Wireless front parking camera with monitor for front bumper protection

[ken830]

WhiteP85... I'll admit that I don't know how far this will lead or whether or not we'll hit a solid wall, but here's what I'm thinking... The four lines in the cable only carry power and ground to the camera model, and a serial differential pair for the video stream from the camera module to the touchscreen CPU. I don't believe there is any communication from the touchscreen CPU to the camera module. Having said that, the Omnivision OV10630 default value on the TIMING_CTRL1D register is 0x40 (assuming the same as the OV10633) which means the mirror bits (bits [1:0]) are both 0 -- mirroring is disabled by default. This means that something in the camera module is setting that register -- and my guess was an onboard microcontroller.... Which brings me to this....

Artsci.. Thanks for the photos! At a quick glance, I see a Microchip PIC microcontroller (PIC16F1829) in a 20-pin QFN package. This microcontroller can be configured with up to two I2C ports on Port B:

SCL1 on pin 8 (RB6)
SDA1 on pin 10 (RB4)

SCL2 on pin 7 (RB7)
SDA2 on pin 9 (RB5)

I've highlighted these in red. Can you visually trace to see which pair goes to the OV10630?

The PIC microcontroller has program and data memory protection. If this is enabled, then, I'm afraid there's not much we can do short of replacing the PIC, but we'll have to somehow duplicate all of the other functions that it serves, which could be nothing or could be many. Before we can even consider that, we should see if the programming pins are broken-out to a connector footprint somewhere.. These are pins 15 and 16, which I've highlighted in green. From the photo, it looks like those two signals go straight to the back-side of the board. Any way to see the back-side of the board to follow those signals?

 

[artsci]

Artsci.. Thanks for the photos! At a quick glance, I see a Microchip PIC microcontroller (PIC16F1829) in a 20-pin QFN package. This microcontroller can be configured with up to two I2C ports on Port B:

SCL1 on pin 8 (RB6)
SDA1 on pin 10 (RB4)

SCL2 on pin 7 (RB7)
SDA2 on pin 9 (RB5)

I've highlighted these in red. Can you visually trace to see which pair goes to the OV10630?

The PIC microcontroller has program and data memory protection. If this is enabled, then, I'm afraid there's not much we can do short of replacing the PIC, but we'll have to somehow duplicate all of the other functions that it serves, which could be nothing or could be many. Before we can even consider that, we should see if the programming pins are broken-out to a connector footprint somewhere.. These are pins 15 and 16, which I've highlighted in green. From the photo, it looks like those two signals go straight to the back-side of the board. Any way to see the back-side of the board to follow those signals?

Sorry for asking what appears to be a dumb question, but I assume the OV10630 is the PCB you marked? Getting the the back side of that board to follow pins 15 and 16 would involve some risk, as it's glued in place in each corner. You can probably see the glue in the photo you marked. I've got to put the camera back together later today for the test today of the current PCB we're working with, so I don't want to risk damaging it. Once I've done the test I'll see what else can be done to flip the PCB you marked.

The only identifying marks I could find and read were on the PIC you marked. Here's how they read:

PIC16
F1829
E/MC
144550

Next to the E/MC line there's a circle with what appear to be E3 inside.

 

[WhiteP85]

@ken830. Good eyes. I did not suspect an on board PIC as the LVDS serializer/deserializer data sheets I read showed examples of both the video data and control lines being converted to serial. I am just a rusty old guy playing being an engineer again. It will be interesting to see if the PIC is protected. Do you think code disassembly is a practical route? If it is the PIC setting the OV10630 then I guess someone with more skill and time than me could also monitor the I2C messages, work out what the PIC is doing, and write and burn some new firmware on this or a new PIC?

If we get this far and manage to work out the settings then it might just be easier to go the third party route or make our own video boards particularly if the camera is just a dumb video pipe with no communication with the console CPU. :wink:

- - - Updated - - -

Sorry for asking what appears to be a dumb question, but I assume the OV10630 is the PCB you marked? .

I think the OV10630 image chip should be on the PCB directly facing the lens. It looks like this PIC is on the other/second PCB?
Mmmm - looking at your pics again I am not sure. It looks like the OV10630 might be mounted behind the lens with the parallel video cable leading to one of the two PCBs. One of the chips looks like it is shielded. Might be the serializer.

