There it is. Marcy-1's first free flight on the golf course.The 900Mhz RC transmitter interferes with the telemetry receiver even though they're on opposite ends of 915Mhz. You need to keep them separated or on different bands, which requires rebuilding the entire transmitter.The notoriously fickle capture compare on the PIC is still extremely hard to get working in the field.

Salute the Mar.

By now, we realized she had a serious roll oscillation which is probably from not having enough weight on the balace beam.

Longest flight.

Last flight, ending in a tree impact.

Marcy-1, warrior copter.THE 900MHZ NIGHTMARE

Tracked down the SI4421 problem to a floating solder joint. The SI4421is electrically identical to the MRF49XA.

Dual 900Mhz receivers for Vika 1 & Marcy 1.

Ancient 72Mhz converter from 3 years ago.

900Mhz converter fits inside the case.

Much better.

900Mhz converter has resistors for all the inputs to prevent overloading.The 900Mhz conversion runs at 9600bps. Manual control now has the full60Hz of the RC transmitter. It was reduced to 50Hz when it was 72Mhz.

After being plagued by broken wires & floating solder joints, the Marcy1 ground station was finally capable of flight.

Vika 1 is also beginning a long hard 900Mhz upgrade. Since the MRF49XAhas only a 16 bit buffer being consumed at 86000bps, getting VikaCopter to output data fast enough requires buffering an entire packet on the flight computer & sending it in 1 shot instead of sending data as it's collected.MICROMAG3 RIDES AGAINSince we didn't win the free $parkfun hibache knife set & no-one knowsif Marcy-1 will work, that leaves digging the old Micromag3 out of thegadget graveyard & fabricating a new Marcy-1 flight computerspecifically for it.A 3rd revision will be necessary. Depending on traffic, these boardscan be fabricated in 1 commute & since we're tax paying renters, thechips are fully recycled.

Spent $30 to find out only Mane Clog Remover, only at Walmart works.Other drain cleaners have additives which prevent the NaOH fromdissolving photoresist. Mind you, you're supposed to use batchpcb.

Etching skills are a bit rusty after the draino experiments.

Size comparison. Went back to single sided boards because they'relighter with our methods.

This one has a simple dipole with choke inductors, no decouplingcapacitors, 1 digital line under the chip & it still works. Filling aground patch & using 3 decoupling capacitors is only necessary formaximum range.

Suspect any inductors above 8nH would do the job but these were whatthe balun had.

Marcy 1 phase 2 with magnetometer grafted on.

& this is the raw & derivative of the magnetometer output at flight RPM. Flight RPM has 144 samples per period. The limiting factor is radio bandwidth, not the magnetometer. It takes 20000bps of radio bandwidth to downlink that. Monocopters take the zero crossings of the derivative to get azimuth & angular rate, but it's not so easy in realtime.
E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones


  • For all the trouble the PCB problems caused its worth it to have a pro make the boards for you. I use ExpressPCB. For $63 I get three high quality plated through hole double sided boards in about 3 days.
    They give you free software for creating a schematic net list and pcb design. Called their MiniBoard Service
    the board is limited to 3.8x2.5 inches. All 3 boards are the same so you can share the cost with a couple friends.
  • Have you tried the Toner Transfer method to make your boards? If not, then you should look into it.
    Is a "commute" a measure of time?
    Or do you use corrosive chemicals generating toxic gas while driving on equally deadly Califoria high ways .
  • Congratulation!
    Photos very impressive, it will my wallpaper :)
    And I'm interesting SI4432 since few weeks but i cant find the time for testing them. Are you suggesting this chips for data transmission?
  • Congratulations on your first flights! For SamarEye, we put one axis of our magnetometer through an op-amp, then use a digital-in on the Arduino as a comparator to get a once-per-rev interrupt. The Interrupt Service Routine stores and resets a timer, which is our phase reference. Because the change in time from rev-to-rev is small, this is an easy way to get azimuth and angular rate.
  • @ Jack,
    :) almost 7-8 pins were lifted off the PIC that i could see from u photo.It seems the copper base it self came off , is it bad PCB quality or too much heat while soldering killed the glue on pcb ? Cool project. :)
  • Developer
    Strap on a small A10-P rocket motor... lite it remotely to get some altitude fast...
  • I find that Red Devil lye works great for developing circuit boards. I believe that it is straight NaOH and comes as a powder that is easy to store. You can just spoon it out as needed. My memory is a bit hazy but it seems like it costs less than $10 for a canister and one canister can develop a lot of PCBs. There might be variants that have additives so you have to be a little bit choosy.

    I should note that you have to be a little careful when mixing up a new batch of developer as the heat of solution for lye is very large, large enough that if you add water to the dry lye it can boil the water and splatter boiling lye everywhere. So mixing requires adding the lye to water slowly enough that it does not overheat.

    For PCB exposure I have found that a florescent tube blacklight works great. Depending on the resist it takes on the order of 45 to 90 seconds to fully expose the board. I also like sodium or ammonium persulfate for etching as it is clear enough to watch the etching process, unlike the totally opaque FeCl.
This reply was deleted.