Hyperblade
-
Posts
51 -
Joined
-
Last visited
-
Days Won
3
Content Type
Profiles
Forums
Events
Gallery
Blogs
Posts posted by Hyperblade
-
-
On 1/20/2024 at 10:50 PM, Adamw said:
It looks like you possibly have more than 1V ground offset there, especially at higher RPM when that fuzz appears on the signal.
Good spotting, yes we do.
We are looking at options for trying to work out where that may be coming from as there doesn't seem to be any settings in ECU to adjust the threshold (just bury head in sand and ignore the problem approach to solving it).
Plug A and Plug B both have sensor grounds (Gnd Out) pins at the moment all sensors are connected to A and B has nothing on it.
I'm assuming internally are they treated the same? e.g. you don't need to hook sensors that are connected to plug B to have sensor gnd from plug b?
-
On 4/10/2022 at 9:43 AM, Adamw said:
Ok, I tested you file on the bench with a simulated engine running etc so full processor load, I can get up to about 13000Hz before it tops out.
So my best guess is at higher frequencies the signal voltage is no longer meeting the thresholds. It may need a stronger or weaker pull-up, but it is hard to guess without a scope showing what is wrong with the signal.
So it's taken me a bit but I finally got the Oscilloscope trace for you...
I had someone fairly knowledgeable do it, and everything on the sensor looks good and as expected and the signal is stable and strong.
I'm curious as to how the S2000 plugin ecu works fine, unless you've got some additional hardware?
-
Here's my saga with same errors
In the end moving the ground away from coils ground solved it for me.
- Hodgdon Extreme and dx4picco
-
1
-
1
-
On XtremeX i'm just doing a bit of rewiring in preparation for going to e-throttle, I want to be able to plug in the E-throttle later and just reconfigure the settings to get it to go.
One of the things required is the E-Throttle Relay, can I set that up ahead of time and just control it as soon as the power comes on e.g. if (true) or if (1=1) so it's always powered for now?
-
4 hours ago, Adamw said:
Ok, I tested you file on the bench with a simulated engine running etc so full processor load, I can get up to about 13000Hz before it tops out.
So my best guess is at higher frequencies the signal voltage is no longer meeting the thresholds. It may need a stronger or weaker pull-up, but it is hard to guess without a scope showing what is wrong with the signal.
Ok thanks for looking into it, will try and get a scope somehow.
-
30 minutes ago, Adamw said:
Duty cycle is just the percentage of high vs low. Does the duty cycle change as speed increases or just suddenly when the signal drops out?
Can you attach your tune and a log.
When looking at it in real time I thought the duty cycle changed with speed in a consistent fashion, but looking at logs that doesn't appear to be the case.
Files here:
https://drive.google.com/drive/folders/1FnPhQe_9f1wKRYdibQIAtl0XKAYnfxF8?usp=sharing
-
@Adamw I swapped in a new sensor as I couldn't get access to an oscilloscope, same issue.
I confirmed the pull up resistor has to be on for me to get any signal, so not sure how the S2000 base map works with it off?
But I did notice something different this time, the Duty Cycle hits 100% on DI2 when it cuts out.
How is that duty cycle calculated? Is it based on a previous max value the input has reached, or is it based on what the ecu supports in the firmware?
-
11 hours ago, Adamw said:
It looks correct enough, I would change a couple of things:
- I would change cut effect to adaptive. This means the cylinders will be cut in a randomised pattern. With constant mode at some specific conditions you can end up with the same cylinders continuously being cut so it is very hard on the crankshaft etc.
- Start Cut TP 100 may need to be a bit higher (often around 60%). This is the amount of cut that first gets applied when the control range is entered. If too small there may not be enough cut to prevent an overshoot condition when engine acceleration is very high (ie when wheels come off the ground or boat leaping out of the water at WOT).
- Cut decay time could possibly be reduced - it may feel slow to recover after a limit at 100ms.
Brillant, thanks for the feedback.
-
My original map had dual RPM limit tables, and I've turned this off (due to the shift light not changing when it's activated when switching between the too.) and I also lowered the limit by 250 rpm as it was not meant to be set that high.
Originally it was 9250 and I've dropped it to 9000 (which i do not want to go over ever) and I'm happy for the soft limit to start applying at 8800.
