ECU repair P1117 O2 heater circuit

Dear all,
Warning - this thread is not for the faint-hearted, it contains details about open heart ECU surgery.

I have been chasing a P1117 ("left side, rear O2 sensor heater high reading") code for a while and have narrowed it down to the ECU (well I think so). I am wondering if any one has delved into the ECU to effect a repair? Or worked out a simple bypass or O2 simulator for the heater circuit.

What I have discovered:
1. The O2 heater is not a critical component, in the old days O2 sensors did not even have heaters. Obviously affects cold start running and emissions though. Otherwise engine runs fine.
2. On the bench, the O2 heater is working fine , I have replaced both front and rear sensors and load tested them with 2A current. Cold static resistance is about 3 ohm. Start up current is 2-3 amp which drops steadily to 1 amp over 1-2 minutes.
3. Harness continuity has been load tested and is fine
4. Relays have been swopped, left to right. The O2 sensors use the fuel pump relay.
5. Fuses are fine
6. ECU connectors were tested, cleaned and inspected. Some slop in the spade-slot connection was remedied. Pin 1 is used for the rear left O2 heater.
7. Measuring the current draw at pin1 of the ECU I see it randomly oscillates from 0.5 to 2.5 amps so the ECU is actively varying the current. The driver side ECU does not do this , the current drops steadily from 2A to 1A as expected.
8. Removing and opening the lid of the ECU, one can trace pin 1 to a transistor BTS132 which is a special sort of MOSFET (TEMPFET) with a built in temperature sensor shutoff. It is a discrete component , and I could easily replace it if failed, but I will try and test it in situ first. Anyone no a good test procedure, I have the datasheet.
9. Everything inside the ECU looks clean , no signs of burning or corrosion

Anyone electronics specialist care to suggest another course of action, maybe a O2 heater simulator? Maybe one for Cribbj?

How the O2 heater circuit works:
12-13V is supplied from the battery via the fuel pump relay to the O2 sensor which is then current controlled and earthed via the BTS 132 in the ECU. I believe there is active feedback control of the current via this transistor (TEMPFET) that has built in temperature control and shutoff (current goes to zero at 160 degrees C).

Seems a shame to throw the whole ECU away for such a small $5 item. Or maybe I have overlooked something.

Regards,
Ron

PS as a by-product of all this ECU connector removal and cleaning, I found that "SLOWDOWN" lights are easy to reproduce if the pin 6 slot is loose.

 

Ron, great detective work, and congratulations for tracing the heater wiring. Have you tried powering the heaters from the ECU harness, ie from the same point that the ECU powers them?

I understand you've bench tested them and they work fine, but don't work when powered from the ECU? However, in your process of elimination, I think you may have left out the possibility of a dodgy harness, or possibly an dodgy intermediate connector, etc.

Out of curiousity, and referring to the Ferrari relay diagram, which set of fuel pump relays powers the heaters? I'm guessing it's relays C & D, and NOT I & L? Have you tried swapping in new relays?
Image Unavailable, Please Login
Image Unavailable, Please Login

 

Dear John,
Yes, I have load tested the harness, by bypassing the ECU completely. I cut the brown wire that goes to pin 1 about one inch from the ECU. From this breakpoint, on the sensor side I load tested the harness and sensor with a 1 ohm resistor connected to ground and see a steady 2 amp dropping to 1 amp when warm, sounds good to me. I then load tested the ECU side with a 12v supply and 8 ohms resistor and see the oscillating current.
That reduces it to the connector or the ECU. I then load tested the connector and sensor (no ECU in circuit) and it was fine.
There remains the possibility that the spade on the ECU is still loose in the slot connector but I dont think so, the current is swinging so fast that only an active device like a MOSFET (with current/temperature sensor) could do this rather than a loose connection. Interestingly pin 1 is the last pin to connect as the 88 pin connector is swung into position from the bottom up, so this connection is the most vulnerable.
Relay D powers the heater and yes I have swopped C and D.
Thanks for jumping in,
Ron
PS in your diagram, 2nd fuse from left is the right bank O2 heater (handles both front and rear), you can remove the fuse and insert a current meter to monitor current changes.

 

Ron, if you're getting that current swing with a dumb resistor in there, then there's probably no hope of getting an O2 sensor simulator to kill the code.

How good are you with a soldering iron?

 

Ron, +1 on what John said and thanks for the info. I can only offer moral support here as I barely understand what y'all are talking about. But, I had a bad capacitor in a $1200 Stealth TT ECU. I knew what it was but didn't trust my soldering skills. Taking it to an electronics guy saved me about $1150.

