Ducati Monster Forum

powered by:

March 29, 2024, 05:28:40 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
News: Tapatalk users...click me
 
   Home   Help Search Login Register  



Pages: [1] 2   Go Down
  Print  
Author Topic: 620 throttle body adjust  (Read 3519 times)
bagodoosh
New Member
*
Offline Offline

Posts: 42


« on: June 07, 2020, 11:53:48 AM »

just an info dump re: 2003 620 59M non-linear throttle body adj.

idle throttle angle is listed as 2.7 by bikeboy.
2002 service manual says 2.3 degrees (section D4, p28).
this post on ducati.ms says first value in TPS row in ECU table.
first value in "TPS legend" and consequently "Fuel Main", etc. is 2.0.

--- re: ECU dump

FTDI driver is REQUIRED.
i used IAW5xReader_v.28 for windows. it sets the baud rate to 64200. this failed on my regular RS232-OBD2 cable, and KeySpan USB-RS232 interface.

with those cables you can still use guzziDiag, JPdiag, etc to read ECU values (rpm, throttle position, CO trim, etc), reset codes, etc. it will pass the adapterTest. IAW5xReader logs in. rite before reading the ECU it tries to set the baud rate. it fails with an error pop-up which is not clear (read unsuccessful?). failed log file looks like this:

Quote
   10573ms Slow init
   13372ms Init Ok
   13521ms Diagnostic started
   13978ms Drawing:           
   13978ms Hardware: IAW59MHW010
   13979ms Omologation:
   13979ms Software:           
   13979ms Tester: LOADF:2.10
   13979ms Date: 2002.08.30
   13979ms IAW59M detected
   14408ms Diagnostic 2 started

i used KKL-USB cable with FTDI driver.
successful log file will indicate Baudrate switched:

Quote
   10294ms Slow init
   13064ms Init Ok
   13196ms Diagnostic started
   13633ms Drawing:           
   13634ms Hardware: IAW59MHW010
   13635ms Omologation:
   13635ms Software:           
   13635ms Tester: LOADF:2.10
   13636ms Date: 2002.08.30
   13636ms IAW59M detected
   14012ms Diagnostic 2 started
   14266ms Baudrate switched
   15950ms 29 8E 29 8F
   15951ms E3 26 69 27
   16144ms Login Ok
 2368176ms Checksum over download: B4B8
 2368274ms Checksum cross checked!

the transfer rate was about 113 bytes/sec. it took 39min to transfer 256KB. i unplugged the 15A blue headlamp fuse.

all that was unnecessary. my ECU.bin was an exact match of the the 2002 bin here:
Impulsive_Duc's wiki page, 28640511A_2002_Monster_620_OEM_IAW59MH010.bin in this repo, XDF file, tunerPro.

run tunerPro. load XDF. load BIN. check TPS legend table, row 1.

--- to do:

i have Innovate Motorsports 3894 kit with dual wideband oxygen sensors.
i am waiting on exhaust bung adapters (1/8" BSP according to bradblack),
..and carbtune pro 2.

i will add more once i start the process.
Logged
bagodoosh
New Member
*
Offline Offline

Posts: 42


« Reply #1 on: July 03, 2020, 08:04:10 PM »

found on YT
Logged
bagodoosh
New Member
*
Offline Offline

Posts: 42


« Reply #2 on: July 28, 2020, 11:07:22 AM »

the work is done. i'll post my info in segments.

the instructions to set the TPS are posted above. here's my take on it:

the throttle body needs to be at minimum position. to do this the following 4 adjustment screws have to be backed off enough to show a gap: right hand (vertical) throttle stop, balance/sync, left hand (horizontal) throttle stop, fast idle.

throttle stops use 2.5mm hex. to reach the right hand screw i had to remove the air box. 1 bolt on the front. 2 hoses. 2 air intake hoses. 2 bolts in the back. service manual says to just lift the air box with the rear bolts attached. i had to unscrew the bolts from the air box. to reach the left hand throttle stop screw i had to undo the fuse box. it's an awkward angle. there's not much depth to the hex bolt. it seems easy to strip the bolt if you don't go in straight.