But i'm not confident that the rev limit is applying correctly, it would be awesome if someone more knowledgeable could double check my setup.
Map attached.
-
5 hours ago, Adamw said:
Well the picture of your setup that you attached shows it turned off hence my suggestion.
You will have to get an oscilloscope on it to see what is happening to the signal at high speed.
Sorry about that, took that then decided to actually write it down so it was searchable in the future for others, have removed that image now.
Thanks for the help.
-
1 hour ago, Adamw said:
Dose turning on the pull-up resistor help?
It was on when i had the issues, doesn't get a signal when it's off.
-
Setup: K20 with a Honda S2000 AP1 Gearbox and G4X XtremeX.
The speed sensor signal keeps dropping out just over 5000 Hz (DI2), consistently, around 130km/hr with current setup.
What is the best way to diagnosis where the fault lies?
Form a configuration standpoint I've checked out the Link S2000 Base map and it has the following
- Driving Wheel Speed Correction: 0.0%
- Slip Filter: 20Hz
- Pull Up Resistor: Off
- Active Edge: Falling
- Freq Filter: None
- Calibration 9600
Where as mine(which has been on dyno at tuner) has
- Driving Wheel Speed Correction: -10.0%
- Slip Filter: None
- Pull Up Resistor: On
- Active Edge: Rising
- Freq Filter: None
- Calibration 12974
Which obviously stands out as being different in the Active Edge, (but could be because the Honda S2000 base map is for plugin ECU), but would that cause it to just drop the signal over 5000Hz? I wouldn't have thought so which leads me to think the sensor may be faulty?
What else can I check?
Video Here:
-
I have the following shift light in my race car, which has green and red led's, using 2 Aux outputs from the ECU.
Currently the Red are set to CEL and the green set to the Link Shift light setup.
However the default link shift light strategy isn't good enough, going from blinking led's to solid just isn't noticeable enough to accurately time gear changes.
I would love a setup where the green LEDs come on then start flashing in the last bit of the rpm, then the red start flashing as warning before the cut.
e.g.
- Shift: Green Solid -> Green Flashing -> Red Flashing
- CEL: Red solid
I figured i could use a table to do some of what I wanted.
This is for green LEDs (ignore 1800, 1850 that's just for testing).
Which I think will work.
However I can't work out how to also have the Red come on and also still perform the CEL function.
Red LED setup (ignore 2000, 2050 for testing)
I was thinking something like this, but then thought that will only show the light at the top RPM as it uses the table above
Is what I want to do possible?
Just thinking out loud, I guess other option is to have all cells set to 100 under the shift point so if CEL turned on then it would show, but would still flash red at shift rpm? Not ideal, but better then nothing?
My other idea was I could potentially show the Red solid, via the conditions of > rpm or CEL = active. But that means it obviously doesn't flash which is what I really want.. So couldn't use the table.
Any advise appreciated from those more experienced.
-
On 10/8/2021 at 8:25 AM, Hyperblade said:
The K20a doesn't run a suppressor.
Thanks all the info and the help, really appreciated.
Just to follow up on this thread, I've left the CAN Lambda grounded to the chassis and it's all been dyno tuned and everything is reliable.
Not exactly the ideal setup I wanted, but nothing else was affected by the noise.
It would be nice if the Link CAN Lambda unit just handled the noise like everything else, e.g. the link ecu which was also ground to the same place on the head which works fine... -
10 hours ago, Adamw said:
Yep I would say that confirms it is likely noise.
The biggest electrical noise connected to the engine would be the ignition coils if there is a large loop area. Does you coil power supply have a supressor on it? Im not sure if the K20s have one factory of not - but if not the factory may have kept the loop short using other methods.
Im pretty sure the K20 coils only have 3 pins, so the the secondary winding actually "grounds" through the 12V supply pin. If there is no supressor, then for the spark to complete its circuit back to the coil, the spark goes out the end of the coil to the spark plug, jumps the gap to the electrode, through the cylinder head, to the ground cable back to the battery, through the battery, back down the 12V wire, through ign switch etc and eventually back to the coil though the 12V pin. With a supressor on the coil power supply this effectively creates a short to ground when a high frequency passes through it so the loop is much shorter.
