Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Electronic Engineer needed to design circuit for switching front and rear cameras

This site may earn commission on affiliate links.
deonb,

I found an evaluation board for $200. It looks like this will this do what you want. The connectors are not compatible though.

SUPRKAR, can you put an oscope on the LVDS signals to get an idea of what the data rate is?

Ok, thanks. So that's for the DS25CP104EVK, which is a 3.125 GBPS chip.


I also found this $50 board, which is based on the SN65LVDS125A, a 4x4 1.25Gbps crosspoint switch.

This board is driven by jumpers, so it means you can directly hook up mechanical switches to the jumpers to do the switching. (Basically to S10 and S11. See page 2-3 in the datasheet below to see whether to bridge/leave open each individual jumper. So e.g. a simple SPDT switch connected to S11 will allow you to switch input 1 vs. input 2 to output 1.)

http://www.ti.com/tool/sn65lvds125aevm#technicaldocuments

sn65lvds125aevm.jpg


Datasheet:
http://www.ti.com/lit/ug/sllu064a/sllu064a.pdf

You can also use a SN65LVDS122EVM instead, which is a 2x2 switch instead of a 4x4. But it's the same price, and the 4x4 will allow future inputs. (E.g. side cameras, or DVD etc.).


However, we need to know whether the data rate of the rear view camera is lower than 1.25GBPS to see if it will work. Isn't there a part number printed on it?

Not a lot of people have a > 10ghz scope lying around, which you'll need. Actually you can use a logic analyzer rather than a scope, which is cheaper, but my logic analyzer is also 1gbps max. You can rent a faster one, but at some point it's more efficient for someone just to try this out and see if it works. It's a $50 board after all, there is no S&H.
 
Last edited:
If Tesla used a different camera they would need to use a different main screen. I doubt they did that. it would be a different part number for a main screen without tech vs with tech.

in reading the TI spec on this board it says " Designed to support signaling rates up to 1.5 Gbps " perhaps this means it will work with lower rates and max out at 1.5.
 
Last edited:
Parking assist lines thoughts.

If there are two cameras (HD and analog) that may explain the delay of parking assist lines on the screen.

Wonder if some of these schemes described on these two "front camera" threads will end up with the front camera image with these overlayed.

There installation would have to be much more precise with those lines. I'm not saying it is sloppy now but the camera would have to be adjustable and/or the lines need to be adjustable. Even more than headlights these things gotta be right. I can image a floor alignment "jig" or a screen overlay template (electronic) to make sure the the camera is pointing the right direction and the lines are true the the sides and end of the car.

All this points to another possible;
"Parking lines are not a feature we can add to already delivered cars" problem. Even if it could be done at the service centers the cost of coming up with a system and jigs for every service center and teh time to install would be so outrageous that it would anger customers.
 
Sign me up, i have access to solder paste and an oven. Lay it out in eagle and i'll give it a go. I've done down to 0402 don't relish hand populating and stenciling below that. I have access to mighty scopes at work. Anyone have pics of accessing the wires in question?
 
Parking assist lines thoughts.

If there are two cameras (HD and analog) that may explain the delay of parking assist lines on the screen.

Wonder if some of these schemes described on these two "front camera" threads will end up with the front camera image with these overlayed.

There installation would have to be much more precise with those lines. I'm not saying it is sloppy now but the camera would have to be adjustable and/or the lines need to be adjustable. Even more than headlights these things gotta be right. I can image a floor alignment "jig" or a screen overlay template (electronic) to make sure the the camera is pointing the right direction and the lines are true the the sides and end of the car.

All this points to another possible;
"Parking lines are not a feature we can add to already delivered cars" problem. Even if it could be done at the service centers the cost of coming up with a system and jigs for every service center and teh time to install would be so outrageous that it would anger customers.

This is an extreme wide angle camera, so perfect alignment is much less a concern. Plus the final alignment can be made with the camera active. So I don't see alignment as a problem at all.
 
Artsci you get any more parts in?

Unfortunately no. The Amphenol folks have not responded yet to my order request and the switch is due tomorrow. But without the
Amphenol cables and connectors the switch will do me no good. So I just have to wait. The cables have to be custom assembled -- I ordered four, one for the long connection from the front camera to the driver's side kick panel, three shorter ones for the connections between both cameras and the touch screen. I've done this so in the event the mechanical switch approach doesn't work I won't have cut up any of the stock cables already in the car.
 
Last edited:
Cross-post from the original thread that started this, in case people are following this now. I'm a robotics engineer and interested in helping. For full touchscreen integration, I think we can use a web app running in the web browser. If you have version 5.0 of Tesla OS, you can access web pages hosted on a local wifi network. The camera switcher can host a web page that you open in the browser to change the camera view. It's a few more steps than a physical switch, but it's a hack that allows us to control arbitrary devices from the touchscreen. In fact, I'm thinking of designing a "universal" device with some general-purpose outputs that can hook up to other customizations in the car. You could set up your own custom web page with all the controls you want, and the hardware would send appropriate electrical signals to any device you'd like in the car. Unfortunately I don't have 5.0 yet, so I can't test out local web page access. Can anyone confirm that it works?
 
This is an extreme wide angle camera, so perfect alignment is much less a concern. Plus the final alignment can be made with the camera active. So I don't see alignment as a problem at all.

The width of the field of view makes it easier but no less critical since the lines can mean car damage while reversing. Just like headlights the final adjustment check should be done while on.

Maybe it's not a big deal, Autobody shops fix damaged back ends of different makes and model cars all the time and many have cameras with parking lines. I guess I just talked myself out of the concern. :)
 
Cross-post from the original thread that started this, in case people are following this now. I'm a robotics engineer and interested in helping. For full touchscreen integration, I think we can use a web app running in the web browser. If you have version 5.0 of Tesla OS, you can access web pages hosted on a local wifi network. The camera switcher can host a web page that you open in the browser to change the camera view. It's a few more steps than a physical switch, but it's a hack that allows us to control arbitrary devices from the touchscreen. In fact, I'm thinking of designing a "universal" device with some general-purpose outputs that can hook up to other customizations in the car. You could set up your own custom web page with all the controls you want, and the hardware would send appropriate electrical signals to any device you'd like in the car. Unfortunately I don't have 5.0 yet, so I can't test out local web page access. Can anyone confirm that it works?

It's an option, but I'd be concerned about the quality and update rate you'd be able to get on Tesla's browser.

The hardware solution allows for more than just a camera, you can also plug in a DVD player, which you can't do with a web page solution.

On the other hand, the web page solution allows for more general purpose extensions to functionality.

We should just pursue both options :).
 
It's an option, but I'd be concerned about the quality and update rate you'd be able to get on Tesla's browser.

The hardware solution allows for more than just a camera, you can also plug in a DVD player, which you can't do with a web page solution.

On the other hand, the web page solution allows for more general purpose extensions to functionality.

We should just pursue both options :).

I think the idea was to use the web interface to handle the switching, and the rear camera pane to display the video.
 
I think the idea was to use the web interface to handle the switching, and the rear camera pane to display the video.

Oh, you're right. Duh.

Wow, that's an elegant solution!

I still think there should be 3 modes, forward, backward and auto (driven from reverse lights), but then the actual mode can be controlled via the web interface.

Ok, I'm feeling more inspired now to participate with my own Model S! Wasn't ever going to add a physical switch to my car, but this is non-intrusive, and fits into the spirit of the car interface.