Discussion New Quadrocopter and Tricopter UAVP updated with GPS and more - UAVX!

[jesolins]

Thanks Narpat.
Greg and I actually do have lives outside of this project too. Lot's of fun ahead
Cheers,
Jim

Originally Posted by narpat007

Hi Jim !
Thanks a lot for the reply on behalf of Greg !
I shall be waiting desparately for the next release !
Jim, my official life does not allow me to devote that much of time and commitment as that of Prof. Greg and you.
I find time only on Sundays and not even on Saturdays as we have Saturday working.
I am heading Network Operations of the entire City of Gurgaon of a big Telecom Service Provider.
Wish you and Greg all the best for the next release !
Thanks & Best Regards
narpat

[dleroi]

I finally bit the bullet and built a proper circuit board for the GPS. It has everything on it, including a MAX232, so that it can be plugged directly into the PC (using the same cable as the UAVX) for checking and changing the configuration while it's still installed in the quad. The battery gets recharged whenever the quad is powered up.

Cheers,
Don

[jesolins]

Wow Don. Beautiful job!
The "arm" is for an LED GPS lock indicator I presume?
Cheers,
Jim

Originally Posted by dleroi

I finally bit the bullet and built a proper circuit board for the GPS. It has everything on it, including a MAX232, so that it can be plugged directly into the PC (using the same cable as the UAVX) for checking and changing the configuration while it's still installed in the quad. The battery gets recharged whenever the quad is powered up.
Cheers,
Don

[dleroi]

Originally Posted by jesolins

Wow Don. Beautiful job!
The "arm" is for an LED GPS lock indicator I presume?
Cheers,
Jim

Hi Jim,

Thanks!

"Arm" is for the arming switch. I have a 2-pin pigtail coming off the appropriate side of the arming switch, then a straight cable between the GPS board and K-19.

The only way I know to detect a lock is to read the sentences, which would require adding a processor. That's actually something I've been meaning to discuss with Greg. That is, having the GPS board process the NMEA sentences and pass on only the desired fields to the FC. This is something I'm doing for another project. The idea is to take the burden off of the main processor.

Cheers,
Don

[ta2cba]

Originally Posted by dleroi

Hi Jim,

Thanks!

"Arm" is for the arming switch. I have a 2-pin pigtail coming off the appropriate side of the arming switch, then a straight cable between the GPS board and K-19.

The only way I know to detect a lock is to read the sentences, which would require adding a processor. That's actually something I've been meaning to discuss with Greg. That is, having the GPS board process the NMEA sentences and pass on only the desired fields to the FC. This is something I'm doing for another project. The idea is to take the burden off of the main processor.

Cheers,
Don

Hi Don,

What a neat job,looks great! I wonder if you had a chance to build and test the hand terminal?
To have a second processor to take some of the burden off the main processor is an excellent idea.Actually I had a bit further dream which I wanted to sound for sometime to Greg and Jim but put myself back as it is a complete remake of the UAVX hardware/software. My idea was to add a second processor and take all the sensors,gyro,ACC,Baro,compass and GPS as well as battery monitor from the main processor and load onto the second processor or so called co processor and let this one to process all sensory data info and then pass the refined data to the main processor via a fast two line interface such as I2C or SPI when asked. This might take the load from the main proc which can just concentrate on rxing the data from R/C and the co processor and concentrate on PID algorithms.

Obviously this will need a new pcb and remake of the software and might be a far fetch. Just a thought

I still feel that as the UAVX is far more advanced then the original UAVP may be it's time for a new pcb layout with all the mods,level translators and PSU's taken on board.Also I think sticking to 16 Mhz xtal is a downside with 18F2620 which can go up to 40,48 Mhz easily just by changing xtal which can considerably speed up the processing speed.A UAVX NG may be!

Cheers,

ba

[dleroi]

Originally Posted by ta2cba

Hi Don,

What a neat job,looks great! I wonder if you had a chance to build and test the hand terminal?

To have a second processor to take some of the burden off the main processor is an excellent idea.Actually I had a bit further dream which I wanted to sound for sometime to Greg and Jim but put myself back as it is a complete remake of the UAVX hardware/software. My idea was to add a second processor and take all the sensors,gyro,ACC,Baro,compass and GPS as well as battery monitor from the main processor and load onto the second processor or so called co processor and let this one to process all sensory data info and then pass the refined data to the main processor via a fast two line interface such as I2C or SPI when asked. This might take the load from the main proc which can just concentrate on rxing the data from R/C and the co processor and concentrate on PID algorithms.

