HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Oct 07, 2012, 04:30 PM
Registered User
Crashed22's Avatar
United States, ID, Boise
Joined Mar 2012
162 Posts
Sure, iif you match them up to this...
Crashed22 is offline Find More Posts by Crashed22
Reply With Quote
Sign up now
to remove ads between posts
Old Oct 07, 2012, 05:09 PM
Registered User
Austin, TX
Joined Jan 2009
741 Posts
I just got another G-OSD unit (with a working GPS unit...my old parts bin one had the GPS RX dead)....

Finally figured out how to get Truglodites version(kudos to Kev) to compile with 0 diffs from his hex, made a few mods in the config, flashed it and drove around a bit with it to test it out....seemed to work pretty well...speed matched the speedo in my car....home arrow was a little slow on the tracking, but worked...distance seemed okay...altitude was off by the end of the trip by 30 feet.. dont know if I have to use the fudge factor or not for that to be fixed...

A couple of questions:
1..When the statistics showed...the max ALT value shown was just the text, "ALT"....anyone know what would cause that?

2.Just to confirm my guess, but home circle with an X means that there is not enough ground motion to determine direction and therefore bearing to home? what are the other symbols (circle in circle...can't remember when I saw that)? a quide to those would be nice...

3. can I easily increase the font size for the top line ( I only use V1, not RSSI, V2 or GPS coordinates)

4. Finally, from the config.h, it implies that there is a % scale somehow for V1....I uncommented the line,SENSOR_BATTERY_PERCENTAGE_ENABLED, but nothing displays..

thanks
Atx_Heli is offline Find More Posts by Atx_Heli
Reply With Quote
Old Oct 08, 2012, 11:33 AM
Registered User
Joined Dec 2011
138 Posts
Quote:
Originally Posted by Atx_Heli View Post
...my old parts bin one had the GPS RX dead)....
...

altitude was off by the end of the trip by 30 feet.. dont know if I have to use the fudge factor or not for that to be fixed...

...

2.Just to confirm my guess, but home circle with an X means that there is not enough ground motion to determine direction and therefore bearing to home? what are the other symbols (circle in circle...can't remember when I saw that)? a quide to those would be nice...
i'll take a stab at some one them. the home circle X means that you are either moving too slow to establish direction, OR you've lost all your satellites. the circle in circle means plane is right above home position IIRC.

the fudge factor is for correcting *absolute* errors plaguing many of these gps:es. having the *relative* height be off by 30 feet, well, that just happens sometimes. the only way one can possibly mitigate it is by making sure that the initial home position be captured more accurately, e.g. by upping the # of satellites necessary. this, of course, will slow down the home position accuring process.

finally, i know this was not your question, but did you bother to tear off and check the wiring harness of your original "dead" GPS module? the reason i am asking is that mine was DOA too. turned out the GPS was fine but the HK monkeys had botched the harness. it was a 5 minute soldering job to fix it.
fpvwannabe is offline Find More Posts by fpvwannabe
Reply With Quote
Old Oct 08, 2012, 01:15 PM
Registered Aircraft Offender
Truglodite's Avatar
Carmichael, CA
Joined Feb 2007
3,508 Posts
I should have removed all the code for the altitude fudge. Setting home calibrates the altitude to zero anyways. So fudging altitude is straight up a waste of time & space, unless you don't set home.

I think you are correct regarding the # of satellites. Zero altitude is stored when we set home, and we have more error with less satellites. I haven't measured the actual errors on my units, but lets just say that with 4sats altitude error is +-30', and with 10sats that goes down to +-3'. If we set home with 4 satellites the stored position might be +30' off. Then later in to the flight we get 10sats... then we land an see the altimeter reads -27' to -33' instead of 0'.

OTOH if I wait for more satellites before setting home (~7+sats), the set position is much more accurate , and after landing minutes later the altimeter still reads 0'.

Bottom line is the more sats you have when you set home, the more accurate your coordinates and altitude will be throughout the flight. Conversely, if you takeoff too early, your aircraft might think it is 30' from where it actually is, and that error will be stored until reboot or reset home.

