VAGCOM Adapter: Questions and Answers

Hello to the PC specialists.

In Maik's adapter, the MAX232 interface chip is powered from the car's on-board power supply, and (as far as I understand) uses the power from the car to drive the RXD input of the PC.

My question: Could this potentially damage the PC and/or the adapter under any worst-case conditions?
I have no idea what the internal circuitry of a COM port looks like inside a PC.
At least, with an adapter that uses optocouplers, the electrical circuits of the car and the PC remain completely separate, which, in my ignorance, seems somehow safer to me.

"If the 'mixing' of power supplies from cars and PCs is safe, I would also build a cheap adapter based solely on transistors, without galvanic isolation, and try it out for myself."

If it works, I'll offer it to Rainer as another option for his website.

Hello Ulf,

I have a bad feeling about this. What's being turned on as a PC is often a laptop. These devices are expensive and use surface-mount components. If the COM port is damaged, that's it (the same applies to onboard COM ports). No one except the manufacturer can repair them. And it's pointless to try and figure out where the problem came from. On the other side, there's a motor control unit, which is also not cheap. Your adapter has at least 100 ohms in series with the ground. This corresponds to a maximum of 120mA and is already a good protection. I have also implemented it safely on the perforated board. A few milliamperes won't break anything immediately. I also don't see any reason why your adapter should inherently cause problems. An optocoupler is, in simplified terms, 'a piece of wire with latency.'

AndyO wrote:

An optocoupler can be simplified to a "piece of wire with latency".

Hi Andi,

... but this runtime can potentially cause significant problems if it's not properly managed.
On my 2.2, the delay between K-Line and RXD was, as mentioned, around 40 µsec, and that was only for the negative K-edge .

Apparently, this issue can be addressed by adjusting the "working point" of the optos.
However, having to adjust three trimmers on an adapter before it (hopefully) works introduces three additional potential sources of error, which, in addition to all other possible bugs, can prevent the successful use of VAGCOM.

My design for the transistor adapter doesn't require any trimmers and should work with any COM port, even those with relatively weak input stages. This makes the idea quite appealing.
Especially since the design is relatively simple in terms of circuit design, making it more appealing to me than the "monster" with comparators that I once sketched out but haven't built yet due to laziness, and which, moreover, doesn't seem to offer any fundamental advantages over the non-galvanically isolated transistor version.

I will definitely check their runtime behavior on the oscilloscope soon.
If you're interested, I can send you the plan via email.

Hello Ulf,

'Please test the acceleration capacitors. I'm getting 6us, which is faster than the specified integrated circuit (9.6us). Let me know when you're ready, and we'll give each other a call.'
I wish you a Happy New Year!

AndyO wrote:

Please test the acceleration capacitors. I'm getting 6us, which is faster than the specified integrated circuit (9.6us).

When you're ready, let's give each other a call. }

Okay, maybe I'll have time for that tomorrow.