USB Consult Disconnect and other errors

Discuss hardware problems here.

Moderator: Matt

Post Reply
Jason84NA2T
 

Posts: 11
Joined: Fri Feb 17, 2006 8:30 am
Location: Michigan, USA
Contact:

USB Consult Disconnect and other errors

Post by Jason84NA2T »

After having an apparently non-fuctioning board in my ECU for the longest time and trying everything, I finally used another USB cable and it seemed to work fine (grrrrrrrr).

Well, at least it worked fine for a little while. I was getting random disconnects of the consult, leading to me having to disconnect the USB cable, plug it back in, and then reconnect. It only seemed to disconnect regularly when I would hit higher load or RPM, the rest of the disconnects were random. I have re-checked all related wiring.

This is what the screen looks like after the consult locks:
http://www.redz31.com/nistune/nter.JPG

This worked and was mostly livable but annoying, until tonight where I now can't even connect to the consult. Now I get a consult reg error as shown below:
http://www.redz31.com/nistune/nter1.JPG

So after reading the other threads here on a similar (the same?) issue, I am no closer to figuring it out. I have another ECU with another NisTune board installed, and another USB cable that I will be testing here when I have the time and motivation.
Russ84NA
 

Posts: 66
Joined: Fri Mar 31, 2006 11:36 am
Location: Ohio, USA

Post by Russ84NA »

Jason,
What version of NisTune are you using? Are you running your LM-1?
I had similar problems with the last version .9081a. Haven't had time to check this current version .9084a that was posted on the 30th.
Matt said that it is supposed to support the LM-1 again.
Jason84NA2T
 

Posts: 11
Joined: Fri Feb 17, 2006 8:30 am
Location: Michigan, USA
Contact:

Post by Jason84NA2T »

Russ84NA wrote:Jason,
What version of NisTune are you using? Are you running your LM-1?
I had similar problems with the last version .9081a. Haven't had time to check this current version .9084a that was posted on the 30th.
Matt said that it is supposed to support the LM-1 again.
Using version .9084a

The LM-1 is not connected, because mine is still completely messed up.

The problem now is that the consult won't even so much as connect...
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

Post by Matt »

I've put together the USB Diagnostics document
http://home.aanet.com.au/nistune/reg.html

However some other things to check:

I've noticed that there are problems some times with the first batch of cables that were sent out. This is something worth looking into. The crimping tool sometimes doesn't give a good enough connection

I should post some pictures up of what to do here to fix it but here is a description. This applies both to the 3 pin connector for the USB and ECU cables
- Use a small screw driver to push in the 3 holes where the connectors clip into the plastic housing, whilst gently pulling on the connector cable.
- Do this one at a time until all three connectors are removed from the plastic connector housing.
- Gently dab some solder on each connector so that you get a better connection

Then clip the connectors back into the connector housing.

The USB cable or ECU cable failing will cause an ECU disconnect. On newer boards I am soldering the connectors. If the ECU cable fails on these revc/d boards the ECU still stop functioning, so cable connections are vital

Unfortunately its one of the things that has cropped up which is why these boards are going through the testing they are

The other things to check are
- Does your USB connector have two 47pf caps and a 0.1uf cap connected to it? This stablises USB noise that can cause problems. If not I can send you the parts to add to your connector

- Different USB cables have been documented in the diagnostics document as causing problems. Bryant identified one of his cables had increased noise originally. If there is not enough shielding from automotive noise spikes you can have problems

Newer boards have these capacitors added and being NVRAM based do not use the romulator, so any hardware problems associated with that device will be elimianated


Also your picture shows patch rev 5. The latest is patch rev 10, please use this with the latest nistune software uploaded several nights ago. Make sure you use the Z31_128_E.adr file as the Z31_256_E.adr is for the newer boards
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

Post by Matt »

Russ - what version was working prior to 0.9081a?

there are two things changing - patch code revision and NIStune revison, so i need to make sure its the NIStune code causing the problem

i'll stick a romulator back in my car and try it out with the same boards you guys are using. currently i've been running the NVRAM version
Russ84NA
 

Posts: 66
Joined: Fri Mar 31, 2006 11:36 am
Location: Ohio, USA

Post by Russ84NA »

I was using .9071a when the LM-1 unit was working. I am not sure which patch code revision, but it was what ever the latest was at the time.
Jason84NA2T
 

Posts: 11
Joined: Fri Feb 17, 2006 8:30 am
Location: Michigan, USA
Contact:

Post by Jason84NA2T »

I'll test the usb cable first by buying a nicer shielded one... neither of them I have used are.

