VAGCOM: Motorcomputer antwortete nicht | Posts 16+

Hello,

I have now completed the schematic and the board.

I will review it again tomorrow, I don't have the energy today.

Nick

Nicklas wrote:

by the way, what do you think about the suppressor diodes on the K/L line and VBat? I think they belong to each pin that goes to or from the car.

Hi Nick

I don't know the characteristics of the diodes, but IMO, everything makes sense in principle if the voltages at K and L are kept within the range between 0 volts and the battery voltage – as long as the diodes don't have a very large blocking capacitance, which would prevent the slew rate from being controlled .

On the other hand, all the "public" replication proposals I am aware of are designed without such measures, and yet I haven't heard of any damage from over- or under-voltage issues when using these circuits .

Hi Nick,
Suppressor diodes have never caused any damage. However, I think 33V is a bit excessive. If so, you should choose a type with a minimum 'clamping voltage' of 16V (which would then be around 18V nominal voltage).
Also, small coils are also worth mentioning. For example, in the 1552, there are approximately 10 uH in the supply, K, and L lines.

Best regards, Rudi

The new HEX-USB+CAN adapter by Uwe Ross uses a similar concept to the Elektor adapter (with uC) and then directly connects to the USB interface.

Best regards, Rudi

Addendum:
As far as I know, the OBD2 software targets the 'standardized' OBD commands and does not take into account manufacturer-specific things, such as customizations. If the information is correct, it is not possible to activate the cruise control with it.

The 'adapter for all cases' from elektor does not run with 100% certainty with vag-com, as it is an OBD-II adapter. OBD-II has nothing to do with our vag-com! Furthermore, the input signals from the vehicle are converted to ASCII!

use ULF's adapter or MAIK's adapter (or buy MAIK's adapter), then you'll have what you need. This will protect your nerves and possibly your stomach!

Now Rudi is faster !!!!! But nevertheless, this adapter will not work with VAG-COM !!!!!!!! While Ross-Trech's concept is similar, it is in no way the same !

Hello Rudi and Jörg,

Thanks for the information!

It just doesn't make sense to me in the whole adapter issue:

According to the general consensus, the communication problems with various MSG devices are caused by fluctuations in the MSG's bit length.

But I absolutely don't understand how Ulf and Maik's adapters are supposed to correct these fluctuations?

Greetings, Jürgen

@juergen

Maik and Ulf's adapters also cannot compensate for these bit length differences! That would also only be possible with a great deal of technical effort!

but: due to optimal sizing, for example, with the adapters, there is a certain degree of flank steepness, and also the RS232 levels are maintained or generated quite optimally. However, it seems that this is not necessarily the case with your Noxon adapter, if I read it correctly. Furthermore, the magnitude of the signal delay in these adapters is relatively small. And these three points, in particular, make it easier for VAG-COM to synchronize the signals coming from the ECU and to correctly retrieve the information!
'Why don't you take the time to implement the adapter threads from Ulf! I know it's a lot of work, but I think your questions will be answered, and you'll definitely notice which weaknesses your adapter still has!'

Hello Jörg!

'Politics is the art of the achievable!' - Seems to be addressing MSG as well!!!

Since I can't modify anything on the MSG, the RS 232 port, or the VAGCOM, it seems that replacing the adapter is the only way to solve the problems, which I have now realized.

I will take a closer look at the Maik adapter, as at least there, the optocoupler issue cannot occur.

Greetings, Jürgen

@juergen

Bravo!

but: as already mentioned in other threads, the MAIK adapters do not have galvanic isolation between the car and the notebook!
It depends on whether you're willing to risk your notebook or your STG in the absolute worst-case scenario (hopefully this will never happen!!!!), or whether you prefer to be on the safe side (i.e., ULF's opt-variant). I think ULF's latest version is on par with Maiks version in terms of reliability and functionality. However, as I said before, both versions have advantages and disadvantages that you can only weigh for yourself!

Hi,

The Elektor adapter really has a uC. The MC33199 is not a uC, but rather 'just' a driver IC.
I believe the 33V for the Sup diodes is standard, and it's used in all our control units. The 6.8nF capacitors at the pins should be sufficient for protecting the 1 and 2 shooters, as well as 3a and 3b. I've noticed that I've forgotten the 6.8nF, and I'll need to order those as well.
16 or 18V seems too low to me, especially for jump starting, although it's unlikely that an adapter is connected in that case. Also, even when switching inductive loads, the energy for the diodes might be too high? And a normal Z-diode wouldn't help either with negative voltages.
Maybe I can still find a few 27V Sup diodes, then I will use those, making sure that the circuit doesn't get destroyed at either 27V or 33V.