Have a look at the example at 8:24 in this Adaptronic video: https://youtu.be/ZTDsm6b69Lk
The K20a doesn't run a suppressor.
Thanks all the info and the help, really appreciated.
-
On 10/6/2021 at 12:13 PM, Adamw said:
Have you got a jumper pack or something you can try powering it from?
Ok, I've managed to get it going consistently and also erroring consistently, was tricky with no jumper pack / spare battery.
The issue is actually with the CAN Lambda ground which was wired up to go straight to the cylinder head (along with coils, ecu grounds etc all in same location.)
After a lot of mucking around I've stumbled onto the fact that when it's ground to the chassis it works fine, but if it's grounded to the head it shows the errors, I changed location on the head to a spare bolt, but it caused the same errors.
So I changed the ground from the engine to the chassis end closer to the battery, but still get the errors when the CAN Lambda is ground to the head.
So does that confirm it's a noise issue? (same wire is used when going to chassis vs head)
Is there any common causes, and solutions for that?
-
8 hours ago, Adamw said:
Have you got a jumper pack or something you can try powering it from?
Thanks for the help.
No I don't, So I tried something different.
I figured I would eliminate the bulkhead connector so direct wired the main feed (from the bussmann stud and ground to another location.
That appeared to work, so I changed the ground back to the original location that still worked.
So figuring I had narrowed it down to perhaps the bulkhead connector being the issue I pulled the pin out on the socket side and put a new cable and pin in (left engine loom side).
I then hard wired the new pin to main stud, and it worked.
I figured at this point maybe the cable from the bussmann had something wrong with it. So I wired in a new one (16awg again)
Errors came back.
So then I went straight from bussmann stud to the new pin and I got the same errors (this previously worked fine).
I then shut the car down, started it back up and the errors have gone away again.
I've measured the voltage at just before the bulkhead connector at 14.37v consistently through all the chopping and changing.
I am getting a 48 Over Voltage error now as well on startup.
The lambda is actually reading a value through most of this after a delay.
I'm left even more confused now, I thought I was narrowing it down, but now I don't think so, it just seems so random.
Here's a picture of the setup that both worked, and didn't (back to back 1 min in between) that's 12awg to 16awg at the bulkhead connector.
Any thoughts on what to try next?
-
13 hours ago, Adamw said:
Yeah, 16 & 33 are nearly always power supply related. It can be either wires are too small so there is significant voltage drop, or significant inductive noise.
If it works fine connected to the main feed then that would suggest it is noise related. Can you connect to the ECU supply? That has extra filtering that will clean things up further.
I'm running from the bussmann box using 16 AWG Tefzel to a bulk head connector (Deutsch), then 16 AWG Tefzel to the CAN Lambda Unit, with 16 AWG ground to the engine head.
I tried connecting the CAN Lambda to the main stud again to double check everything and I'm getting similar errors
Version 1: Alternator/Battery > Stud > CAN Lambda (Capacitor just before plug)
- 48 APE Over Voltage
- 34 Open Circuit RE-IPE
- 16 Heated To Long
- Repeat 16/33
- Then Lambda works (with error 16 Heated To Long still showing)
So I then tried going to the feed to the ECU and got the same errors.
Version 2: Alternator/Battery > Stud > Bar > Relay 1 (to ECU as well) > CAN Lambda (Capacitor just before plug)
- 48 APE Over Voltage
- 34 Open Circuit RE-IPE
- 16 Heated To Long
- Repeat 16/33
- Then Lambda works (with error 16 Heated To Long still showing)
Does the fact that the lambda is working eventually, I assume when it's naturally got up to temp or using less current to get there, suggest a power supply issue (voltage drop?), rather then a CAN Lambda Unit/Sensor issue?
If so what's the best way to verify that?
-
Struggling a bit with this one and need some advice.
Initially I was getting the following errors on the CAN Lambda
- 16 heated to long
- 33 open circuit ape-ipe
Searching across the internet, it appears that it can be a broken sensor or noise on the line.
I run a Bussmann distribution box with a main 12v feed from alt/battery to a stud which is connected to a common bar which then goes out to relays/fuses.
e.g. Alternator/Battery > Stud > Bar > Relays/Fuses
Initially I had one relay feeding the Coils and the CAN Lambda (switched on ignition on).