the TPS adapter is wired as follows:
looking at the adapter pins head on; lock clip on top.
pins LEFT to RIGHT: 3 2 1
1 - Green/Black = Ground, goes to ECU pin 20
2 - Green/Brown = V+ (5.00v), goes to ECU pin 32
3 - Orange = TPS, goes to pin 3

to measure TPS, you have a few options:
a) supply ground & spot-on 5.00 volts to pins 1 & 2, measure pin 3
b) measure at ECU, with bike on (to supply 5v)
c) measure at connector, with bike on (to supply 5v)

i found C to be the easiest. take a wire tie (thin wire). remove the cover. cut to about 2in. insert wire into pin 3. reattach TPS connectors. it doesnt have to fully click. just enough to make contact. attach alligator clip to exposed wire. make sure the wire does not contact body/ground.

you need an accurate volt meter. i have a UT61 that i had calibrated to 5.000 volts.

the TPS needs to read 150mv +/- 15mv. that's 135-165mv. adjust TPS until close to 150. it will drift slightly when you tighten the bolts.

when you tighten the TPS bolts, left bolt lowers the voltage. right bolt increases it. the trick is to go back and forth between the bolts in small increments so you don't drift off from your set value.

my throttle angle at this point was 0.0 (using guzziDiag). the ducati.ms post says to reset the TPS here. i did. the throttle angle reset to 1.9 degrees. i don't understand why bradblack suggested this step. this is inaccurate. 2.0 degree should be (2.0 * 105 + 150) = 360mv, not 150mv. it is harmless since another TPS reset is done later.

after this is done, the throttle adj screws need to be tightened until TPS reads 360mv. 360 = 150mv + (2.0 * 105). 105 is the mv/degree for PFC3 TPS sensor. 2.0 is the idle TPS value according to the ecu map. 150mv is where the TPS was locked in. it could be between 135-165mv. let's just say it was 150. (mine was 151mv)

i used the left hand throttle stop to adjust the throttle to 360mv. right hand throttle stop and fast idle screws were tightened until the voltage increased a bit. then i backed off the adj screw. mind you the resistor is not super accurate. the value will increase. you back off the bolt and the value will not decrease. so you open/close the throttle valve gently and it'll land somewhere different. it takes patience. ultimately after the 3 screws are back to normal, you should still be ~ 360mv. normal meaning to turn them as far as possible without the TPS voltage going over 360. i left the balance/sync bolt alone because it'll be modified during the sync procedure.

TPS reset using guzzi diag. TPS voltage should read 360mv @ ~ 2.0 degrees. mine was 2.1.
clear ECU fault codes.

put airbox back in. reattach fuse box. bike is ready to be balanced and tuned.


attach carb tune pro 2. start bike with air bleed adjustment screws fully turned in on both throttle bodies. use the fast idle knob to keep the bike running. use a flat head screwdriver to adjust the balance/sync screw until the two sides are equal. OCD warning: they will never truly equalize. as you increase throttle, they will go out of sync. then settle back down differently. they will also read different over time as the bike warms up.

if you're inside the garage, use a fan to blow the exhaust gases away from you. the fan will also cool the bike a little.

turn out (loosen) the air bleed screws on H & V cylinder until RPM reads 1200 +/- 100 AND the vacuum reads equal on both cylinders. back off the fast idle knob.

my TPS angle, after i equalized the throttles, read 2.7 degrees. i am not sure if another TPS reset is warranted here. i did not. the balance screw opened the throttle up from 2.1.

end product here: TPS angle 2.7 degrees, idle RPM 1200, throttle bodies synced.