 

[artsci]

@ken830. Good eyes. I did not suspect an on board PIC as the LVDS serializer/deserializer data sheets I read showed examples of both the video data and control lines being converted to serial. I am just a rusty old guy playing being an engineer again. It will be interesting to see if the PIC is protected. Do you think code disassembly is a practical route? If it is the PIC setting the OV10630 then I guess someone with more skill and time than me could also monitor the I2C messages, work out what the PIC is doing, and write and burn some new firmware on this or a new PIC?

If we get this far and manage to work out the settings then it might just be easier to go the third party route or make our own video boards particularly if the camera is just a dumb video pipe with no communication with the console CPU. :wink:

- - - Updated - - -



I think the OV10630 image chip should be on the PCB directly facing the lens. It looks like this PIC is on the other/second PCB?
Mmmm - looking at your pics again I am not sure. It looks like the OV10630 might be mounted behind the lens with the parallel video cable leading to one of the two PCBs. One of the chips looks like it is shielded. Might be the serializer.

I've now put the camera back together and thankfully it works fine. I have to make a new PCB board and as soon as I do the test will begin. Hopefully I can complete all of this before it gets dark.

After I do the test this I'll take the camera apart again and risk removing the PCB that's glued in place. I don't think it will be that difficult, as the glue looks like some kind of Goop-like material.

 

[Akikiki]

artsci, you are the Indiana Jones of reverse mirrored auto cameras. Thanks for the time and efforts for the benefit of all the rest of us.

 

[artsci]

artsci, you are the Indiana Jones of reverse mirrored auto cameras. Thanks for the time and efforts for the benefit of all the rest of us.

I'm just having fun getting this done.

 

[nightau]

Please don't laugh at me:
is it feasible to mount the front camera at 90 degrees, and put a mirror/prism in front of it, like a periscope thingy? see pic below

I'll go back in hiding now...

 

[WhiteP85]

I like your illustration. Not to rule anything out I grabbed one of my daughters' hand mirrors and surreptitiously wandered into the garage, held up the mirror and peered in the back window at the display. I held it at a 90 degree angle and as close as possible. That camera is very wide angle so I could see the whole of the 6" diameter mirror and my hand. Bottom line is the mirror would have to be very big. Good thing the garage door was closed or my neighbors would have had even more reason to think me crazy inspecting my new car for defects with a hand mirror.

 

[artsci]

Please don't laugh at me:
is it feasible to mount the front camera at 90 degrees, and put a mirror/prism in front of it, like a periscope thingy? see pic below
View attachment 33531
I'll go back in hiding now...

I don't think this is practical given where the camera has to mount.

IMHO we're going to a great deal of trouble to set the horizontal image when it really is not that critical to the purpose of the camera: protecting the front bumper from damage when parking head in. I'm really not sure I want to take any more risks taking apart my spare camera yet again for that purpose. For those to whom this is really important perhaps you may want to buy a spare camera for experimentation.

I'm about to mount the camera on the front bumper for my next round of testing. If it goes well I'll be satisfied that we have a good solution pending final revisions to the PCB design. I understand that there are some people who won't be satisfied until the horizontal image is correct but that's probably a solution that's going to take some time and a great deal of difficulty. It's technically way above my head.

It appears that the easiest and best solution for the horizontal issue is another camera that will work with the touchscreen and maintain the image quality. I'm certainly for that. If someone finds such a camera I'm game.

 

[WhiteP85]

IMHO we're going to a great deal of trouble to set the horizontal image when it really is not that critical to the purpose of the camera: protecting the front bumper from damage when parking head in. ....
It appears that the easiest and best solution for the horizontal issue is another camera that will work with the touchscreen and maintain the image quality. I'm certainly for that. If someone finds such a camera I'm game.

I am of the same opinion. Camera reversal is a nice to have for me too. My first priority is to help you with the switch PCB. Looking forward to your test.


Still happy to pursue the other option but so far it is looking like too much effort. We have learned a lot though and as we uncover more we may discover a viable option. I will still make some calls next week.

 

[nightau]

I like your illustration. Not to rule anything out I grabbed one of my daughters' hand mirrors and surreptitiously wandered into the garage, held up the mirror and peered in the back window at the display. I held it at a 90 degree angle and as close as possible. That camera is very wide angle so I could see the whole of the 6" diameter mirror and my hand. Bottom line is the mirror would have to be very big. Good thing the garage door was closed or my neighbors would have had even more reason to think me crazy inspecting my new car for defects with a hand mirror.

