TS losing comm with Microsquirt V3

[Whitbread]

Had this posted over on Microsquirt.com but there doesn't seem to be much traffic over there. If someone here can help me out, I'm ready to pull my hair out. I'm not a computer virtuoso by any means, but I'm not a newbie either.

Hardware - Win 7 32 bit
DIY Autotune USB/serial adapter with correct driver
V3 Microsquirt with 3.83a code loaded and matching .ini file
V2.6.01 TS registered, also tried a 2.21 version of TS
Tried setting com port to 6 and 3 in device manager and TS, same result either way. Restarted laptop after changes in device manager.
Tried 2 cheap adapters and the one from DIY.
Tried all 3 usb ports on my laptop including the one on the back that's never used so it can't be worn out

Long story short, had a cheapo usb/serial adapter and everything was peachy. TS and MS talked to each other happily for last 4 weeks as I've been testing things and setting up my file in TS. 4 days ago, suddenly started having communication issues with no software or hardware changes. Was suggested to get a new usb adapter. Got the new USB/serial adapter from DIY the other day. Installed the driver it came with but still no love. Tried re-bootloading the code multiple times to no avail either. Fuel pump relay is cycling on/off/on/off/etc when you key on, it will establish a connection and create a difference report, but a "serial data connection is lost" message pops up after attempting to send tune values to ECU, then the fuel pump relay stops cycling. If I hit "keep controller values", I'm able to change ecu type from 255 to 2 and hold communication for as long as I want. I can then do a restore and burn the correct tune values to the ecu. Once I cycle power to MS, it will not connect again. It will put up the bar graph of "reading ecu data", get to maybe 50%, and then the serial data error message pops up.

Here's what port checker says


After doing the detect and accept for the port settings, it still fails when you hit the test button.




Here's what I'm seeing now when it generates the difference report; the tach jumps between 0 and 8K and fuel pump relay cycles accordingly after re flashing firmware. It will sit here all day like this, as soon as I hit send tune values to ECU, it get about 25% across the bar graph and then communication drops. I have everything unplugged except for the main power relay and the fuel pump relay.





I downloaded a really old version of TS and finally got the serial test to pass! However, it still only held communication for a brief second.



Same old message here


At this point I was leaning towards something is screwy in the MS unit itself. A brand new Microsquirt V3 arrived today and I still have the exact same problem. If anyone can offer some insight, I will be indebted forever! I've attached debug and tune files.

[Whitbread]

Here's the debug file, hit the attachment limit.

[Whitbread]

I was playing around tonight and discovered some odd things.

After re flashing firmware, I went/can go into TS and change ecu type to 2 and reload my tune settings. I can stay connected with TS as long as I want and everything works perfectly. Even put my air drill on and cranked the motor over, coils and injectors are firing properly. https://www.youtube.com/watch?v=1aZsAKvk9jk

However, after cycling power to MS, I can't connect with TS and if I try and spin it over with the air drill again, there's no spark or injector pulse. I backprobed the 12v+ wire in the ampseal connector and the ecu does not loose power when it goes into this no communication/spark out mood. It's grounded directly to the battery 3" from ecu so it's impossible to loose ground. Battery was at 12.64v.

At least this rules out my laptop and any serial adapter problems and MS is theoretically ruled out also. So if the ecu isn't loosing power and still exhibits this behavior with every single thing disconnected electrically other than 12+ to MS, what possible explanation is left that everything works peachy until a power cycle then nothing works?

Does this ring any bells to anyone?

Thanks again!

[jsmcortina]

For better support you may want to consider installing the MS2/Extra firmware instead - available from the Downloads page.

James

[Whitbread]

For better support you may want to consider installing the MS2/Extra firmware instead - available from the Downloads page.

James

So you'd suggest the ms2 firmware over the micro squirt 3.830a code in the ecu? Any particular reason?

[jsmcortina]

So you'd suggest the ms2 firmware over the micro squirt 3.830a code in the ecu? Any particular reason?

More widely used, greater userbase and support knowledge. More firmware features. Easier tuning interface. Serial comms known to work well.

I am biased though.

James

[24c]

...I am biased though...

Understandably you are! ....but maybe you could add manuals are better too... when people can find them.

Mike

[LT401Vette]

when people can find them.

24C, haven't seen you around in a spell...
In the latest Extra firmware James's new pdf manual is built in and quite comprehensive. So yes that should be added as a reason to switch.

Whitbread,
The tach cycling from 0 to 8000 is what the B&G firmware does until you set ECU.
If you are dropping comms after sending your old restore point, it seems there is some bad value in the msq that is causing a bad problem.
I have seen the can_id set to something else, that will break it, but in your posted msq that is fine.
baud looks fine too.

If you aren't going to upgrade to MS2Extra firmware, we can try to reproduce it here with your msq.

[Whitbread]

when people can find them.

24C, haven't seen you around in a spell...
In the latest Extra firmware James's new pdf manual is built in and quite comprehensive. So yes that should be added as a reason to switch.

