Hi Rob.
Every part of metal that is exposed to seawater will consume Volts/amps/watts.
Keep as little as possible exposed. Put heatshrink or varnish on them.
So yes you can also use them for lights.
Regards Antony.

I did a test fit of every thing into the housing, I will not be able to shorten the tube by much, so now considering having the batteries in a separate box below the main housing, something like an otterbox which are rated to 100ft. Hmmm things to think about.

well i had pretty much made up my mind that i would go with the otterbox setup but after pricing lipo batteries even from hobbyking delivery to UK makes it a costly option, plus then the otterbox, so for initial tests i'll use two 2200mAh 11.1v lipos in parallel. just using the two means I should be able to stack them above the control circuit and i'll be able to shorten the body tube to a length i like.
You probably notice am 'winging' it a bit, would be nice to have a clear plan and stick to it I have loads of plans for a version 2, but it would require careful planning and a budget

Rob

The budget is part of the fun.

been chipping away and trying to progress with the project, made up an inner frame to hold the electronics in place.
I am going to use a magnet switch to turn the electronics on/off via a relay without having to open the hull.

Hi Rob.
I have used this circuit in one of my RoV's with success.
It's a Latching Relay Circuit.
From http://www.eleinmec.com/article.asp?24
This circuit uses NO power when not energised.
I used this circuit to drive an 100 amp relay. Which is the main power.
Regards Antony

well just thought I'd give an update for anyone that was following this build I know its annoying when threads just die out.

What happened?? well everything was great until i tried to test my 1wire communication over the umbilical, it wouldn't work!!! tried boosting signal etc but i was stumped and deflated.

What next, well I plan to recode the comms with rs232 protocol as its tried and tested but just need to get motivated again, assembly programming can be a chore

you never know, maybe this short update will be enough to spur me on

Rob

well let me try to encourage you...
great build! I especially like your brass feed throughs.
I have just started my 1st rov build and I'm already planning a version 2.
funny how that happens.
get yours up and running and then post more pics.
people are trying to live vicariously here!

I just noticed this thread and have quickly scanned trough it . Very nice build.

Rather than RS232 for the comms, what about RS485? I've done a lot of 2-wire 485 using Sipex SP485 chips. You can get decent baud rates and it's reliable over long (low-spec) twisted pair cables. A 4 wire system (2 SP485 chips at each end) would just drop in replacing the RS232 chips and not cause any protcol problems.

I've done a lot of Dallas 1-wire support and had it working over 30 meters or more, but it's difficult to get reliable and changes in the cable cause problems.

I also do a lot of PIC assembly programming. Many years experience with 16 series and now also using 18F chips; nothing bigger. I've got a lot of code/projects created using the Cross32 assembler on Win-XP over the years. I never really used MPLAB except to drive the programmers. I'm now using MPLAB-X on LInux; big learning curve, rubbish programmer support, but I like it.

Paul

Hi Paul

Yes i will probilily use 485 chips to send over the umbilical, must say am really struggling to motivate to finish programming as I only have a basic knowledge. The chips am using are 16f628a in the surface controller and 16f844 ( i think) in the rov. Assembly is turning into such a chore to work with.
Can I ask a question, I want to send 8 bytes down to the rov and receive 3 bytes back, what is the best way to do this? can i do it with 3 wires tx/rx/gnd? The data transfer is repeated.
I dont want to have either pic sitting waiting on each other for data as the surface controller is also talking to a ps2 controller and the rov controller is driving servos.
Any tips welcome

Rob

Originally Posted by robe_uk

Hi Paul

Yes i will probilily use 485 chips to send over the umbilical, must say am really struggling to motivate to finish programming as I only have a basic knowledge. The chips am using are 16f628a in the surface controller and 16f844 ( i think) in the rov. Assembly is turning into such a chore to work with.
Can I ask a question, I want to send 8 bytes down to the rov and receive 3 bytes back, what is the best way to do this? can i do it with 3 wires tx/rx/gnd? The data transfer is repeated.
I dont want to have either pic sitting waiting on each other for data as the surface controller is also talking to a ps2 controller and the rov controller is driving servos.
Any tips welcome

Rob

I used the 16F628 a long time ago; now I mostly use the 16F684 and 18F2525. I don't find assembly too difficult to follow, but I have a structure I've been using for over 10 years and write a lot of comments.

I'd recommend a 5-core comms; gnd/D+/D-/D+/D-. And I guess you will need some power, or is it battery on-board? The two D+/D- pairs are for TX and RX of RS485. You can do half duplex over one pair but that involves a bit more complexity in direction control.

For this type of thing I put the data into a simple packet with header and checksum. Have a couple of buffers in RAM (TX and RX), 12 bytes each should do and a few bytes per buffer to track byte count, buffer position and checksum. Interrupts and a service routine to each RX and TX. Some flags to indicate status; TX or RX in progress, complete, failed etc.

With 4 wire comms this can TX and RX at the same time. The small PIC UART is not double buffered so you have to save the received byte pretty quickly or you need gaps between bytes transmitted.

I'll have a look through the archives and find some example code to post.

I haven't used the UART/232 to connect PICs before but have used it to connect to PCs. I have connected between PIC chips using a clock/data arrangement, which is much faster but not suited to long distance. But it should demo the comms buffer arrangement and interrupts.

This is a PIC LCD driver that I used to use with HC11 chips. I haven't made any in long time but I'm currently porting the old HC11 code to connect a LCD and buttons to a PIC18F2525 board.

Paul

Here is some example PIC16F876 RS232 code using buffers and interrupts. It should convert easily to any other PIC16F chip that has a hardware UART.
There are 4 files that you can view or save.

http://www2.pmb.co.nz/blog/work-misc...-code-example/

Paul

Thanks for the info Paul, I have some reading to do.

Rob

Hi all,
Well I've been busy working on the ROV again and I have some good news, finally been able to get the comms working over the umbilical, still using the one wire protocol although I have slowed it down slightly which is not a problem as there is only about 10bytes being exchanged at a time.
Also used a bus extender which has allowed me to transmit over 50meters.
Also coded in timer interrupts into the data exchange routines so that if noise causes the comms to fail the controllers will not get stuck in a loop and hang up. I have a LED on the ROV control which shows if there was a comms error and it does light from time to time.

So as you can imagine am quite happy now that the basic controls are working and the hardware is pretty much complete. BUT, always a but, there is always loads to do;
1) need to code in some fail safes into the ROV controller so that if there is a total loss of comms the ROV will surface.
2)Some form of leak checking
3)lights
4,5,6..........

Also need some form of video signal boosting as the signal is distorted after going through the umbilical.


So just letting you all know the projects not dead yet

Rob

I have completed another step, I filled up or recycle wheelie bin with water, about 4ft deep and put the rov in, it didn't float level, heavy at the back with the brass feed throughs so I cable tied some lead to the front, once leveled the rov sunk slowly to the bottom and I used the vertical thrusters to rise it up, the rov rose nice and level and the vectored thrusters also worked a treat.
I will have to add some flotation to the rear rather than weight to the front and I think the rovshould float then.
Sorry no videos but will do for next test,

Rob