So if you do takeoff too early, you can fly around and wait until the sat count goes up, then land and reset home to get more accuracy. I found the GOSD on my quad can quickly find 8-10sats when I'm hovering at least 10' off the ground (trees, buidings, power lines, etc out of the way). So I frequently make use of this "fly to find sats, land, and set home" routine. Not sure with planks though... if the constant motion might reduce how many fixes you can get?

Kev

Incidentally, this same low sat count error affects LAT/LONG coords too, but to a lesser degree than it affects altitude.
Truglodite is offline Find More Posts by Truglodite
Last edited by Truglodite; Oct 08, 2012 at 01:28 PM.
Reply With Quote
Old Oct 08, 2012, 02:07 PM
Registered User
Austin, TX
Joined Jan 2009
741 Posts
Quote:
Originally Posted by fpvwannabe View Post
but did you bother to tear off and check the wiring harness of your original "dead" GPS module? the reason i am asking is that mine was DOA too. turned out the GPS was fine but the HK monkeys had botched the harness. it was a 5 minute soldering job to fix it.
the OSD and GPS was on my very first FPV over a year ago...a SlowStick that I had a fly away (noobie error...let it get out of video range)....I recovered it 7 months later when a kid playing ball found it....everything electronic still worked after that time in the winter rain and cold! except my GPS module...

I have played with the code for a bit now and have the answers to my questions...I modified Trog's code (who could resist)...

1. I put the sat on the upper far right...I only show one voltage for now...played with increasing font for voltage, but you have to do it for the entire line...then the GPS lat readout is bigger than the long ....wish I could selectively just increase the voltage printout...anyone have an idea how to most easily do that without screwing up the line 2?
2. I centered the GPS readout in the top lines and made it appear only on low alt alarm....tried to do it in line 4 when stats arent showing and low alt alarm occurred, but while the lat printed fine, the long was not printed...just the blank inverted background for it...and it screwed up the FT text...couldnt figure out why
3. I changed the home symbol to show the X whenever its not home set or bad gps...not when its below min speed in conf.h...
4. I made the home arrow active ANYTIME home is set and the "at home" circle to overlay the arrow when within 10 feet

this suits me better....might need to up the # of Sats for homeset from 4 to 6 like Kev suggests...I really dont want to have to push the button
Atx_Heli is offline Find More Posts by Atx_Heli
Reply With Quote
Old Oct 08, 2012, 04:01 PM
Registered User
glue hand brian's Avatar
United Kingdom, England, Uxbridge
Joined Jan 2012
885 Posts
Help!! i've been trying to flash my g-osd for a good few hours, with no joy. all i'm getting when plugged in is white lines and a slow flashing red light on the g-osd. 1st i tryed with avr studio but i've got the atmega88PA and couldn't figure out how to add the 88PA entry to avrdude.conf.
so gave up after a couple of hours.. then read about someone using eXtreme Burner, so gave that a go, seemed to work, said it worked, but no change.i installed the hex file and the epp file there's nothing else i should have done is there?
as you might guess i havn't a clue about programming
glue hand brian is online now Find More Posts by glue hand brian
Reply With Quote
Old Oct 08, 2012, 11:16 PM
Registered User
Austin, TX
Joined Jan 2009
741 Posts
is this becuase line three and four are larger?

getting artifacts in line four from line three and sometimes line three flashed breifly below line four even.....anybody know why? line 3 and 4 are the only ones with bigger font sizes....could this be the issue?
Atx_Heli is offline Find More Posts by Atx_Heli
Reply With Quote
Old Oct 09, 2012, 12:24 AM
Registered User
Joined Jul 2008
253 Posts
Quote:
Originally Posted by glue hand brian View Post
Help!! i've been trying to flash my g-osd for a good few hours, with no joy. all i'm getting when plugged in is white lines and a slow flashing red light on the g-osd. 1st i tryed with avr studio but i've got the atmega88PA and couldn't figure out how to add the 88PA entry to avrdude.conf.
so gave up after a couple of hours.. then read about someone using eXtreme Burner, so gave that a go, seemed to work, said it worked, but no change.i installed the hex file and the epp file there's nothing else i should have done is there?
as you might guess i havn't a clue about programming
which avr studio you were using? because white lines sounds like avr studio 6 problem I've and many others encountered.
dziedrius is offline Find More Posts by dziedrius
Reply With Quote
Old Oct 09, 2012, 02:57 AM
Registered User
glue hand brian's Avatar
United Kingdom, England, Uxbridge
Joined Jan 2012
885 Posts
Quote:
Originally Posted by dziedrius View Post
which avr studio you were using? because white lines sounds like avr studio 6 problem I've and many others encountered.
avr studio 5.1, also tryed extreme burner, did see avr studio 6 was going to give that a go, won't bother now, saved me another headache.lol
glue hand brian is online now Find More Posts by glue hand brian
Reply With Quote
Old Oct 09, 2012, 04:51 PM
Registered User
Austin, TX
Joined Jan 2009
741 Posts
Incorrect gps data fixes being used?...result display and stat errors