next, exhaust tune/CO trim.
Logged
bagodoosh
New Member
*
Offline Offline

Posts: 42


« Reply #3 on: July 28, 2020, 04:59:18 PM »

the Innovate Motorsports 3894 kit comes with 2 bosch LSU 4.9 oxygen sensors.
the threads on the O2 sensor is M18x1.5. data sheet can be found here.

the exhaust bungs are 1/8 BSP. an adapter is required to connect to the O2 sensor. in ducati's video they use a long copper pipe with what looks like rubber hose.



my first attempt was to have something similar. attach a brass adapter, 1/8 BSP on one side and 6mm barb on the other. attach rubber hose to the brass adapter to route the gases to a container with the O2 sensor attached. the container needs some restriction on out flow to build some back pressure and reduce contamination of fresh air.

brass adapters are fairly cheap from china. i bought 5 for $6 on ebay.

the kit comes with 2 bungs that measure .998in. i used couple foot long of 1.5" PVC pipes to construct the following:



i drilled the end cap and enlarged the hole to .98 using a lathe. i press fit the bung into the end cap and created a seal from the inside using hi-temp RTV (black).

side note: i tried drilling the end cap with a 1" paddle bit. that was a disaster. the hole is too large, not circular and not smooth. the drill bit scraped more than cut.

O2 sensor on one end:



gases enter perpendicular to the O2 sensor and exit thru the opposite end:



the exit port is too large, but there are other issues.

this design was a failure.
  • brass adapter attached to the exhaust gets very hot. the rubber hose that is attached to it will melt. solution was to use hi-temp silicone hose rated at 400F.
  • oxygen sensor itself gets hot during operation. the internal temperature is about 600C. the outside gets around 140F. PVC releases bad chemicals at this temperature. it will also get very soft. the O2 readings were not reliable. it would measure extremely lean (50+ AFR) once the PVC would get hot.

an all metal adapter is ideal.
Logged
bagodoosh
New Member
*
Offline Offline

Posts: 42


« Reply #4 on: July 28, 2020, 05:50:52 PM »

there is NO ready-made adapter that is M18x1.5 Female on one end and 1/8 BSP Male on the other.
bosch datasheet indicates that the O2 sensor needs about 28mm from the threads to the tip. in addition to M18x1.5 threading the adapter needs to have sufficient depth to house the O2 sensor.

one option is to connect 2 adapters together, i.e.
M18x1.5 F to M10x1 M (gtm21_m18x15f_m10x10m) + M10x1 F to 1/8 BSP M (gbm2p_02gm_m10x10f)
M18x1.5 F to 1/2 BSP M (gtm2p_m18x15f_08gm) + 1/2 BSP F to 1/8 BSP M (pbp21_08gf_02gm)
M18x1.5 F to 1/4 BSP M (gbm2p_04gm_m18x15f) + 1/4 BSP F to 1/8 BSP M (pbp21_04gf_02gm)

the cost is at least $35 per adapter. i am not sure if the M18 depth is sufficient. i could not find any specs.

i decided to create my own adapter. design:



there's an aluminum outter housing for the O2 sensor. it needs to be tapped with M18x1.5. the other end there's a brass piece that's press fit into the aluminum and held against rotation by a set screw. the nipple is 9.6mm in diameter with a 6mm hole drilled. it'll require a 1/8 BSP die to create the threads.

the threads on the exhaust bung end. you can't screw the adapter all the way in. the thread depth has to be 8.5mm maximum so that the bung will mate against the brass piece and act like a washer.

some foolery on the lathe, and the finished adapter:





verify that both adapters measure the same. there has to be zero doubt that the difference in O2 sensors is due to the setup.

i am curious if this setup is ideal or whether it requires flow via a small exhaust port. i think exhaust pulses create some vacuum so fresh oxygen may get sucked back into the adapter.

i came across "oxygen sensor extenders" which are basically this:



they move the oxygen sensor away from the direct flow of the exhaust gases and restrict the flow. they're used in car applications where the catalytic converter is removed. i am not sure how much exactly it affects the reading.

the optimal AFR range is fairly large. the absolute reading is not as important as the relative value between the two cylinders.