Obviously this will need a new pcb and remake of the software and might be a far fetch. Just a thought

I still feel that as the UAVX is far more advanced then the original UAVP may be it's time for a new pcb layout with all the mods,level translators and PSU's taken on board.Also I think sticking to 16 Mhz xtal is a downside with 18F2620 which can go up to 40,48 Mhz easily just by changing xtal which can considerably speed up the processing speed.A UAVX NG may be!

Cheers,

ba

Hi ba,

Thank you.

I have not yet built the terminal. I have a few other projects in the pipeline that I have to complete, first.

I (and others) have also thought it a good idea to have a processor dedicated to the sensors and I think it could be done without requiring a new main board, rather a daughter board for the new processor and sensors. The daughter board would communicate via one or more of the existing sensor connectors.

We should discuss this further.

Cheers,
Don

[ta2cba]

Originally Posted by dleroi

Hi ba,

Thank you.

I have not yet built the terminal. I have a few other projects in the pipeline that I have to complete, first.

I (and others) have also thought it a good idea to have a processor dedicated to the sensors and I think it could be done without requiring a new main board, rather a daughter board for the new processor and sensors. The daughter board would communicate via one or more of the existing sensor connectors.

We should discuss this further.

Cheers,
Don

Hi Don,

Actually you are right,there is no need for a new main board at all.A daughter board to accompany all sensors, level converters,GPS, with a new co-processor and also may be a decent 5 volt and 3.3 volt supply can be a perfect upgrade for UAVX without throwing the main board we used to fly which can be piggy back on top of the main board with a couple of spacers. Sure everything can be tapped in on the main board from old sensor points like for I2C for main proc-co proc comm,+5V and 3.3V.

Don I think an upgrade like this would be a significant improvement to the existing UAVP/UAVX hardware.This can also wrap up some of the mod issues also,like the PSU's,filtering to yaw gyro etc. I will be more than happy to help to develop this upgrade if Greg, Jim and all agree to give it a go.

Cheers,

ba

[jesolins]

Gents,
I love your guys enthusiasm and ingenuity! Ba, Greg has started re-writing (or might have already finished) the code for 40Hz to take advantage of the faster 18F2620 processing. He wanted to get the new features of his code matured before he changed went to that version it as it will then require the replacement of the $1 16Hz crystal to a 10Hz. A simple mod for most of us.
The daughter boards seem to be the way most are going, but Greg's focus todate has been to keep things as simple to mod from the original UAVP so that others can easily follow. We can certainly make a daughterboard with that in mind, but the beauty of what he has accomplished now is that no other quad has come up with the features he has implemented with his efficient coding on one board. If we can continue the KISS method with the board(s), it will remain the best flying and most motor/ESC/device/Tx/Rx/component adaptable quad for the price. Keep those great ideas coming!
BTW, here is an interesting IMU mod done by making the acc, and gyro boards fit to a square with a connector. I'm not sure it would be any better than bridging the boards with balsa to keep them rigid, or where the best place to mount it would be, but it looks worthy

Cheers,
Jim

Originally Posted by dleroi

Hi ba,
Thank you.
I have not yet built the terminal. I have a few other projects in the pipeline that I have to complete, first.
I (and others) have also thought it a good idea to have a processor dedicated to the sensors and I think it could be done without requiring a new main board, rather a daughter board for the new processor and sensors. The daughter board would communicate via one or more of the existing sensor connectors.
We should discuss this further.
Cheers,
Don

[ta2cba]

Originally Posted by jesolins

Gents,
I love your guys enthusiasm and ingenuity! Ba, Greg has started re-writing (or might have already finished) the code for 40Hz to take advantage of the faster 18F2620 processing. He wanted to get the new features of his code matured before he changed went to that version it as it will then require the replacement of the $1 16Hz crystal to a 10Hz. A simple mod for most of us.
The daughter boards seem to be the way most are going, but Greg's focus todate has been to keep things as simple to mod from the original UAVP so that others can easily follow. We can certainly make a daughterboard with that in mind, but the beauty of what he has accomplished now is that no other quad has come up with the features he has implemented with his efficient coding on one board. If we can continue the KISS method with the board(s), it will remain the best flying and most motor/ESC/device/Tx/Rx/component adaptable quad for the price. Keep those great ideas coming!
BTW, here is an interesting IMU mod done by making the acc, and gyro boards fit to a square with a connector. I'm not sure it would be any better than bridging the boards with balsa to keep them rigid, or where the best place to mount it would be, but it looks worthy

Cheers,
Jim

Hi Jim,

I think UAVP is having a giant leap fwd with Greg and your amazing work.If we can carry it one step further with this daughter board add on with versatility and simplicity, as you pointed out UAVX will be the ultimate adaptable and versatile quad to fly.