To anyone familiar with GPS messages and/or this code (callelj or truglodite?):

Looking thru the code I see that in gps-common.h the current gps data is taken as valid and placed in the last valid data array anytime the message has a valid checksum.....is this correct?....as I understand it, the checksum is only to assure that the message from the GPS RX is actually received by the OSD correctly...not that the data is correct itself...

I think this would lead to lots of errors....shouldn't the data be considered valid only when the FIX is either 1 (2 dimensional fix) or 2 (3d fix)? probably only on 3D fix since we use the altitude data....

this came up since I was getting occasional momentary big negative altitudes displayed...and I also noticed that the sat symbol blinked (indicating a bad fix) at the same times...

it would be a simple code change to make the test for valid data be FIX=3 or even FIX not zero...shouldnt this work and make the integrity of the data a lot better?
Atx_Heli is offline Find More Posts by Atx_Heli
Reply With Quote
Old Oct 09, 2012, 07:47 PM
Registered User
Crashed22's Avatar
United States, ID, Boise
Joined Mar 2012
162 Posts
Quote:
Originally Posted by Atx_Heli View Post
getting artifacts in line four from line three and sometimes line three flashed breifly below line four even.....anybody know why? line 3 and 4 are the only ones with bigger font sizes....could this be the issue?
I'm getting the same thing, and the stats don't show at the end, with the same issue- Kev, any idea? I prefer to use yours over Jeorn's since I can use analog 3 for rssi without a buffer. Analog 2 bounces too much, without a buffer. Thanks- your setup is awesome...
Crashed22 is offline Find More Posts by Crashed22
Reply With Quote
Old Oct 09, 2012, 08:26 PM
Registered Aircraft Offender
Truglodite's Avatar
Carmichael, CA
Joined Feb 2007
3,508 Posts
Quote:
Originally Posted by Crashed22 View Post
I'm getting the same thing, and the stats don't show at the end, with the same issue- Kev, any idea? I prefer to use yours over Jeorn's since I can use analog 3 for rssi without a buffer. Analog 2 bounces too much, without a buffer. Thanks- your setup is awesome...
Thanks, I noticed the same artifacts with mine, but it wasn't horrible so I ignored it. OTOH flying with it can be annoying, especially when a white line flashes across line 4 (have you guys seen that one? happens randomly every minute or so).

I don't know for sure why this happens, but I'm highly suspicious it has to do with lines 3 & 4 overlapping. At least to me, they look close enough to where the backgrounds could overlap. I'll try redoing the layout with more space between those lines 3 & 4 and see what happens. If you want to give it a shot yourself, the lines to edit are in config.h, line 196 or so:

Code:
//font sizes for each line
//#define TEXT_TRIG_LINES_LIST 25, 50, 135, 180, 210, 235 // Must be rising line numbers!
#define TEXT_TRIG_LINES_LIST 25, 40, 135, 155, 170, 235 // Higher number = lower on screen, Must be rising line numbers!
I have lines 3 and 4 set to 135 & 155 respectively. Maybe setting line 3 to a smaller value like 130 would get rid of the problem (this puts line 3 higher on the screen). I want my units to be as close as possible though, and if they start looking like separate items on the screen I might nix the units altogether.

Of course, this might not have anything to do with the line spacing. IIRC, a while back Carl found that lines were being drawn before they were fully processed, which created some similar but more dramatic artifacts. He fixed the code by calculate certain lines a earlier. Such is the path of squeezing blood out of a tiny chip. If it's something like that causing the problem, I'll have to live with it until a good enough programmer comes along to fix it... I'm clueless about balancing time/space on a uC.

