M150, Jenvey actuator, on Radical SR8 with odd sensor data

[parsonsj]

I've been working to get the DBW on my Radical SR8 project finished and the car to the track. There are two problems that I've run into that I can't seem to figure out. The servo sensor is a Penny+Giles DP280 (two sensors packaged together with common 5v and 0v), and the actuator is a Jenvey ETA2. The sensor is mounted to the servo.

1. At full throttle the main and tracking voltages diverge beginning at around 4k rpm, and by 8k rpm are more than 10% different, causing a sensor fault. I can't see how that's even possible, but it obviously is. See the first screen shot.

2. At idle, the throttle oscillates, at about 5 Hz. See the second screenshot. I isolated the main and tracking voltages for the DBW and referenced it against the Servo Diagnostic time. The car is idling with a throttle aim of 6.2 %. For the most part it’s managing ok. Now look at the main raw voltages. See how it has periodic spikes? But not the tracking sensor. What could cause that? Ignition EMI? I can't see that occurring at 5Hz... but..

Any ideas out there?

MainvTracking.jpg (237.73 KiB) Viewed 11019 times

IdleSpikes.jpg (206.52 KiB) Viewed 11019 times

[Stephen Dean]

Hi,

Can you send the actual logging through showing this thanks, you can send it to support @ motec.com.au if needed.

[David Ferguson]

What does your Throttle Servo Bank 1 Position Sensor Tracking Linearisation Table look like?

Any chance you don't have the Throttle Servo Bank 1 Position Sensor Main Voltage Reference and Throttle Servo Bank 1 Position Sensor Tracking Voltage Reference set to the same/correct reference voltage? Note, it should not be absolute with that sensor (assume you meant SRH280DP or TPS280DP) as the output is ratiometric to the voltage for the 5V version.

What resources are you using for the Throttle Servo Bank 1 Position Sensor Main & Tracking?

Have you tried re-calibrating the Main Offset & Scale with the sensor both cold and warm to see if there is a difference?

[parsonsj]

Stephen, I've sent the dyno pull log to you @ support. But the idle spike log is too big for email. Is there a public drop box I could use or something similar?

David,

The Linearization table is a simple 0/0, 100/100 table. The voltage references are the same. The Penny+Giles sensor has 4 wires: 5v, 0v, signal1, signal2. Both are set for 5v C.

I'm using AV1 for main, and AV4 for tracking.

I've not tried calibrating Offset/Scale hot. That's an experiment I can try tonight if you think it's worth a shot.

[parsonsj]

I just completed an experiment: After a conversation with my dealer/tuner last night, I patched into the connector for the servo main/tracking sensor, and routed it away from the engine and over to the front of the car where I had sensors connected for steering angle, brake pressure, etc. I connected the servo sensor patch cable into that loom, reassigned, and recalibrated. The voltage spikes are reduced by more than half, and no longer cause any errors. Not willing to take that as the fix, I put it all back. Voltage spikes came back. So I went back to the patch cable. Voltage spikes go away.

Next I took the car around the block this morning and did a WOT run (I nearly ended up in the trees -- won't be doing that again). The main / tracking divergence went from .7% at 4k rpm to 1.7% at 7k rpm. Well within tolerance (before was 7.5% to over 10%).

So two more questions:

1. I think that my main loom must have some significant noise in it. Both cam and crank sensors are in the loom,
EDIT: this is incorrect! The cam and crank sensors are in a different loom with a different bulkhead
and both are magnetic sensors. They both are wired with M27500 shielded cable, but I have a bulkhead where the shielding isn’t present. Is there a best practice for finding the source of the noise before I tear into loom?

2. Should I put something in the linearization table to provide for the 1 to 2% difference between the main and tracking values?

Thanks again!

[Stephen Dean]

Hi,

Do not change the translation table, resolve the issue!

Noise is induced into a loom, unless there are loose connectors splice in the loom.

I would look at the other sensors that share the same 5V+ and 0V feeds, to see if they are showing similar behaviour and not excessively drawing on the 5V+ feed.

The shielding should start at the sensors, be passed through the bulkhead connector and then drained to a 0V feed (preferably the same as that used with the sensor) at the ECU end.

[parsonsj]

Stephen,

My cam and crank sensor are shielded as you describe. A review of my looms and documentation also reminds me that the cam and crank sensors are in a SEPARATE loom with a separate bulkhead from the throttle servo loom and bulkhead. The throttle servo loom includes the servo bridge, coil 12v/ground/ECU 0v/drivers, injector 12v/drivers and fuel temp sensor. This noise theory is making less sense than ever.

I can't see any evidence that any other sensors are affected in a similar way.

But there's no doubt of the data above, with a 5Hz spike at idle and a slowly rising throttle servo main sensor voltage when the servo itself isn't moving.

I'll keep digging...

[Stephen Dean]

Can you get hold of another sensor? I had another person with a similar issue contact me yesterday and it turned out the the sensor itself was faulty, they swapped the sensor and all of the issues went away.

[parsonsj]

I've tried two sensors (one Variohm, one Penny+Giles). Both have the same symptoms. I wish it was the sensor, but I don't think so.

On new thing I've found that might be relevant is that my bulkhead packaging of the loom has the coils' 12v supply right beside the main signal wire of the servo sensor. I just can't see how that would translate into a 5 Hz signal, but maybe...