Scopes
 

It is my guess that there are a number of electronics people here. At least there are a number of computer people here.

I have recently been involved with a CAN repair on a ML. It comes after a couple seminars heavy with CAN discussion. My case and all the training says I don't need to know what it is saying establishing the network is mostly a multimeter task.

But I wish to see and possibly identify which controller is talking by his signature bit. I'm sure it is not being done to fix cars but who knows what techniques can be developed till one has all the tools to view.

I posed the question on a tech site over the weekend and it will be seen by some real geniouses, but I'm sure the answer is not to a automotive question so I thought I would ask it here.

I am going to try to post a scope pattern that I took from a properly functioning CAN on an early ML. The scope I am using is state of the art automotive but I am at the lowest time frame 20us. I want to stop that pattern till I can see all the digets and then I want to scroll through enough that I could do something and hope to capture a response.

One of the excellent replies I got was to approach the problem differently and I was directed to a site that I could get a CAN reader, some sort of developmental device.

I got this idea while brousing Tektronics Sales sites. I found I didn't know what I needed but I did find a company that rents these expensive scopes by the month. So I am thinking of doing a project. Anyway here is the pattern I wish to expand on:

I looked through a number of saved scope [atterns to find the CAN one. I don't know if anyone is interested but there are some good stories.

I am going to post another one just to show one aspect of a good automotive scope. There are many.

Anybody know whats wrong here? This is secondary ignition on a 4 cyl.

Don't really know what I'm looking at, but many years ago a tech showed me a pattern on my 4-cyl VW engine that looked much like that...said the other spikes were the result of excessive resistance from bad plug wires...

...am I close?;)

Try looking at it this way:

The activation voltage for all cylinders is the same probably because it is a result of jumping the rotor gap (highest form of resistance). The distinct characteristic is the flat firing line on the one cylinder.

The problem was a plug wire shorting to the head.

I see now...

...Would a scope be beneficial to us shadetree diy'ers like myself? I'd like to solve some of those knotty electrical problems that codes alone don't tell you...

...or is it cost-prohibitive? I would like to get a salvage one from a going-defunct shop if possible?

Check this one out. It represents a graphic illustration of the secondary air test. The first part is an O2 sensor in closed loop.

It drops to minimum voltage immediately as the airpump is engaged. The engine management controller wants to see less than .04v. In this case it only made .135v. A new O2 sensor was the cure.

I was examining Initiation sigs from a scanner to OBD1 in an attemp to make a HH digital tool and used a Tek TDS-3012.
Small Phos color Digital design/de-buger type scope with time frames of 5 Gs/s..
It also has a Single signal capture feature for each trigger ..so I was able to get each pulse time and amplitude sq. wave configeration, but i was not able to break the coding sequence they use...

I have access , if need be..

I have two large pdf documents from Bosch documenting the arrangement of the bit sequences. There is an arbitration section that I presum muust indicate the controller but it may just prioritize messages by content.

Just thought I would take a larger look. It would probably be a lot like looking at brain waves by an amateur.

Do you have both CAN lines superimposed on one another? One is supposed to drive high and the other low from their recessive state of about 2.5 volts.

I am assuming they are for the moment, and that the bit transfer rate is 500 Kbps. This would give a bit time of about 2000 nsec.

From my knowlege of CAN...

Each message has a Start of Frame (SOF) and an End of Frame (EOF). The SOF marks the beginning of a message. The SOF consists of a sigle dominant bit after a predetermined idle state. The EOF is an idle period of seven bit times after the message, therefore unseen. The logic '1' bits are the driven states, fluctuations going positive and negative, that are evident in the scope trace.

There is a standard frame and an extended frame. The standard frame consists of an 11-bit identifier, Remote Transmission Request (RTR) bit, a 6-bit control field, up to 8 bytes data, 2-bit Cyclic Redundancy Check (CRC) field, and an in-frame acknowledgement field. The RTR is used to id the frame as a data frame, The control field gives the data length.

The in-frame acknowledement consists of 2 bits. The first is always passive and the second is driven by any listener that correctly receives and qualifies the message. Qualification is by use of a mask and the CRC. First, the listener checks a couple of internal masks with that of the senders identifier. If the message is accepted, then the CRC is checked. If both these checks pass, then the buss is driven during the second bit of the acknowledgement field. Transmission of this bit may originate from MORE THAN ONE MODULE. It is therefore impossible from viewing a message to determine who acknoledges the message.

It is possible to determine who sends the message. Remember to ignore the first dominate bit, the next 11 bits are the identifier (the node ID of the sender). Looking at the trace that is driven low, we have (after the SOF) a binary 01010111111 or a 2BF hex as the sending node identifier. If I look at the top trace I get 2FF hex. These traces don't look like they come from the CAN buss at the same time. It would help me to see them separate.

I don't know if any of the two are valid IDs or if MB uses this CAN standard. But I hope it helps in any case.