[RayeR]

I am thinking it is excessive capacitance.

I don't think so. How long is the cable? It seems that driver output impedance is in kiloohm range, when you load it by a 100pF capacitance (~1m of some coax cable) it would make RC with tau 100ns and clock period is 4us so it shouldn't ruin it. I would take 3 thin wires and wound them directly on GPS contacts before pushing to hotshoe and hook them to probes.

If I do manage to figure it out my next project will be a device to allow me to dial in a compensation for the magnetic compass error.

Why do you need this? I read in manual or somewhere else that there's calibration item in menu. It use GPS data when you walk to get north direction and then it measure the offset and do compensation for you automatically

[lister6520]

I don't think so. How long is the cable? It seems that driver output impedance is in kiloohm range, when you load it by a 100pF capacitance (~1m of some coax cable) it would make RC with tau 100ns and clock period is 4us so it shouldn't ruin it. I would take 3 thin wires and wound them directly on GPS contacts before pushing to hotshoe and hook them to probes.

I'd rather wait for the cheap cable to arrive and replace its cable rather than risk damaging the hotshoe or GPS. I don't have problems with the flash, it's only the GPS that doesn't work with the cable, although I do have a feeling that the flash is actually less reliable when used with the cable. I am thinking the only possibility is that the GPS has weaker outputs and/or less sensitive inputs than the flash.

Why do you need this? I read in manual or somewhere else that there's calibration item in menu. It use GPS data when you walk to get north direction and then it measure the offset and do compensation for you automatically

That's not enough in some situations unfortunately. I had started a thread on that subject:
https://www.pentaxforums.com/forums/6-pentax-dslr-discussion/241072-astrotrac...libration.html

[RayeR]

I am thinking the only possibility is that the GPS has weaker outputs and/or less sensitive inputs than the flash.

Maybe we get surprise if it communicates at higher speed or different protocol than flash... I think that all inputs and outputs are protected by some resistors and ESD protection as it's exposed to user it must withstand some ESD testing and shorts... But there's no time to hurry. I'm thinking if there's available some cheap hotshoe connector or spring contacts that can be solderen on a PCB or some cheap cable that can be cut..

That's not enough in some situations unfortunately. I had started a thread on that subject:
https://www.pentaxforums.com/forums/6-pentax-dslr-discussion/241072-astrotrac...libration.html

Aha, local fields, I understand... If you can direct camera to right north and if it is possible to display compass measurement on LCD, then you can place a small magnet around in the right location to compensate the deviation. It's not cool but should work. Maybe Shodan will find a function that parse the compass data and one day will be possible to fix it via menu...

[Shodan]

One of the reasons I started this was to hack up my own GPS module...I then got distracted with cutting my own code.

I'm knee deep in the hardware set up code at the moment. There are GPIO and I2C drivers in the codebase. If you can work out the comms it should be possible to intercept the raw data, call our code and change the values. This is a bit of a way off at present but conceptually very simple.

[patarok]

I was all set up to start figuring out the GPS protocol but I've hit a snag which is going to delay me somewhat. For the flash I have used a P-TTL extension lead which I slightly modified to tap into the connections but for some reason the GPS is not communicating at all with the camera when I connect it via the extension cable, even before I have conencted anything to the tapped connections. I've tried also with a second splitter cable and it doesn;t work either - seems it likes to be connected directly to the hot shoe.

I am thinking it is excessive capacitance. The thing is I don't to ruin the two cables I have beyond the minor intervention I made to tap into the connections. I've just ordered a cheap one from Poland on eBay which I'll then dismantle and rebuild wiring the two ends together with very short lengths of single wires. Hopefully that will work.



Is there anyone who has an O-GPS1 and a P-TTL extension cable who could try see if the GPS can actually work on an extension cable, maybe I am just unlucky and the specific cables I have are the problem?



If I do manage to figure it out my next project will be a device to allow me to dial in a compensation for the magnetic compass error. The idea would be that a PIC processor would sit in between the GPS and the camera passing on most of the data unchanged but then when it recognises a compass reading it would add an offset to it and send on the modified version instead of the original. Well that's the theory at least. Meantime I'm still hoping Pentax would include such functionality in the next firmware update and saving me a lot of effort.

I ll soon have a P-TTL Hotshoe Adapter Cable in my Hands... I will try the GPS on it anyways, because i am also looking forward to dismantle that adapter, because i will try to build a Hotshoe stage, so i can hook up gps and still use built in flash.
My first guess when i was reading about the problems you have with hooking up gps to a p-ttl adapter was:

Maybe it has to do with the conduction of the p-ttl cable respectively the resistance of the cables.

[lister6520]

......
Aha, local fields, I understand... If you can direct camera to right north and if it is possible to display compass measurement on LCD, then you can place a small magnet around in the right location to compensate the deviation.
...