CO to AFR table here. formula is roughly -.377C+14.51, where C = CO%

CO% recommendations:
ducati service manual: 1.5% - 6% (13.9 - 12.3 AFR), cyl difference < .5% (.2 AFR)
bikeboy: 4-5% (13-12.6)
ducati video: 3-6% (13.4-12.3)

higher AFR = better gas mileage, less power, sluggish low RPMs
lower AFR = more power, better response, worse gas mileage
too high AFR = lean pop on engine braking (decel pop)
too low AFR = terrible gas mileage, sluggish

ducati and bikeboy both recommend sacrificing throttle balance to achieve similar fuel burn on both cylinders. that means to match the two cylinders' AFR, vacuum will no longer be synced.

up next: tuning..
Logged
DarkMonster620
According to some, a bottomless pit eating machine
Post Whore
******
Offline Offline

Posts: 37932


Enjoying the ride


WWW
« Reply #5 on: July 28, 2020, 05:54:44 PM »

What about something like this, https://cdn.shopify.com/s/files/1/1764/8111/products/O2_Installation_Block_720x.JPEG?v=1527711360

Logged

Carlos
I said I was smart, never that I had my shit together
Ducati is the pretty girl that can't walk in heels without stumbling. I still love her.
"When you have eliminated all which is impossible, then whatever remains, however improbable, must be the truth."
DarkMonster620
According to some, a bottomless pit eating machine
Post Whore
******
Offline Offline

Posts: 37932


Enjoying the ride


WWW
« Reply #6 on: July 28, 2020, 05:57:46 PM »

I used Brad's setting for my TP and CO and bike ran like it should have from factory, I had Hi flow airfilter and modified airbox with hi mounted Termis

for CO I used the Gunson Gas Tester   http://www.gunson.co.uk/search/gas-tester
Logged

Carlos
I said I was smart, never that I had my shit together
Ducati is the pretty girl that can't walk in heels without stumbling. I still love her.
"When you have eliminated all which is impossible, then whatever remains, however improbable, must be the truth."
koko64
Post Whore
******
Offline Offline

Posts: 15650


« Reply #7 on: July 28, 2020, 07:19:14 PM »

Thanks for sharing  chug.
I have a Gunson gas tester. Make sure you let it and the bike warm up for accurate readings.
Logged

2015 Scrambler 800
2004 GSXR 750
greenmonster
Hero Member
*****
Offline Offline

Posts: 1282


« Reply #8 on: July 29, 2020, 03:15:21 AM »

Tutorial material?
Logged

M900 -97 
MTS 1100s  -07
koko64
Post Whore
******
Offline Offline

Posts: 15650


« Reply #9 on: July 29, 2020, 03:29:04 AM »

Good idea.
Logged

2015 Scrambler 800
2004 GSXR 750
Orange16
New Member
*
Offline Offline

Posts: 44


« Reply #10 on: July 29, 2020, 03:43:44 AM »


my throttle angle at this point was 0.0 (using guzziDiag). the ducati.ms post says to reset the TPS here. i did. the throttle angle reset to 1.9 degrees. i don't understand why bradblack suggested this step. this is inaccurate. 2.0 degree should be (2.0 * 105 + 150) = 360mv, not 150mv. it is harmless since another TPS reset is done later.



you're doing a non linear tps adjustment.  you don't, at any point, use the electronic reset feature in the diagnostic tool.  that's only for linear tps.
Logged
bagodoosh
New Member
*
Offline Offline

Posts: 42


« Reply #11 on: July 29, 2020, 03:02:49 PM »


that's the TechnoResearch TR200047 O2 Installation Block. it costs $142 USD per adapter!  Shocked you'll still need the 1/8 BSP to barb adapters and silicone hose. it kinda looks like it has a one way valve to direct exhaust gases out. it would work. at that price range, you're better off having a local shop create the adapters for you. no? can you find it cheaper?

