[rusEfi] 1990 Geo Storm GSI - Advice/Suggestions? #18

[kb1gtt]

I expect this is the VR. While it looks a bit different than most of the VR's we have seen, I believe it is still a VR.

VR.PNG (216.76 KiB) Viewed 17077 times

I'm not sure how much I agree with the VR resistors in the wiki picture. I would generally expect better CMR if you did not connect that one side to GND. You have a blinking light, which is good. If you end up with noise issues, then we can look at these resistors. For now it appears your issue is that the signal is making it to the STM, but not showing up in your logs. I think this is a software issue.

[AndreyB]

Blue box with green center - just one trace away from "CAM 1st", I am sure it would have continuity as well. Very strange that the LED is blinking but the software does not register.

Have you tested just CAM channel with the battery or both channels the same way? Do you have an oscilloscope?

[Spine]

I attempted the test based on the link provided for VR input.

I meant to state that I attempted with a AAA battery on the CAM SIG IN and CAM NEG using both poles (+ bat to CAM SIG IN and -bat to CAM NEG, then reversed battery)

No, I don't have an oscilloscope. I wish I did.

[Spine]

Another thing I noticed, when using the engine sniffer it will have a popup that shows which trigger, but the physical pin is blank.

****Begin Popup****
Channel Trigger #1
Physical Pin:
****End Popup****

Is this something of importance? This popup is on the engine sniffer tab when I leave the mouse in one spot. It is not registering that Trigger #1 is attached to a Physical Pin - based on the popup.

[AndreyB]

No, this is not an issue since that's just this tab waiting for data and not yet receiving any since it depends on trigger events which we do not have. I will make this less confusing under https://sourceforge.net/p/rusefi/tickets/280/

your log has

Sun Apr 10 23:56:37 MDT 2016<EOT>: postMessage EngineState: primary trigger input: PC6
Sun Apr 10 23:56:37 MDT 2016<EOT>: postMessage EngineState: secondary trigger input: PA5

so I am not worried about the settings. Can you please do the battery test on the CRANK pair of wires just to see if the other channel has the same behavior? That's probably the last test before sending the board back to me

dso quad is the baby oscilloscope of choice on this forum - http://www.ebay.com/itm/DSO203-Nano-4-Channel-Mini-V2-Quad-Pocket-Digital-Oscillo-scope-/301457393871

[Spine]

Ok, so I did both CAM and CRANK again. Switching battery + and - on each highlighted post. I noticed that the CAM and CRANK light would flash when + was on CAM/CRANK SIG IN, but did nothing if + was on CAM/CRANK NEG (while touching - to corresponding post). Hard to see my notes in the picture - zoom in to the CAM/CRANK inputs.

cam-n-crank_sensor_test.jpg (3.2 MiB) Viewed 17056 times

MAIN_rfi_report_2016-04-11 11_27.csv

(442.01 KiB) Downloaded 639 times

[AndreyB]

Same nothing. I cannot explain since I did the battery test prior to shipping the board.

Mon Apr 11 11:28:51 MDT 2016<EOT>: postMessage EngineState: trigger#1 event counters up=0/down=0
Mon Apr 11 11:28:51 MDT 2016<EOT>: postMessage EngineState: trigger#2 event counters up=0/down=0
Mon Apr 11 11:29:03 MDT 2016<EOT>: postMessage EngineState: trigger#1 event counters up=0/down=0
Mon Apr 11 11:29:03 MDT 2016<EOT>: postMessage EngineState: trigger#2 event counters up=0/down=0
Mon Apr 11 11:30:37 MDT 2016<EOT>: postMessage EngineState: trigger#1 event counters up=0/down=0
Mon Apr 11 11:30:37 MDT 2016<EOT>: postMessage EngineState: trigger#2 event counters up=0/down=0

There is another thing we can try but you have to be really careful and I think you should only do it if you have a female-female jumper wire. A male-male jumper wire is too much risk of shortening wrong pins and burning the green board - the green board is pretty fragile once you start applying voltage directly to it.

1) detach the green board from the red board
2) take a female-female jumper, put one side on the +3v pin, and touch PC6 with the other end of the jumper wire a couple of time while console is connected.

Same triggerinfo test after you apply 3 volts to PC6 a number of times - now with digital signal applied right to discovery pins instead of applying unprocess signal to the MAX9926 input pins.

[Spine]

So, here's what I did...
I created a jumper via the plug here (only stripping the yellow wire):

IMG_20160411_124016.jpg (2.09 MiB) Viewed 17050 times

I then attached is so the yellow was on the 3V.
Connected it to the console and did:

reset_trigger
triggerinfo
**did jumper test with yellow wire to PC6 15-20times**
triggerinfo

I also noticed something on the green board when I was looking at it. It is in the picture below (bottom right), along with the two pins I jumped.

direct-jumper test.jpg (3.35 MiB) Viewed 17050 times

[AndreyB]

What was the output of triggerinfo after the jumping the yellow wire?

The red area is the remaining of the audio jack which is removed to get the case cover close, too tall as is.

[Spine]

Bah - I forgot to include the log file, sorry. It's right here:

MAIN_rfi_report_2016-04-11 12_27.csv

(489.52 KiB) Downloaded 628 times

[AndreyB]

Wow, wow:

FATAL error: PB9 req by i3 used by i1

somehow this was not in the previous log files?! I wonder if that's our issue. Please run resetconfig command and reset and reconnect (this would reset the pump and fan relay setting which I hope is fine for now), and try the same yellow wire again? I am really hoping that's our issue.

[Spine]

Will do in a second. Remember we switched the pump output to 12v hi and moved ignition4 to where pump used to be. That's the only change I've made besides the different trigger settings - which is now back to 1+1 (I believe).