Whitbread,
The tach cycling from 0 to 8000 is what the B&G firmware does until you set ECU.
If you are dropping comms after sending your old restore point, it seems there is some bad value in the msq that is causing a bad problem.
I have seen the can_id set to something else, that will break it, but in your posted msq that is fine.
baud looks fine too.

If you aren't going to upgrade to MS2Extra firmware, we can try to reproduce it here with your msq.

Thank you very much for the insight! I'll try creating a new project first, if that doesn't do it, then I will load the ms2 code tonight into the micro and create a new project with that to see if that cures it.

I have a quick question for an expert here, if you watch my video you'll see there's a sizeable delay between when I start spinning the motor vs the coils sparking. Is there a way to change that so after the first time the missing tooth passes the Vr sensor, the ecu begins firing coils/injectors? This motor is kick start and I can only get probably 3-5 revolutions per kick with the spark plugs in. The kicking rpm is well over 200rpm, just doesn't last very long Haha.

[jsmcortina]

I have a quick question for an expert here, if you watch my video you'll see there's a sizeable delay between when I start spinning the motor vs the coils sparking.

Not sure why that takes so long.

I've run a pull-start lawnmower engine with 36-1 (using MS2/Extra) and it seems to fire up ok?

James

[Whitbread]

You guys rock! I owe you all a round at the local pub. Uploaded the MS2 microsquirt code with matching INI for TS and we're in business! Can't believe it was that simple this whole time!

[Whitbread]

Here's the base tune I just finished writing, if you guys could let me know if anything sticks out of place drastically let me know. Thanks again for all the help! I have spark while spinning over with the air drill so it looks like we're getting somewhere. Need to bust out the timing light to get the ignition offset perfect.

[kjones6039]

Ok.... you have me curious now! What is this 2 cylinder engine, that displaces ~22 cubic inches and turns 9000 rpm? I am also seeing some rather strange numbers in the ReqFuel calculations. Are those intentional?

Ken

[Whitbread]

Ok.... you have me curious now! What is this 2 cylinder engine, that displaces ~22 cubic inches and turns 9000 rpm? I am also seeing some rather strange numbers in the ReqFuel calculations. Are those intentional?

Ken

Built to the hilt yamaha banshee. Tables should go to 11K, not 9Krpm. Should make about 70hp on premium and 85hp on e85. Need to get bigger injectors or add secondaries before I can put the 'shine in the tank though. Here's the full build thread http://bansheehq.com/forums/index.php?s ... 855&page=1

[Whitbread]

Have a quick question regarding trigger wheel #1 offset. My tone wheel runs CCW, all the manuals I'm finding always talk about CW. I just want to double check that I'm putting in the right values here. Going from the front edge of the #1 tooth, I'm 50* past it right? With that offset value in TS it's spitting fire balls and kicking back at me telling me the timing is not correct. At 310* it's much closer to firing but still quite a ways off and spitting flames.

(that's #1 in red sharpie, I know it looks like a 2 in the pic)

[24c]

Hi Whitbread,
Just seen you link, didn't realise Banshee was a two stroke!
Have you got a side view that shows the sensor and the missing tooth wheel position, and a position for TDC? Looks like the timing wheel is fixed (welded??), so you either move the sensor or IMO remove the key in the taper of the flywheel, rotate the stator to a better position, and retighten.

Mike

[Whitbread]

Hi Whitbread,
Just seen you link, didn't realise Banshee was a two stroke!
Have you got a side view that shows the sensor and the missing tooth wheel position, and a position for TDC? Looks like the timing wheel is fixed (welded??), so you either move the sensor or IMO remove the key in the taper of the flywheel, rotate the stator to a better position, and retighten.

Mike

Yup, it's a wailing demon. Not many things are as dr jekyll/mr hyde as a built banshee.

I'll snap one tomorrow up at the shop that's more sideways. The VR sensor is at 5 o'clock. TDC on #1 is ~48 degrees after the front edge of the #1 tooth in the direction of crank rotation.

Yes, the timing wheel is pressed and tack welded on the flywheel and the flywheel is keyed to the crank. Why would I want to move it? Isn't the whole point of the tooth offset function to tell MS where #1 TDC is regardless of actual wheel/tooth position?

[jsmcortina]

Why would I want to move it? Isn't the whole point of the tooth offset function to tell MS where #1 TDC is regardless of actual wheel/tooth position?

Indeed.
The only time you might need to move it is if the missing tooth region happens to end up where there's a big change in engine speed during cranking. A tooth log during cranking will show if this is the case.

James

[Whitbread]

I got it running after using a timing light to trial and error find the correct #1 offset number for TS. Pig rich, but running on both cylinders. Thank you again everyone for the help! I'm sure there will be more questions.

https://www.youtube.com/watch?v=anPTM-O ... e=youtu.be

[24c]

Hi Whitbread,
The camera shake & sounds suggests you started it from a kick start, and it finally caught on the third kick... did you?
Never mind the pig rich, at least you've got a cold start setting!

Mike