Quote:
Originally Posted by Atx_Heli View Post
To anyone familiar with GPS messages and/or this code (callelj or truglodite?):

Looking thru the code I see that in gps-common.h the current gps data is taken as valid and placed in the last valid data array anytime the message has a valid checksum.....is this correct?....as I understand it, the checksum is only to assure that the message from the GPS RX is actually received by the OSD correctly...not that the data is correct itself...

I think this would lead to lots of errors....shouldn't the data be considered valid only when the FIX is either 1 (2 dimensional fix) or 2 (3d fix)? probably only on 3D fix since we use the altitude data....

this came up since I was getting occasional momentary big negative altitudes displayed...and I also noticed that the sat symbol blinked (indicating a bad fix) at the same times...

it would be a simple code change to make the test for valid data be FIX=3 or even FIX not zero...shouldnt this work and make the integrity of the data a lot better?
Excellent find! If this was fixed, the statistics page would actually be worth looking at. As is, I think I've only had one flight where the statistics made any sense. It sounds like you have a lot more programming skills than I. So I'll follow along and try to help you and Carl out where I can. A better altimeter would be extremely beneficial to my quad.

Kev
Truglodite is offline Find More Posts by Truglodite
Last edited by Truglodite; Oct 09, 2012 at 08:38 PM.
Reply With Quote
Old Oct 09, 2012, 09:02 PM
Registered User
Crashed22's Avatar
United States, ID, Boise
Joined Mar 2012
162 Posts
Quote:
Originally Posted by Truglodite View Post
Thanks, I noticed the same artifacts with mine, but it wasn't horrible so I ignored it. OTOH flying with it can be annoying, especially when a white line flashes across line 4 (have you guys seen that one? happens randomly every minute or so).

Code:
//font sizes for each line
//#define TEXT_TRIG_LINES_LIST 25, 50, 135, 180, 210, 235 // Must be rising line numbers!
#define TEXT_TRIG_LINES_LIST 25, 40, 135, 155, 170, 235 // Higher number = lower on screen, Must be rising line numbers!
I have lines 3 and 4 set to 135 & 155 respectively. Maybe setting line 3 to a smaller value like 130 would get rid of the problem (this puts line 3 higher on the screen). I want my units to be as close as possible though, and if they start looking like separate items on the screen I might nix the units altogether.
I did see that as well, thought it was my camera as I was flying in the sun this morning I'll play with lines a bit and see what I get- Thanks
Crashed22 is offline Find More Posts by Crashed22
Reply With Quote
Old Oct 09, 2012, 11:13 PM
Registered User
Crashed22's Avatar
United States, ID, Boise
Joined Mar 2012
162 Posts
Changed line 3 to 130, and so far looks good. Can't get sat lock in my house, but will give it a go in the am. Was getting ghosting on the home not set, and that is now gone, so that may just do it. Thanks!
Crashed22 is offline Find More Posts by Crashed22
Reply With Quote
Old Oct 10, 2012, 12:48 AM
Registered User
BigJimNZ's Avatar
Auckland New Zealand
Joined Sep 2010
1,294 Posts
Just opened up my other camera to reprogram. Thought you might appreciate this CL-OSD placement
Right in there with the camera and ubec for the GPS

Name: ImageUploadedByTapatalk1349848028.595928.jpg
Views: 190
Size: 214.6 KB
Description:
BigJimNZ is offline Find More Posts by BigJimNZ
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Build Log Super OSD ($90): Open Source Graphic OSD: Vario, 6xADC, Games, Datalog, Dual Video tom66 FPV Equipment 1928 May 24, 2014 11:07 AM
Discussion Open Source Remote Control System (OSRC) Gizmoman31 Radios 152 Apr 24, 2012 06:01 PM
Discussion Open source OSD for FPV UFO_MAN FPV Talk 75 Sep 07, 2011 02:03 PM
Question Open Source Octo Design Available...Looking for someonw who can cut a copy for me... Schatham421 Multirotor Talk 14 Apr 12, 2011 05:22 PM