What do you have to lose?

Regards.....Mark

 

Dear John and Mark,
My soldering skills are pretty good, I have done a lot of DIY electronic projects. Even high voltage ones - direct drive amplifiers for electrostatic speakers keeping one hand in my pocket!
I did a current load test on the MOSFET (TEMPFET) last night - I used a 6V battery to switch on the gate of the FET and loaded the drain-source with 1.5A and the current was steady as a rock, so the power side of the device looks good.
Leaves me 2 possibilities - the gate logic driver IC is switching the device on and off rapidly (too difficult to solve? capacitance issues?) or the pin 1 slot in the connector is bad, broken or burnt inside. I will replace this connector or just do a hard wire bypass and retry the ECU in the car.
Pin 1 being the last to connect makes me suspicious, the slot opening was a bit on the large side , see picture - top row right corner is pin 1 , - I have now closed it up a bit.
Thanks,
Ron
Image Unavailable, Please Login

 

Ron, let's hope it's the lack of tension on that female connector. It does look dodgy.

Had a similar issue on an aftermarket ECU installation, only the problem was an undersized male pin on the ECU itself, and not spreading contacts on the harness side.

I pulled a single strand of copper from a 16 or 18 gauge single core conductor and wrapped it up and over the undersized male pin, then put the harness connector back on and voila, my intermittent issues disappeared.

Might be worth a try, even if only for testing.

 

Maybe I'm dense but it sounds like the mosfet is part of the circuit that detects if the heater in the O2 sensor is working correctly. The workshop manual (I'm looking at the 456M manual) and your comments above seem to indicate that the O2 heater is powered off the fuel pump relay and "For temperatures over 572 degrees F the resistance value becomes such that the "electric" heating effect is virtually cancelled out." How is the fuel pump relay connected to the O2 sensor? Is it connected directly or is the mosfet driving the heater wire? Or something else?

 

I looked up the datasheet for the BTS132 mosfet to see if it provided any clues. It would be very interesting to know if the O2 heater is driven (connected to) the mosfet and only connected to the mosfet. After looking at the datasheet I'm guessing that the O2 heater is driven by the mosfet. My guess is that the temperature sensor built into the mosfet is there to protect the mosfet by limiting the amount of current that passes through the mosfet (the mosfet increases its internal resistance as it gets hotter which would limit the amount of current that passes through it). I.E. I think the temperature sensor in the mosfet is designed to protect the mosfet in case the wire driving the O2 heater shorts out or the O2 heater circuit itself shorts out.

If you're seeing the output of the mosfet is cycling on and off (even when the output has been disconnected from the O2 sensor and wiring harness) then the next step would be to look at the gate and see if it is cycling. If the input (gate) is cycling then the mosfet might be ok and the problem is upstream of the mosfet and we'd have to see what is driving the mosfet. If the voltage on the gate is stable then the mosfet is probably defective.

The easiest thing to do might be to replace the mosfet and see if that solves the problem.

Please keep us informed.
Jim V.

 

Dear Jim,
Thanks for your interest. The circuit flow is as follows, 12v from the battery goes to the fuel pump relay, if the ignition is on and engine running then the relay is closed and current flows to the right hand fuel pump and, via the fuse 2, to the two O2 heaters (front and rear). One branch flows to front O2 heater and then to pin 30 of the ECU. The other branch flows to the rear O2 heater and pin 1. Pin 1 and pin 30 each have a separate mosfet that connects the current flow to earth, the on resistance of the mosfet is very low (0.065 ohms in datasheet) so only a small voltage drop to earth across the mosfet.

Thus I would say the mosfet connects the heater current to earth rather than the mosfet drives the heater, just a different perspective. The heater is a resistive load to the mosfet drain. The effective heater resistance (dynamic in circuit) is about 6 ohms when cold (2A current) and greater than 12 ohms (1 A current) when warm. The static cold value is 2.5-3.5 ohms.

Yes, the temperature sensor is just a fail safe device that switches the mosfet off at 150 degrees C (very high current). I was expecting to see a current sense resistor as a feedback control device but there is nothing in the drain source path so I guess the mosfet already does that. The Ferrari high/low heater pcodes are set to trigger at a high current of 2A and low current of 0.2A, so I wonder how the ECU senses that. Also I wonder if the heater is always on at a low current of say 0.5A when warm or whether it switches off entirely.

A mosfet is often used as a variable resistor to control the current flow and has a constant current region of operation. Not sure how the gate circuit is arranged, it could be as simple as a simple switch on voltage of 4-5V.

