VAGCOM Adapter: Change/Info on Optocouplers

Hello everyone,

To get started with VAGCOM, I downloaded the wiring diagrams from Noxon and Stautz for connecting the diagnostic adapter.
They showed me (unfortunately) that it's probably hopeless to simply connect the simple cable from my pocket-sized diagnostic tool to the PC and hope for a working connection to the car.

However, it sometimes seems that the adapter is designed in the "whatever's available at the moment" style.

The L-line is driven directly from the optocoupler output with a pull-up resistor of 10 kΩ, while the K-line is driven by a small transistor, which in turn (following the Darlington principle, i.e., without phase inversion) is controlled by the optocoupler.
The amount of overdrive applied to the driver transistor, with a base resistor and a collector resistor, both equal to 1 kOhm, almost seems grotesque to me.

What would be the disadvantage of simply eliminating R5, R6, and T1 for simplicity, and connecting the K-line to the optocoupler output in the same way as the L-line, using the pull-up resistor R7? The current load on the optocoupler would remain practically the same, at approximately 12 mA.
If the transistor stage introduces a timing delay, I believe it could be just as easily achieved by using a suitable small capacitor in parallel with the input of the optocoupler.

The reverse-voltage Schottky diodes on the PC-side opto inputs seem excessively oversized, as they are rated at 1N4004 (400 volts / 1 ampere), while they are only loaded with approximately 11 mA at an input voltage of -12 volts (does the COM output of the PC even operate with negative signal voltages?), and blocking voltages higher than 12 volts are unlikely to occur.
Why not use cheaper, more basic components like the 1N4148, which still have almost a tenfold safety margin with a 100V/100mA rating?

I read in this forum that using a 1 kOhm resistor as a pull-up (R7) can sometimes cause problems, which can be resolved by replacing it with a 10 kOhm resistor. However, this reduces the high level of the K-line by approximately 0.8 volts due to the load from the base resistor of transistor T2; with a 1 kOhm pull-up resistor, the loss is only about 0.1 volts.
What is the actual safe threshold for detecting "High" as indicated by the on-board electronics?

I purchased three PC837 optocouplers (35V / 50mA). Is there anything that would prevent me from using them in a VAGCOM adapter?
What specific parameters (besides maximum voltage and current for the output, and insulation voltage between the LED and the transistor) are important for the operation of optocouplers?

What does the % figure for CTR (Click-Through Rate) or conversion rate actually indicate?

Does anyone know of any links with beginner-friendly information about optocouplers?

Hi there,

To be honest, I've never actually calculated the circuit myself.
I built these based on Jeff's PCB design (http://www.planetfall.com/~jeff/obdii/schematic_b1.pdf), and specifically followed the schematic with the following modifications:

diodes 1N4007 (yes, yes... from the electronics kit). I wouldn't use 4148; 100V spikes at an interface aren't rare or valuable.

transistors: BC546, used as a substitute because the original was not available, and the values are almost the same.

opto: ILQ5 (quantity: 4), alternatively 3x CNY17-3 (same values as original).

The circuit itself is completely non-critical at the data rates being used.

The most important parameter for optocouplers, besides the breakdown voltage (which is probably less critical here), is the transfer ratio. This is roughly comparable to the current gain of a transistor.
Therefore, use CNY17-3; the standard CNY17 doesn't have a high enough current transfer ratio, and there's a possibility that the signal levels won't be sufficient.

I've built the interface twice using jumper wires on a breadboard. It works 100% in all the vehicles I have available (Octavia, A4, A4, Seat Ibiza, Golf-II).

CU Gremlin.

Gremlin wrote:

The most important parameter for optocouplers, besides the breakdown voltage (which is likely less critical here), is the transfer ratio. This is roughly comparable to the current gain of a transistor. Therefore, use CNY17-3; the standard CNY17 doesn't have a high enough current transfer ratio, and there's a possibility that the signal levels won't be sufficient.

Hi.

Thank you for your explanations.

I'm going to have some fun(?) and build this thing according to my ideas.

Okay, I'll measure the initial levels "in a dry run" and, if necessary, adjust the resistors so that "low" is truly "low."

For example, I find the way transistor T2 is controlled, with only 140 kΩ from the K-line, to be somewhat crude.

As long as the K-line has more than 0.6 volts, the optocoupler for the direction Auto --> PC will still be slightly activated, which (depending on the coupling factor ) could confuse the PC as a receiver.

I suspect that the tip mentioned in this post, to replace R7 with a 10k resistor, is based on the idea that the onboard electronics are more likely to reliably maintain a voltage below 0.6 volts with a 10k resistor than with a 1k resistor.

If you simply add a resistor in parallel with the base-emitter path of transistor T2 (e.g., 15k ohms), the high-low transition will switch cleanly towards the positive supply voltage at approximately 6 volts.

Quote:

Diodes 1N4007 (yes, yes... from the electronics kit). I wouldn't use 4148; 100V spikes at an interface are nothing special and not rare...

It might be true, but the diodes I'm referring to are connected in an antiparallel configuration to the opto-LEDs, and these opto-LEDs would have to short-circuit or convert voltage "peaks" (for the 1N4148 diodes) of over 2 volts into short pulses.
Therefore, to me, 1 N 4004 still seem like atomic bombs against sparrows .