To update the patch code I can just copy my maps and tables into the new stock binary you include?

The rest I can try or check when I have some time. Thanks.
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

Post by Matt »

with the latest install

1. load the 16K address file (Z31_128_E)
2. then load your modified binary as the main image
3. then load one of the Patch Rev10 binaries included as the comparison image

using the mapcopy feature you can copy your maps into the comparison image and save that.

now release your remapped patch rev10 image as the main image and upload that to the romulator

im going to check the USB timeouts set as this may be causing some dropouts. i noticed this last time in the car and it may be affecting you guys

my main aim at this stage is to get everyone who has currently got boards completely up and running. i need to find some time to run through this in the car again, so will get a RevB board install up and running for next week and see how it runs and where i can make things better
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

Post by Matt »

you will have to copy the maps 'manually' using COPY/PASTE since the mapcopy doesn't work in that version

just found out about it tonight and i have fixed it. will be in the next release in a few days time. currently sorting out the romualtor issues with uploads/timeouts
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

Post by Matt »

just been chasing up this issue last night

I've been running 88ss_nt.bin on the bench with the RevB board and romulator. The load is around max and RPMs around 6000rpm

There is no dropping in USB communications at this range. This means that it is likely that the cause of the problem is not firmware related/software related but electrical noise issues from the car

I then added a spark plug to the setup to try and induce noise on the bench but still stays connected.... so I need an actual engine for this problem to occur...

I have had similar reports on a CA18 using the RevC board which around high load/revs does drop out. The tester will be getting some logs from the dyno session today when it drops out

The best thing I can do at this stage is acknowledge the errors in USB communication and keep the connection going as having the connection drop out is annoying.

Other option is an oscilloscope on the lines and see if I can filter the USB data lines further

Please let me know if you have the capacitors installed on your USB connector board. Those are important
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

NIStune bug checking

Post by Matt »

did some testing in the Adelaide hills tonight as usually most of my testing is on the bench, but noise issues are different

im trying to reproduce the disconnect problems you guys are seeing. I tried two different boards RevB and RevC with romulator... mainly RevB since that is what you guys have

The RevC board pictured below is just a smaller version with the USB capacitors on the board rather than the USB connector miniboard

I had no disconnects whatsoever... and I tried my CR31 image and then the 88 NA Federal image and then the 88 SS image. These all seemed to run my RB30 engine okay (suprisingly)

I jumpered the ECU to the chasis as would normally be the connection when the ECU is permanently installed.

I pushed the engine through high load revs bouncing off the limiter at 6200 rpm and still no disconnect

So I am highly suspect there is a noise issue through the USB cable either due to the cable itself or the vehicle. For more assistance I will need your debug logs which are saved in the NIStune logs directory

Attached are logs from my runs, pics of the install and my debug log showing the different uploads

I found that using the shortcut keys is very handy whilst on the road, and that using CTRL-TAB will switch between internal windows inside NIStune, and using ESC to close windows you no longer need.

The logging function keys need to be setup, something like P for play and R for record so you can press F4 to connect to consult, F7 to log and then R to start recording. I will do those shortcut keys for the logger in the next version
Attachments
nt_pic1.JPG
(25.01 KiB) Downloaded 7685 times
nt_pic2.JPG
(50.36 KiB) Downloaded 7685 times
nt_pic3.jpg
(60.5 KiB) Downloaded 7685 times
nt_pic4.jpg
(77.81 KiB) Downloaded 7685 times
nt_pic5.jpg
(109.9 KiB) Downloaded 7685 times
nistunelog.zip
NIStune debug log file
(258.03 KiB) Downloaded 253 times
nistune_2006-11-14_2313.csv
(91.79 KiB) Downloaded 226 times
nistune_2006-11-14_2306.csv
(203.42 KiB) Downloaded 236 times
nistune_2006-11-14_2257.csv
(145.55 KiB) Downloaded 224 times
Jason84NA2T
 

Posts: 11
Joined: Fri Feb 17, 2006 8:30 am
Location: Michigan, USA
Contact:

Post by Jason84NA2T »

Matt,

Sorry for the slow reply, I sent this in an e-mail as well.

I checked the USB board today, it does have the capacitors on it.

I have yet to try updating the patch code. I will do that ASAP when I get home and have access to the car.

I still have yet to try another usb cable, this one is cheap though.

attached to the e-mail I sent you is all of the logs from that folder... I know I was logging one time when it happened at least, but I don't know which log it is and have not figured out the player yet.

Also noted that most of the time it freezes is when I am also maxing out the MAF voltage up near 5.1V and a tp of 101.

