I tried your app, I get an error - 'Missing MFC70.dll, reinstall the application' or something very similar...
I tried your app, I get an error - 'Missing MFC70.dll, reinstall the application' or something very similar...
And this was supposed to be the *easy* part of Windows software development. I updated the post with the .zip attached to have a statically linked .exe . As long as the MFC lib is the only one missing it should be ok now , even though the binary is now like 4x the size of what it used to be.Originally Posted by ShelGame
I'm going to try and put the 22v to the bootstrap and then power the rest of the smec in a minute.
OK, I tried it. No ID.This is on a known '88 2.2 T1 computer. I can read the call off it with D-Cal, so I know it's working right.
OK, well I realized as I was going to bed why it didn't work (or, at least one reason why it didn't work) - COM1 can't talk at 976 baud (or 62500 for that matter) because it's a standard PC serial port. It'll take some figuring, but I'll try to plug my FTDI (USB to TTL) cable into my bench rig, bypassing the serial converter board, and see if it works that way.
I wonder, though - does your software needs to do anything special to work with the FTDI chips?
I've used the other logworks software and winlog with the FTDI chips @ the LM baud rates and it was fine. As long as your OS has all the drivers loaded properly, the USB device should be fine.
Even if the program gets the ID back successfully, it won't know what to do with it and probably still fail. I still have to write the function to interpret the ID it's receiving back. I'm more interested in the portmon capture of the serial transactions, so just capturing that and sending it to me will be enough.
I couldn't get my smec into bootstrap last night. I need to re-check my connections and see what the deal is later on today. I really should wire this thing into a breadboard and stop messing with the electrical tape and twisted wires.
I'm curious how are you determinig which ram map from the SMECID? I've exported all the cal ID's sorted by year and T1/T2, should it just be a matter of taking the year and T1/T2 item and mapping it into the 1-15 different types?
Also, for the SMECID what does the leading '_' signify? Any digit plus the following 3 is a match or will it be a leading 0?
Yep, that's what I did. I looked at the binaries I had and if the data locations were the same, I assumed the RAM locations were as well. I found 15 different RAM maps and categorized the binaries as belonging to one of these. Now, I wasn't really able to confirm that for the 85/85 LM stuff. I think they actually may have a bit of variation in those years for the RAM maps.
I'm not super worried about the accuracy of the maps at this point, since you can change any of the addresses you want after picking a map. Plus, if anyone knows those maps, and wants to share them, we can easily update them.
What about the leading '_' is it a wildcard type character?
You mean in the cal ID? That's because the CalID is stored as 2-bytes. It's always something like A123. Where the '123' matches the last 3 digits of the P/N. but, the 'A' is an unknown. I don't understand the significance, other than a placeholder. For the binaries I have access to, I used the actual letter/number. For the others, I put '_' to note that I didn't know what the value was. I actually don't use that in MiniDash at all. It's only in the Excel file...
Is pin 31 (RX) always held @ 5v? For some reason my smec is holding that pin at 5v almost always. The RX led on my schmartboard serial module is almost always lit, it flashes off on occasion. Is this normal or is my smec shot?
Hmm, I never noticed. But, that doesn't really sound right...
I should say it flashes when I just remove the ASD 12v on startup. Otherwise it seems like it's on solid. I know the same module on my LM flashes non-stop, to the point where I can't tell if it's on more or off more. But this one on the smec seems like it's on solid. That and I noticed that the CEL comes on, goes off, then stays on and it's always very dim.
Anyways, did you manage to get the usb->ttl converter hooked up? I still have the autoid function to write, so there's no huge hurry, but I am curious to see what the serial comms look like. Maybe I'll see what a j/y around here will charge me for a smec.
No, I realized what a pain it would be to open up the box on my bench rig. I'll try to get something from it this weekend...
I ran the latest driver on the minivan and got no Dll errors but No Auto ID. This was done on the regular COM1 though one of Chad C's serial converters. I was going to try the old '89 baudmod cal but if the driver is using 976/62500 then no cal will work. I have a rs232 to usb adapter to try out. How about adding the low/high baud to the driver as parameters?
It's not going to get the autoid back successfully because the function to actually do the ID is just a stub that returns false. So even if the driver gets an ID back it's going to fail. I really just want to see what a run with portmon looks like. If you want to try and do the autoid with portmon watching the serial port you're running it on that would be the most help to me right now.
The driver is using the 976/62500, 7812/62500, or 976/7812 baud rates depending on which ecu is selected (or being tested). I rewrote a bunch of the stuff I had hard coded in (which the baud rates were in the autoid .exe above), and I can break the rates out pretty easily with a couple more edit boxes. Unfortunately, I won't be in a position to work on it until after I get back from vacation on the 25th.
If you want make a run with portmon and send the portmon, that'd be great. If you want a recompiled version of the original LM driver that won't complain about the missing .dlls I can do that too, since it only takes a couple of clicks to produce it.
OK, I've managed to get the bench smec working with the DLP-TXRX-G usb adapter. I think the 5v on the rx line is right when I look at the schematic for the serial adapter, but being able to check a box on the dlp device to have it invert the signal makes things 100x easier.
I can see the SMEC echo back the ID so now I've got to make sure I'm pulling the correct ram map based upon the ID. Now I can get on with the real work.
Thanks Rob and jpcturbo for all your help with this so far. I'm finally excited about this project again.
I have to take that previous post back. After closer inspection of the SMEC's responses it became clear that I had a short between the tx and rx lines on my adapter, and I was just reading back everything I sent. Back to the drawing board. I'm sure I'll get it eventually.
Ok, I now have the bootstrap working properly. I must not have been hitting the pin with enough voltage to overcome the zener. If anyone wants the perl scripts (or I can perl2exe them) to bootstrap a smec and suck the cal off of it, LMK and I'll post them. The reason they're useful is that you'll be able to bootstrap a smec with a USB adapter since DCAL only works with real serial ports. The reading of the cal is a little slow since it waits 1s for every 256 bytes, but the scripts serve their purpose for troubleshooting.
Also, I'll post detailed instructions on how to configure the DLP-TXRX-G adapters for external use in a few weeks (just need to connect tx and rx and set inversion with MPROG). I'd like to confirm it works on my CSX before posting. Also, this means people won't have to solder them into the LM anymore (unless they want to).
I'm going to get some wiring diagrams for the 88/89 cars this weekend and try to get my bench rig to more fully simulate a in-car situation. I'd love to get something like the megasquirt stimulator going, but it'll take me *forever* to figure out how to simulate the HEP signal.
Risen,
What wiring diagrams do you need? I have all the 85-93 wiring FSMs and I can make scans.
84 Laser XE TII A555 89 Caravan SE TIV A568
Thanks for the offer. I'll definitely keep that in mind. I actually got a set of '89 manuals from Amazon for like 60$ shipped so I'm good at this point. I thought I had a set, but it was for '85. The plugin for the smec cars is actually close to done. I've got it out to a couple of people for testing purposes right now. It won't contain autoid, but it will work for SMEC/SBEC. Once I'm satisfied I'll release it for wider consumption.
If anyone is somewhat knowledgeable about software development, has a smec/sbec based car, and a serial or usb adapter and would like to test it PM me.
Posting Permissions
Reply With Quote