I have ordered a new BTS132 and will first try a hardwire bypass of the suspicious, pin 1 female connector.
Thanks,
Ron

 

Ron, So many of the female pins in the plug look like this! I recommend you either source new pins yourself or SRI Gold kit. I have had extemely good luck with the latter.
Image Unavailable, Please Login

 

Hi Ron,

There is frustratingly little technical data online for the resistance specs of O2 sensors. It would have been great to see the datasheet calling out the minimum cold resistance and maximum hot resistance on the O2 sensor.

Above you write "The Ferrari high/low heater pcodes are set to trigger at a high current of 2A and low current of 0.2A, so I wonder how the ECU senses that. Also I wonder if the heater is always on at a low current of say 0.5A when warm or whether it switches off entirely."

The info that I could find online indicates that the O2 heater is expecting to be driven with a pulse stream (not a steady voltage). This would indicate that if you connect a scope to the gate or source of the mosfet that you'll see a (more or less) square wave.

The datasheet for the Infineon BTS132 indicates that the internal resistance is typically 0.55 ohms / 0.65 ohms maximum. They don't spec a minimum resistance; let's guess that it is 0.5 ohms.

If the 0.5 ohm guess is correct then they could detect less than 0.2 amps of draw by looking for less than 0.1 volts on the mosfet's source (0.1 volts/0.5 ohms = 0.2 amps).

They could detect more than 2A draw by looking for more than 1.3 volts on the mosfet's source (1.3 volts/0.65 ohms = 2A).

An in-circuit test would be to measure the voltage on the mosfet's source and see what voltage you see. My guess is that there are couple of voltage detection circuits attached to the mosfet's source to detect high/low voltage and trigger the P code.

P1117 - O2 sensor heater high reading would indicate that there is more than 1.3 volts on the source. If the O2 sensor heater is what supplies the voltage to the mosfet' source then a bad connection across the connector would seem like it would either disconnect the 02 sensor from the mosfet (dropping the source voltage to zero) or increase the resistance seen by the mosfet also causing a low reading alarm.

I love a good Sherlock Holmes mystery. I can't wait for you to write the next chapter.
Jim V.

 

Jim,
I will try some more tests tonight. By the way, the Siemens datasheet I have says Rds=0.055 not 0.55 ohms. The pulse drive possibility is interesting so maybe I should remeasure AC voltage change across an inline test resistor using my Fluke scopemeter not my cheapo current meter since it will not handle pulses. Funny, the driver side sensors show a steady reading but the passenger side does not.
What was it Sherlock Holmes said "When you have eliminated all that is impossible, then whatever remains, however improbable, must be the truth."
Regards,
Ron

 

FBB,
Can you explain how these female pins stop working? Is it lack of spring tension, corrosion or what?
Thanks,
Ron

 

Ugh - you're right. The datasheet does show 0.065 max resistance. Here's a link:HTTP 301 This page has been moved

That changes the math to less than 0.01 volt on the source means less than 0.2 amps.
More than 0.13 volts on the source means more than 2 amps.

 

Jim,

Well I hard-wired pin 1 with a flying lead and put an oscilloscope on an inline temporary 0.1 ohms resistor on the pin 1 (=mosfet drain). Indeed I see regular, rectangular pulses of 50mV and 60 millisec duration, so the ECU is indeed driving the O2 heater with a pulsed current. That explains why I see fluctuations in current with a simple multimeter.
Putting the simple multi-meter inline to measure "DC" current I see oscillations as before but now the max. current is only 2.0A not 3.5A, so I am hoping the bypassing of the loose female connector will stop the P1117 code.

The max. current drops from 2A to 1.7A over a few minutes so it looks like all is working.

Followup questions , will the old P1117 codes remain stored in the ECU forever, or disappear if switch battery switch off for a while. Will a generic OBDII scanner remove them. If no new P1117 codes are triggered will CEL light stay out and OBDII readiness signals appear?

Need a number of drive cycles before I will find out.

Conclusions so far:
1. mosfet is OK, just need to be aware it uses pulse drive
2. simple multimeters will show an oscillating reading 0 to 2A due to the pulses
3. pin 1 and pin 30 female connectors need to be checked for tightness or replaced or bypassed.

Thanks,
Ron

 

Just noticed that "spreading pins" is also mentioned in this thread about washing the car:
http://www.ferrarichat.com/forum/456-550-575/414415-watch-550-washing-guys.html
and FBB tells the full story here about loss of spring tension in the female clip
http://ferrarichat.com/forum/technical-q/312756-sri-gold-connector-pin-kit-strikes-gold.html
Regards,
Ron

 

Hi Ron,

There may be ways to reset the Check Engine light with battery disconnect and / or key cycles but I don't know them. Hopefully others will chime in.