Version 1:
Alternator/Battery > Stud > Bar > Relay 3 > Coils/CAN Lambda
On rereading the suggested setup I changed it to have a dedicated relay just for the CAN Lambda.
Version 2:
Alternator/Battery > Stud > Bar > Relay 10> CAN Lambda
Alternator/Battery > Stud > Bar > Relay 3 > Coils
I still got the same errors.
However I temporarily connected the 12v for the CAN Lambda to the main stud on the bussmann box bar (where the main feed connects to (so before the bar) and the errors disappeared (did this twice to make sure).
Version 3:
Alternator/Battery > Stud > CAN Lambda
The suggested solution for that is the alternative wiring diagram from the Link CAN Lambda Manual where it mentions using a 22uf 35v Electrolytic Capacitor
So I've installed that
Version 4:
Alternator/Battery > Stud > Bar > Relay 10 > CAN Lambda (Capacitor just before plug)
And this is where it gets weird.
I started the car and had it running and it was still showing error 16 Heated To Long then at the 0:13s mark it started working giving a reading while still showing error 16.
Decided that was a bit odd and so stopped the car and started it again with video recording.
I then proceeded to get (in order)
- 42 MES Over Voltage
- 34 Open Circuit RE-IPE
- 16 Heated To Long
- 33 Open Circuit APE-IPE
- Repeating Errors 16/33
- Then at the 1:56 mark it stays on 16 Heated To Long, but again starts working and giving a lambda value.
Here's a video:
I'm very confused, it's a brand new loom I have done myself, so there is a high chance of have screwed up somewhere, but I don't know where to start diagnosing this as it's working but not working and I would have thought moving to it's own relay off the common bar would have fixed the issue on it's own.
I would have said the sensor was broken, except it seems to be working...
Engine is a Honda K20a (Euro R) with Link G4X XtremeX in a 84 Toyota Starlet
-
Simon's comment here:
It looks like a solution to this is to shift where the driver is located.
If you move the ftd2xx.dll file from the C:\Link G4X\PCLink G4X\Link USB Drivers\i386\ftd2xx.dll folder to C:\Link G4X\PCLink G4X\Link USB Drivers\i386\ftd2xx.dll
Didn't give the correct path, but I managed to work it out.
Copy the following dll
C:\Link G4X\PCLink G4X\Link USB Drivers\i386\ftd2xx.dll
to
C:\Link G4X\PCLink G4X
And PCLink will now open.
-
On 2/15/2021 at 9:34 AM, Adamw said:
Oh, be aware also, Im pretty sure the K24 crank sensor pinout is different than the K20, so you need to research that yourself.
That's correct they are different.
Also the K20a from the Accord Euro R (CL7) runs the K24 connector and pinout, I would guess the FD2 K20a might also as it's based of the CL7 engine originally.
Taken from the Haltech K20 page...
-
13 hours ago, Adamw said:
You will have to use an older version such as 6.18 which doesnt require the driver.
That's very annoying when I'm just about to set up the firmware for a new ecu which will be on the latest version...
Maybe in future versions of PCLink the dll check could be done when you try to connect to an ECU.
-
17 minutes ago, Adamw said:
You dont need to uninstall anything in my experience. Just plug-in the ECU so the driver can finish installing.
I don't have the option to plug in an ecu at the moment, so how do I get around this issue so I can load PCLink?
-
I've just encountered the issue raised in this locked thread.
The suggested solution requires an ECU to be connected to uninstall the driver.
To fix this issue plug in a G4X ECU, uninstall the ECU driver via the device manager and then reinstall the driver using the driver package installer (PCLink install folder->PCLink G4X->Link USB Drivers->LinkUSBDrivers_X64.exe (or LinkUSBDrivers_X86.exe for 32bit ssytems)).
However that's not possible for me at the moment with the computer having the issue.
What other options are there to resolving this issue?
How to Diagnosis a Potentially Faulty VSS?
in G4x
Posted
Yes, grounded to sensor ground.
Ok can give that a go.
The other thing we were thinking of trying is running a single sensor ground wire back from the bulk head connector to Plug B (so not being spliced into the other sensor grounds that go into plug A)
Currently the VSS is running off 5V, is it worth trying it at 12v and turning pull up off?