I feel that the daughter board can still be kept in KISS format with minimal connection to main board. Ofcourse it will initially put some extra work on Greg to extract all sensor subroutines and develop the software for the daughter board processor but I think that will be a one time effort as the daughter board firmware will stay unchanged after released as long as no sensor additions. pretty much the sensors that can be added are settled at this stage, while the main processor can continue to evolve to perfect the flight attitude characteristics.

Here is another idea! How about if we keep the main board untouched as all the sensors etc are already on that board and treat the main board processor as the co processor for the sensor and make it responsible for all the inferior work to be done such as sensor reading,RX reading,keeping up with the GPS etc. and device a small daughter board with a PIC18F2620 as the main flite processor which can be hooked to the I2C lines of the main board plus a +5V and gnd,4 connections altogether. The only problem with this idea is that it won't be as versatile as having a daughter board besides co processing also addressing all the powering and filtering mods we are conducting on the main board.

Cheers,

ba

[dleroi]

Originally Posted by ta2cba

Here is another idea! How about if we keep the main board untouched as all the sensors etc are already on that board and treat the main board processor as the co processor for the sensor and make it responsible for all the inferior work to be done such as sensor reading,RX reading,keeping up with the GPS etc. and device a small daughter board with a PIC18F2620 as the main flite processor which can be hooked to the I2C lines of the main board plus a +5V and gnd,4 connections altogether. The only problem with this idea is that it won't be as versatile as having a daughter board besides co processing also addressing all the powering and filtering mods we are conducting on the main board.

Hi ba,

One of the problems I have with leaving the sensors on the main board is that there are at least 3 different hardware/software/power supply schemes and lots of compromises to accommodate the various sensors -it's ugly.


@Jim, I like that IMU concept!

- Don

[ta2cba]

Originally Posted by dleroi

Hi ba,

One of the problems I have with leaving the sensors on the main board is that there are at least 3 different hardware/software/power supply schemes and lots of compromises to accommodate the various sensors -it's ugly.


@Jim, I like that IMU concept!

- Don

Hi Don,

You are so right with this.I think the best approach would be a daughter board which can also simplify this different power supply schemes if the +5V and 3.3V regulators can be included also on that board.carrying over just the lipo power with a decent filtering can sort out the problem.

@Jim,

I also like that IMU.What a neat work! Here is another robust and compact approach for IMU.

http://www.sparkfun.com/commerce/pro...oducts_id=9431

Cheers,

ba

[jesolins]

Ba,
A flat IMU is even better in my book. Nice!
Cheers,

Originally Posted by ta2cba

...
@Jim,
I also like that IMU.What a neat work! Here is another robust and compact approach for IMU.
http://www.sparkfun.com/commerce/pro...oducts_id=9431
Cheers,
ba

[ta2cba]

Originally Posted by jesolins

Ba,
A flat IMU is even better in my book. Nice!
Cheers,

Jim you are right, a flat IMU is perfect for fixing and rigidity. I ordered one of these to play with it to see it's performance against ADXRS610.This IMU is relatively cheaper altogether when compared to the gyros and acc used in UAVP,I wonder the performance is satisfactory also.

Cheers,

ba

[jesolins]

Ba! Great! Looking forward to the report. So you are using the ADRXS610's that the MK uses? I thought most UAVP folks were using the MLX90609's now? I am still using the ADRXS300's.
Cheers,
Jim

Originally Posted by ta2cba

Jim you are right, a flat IMU is perfect for fixing and rigidity. I ordered one of these to play with it to see it's performance against ADXRS610.This IMU is relatively cheaper altogether when compared to the gyros and acc used in UAVP,I wonder the performance is satisfactory also.
Cheers,
ba

[RCvertt]

Originally Posted by ta2cba

Jim you are right, a flat IMU is perfect for fixing and rigidity...

If these sensors work out then there is an added benefit that the sensors are even less expensive if purchased on separate breakout boards, for those that want to go the cheaper cube route.

Best of both worlds possibly.