you're doing a non linear tps adjustment.  you don't, at any point, use the electronic reset feature in the diagnostic tool.  that's only for linear tps.

that's what bikeboy said as well. let's just say in the bike's history the TPS had already been reset.
in the ducati.ms post i linked earlier, it says:

Quote from: Impulsive_Duc
after talking it through with my local Ducati expert (Mike @ Nichols), and going over your information, we came up with a number of ways to properly fully reset a nonlinear TPS -- even if every setting is borked.

i followed that procedure from the ground up just to make sure everything ist gut, ja!
Logged
DarkMonster620
According to some, a bottomless pit eating machine
Post Whore
******
Offline Offline

Posts: 37932


Enjoying the ride


WWW
« Reply #12 on: July 29, 2020, 03:09:01 PM »

I used my Gunson, with a 'Y' adapter I had made so I was able to measure both cylinders as an average . ..  then I would do one by one . . .

BTW, was just giving an idea, you can copy it
Logged

Carlos
I said I was smart, never that I had my shit together
Ducati is the pretty girl that can't walk in heels without stumbling. I still love her.
"When you have eliminated all which is impossible, then whatever remains, however improbable, must be the truth."
bagodoosh
New Member
*
Offline Offline

Posts: 42


« Reply #13 on: July 29, 2020, 04:54:35 PM »

the service manual and bikeboy state to set the CO levels with the airbleeds, and to only use the software CO trim if it is out of range. the airbleeds also change the idle RPM. the first time i tuned the bike, i leaned it out so far that the bike was idling at 2500 RPM. what is the procedure to set the idle speed then? do you set the CO trim to zero first?

my approach was the following: use airbleeds to set the idle speed @ 1200, and use software CO trim to adjust the burn rate (AFR).

LM2 power requirement:

the O2 sensors require a decent amount of power to remain heated. LM2 comes with a cigarette lighter attachment. a car battery can be used, but i wanted an instant AC/DC solution rather than worry about car batteries.

LM2 manual states the power requirement as "Dual Channel 8-14 Volt / 4 A (max, 2 A nominal)".

initially i used a 12V, 5A rated "medical power supply":



i encountered a problem where sensor1 would produce an undervoltage error, and go into warmup mode. this would result in about 20 seconds of no data. either the power supply is insufficient, broken or LM2 is more power hungry than stated.

i switched my setup to two 12v, 4.15A intermec power supplies connected in parallel:



i have not experienced the power failure since.

there are a few options here. you can use a bench power supply, or any laptop power supply that is 72W+ and then use a buck converter to deliver 12V. in general i don't trust chinese specs on cheap brands, especially with higher rated power supplies.

warm up:

the ECU has two tables associated with fuel adjustment and "warm up".

one is Engine temp:



graphed in Fahrenheit:



the adjustment factor of 1 is at 75C or 167F. any hotter and the ECU is pulling fuel (leaner).

the second one is called Fuel warmup:



my understanding: rows are engine temp in C. columns are marked "seconds" in the XDF file. i think that is incorrect. that would mean the correction factor of 1.0 at 75C would happen after 2500 seconds (42 minutes?). i doubt it. bikeboy states:

Quote
One very important thing to know about the 1.5M ECU is that it runs rich for the first 3,000 rotations every time you start the engine after turning the key or kill switch off and then on.

this is a 59M ECU but it could still apply. the columns could be number of rotations. it seems weird to have columns for 4,8,16,32,... rotations. why such precision at early stages that the bike turns on?

cold start:

here's the graph of a cold start. the cylinders are synced, however they have different AFR. you can see sensor1 (blue) going out of service due to the power issue stated earlier.
right around 600 seconds (10min) is when the engine temp reached 65C.

i recorded the engine temp manually by setting a timer and writing the number down on paper. later i matched my timed recording with the time stamp in the O2 log.