What I've had great luck with on my 2000 456M GT is a Bluetooth OBDII reader paired to both my Android phone and an Android tablet. The one that I'm using is a "BAFX Products - Bluetooth OBD2 scan tool". The ASIN number on Amazon is B005NLQAHS. I use software called "Torque" on the Android devices. This software can read the fault codes and reset them. It can also log some real-time information. I used it recently to track down an intermittent failure on one of the four O2 sensors. It's pretty cool being able to see what they are reading in (mostly) real-time and import logs of the data into Excel to view the data offline. It is also nice to be able to reset the error code at will so I can see how often the failure is happening.

Good luck,
Jim V.

 

Ron, my 456 is pre-OBDII, so can't comment on what it takes to reset that code on a Ferrari. Assume a generic scanner will work. You may know this, but in Houston and surrounding counties you're allowed 2 non-continuous monitors to be "not ready" on a MY 2000 vehicle to pass emissions. It can take a lot of miles, and gas, to get them to set "ready" after you clear the codes. Ask me why I know.......

Regards......Mark

 

Basically if the pins spread you loose reliable contacts = poor or no electrons flowing = bad running

This is a really big issue everyone is ignoring because they think their cars run OK. I have one 550 with gold kit and 60kmiles and one 550 concourse level garage queen with 20kmiles with OEM pins bent to enhance contacts. What is making the old tired car run like a scaled cat and the garage queen just OK? Is is all in the connections? Or am I that great a mechanic that I can take an old 550 be, anal, time the cams biotchin and have a better running car? I'm not that arrogant to think I am that good. I think it is superior connections that make the difference. Is it gold connections? I have no clue but I bought a 2nd SRI gold kit for my garage queen. The job is too labor intensive to want to do again. I could source alternate gold pins or OEM tin pins "cheap as dirt" but I went with SRI pins because my old 60K mile 550 will make a believer out of you.

 

Dear FBB,
I have been catching up on your old threads on the SRI kit, and I am now definitely a believer in the need to do something about those loose female clips. I realize now that trying to coax more life out of the old clips by bending them in is unlikely to be successful.
I also wonder if all those attempts to cure "slowdown" lights and cat ecus issues may be more to do with "loose female slots" !
Also you and Dave Helms mention that more than 5 disconnect/reconnect of the ECU connector can cause failure, my time is up since I have reached 5.
The hardwire bypass is just a work around until I can do the complete SRI kit.
Regards,
Ron

 

Jim and Mark,
The Bluetooth scanner price looks good, can you read real-time data when the car is in motion as well? My scanner is meant to read real-time (at idle) but the numbers look wrong, Ferrari proprietary interface?
Taz has mentioned the key cycle sequence a number of times , and I have somewhere the proper drive cycle instructions to get the CAR AIR EGR etc monitors ready.
Mark, yes I am aware about 2 not ready is OK for Texas inspections, I have just reached that condition after 3 *20 min runs today. Do you know if you can pass with a pending code (amber light on scanner) as long as no MIL or CEL is lit? The P1117 is still pending even after clearing with the scanner maybe an old pcode in the ECU long-term memory.
Regards,
Ron
PS Is Taz sick he is usually one of the first to jump in? I saw him lurking on the F50 threads.

 

Yes the OBDII reader can read and display some real-time data when the car is in motion. It can also log the data so it can be analyzed later. Some of the data that can be displayed and logged is: the time, engine coolant temp, engine RPM, engine load, O2 bank 1 sensor 1 voltage, O2 bank 1 sensor 2 voltage, O2 bank 2 sensor 1 voltage, O2 bank 2 sensor 2 voltage, CO in g/km, fuel flow, speed as reported by the ECU, throttle position and voltage at the OBDII connector.

 

According to Dave, Ferrari spec'ed out a ridiculously small number of use cycles like 10 IIRC! Something does need to be done. Spread pin in the ECU plug should be screaming red flag to anyone who has ever put two wires, a battery, and a lightbulb together. The general owner has not realized how bad this really is. Keep breathing and one day your hair is gray. If you got gray overnight you would freakout. There are several possible solutions and Dave has explored several. His gold kit has served me well and I put my money where my mouth is. I beta tested the first kit and bought a second one. There are many non-believers. Time will tell who has the right long-term solution. So far gold kit people are happy. IMO a gold kit or refresh of all connections does not go far enough! Fuseboxes are another weak link and close look needs to be done dealing with high current drain systems like the cooling fans and how they are wired.