack (diagram on parallax page for it) or if I just selected the first small crystal that seemed to meet the specs roughly. In any case the part is HCX-5SB at least that is the range it comes from.

I definitely stole the idea of the array of pull up resistors for programming from Phil!

Graham

Graham Stabler
08-15-2011, 04:49 PM
May be of some interest, or at least someone can tell me how bad it is.

84067
84068

Phil Pilgrim (PhiPi)
08-15-2011, 04:57 PM
Graham,

I know you've been studying flying-insect kinematics for a long time. Is this the inevitable next stage? IOW, do the dual H-bridges drive the wings of a radio-controlled moth? :)

-Phil

prof_braino
08-15-2011, 05:15 PM
This thing is supposed to go in flying machines if I work out the kinks.

Would this be suitable for something like Festo robo bird that erco posted?

http://forums.parallax.com/showthread.php?133653-Robo-Bird

I've been wanting one of these since I was ten years old, but never got the wing motion right. Seems so simple now that someone else has done it.

Lawson
08-16-2011, 06:00 AM
My only major nit-pick is that the prop and Fet drivers should connect to power ground directly under the FETs and nowhere else. This will insure that transient voltages on the power ground trace have a minimal effect on the FET gate drive voltage. Worst case, you might be able to fry the gate drive transistors. Looks like you're driving the H-bridge in a CMOS inverter configuration, I assume you've picked the FETs so that shoot-through (both FETs half on shorting the power rails) current is low? (unlikely to be a problem with most FETs and 3.3v on the gates) Any reason you're not direct driving the FETs with a prop pin? It's likely to switch faster, and work even with variations in the input voltage.

Lawson

Graham Stabler
08-16-2011, 10:05 AM
Phil, not a moth but a giant hoverfly.

Prof, nothing like that. BTW, what Festo did is very similar to Holst and Hertzog, check the link at the top of this page: http://www.indoor.flyer.co.uk/buzzard.htm This page shows my attempt at a model with actively twisting wings a la Holst. Notice that I also tested mine on a ring that goes around and around like Festo did :) I did this 6 years ago.

Lawson, not sure I understand the grounding issue but seems to work. The high side mosfet needs the battery voltage to turn it off hence I can't drive from the propeller pin at 3.3v.

Graham

Lawson
08-16-2011, 03:47 PM
Lawson, not sure I understand the grounding issue but seems to work. The high side mosfet needs the battery voltage to turn it off hence I can't drive from the propeller pin at 3.3v.

Graham

The grounding issue is something I've seen in other higher power H-bridge designs. Basically, the power ground trace acts as a tiny inductor so when the input current changes due to the fets switching, a voltage is induced along the trace. In higher power/speed situations this voltage can get high enough to fry the gate or driver if logic ground isn't pulled from under the Fet gates. (in your case, power and power-ground are tightly paired which helps a lot by lowering inductance)

For the Fets I'd have to look at the Idd-Vgs curves to be sure, but the p-fet may be off enough with -0.9v on it's gate (fresh charged Li-po) to work fine. Similarly, with a dead Li-po and 0.1 Vgs the p-fet will just be more "off" than usual. I've also had good success with a capacitive level shifter to drive fets. A capacitor goes between the prop pin and the Fet, then two resistors and a diode are used to define the start up state and voltage range of the level-shifter's output. (ironically this likely takes more space than your solution...)

Lawson