it looks like i started adjusting the airbleeds around 400 second mark. 680 second mark looks like a restart. any time the bike is turned off the AFR starts climbing (lean). right after restart, the bike runs rich for a while. 750 seconds i may have used the fast idle knob to increase the rpm.

restart 1:

here's a restart after i adjusted the cylinders to be ~.2 of each other. the engine temp started at 87C (189F).
here it seems like the bike leans out after a minute and then settles back down after another minutes.



restart 2:

here the bike was at operating temperature. it looks like it starts out rich. leans out about a minute in. and then settles down at the 2 minute mark. sensor1 (blue) had loosened ~ 60sec mark at the exhaust bung, so it was reading leaner.



restart 3 & 4:

i have two restarts here
restart 3: CO trim = 6, red = O2 sensors, green = engine temp
restart 4: CO trim = 4, blue = O2 sensors, yellow = engine temp



if i was to conclude anything, i'd say wait minimum 200 seconds on a warm restart before making adjustments.

making adjustments:

here things got a bit interesting/frustrating.

1) i was only able to make CO trim adjustments in guzziDiag. I tried JPDiag and IAWDiag, but the settings wouldn't hold.
2) when you modify the CO trim, the ECU (or guzziDiag?) slowly ramps up/down the fuel. it takes about 20 seconds. then you have about 10 seconds to read the value. at 30 seconds it rapidly restores back the previous value. i.e. you start at CO=5, increase CO to 10. after 30 seconds, your actual CO value is 5, even though the software will show 10.
3) to set the value, you need to [OK] the value during the 30 second "test mode".

a visual demonstration:



CO is changed from 6 to 3 (red), and 6 to 4 (green). notice how the AFR remains flat only from 20-30 second mark.

example 2:



effect of CO trim on AFR:

in the previous 2 examples, we see that a change in CO from 6 to 4 results in a change of .3 AFR (.15 AFR/CO). 6 to 3 results in delta .5 AFR (.17 AFR/CO). 5 to 0 results in 1.15 (.23 AFR/CO).

here's another experiment:



yellow - 0.14 AFR/CO
orange & blue - 0.16 AFR/CO
green - 0.18 AFR/CO
red - 0.19 AFR/CO

finale:

i set the CO trim to 5 to end up with an AFR of 12.8x. the bike felt sluggish at low RPM and i was getting pretty bad decel pop. i increased the CO trim to 7. there's a noticeable power increase at low RPMs. it rides well. no decel pop except after a long trip when the engine temp is >200F. there's a hint of it on short trips with engine temp < 190F. with CO trim 7 i've gotten an average of 55mpg on 3 tanks. i will increase the CO trim to 8 to see whether the decel pop is eliminated.

is that a cheap method to tune the bike? decrease CO trim until decel pop occurs at engine temps >200F, then increase until it's eliminated.
Logged
bagodoosh
New Member
*
Offline Offline

Posts: 42


« Reply #14 on: July 29, 2020, 05:45:54 PM »

I used my Gunson

how do you like the gunson?

i looked into the gunson G4125. it costs about $250 shipped to US. the innovate 3894 dual O2 kit is about $330. the basic kit with 1 sensor is $245. the gunson seems to take 1 sample per second. the LM2 display will update once a second, but if you log to SD card, it will record about 12 samples/second. the gunson user manual states to calibrate to 2.0 which is the "background level for CO2 in the atmosphere". i never understood if the device measures CO directly or indirectly via CO2. CO in fresh air should be 0 ppm. i don't understand the significance of 2.0 on their device.

there's a downside with the LM2 in that it will only accept older SDSC cards v1.00 (up to 1GB) and v1.01 (2GB & 4GB) formatted in FAT16. the log files are proprietary binary format. 7.5 min log file is 170KB. 1GB is plenty of space. logfiles can be exported to CSV via their software LogWorks 3.
Logged
Pages: [1] 2   Go Up
  Print  
 
Jump to:  


Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Simple Audio Video Embedder
Valid XHTML 1.0! Valid CSS!
SimplePortal 2.1.1