I actually expended quite a lot of effort trying to solve it that way but found that while it was easy to get the display to show North when pointed to the North as soon as I turned the camera or even changed the elevation the amount of compensation changes. I tried positioning the magnets and/or steel rods fixed relative to the camera and/or fixed to the tripod, even tried an electromagnet sort of thing in the form of a large frame but nothing that I did could give me a constant and predictable correction which stayed correct as I pointed the camera away from the North star.

Anyway, I should be receiving the new (and cheap) cable sometime this week and will probably find some time in the weekend to start working on it. I am quite sure the protocol will be completely different from that of the flash - I am thinking it will be straight NMEA ASCII strings except for some initial response to the camera interrogating the hotshoe. If so it should be quite easy to devise a PIC to alter the values within the relevant strings.

[RayeR]

I actually expended quite a lot of effort trying to solve it that way but found that while it was easy to get the display to show North when pointed to the North as soon as I turned the camera or even changed the elevation the amount of compensation changes.

Hm, thats damn tricky. But this mean if you compensate compass via software in one direction it will not match in other direction, the same issue... only if you could bypass compass at all and measure azimuth manually and feed this fixed data...

I am thinking it will be straight NMEA ASCII strings except for some initial response to the camera interrogating the hotshoe. If so it should be quite easy to devise a PIC to alter the values within the relevant strings.

But it seems not. Shodan told he didn't find any trace of NMEA parser in FW. I aslo didn't find anything like that (strings gga, gsv, ggl, rmc...) I think almost all devices that process common NMEA have this strings visible in FW. Here are all strings containing "gps":
C:\GPS, GPS_DAT?.TXT, GPSMMII, GPSWCPVQ, GPS DATA FULL.
and one interesting string
"%02X:Calc Astor trace data."
can you see the typo? Astor instead of Astro. Or was it intended obfuscation?

---

Last edited by RayeR; 03-24-2014 at 05:25 PM.

[Shodan]

There are other popular GPS protocols - SiRF for instance is binary.

"%02X:Calc Astor trace data."

Hmm I can't find this in the strings window of IDA Pro. Was this from a K-30 firmware or DSP dump? If so which version and do you have an address?

[RayeR]

Hmm I can't find this in the strings window of IDA Pro. Was this from a K-30 firmware or DSP dump? If so which version and do you have an address?

It's in decrypted FW 1.06, do you dasm 1.05? I can tell offset later.

[Shodan]

interesting. I've been working on 1.5.

[RayeR]

FW 1.06 - FWDC215B.DEC, offset: 00C02174

There looks to be some code but I haven't RE any of it yet. From a quick look flashing the lens rom should work.

Sadly dumping the lensrom with a K-30 (and K-3 as well) doesn't work (post #91/#92). It should be writing a file called LENSEEP.LNS on the SD card but it doesn't for some reason. I wonder if there is another prerequisite to meet other than just pressing the button on the menu item.

Can you understand what is happening there? Is that hardcoded disabled somehow?

[Shodan]

Attached is the code for both functions. If you look closely you'll see that the function that should do the work actually just returns 0.
As such loading and storing the lens rom doesn't work. I suspect it only works in debug builds. You might get lucky - you could try all the firmwares from 1 - 1.6 and see.

---------- Post added 03-26-14 at 09:24 PM ----------

While i'm on it. A donation update!

We are currently over 2/3 of the way to another K-30. I've been checking eBay a bit and there is a faulty model for sale (dust on sensor) that's a bit cheaper. If anyone is feeling generous this might be a good time to donate. When we get close enough I'll throw down a few $$$ of my own so we can speed this hacking along a bit.

[vrphoto]

Attached is the code for both functions. If you look closely you'll see that the function that should do the work actually just returns 0.
As such loading and storing the lens rom doesn't work. I suspect it only works in debug builds. You might get lucky - you could try all the firmwares from 1 - 1.6 and see.

---------- Post added 03-26-14 at 09:24 PM ----------

While i'm on it. A donation update!

We are currently over 2/3 of the way to another K-30. I've been checking eBay a bit and there is a faulty model for sale (dust on sensor) that's a bit cheaper. If anyone is feeling generous this might be a good time to donate. When we get close enough I'll throw down a few $$$ of my own so we can speed this hacking along a bit.

Small donation made, looking forward to the results of pretty awesome work ...

[thibs]

Attached is the code for both functions. If you look closely you'll see that the function that should do the work actually just returns 0.
As such loading and storing the lens rom doesn't work. I suspect it only works in debug builds. You might get lucky - you could try all the firmwares from 1 - 1.6 and see.

---------- Post added 03-26-14 at 09:24 PM ----------

While i'm on it. A donation update!

We are currently over 2/3 of the way to another K-30. I've been checking eBay a bit and there is a faulty model for sale (dust on sensor) that's a bit cheaper. If anyone is feeling generous this might be a good time to donate. When we get close enough I'll throw down a few $$$ of my own so we can speed this hacking along a bit.

Shodan, I really have not a single euro anymore due to new laptop incoming.
I still intend to donate some ASAP. Keep up the good work. Thanks mate!