Jason
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

Post by Matt »

Okay another update on this problem...

This is the setup I was using when it was working with no disconnect issues

Compaq P266
RB30 ECU with Rev2 (with USB caps installed) and RevC boards (onboard caps but no cap between GND and Shield)
Romulator installed into both boards
Power connection via 3 pin mains connector to 12V-240V power inverter
Jumper cable from ECU chasis to vehicle chasis

The vehicle chasis ground went through the ECU chasis which connects to the USB shield via the connector. This then connects to the PC chasis and then to the Power Inverter via the 3rd ground pin. I think (need to check tonight) that this connects back to the loom ground in the cigarette lighter


Today I got 5 disconnects within 15 minutes on the way to work (along with the Handle error) using a different setup

Dell Latitude P266
RB30 ECU with Rev3 board (onboard caps but no cap between GND and Shield)
NVRAM installed into board (since my Dell causes romulator to lose data on power on and serial connect)
Power connection via 2 pin mains connector to 12V-240V power inverter


The ECU chasis is floating (not connected to vehicle chasis), and this is connected to the USB shield and then the laptop chasis. There is no ground connection from the laptop to the cigarette lighter. Therefore the USB shield is connected to a floating ground.


I think this floating shield issue is causing the data +/- lines in the USB cable noise which causes the disconnects. The compaq setup seems to be the ideal solution in terms of hardware, but I need to get it working on the non grounded dell setup. Normally ECU is installed in vehicle chasis so that and the USB cable would be grounded


Things I'm going to check -
- Grounding the ECU using jumper cable
- Does installing 0.01uf cap between USB shield
- Does grounding the laptop chasis to cigarette lighter ground

Other things I noticed -
- Disconnect did not occur during high load / revs today. Will double check logs tonight when they did occur
- Handle error occured each time and required disconnect/reconnect. Software not releasing handle correctly. This must be fixed before attempting any retry strategy
- Once handle problem fixed, disconnect/reconnect in the background if this problem occurs will not be noticed by user (excpet for milisecond loss of data during reattempt)
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

Post by Matt »

This is an old thread but problems are still evident with some vehicles with the USB consult connection so I have been working on it

Some people have problems whilst others dont.

From my investigations last time (about 7 months ago), I found based on two laptops used, that the one with the GROUND pin connected to my 12-240V converter which plugged into my cigarette lighter had NO disconnections

I also found with the second laptop which was disconnecting, the problem was a lot less when the ECU was be grounded properly to the vehicle chasis (ie put back in the kick panel) or jumper cabled to the chasis

So I suggest that if you have USB disconnection issues that you ground the chasis of your laptop (either shield from RS232 or USB connector) to the vehicle chasis

For finding where noise comes from (coil, alternator etc) I need to find time now to put the scope back on the vehicle and measure USB noise under certain conditions and see what we can do to minimise this so you guys dont get USB disconnects



I have also spent several nights looking into this issue with the software handling when it happens, and have a software solution. When the USB device detects errors, it will attempt 5 more retries. If the hardware is not responsive, then we disconnect the USB device, reenumerate it and then reconnect.

That all happens in the background, and I've written it so that it is quick, transparent and logging is not stopped on USB disconnect.

Similar code change will soon flow through soon for wideband reconnect also (ie when ignition turned on and accessories lose power, the wideband connection gets dropped and requires manual reconnect)
Matt
Site Admin
 

Posts: 8961
Joined: Sun Jan 29, 2006 1:45 am
Location: Adelaide, Australia
Contact:

update...

Post by Matt »

Took some photos of noise coming from NIStune USB port when not connected and there is significant noise the higher the RPMs which causes disconnects on some peoples vehicles.

That noise comes from the engine and causes a pattern of noise to be displayed on the scope. The higher the RPM the more the noise

Now need some more time to find out exactly how it gets into the lines and the best ways to eradicate it. I found as per my previous testing that the grounded laptop coped better with the noise (AC supply is 3 pin built in)

The noise occurs over both USB D+ and D- lines, but usually the noise matches both lines so it doesnt affect data. However when there is a spike in one line but not the other, then you get data errors

A lot of RPMs, hence a lot of noise, and a lot of spikes causes NIStune board to disconnect

Bench testing was performed on 0.9010 in XP and seemed to work. However testing in Win98 on the car, restarting the device wasnt effective enough so I will have to add some more code for recovery which hopefully I start getting time to do tonight and test on Win98 on the bench

I'll post some pics soon of the noise that you see coming from the engine. It is most likely feedback from the coil (ignition) system. Electrical induction causes a lot of noise
Post Reply