[AndreyB]

Let's see if this improves anything while the green board is all by itself. Was the red LED between the black and blue buttons ever constantly ON?

I wonder if the issue was always there but it was just not too visible or if we got it recently somehow.

[Spine]

Doing "resetconfig" didn't do anything. It kept showing the same error over and over. I didn't do the triggerinfo/reset_trigger part. So, here's what I did:

resetconfig
soft reboot (black reset button)
**saw messages kept going**

resetconfig (via button on right, and again with "go" button)
cold boot (pulled both usb cables, then plugged them back in power first, then console)
**saw messages started and kept going**

Attempted resetconfig again via both button and manually typed then go button
held reset button for 15sec then released.
** com8 reconnected
** messages continued to go.

MAIN_rfi_report_2016-04-11 13_21.csv

(527.14 KiB) Downloaded 608 times

[AndreyB]

Where is different way to reset everything: on the rusEfi console welcome screen there is a button to erase chip - before connecting. Please try erasing chip and then press "program firmware/debug" to get the firmware back on the chip. I believe this would wipe out the faulty configuration.

Also there is an option to connect with TunerStudio and set the conflicting pins to UNASSIGNED.

[AndreyB]

Still, somehow there were none of these "FATAL error: PB9 req by i3 used by i1" messages in most of the previous log files.

[Spine]

Both the blue and red leds flash (the ones between the blue and black buttons).

I'll attempt to erase the chip...give me a second...

[Spine]

Here's the output for erasing the chip:

chip-erase001.png (45.28 KiB) Viewed 17051 times

Do you want me to program "debug" or "release"?

[AndreyB]

Debug

[Spine]

Ok, so I programmed with Debug. Here's the output:

debug-programmed001.png (42.8 KiB) Viewed 17022 times

I attempted the trigger input again via the jumper wire. Good news and bad news. There is no trigger#1 (PC6) response. However, there is trigger#2 (PA5) response. Here is the attached log file. BTW, I had to cold-boot the board a few times before it was going consistent without dropping com port connection.

MAIN_rfi_report_2016-04-11 14_19.csv

(1.64 MiB) Downloaded 553 times

[Spine]

Is there a way to "switch" the crank and cam inputs? I only have a cam sensor on this engine.

[AndreyB]

Hurrah!

Mon Apr 11 14:29:57 MDT 2016<EOT>: postMessage EngineState: trigger#1 event counters up=0/down=0
Mon Apr 11 14:29:57 MDT 2016<EOT>: postMessage EngineState: trigger#2 event counters up=11/down=12
Mon Apr 11 14:29:58 MDT 2016<EOT>: postMessage EngineState: trigger#1 event counters up=0/down=0
Mon Apr 11 14:29:58 MDT 2016<EOT>: postMessage EngineState: trigger#2 event counters up=13/down=14
Mon Apr 11 14:30:06 MDT 2016<EOT>: postMessage EngineState: trigger#1 event counters up=0/down=0
Mon Apr 11 14:30:06 MDT 2016<EOT>: postMessage EngineState: trigger#2 event counters up=15/down=15

"cam" and "crank" are just the labels on the board, that's just a unified recommended way to wire it. You can move to the other set of wires and you can make PA5 primary input and PC6 secondary at the "Engine->Trigger configuration" menu.

Let's put the green board on the red board and do the battery test before we change any configuration.

So I need to improve fatal error handling since we've lost two days until we've noticed the fact that we have a fatal error. https://sourceforge.net/p/rusefi/tickets/281/

I know connectivity is unstable at the moment I am looking into this already. https://sourceforge.net/p/rusefi/tickets/277/

[Spine]

Ok, now that's strange. I'm 100% positive I didn't hit the thing over 5,000 times!!!

Well good news, both CAM and CRANK showed input. Here's my log file...

MAIN_rfi_report_2016-04-11 15_06.csv

(349.48 KiB) Downloaded 390 times

Also, it seams much more stable? It didn't drop multiple times when doing the test. In fact, it didn't even drop once.

[AndreyB]

5000 is totally expected with our poor connection between battery and wire. Lets try cranking? Can we disconnect fuel pump etc just in case?

[Spine]

Yes, as soon as I'm outta this S4!T storm at work I'll get it back in there and do more testing.

FYI, the support here is awesome!

[Spine]

YES!!! It shows RPM now!!! Fantastic. So, how do I wire in a 2post coil? I'm only going with 1 coil for now. The coil I'm using is an Accel 140001 from my old race car.

This log is from the car cranking over.

MAIN_rfi_report_2016-04-11 19_15.csv

(148.13 KiB) Downloaded 564 times

[AndreyB]

Do you happen to have a junk yard close by? For a two wire coil you need an ignition module. See http://rusefi.com/wiki/index.php?title=Manual:Hardware_Ignition see J701 but you need the connector. We still need to figure out an ignition module which goes with a connector.

Speaking of your trigger signal, looks like all four teeth are the same size? How does OEM look on your car - is it single injector or simultaneous injection? I do not understand how they which cylinder is which. It's OK for ignition as look as the distributor wires the coil to the correct spark.

[Spine]

OK, got one problem. When I change the output for the fuel pump relay to PE10, it stops registering RPMs. The lights flash like they did before, I also get an unstable connection, ie; it keeps disconnecting and reconnecting. I had to wipe the firmware and reflash to get it stable and read again.

[AndreyB]

Please grab fresh http://rusefi.com/build_server/rusefi_bundle.zip - I've improved FATAL handling in rusEfi console application a bit. You need to see the error message and just fix the configuration via TunerStudio, you do not need to start from scratch.

[Spine]

OEM has 1&3 together then 2&4 together. It also has a "driver pair" for them. Check the wiring schematics on page one. That's all I know.