This is more of what I had in mind, and it's pretty small in size:

 

[ken830]

Thanks for all the effort Artsci. If I didn't have a 9-month old, I would have the time to consider pursuing this myself. Actually, I would like to know exactly what video signal is on the cable itself. Then we can create a new camera system that can just feed in whatever we want. This will be the way to go to install a custom 360-degree video feed and/or monitor the rear facing seats.

 

[artsci]

I am of the same opinion. Camera reversal is a nice to have for me too. My first priority is to help you with the switch PCB. Looking forward to your test.


Still happy to pursue the other option but so far it is looking like too much effort. We have learned a lot though and as we uncover more we may discover a viable option. I will still make some calls next week.

I didn't get as far as I had hoped with the test. I did make up the new PCB but it still has to be drilled and the components mounted. I won't finish that until later tonight and won't be able to do the test until next weekend. Tonight I have to move on to the new 3D printed version of the lighted T. I have six testers waiting for this version and it's far easier to set up and mount than the version now on my car.

I also mounted the camera where I think it will be ideal, but I had a major problem running the cable from the camera to make the connection to the cable that runs to the PCB. The camera is mounted (temporarily) on the grille and chrome cross bar under the nose cone. I had to use a Dremel tool to make some cuts to the grille so I could mount the camera flush to the grille and cross bar and run the camera cable behind the grille. But in the time I had I couldn't find a way to run the cable behind the grille and steel crash bar up into the area where I could make the connection. I think I may have to remove the bumper to do this properly. There doesn't appear to be a large enough (or any) opening to get the cable behind the crash bar into the upper area of the nose cone where it would be easy to make the connection.

If anybody has done this successfully or has any ideas I welcome them as I would prefer not removing the bumper to do this.

Here are a few photos of the mounted camera. The cable is the left of the camera (it's mounted that way). My current nose cone is installed with the Tadi Bothers wireless camera that started this whole business. It's just over the Tesla camera. When I get the this all working properly that nose cone will be replaced with one I have standing by, and the Tadi camera will be gone.

 

[WhiteP85]

Worth a shot. You may be on to something: amzn.com/B00DN2U0YI - so I ordered one. Thanks for the idea. Will post what I find when it arrives.

(My neighbors would not even see that.)

 

[artsci]

Worth a shot. You may be on to something: amzn.com/B00DN2U0YI

(My neighbors would not even see that.)

Just bought one to try out, although it may be too small.

 

[WhiteP85]

Thanks for all the effort Artsci. If I didn't have a 9-month old, I would have the time to consider pursuing this myself. Actually, I would like to know exactly what video signal is on the cable itself. Then we can create a new camera system that can just feed in whatever we want. This will be the way to go to install a custom 360-degree video feed and/or monitor the rear facing seats.

There is a chance that Tesla used a MAX9259 as described here: Application of Serializers with Camera Chips - Application Note - Maxim

 

[nightau]

Worth a shot. You may be on to something: amzn.com/B00DN2U0YI - so I ordered one. Thanks for the idea. Will post what I find when it arrives.

(My neighbors would not even see that.)

I truly appreciate your hard work & expertise, WhiteP85, artsci, ken850. I'm just trying to offer a "simple" man's perspective.

BTW, I'm a firm believer that cars in future will all have front video recording built-in. Not for parking assistance, but kind of like black boxes in planes, to offer information & protection in case of accidents & disputes. Some Asian countries already have 3rd-party front-facing vid cams in 90% of cars & taxis, and I believe Taiwan is considering making this a legislation for vehicle owners.

 

[dsmith2189]

Tonight I have to move on to the new 3D printed version of the lighted T. I have six testers waiting for this version and it's far easier to set up and mount than the version now on my car.

Can I volunteer to test the future lighted T with the forward looking camera?

 

[artsci]

Can I volunteer to test the future lighted T with the forward looking camera?

The lighted T testers have been limited to six and set for a few months. As for the camera, we won't have testers for that, as installation will be plug and play. Not so for the lighted T.

 

[cybrown]

I'm working with an Omnivision camera chip at work and part of my job is to develop firmware for setting the registers, talking to the camera, etc. I agree with the previous posters that the Pic is probably setting the registers over the I2C bus. We wouldn't have to actually intercept these commands though - we could simply issue additional commands from a separate microcontroller once the Pic is done its initialization. However, this would probably be on a one-off basis, since it would require soldering leads to the PCB, as far as I can tell. I'd love to get my hands on a camera to experiment alongside you, artsci. How can I get one? You just asked a service center to order one for you?