I have another question: which vehicles require a CAN interface? (cannot be read with K-Line Audi A?)
How about we build a CAN adapter? Maybe one with USB.
I have previously used an Atmel 90S8515, a serial interface (MAX232), and 32kBytes of SRAM because serial interfaces are quite slow for buffering, and a SJA1000T with an 82C250. I can also exchange data between two such boards, and it should be possible to build a CAN adapter. However, we would then also need a new PC program that is compatible with the CAN adapter.
To see the hardware:
http://home.vrweb.de/~Alex_Weiss//Atmel_CAN/Atmel_CAN_back_web.jpg
http://home.vrweb.de/~Alex_Weiss//Atmel_CAN/Atmel_CAN_front_Web.jpg

Greetings

@ joergs

I have disconnected the Maik adapter galvanically.

Greetings

@nicklas

'I know that! And it makes sense! I just wanted to show Jürgen again, in person, the pros and cons of each option, because I've been reading between the lines that he has too many problems with the Optos, which is completely ridiculous!'

Due to the CAN interface and a tool for reading: Do you have a specification for the request and response protocols? If so, I would be confident in building something like that! I would also be very interested in it myself, as my new Passat probably has a CAN bus.

@ Nicklas

The Elektor adapter is also too expensive for me!

@ joergs

'I don't have a laptop either, Miss!' (Joke)

But now, seriously:

I am fully aware of the advantages of galvanic separation, which is why I also built my own Noxon.

My thoughts regarding: Maik-Version was:

If the cause of the upcoming problem is Bit-length variations and skewed signal slopes then make the problem even worse, as optocouplers exacerbate them.

That Optos doesn't have any problems with transmission speed is perfectly clear to me (see my fax example).

I understand that Ulf and Nick's Opto versions are better than Noxon, but if my MSG deviates significantly from the norm (in the worst case scenario), I believe the Maik version will still work the best.

Greetings, Jürgen

Juergen wrote:

I understand that Ulf and Nick's Opto versions are better than Noxon, but if my MSG deviates significantly from the norm (in the worst case scenario), I believe the Maik version will still work the best.

Hi Jürgen

You're probably right in principle.

If you're interested, I can also send you other circuits:

- einen Opto-Adapter mit Slew-Rate-Nachbrenner für RXD, wodurch zumindest dieses Problem nicht mehr auftritt (Signalverzögerungen von 5usec bleiben jedoch bestehen) - funktioniert prima bei mir, aber bisher hat noch niemand danach gefragt.

- a fully discrete circuit without an optos, likely with comparable performance to the Maiks adapter. A previous version of this circuit also worked perfectly for me.
Unlike "IC-Black Boxes", I can calculate the potential leakage current paths and, consequently, the worst-case operating scenarios there. This is why I find it more appealing than MAX or similar circuits.

Hi,

I have now finished the adapter, and it is working perfectly. I haven't tested it in my car yet, but it works flawlessly with an ECU and test software, up to 57600 Baud. The test software cannot go any higher.
I will test it in my car tomorrow, and then I will link the wiring diagram and pictures.
@ ulf: I see things differently regarding the adapters. I also built my adapter without the MC33199, but I think that anything connected to the serial interface should have a level converter (MAX232 or similar), or alternatively, a positive and negative power supply, as that's what the standard at least states, and it should then work with all PCs.
Especially when a voltage interface is only used for safety and is therefore designed to limit current flow, you should also avoid drawing current, especially for the OK signals. I also understand that 90% of computers can handle it.
Furthermore, with MAX 232, the setting of the OK current, which can vary significantly from device to device, is no longer required.
If there's no criticism of your adapter, I think it's great, but that's just my opinion about the interface.

Greetings

Hello @ all,

I'm going to tell you what I'm planning!

m.d.B. for constructive criticism:

I will implement the adapter design according to the following changes proposed by Maik:

First, I'm leaving the LEDs and their power supply out. Remove the resistors (I can't see them glowing in the case anyway, and they have no effect on the function).

Then replace the 7404 with a 7414 (inverter with Schmitt trigger to achieve the maximum signal transition rate).

Furthermore, I will protect the K and L lines using a 16V Z-diode and a 300 Ohm resistor on the STG side to prevent overvoltage (sometime).
(I have once measured SPG peaks of 40V)

If the signal slopes from the STG output are indeed as steep as claimed, then, in my opinion, it should also be possible to influence the bit length using the potentiometers in the transistor's base-divider circuit (even if only to a limited extent, but perhaps that's enough).

I'm looking forward to hearing your opinions!

Greetings, Jürgen

by the way: was bedeutet eigentlich 'AFAIK'?