TSB EG011-03 Part 3
The car featured on this page is the same car featured on the page documenting our first encounter with P3101 on a first generation Prius. We got a call from our customer saying the Master Warning light, Check Engine light, and HV light had all come on again, and the car had low power. All three lights were staying on even after turning the car off and back on again.
I now recommend that Prius owners try cycling the key or power switch off and on before calling for a tow. If the master warning light goes off, drive it in. If it stays on, call for a tow.
Anyway, when the car was towed in, I was expecting to find the same codes as it had 5 months ago, and I was dreading trying to find an intermittent fault that acts up once every 5 months. I was pleasantly surprised to find a new code, in addition to the two codes I expected. The new code was P1128, throttle control motor lock malfunction. There was also a P3101, engine system malfunction, and a P3190, poor engine power. The P3190 is similar to the P3191. Both codes mean the gas engine is failing to contribute enough (or any) power.
The main difference between the two codes is that P3191 will set when the engine is in the start mode (meaning the engine is being "cranked" by the electric motor), whereas P3190 will set once the engine has already started (meaning the engine lost power or died while driving). Obviously a throttle that won't move will cause a lack of engine power, so I knew I could safely ignore the P3101 and P3190, and concentrate my efforts on figuring out what was going on with the electrically actuated throttle body.
On Toyota cars, each control unit (engine, hybrid, ABS, etc.) must be accessed separately. I started with engine control unit (ECU). You may notice that P1128 is preceded by an asterisk. The asterisk means there is freeze frame data available for time when this code set. There is no freeze frame data for the P3190. The freeze frame data basically said the car was not fully warm, and was not moving very fast. I didn't see any smoking gun in the freeze frame data, so I move on to check codes in the HV ECU.
The HV ECU stored the P3101. I starting to think we'll be seeing a lot more of this code as the first generation Priuses get older. I'm beginning to suspect that any code set in the engine control unit will cause a P3101 to set in the HV control unit, but I don't know this for a fact yet.
I checked EXCLUSIVE INFO3 to make sure this Prius didn't think it ran out of gas, just for the heck of it. I was pretty sure the stalling problem was going to be related to the throttle motor code, but it never hurts to spend a second or two to just to make sure "what you think" and "reality" are in agreement.
I went ahead and cleared the codes in the ICE ECU & the HV ECU. The car started and ran normally with no warning lights
Following the flow chart for diagnosing P1128, the first step is a scope check of the throttle motor output signal. Thanks to quick access flip-down glove box door, gaining access to the ECU is a breeze. The simplified wiring diagram on the flow chart is a little confusing. It shows only 3 terminals: M+, M-, GE01, all on the same connector, E7. A person (well me at least) might wrongly assume that testing should be done with the ground lead on GE01, and the positive leads on M+ & M-. You'll have much better luck using E1 (pin 17 on connector E8) for your ground.
I should mention that the P1128 code must be cleared before checking the M+ and M- signals. The computer will cease attempting to control throttle opening if a code P1125 or P1128 is stored. This is done as a safety measure -- without a signal, the throttle spring will pull the throttle valve closed so the car does not go racing off.
The blue trace is M+ and the red trace is M-. This capture was taken at idle, and I originally thought it was normal, despite the fact it disagrees with the "good" example in the factory manual. In the example, both M+ & M- 0 to 12V square waves. M- is a mirror of M+: when M+ is high, M- is low, and when M- is high, M+ is low. However, I was pretty sure something was lost in the translation. I thought the pattern above had to be correct because the throttle servo was working normally while I had the scope hooked up. It certainly wouldn't have been the first error in a factory manual.
In this picture I have just blipped the throttle with the gas engine running while in "drive". The gas engine accelerated and decelerated normally. M+ is still the blue trace, and M- the red trace. The volts per divisions are mismatched from channel A to B. If they were set evenly, both square waves would be the same height.
I captured this while driving on a relatively flat street at about 25 MPH and with my foot off the gas. M+ is a solid 14 volts (the trace is high on the picture and disappears under the 1980 date stamp tag) and M- is pulsing between 0 & 14 volts. I figured this was a normal for pattern for deceleration.
Next I check the throttle body. It's the next step on the flow chart if we interpret it literally -- the signal to the throttle body was NOT "OK" as they defined it, so they want us to check the throttle body next. Surely there must be something wrong with it if it's not doing what it's not told to do :)
Basically, we're looking for anything that could cause the valve to stick. With the key off, there is no signal to the throttle control motor, and the valve can be moved by hand. It should move freely without a lot of force and should not stick or bind. This one feels fine. If you notice an oil puddle under the throttle valve, especially after the car has been sitting a while, don't worry. Hobbit reports that this is normal for all generations of Prius.
This is a close up of the throttle valve. There's a little build up in the cylinder where the disc sits at idle, but not enough to cause any problems. This throttle body was cleaned 5 months ago when this car had the apparent "out of gas" condition, and it still looks pretty good.
I also made sure the air filter was sealing tightly and there was no sign of debris or scratches that might have been caused by a foreign objects floating around in there.
This is the servo, motor, actuator, whatever you want to call it, that moves the throttle valve. The two wires in the picture are M+ & M-. I'm fairly sure the wiring between the computer and the servo can not be the cause of a P1128, because an open circuit seems to set a P1125 instead. I thought an intermittently bad TPS (throttle position sensor) might cause the code (for example -- the computer commands the throttle to move, but the TPS erroneously reports the throttle did not move so the computer thinks the throttle is stuck or the motor is broken), but once I realized the TPS is redundant (there are two throttle position sensors in one housing), I knew that there was very little chance of both failing simultaneously.
The IACV (idle air control valve) used on many Camrys and Rav4s often has problems with sticking. There's no IACV position sensor, so there is no code associated with failure. The car just dies at idle when the engine is cold. There is an "active test" on the Toyota scanner, where we can tell the computer to tell the IACV to open and close over and over until it gets stuck. This is very handy for diagnosis. Unfortunately, there is not a similar active test for the throttle valve motor on this generation of Prius. Bummer.
So, what to do? Tell the customer to go driving around until their car dies on them, then hope it stays stuck during the tow back in so we can confirm the problem when it arrives at our shop again? And if the jostling during the tow unsticks the valve, what then? Send them out to stall on the road again? In this case I'm fairly confident the ECUs assessment of the situation, and I think the throttle control motor will fix the problem. So confident I'm willing to risk about $1000 of my customer's money (and a proportional amount of customer's wrath if I'm wrong) and install a new valve without a 100% confirmation of its failure. If I blew it, you'll be able to have a laugh at my expense in Part 4.
A few days later .....
The hybrid trainer, Jack Rosbro, was kind enough to ask a tech he knew to give me a call about the Prius in Part 2. After talking about the Part 2 Prius, he also had some insight into this Prius's troubles. He said the throttle servos did fail regularly, and cleaning did not provide a permanent fix. Replacing the servo sounded like it was the right call. Phew! What a relief.
One week later.....
Well, our customer's car is still on the road with no trouble reported, but one of our other customers, who knows quite a bit about electronics, read this page and sent an email gently asking how sure I was that the manual was wrong.
He said, "In electronic systems we engineers will sometimes use a double-ended push-pull circuit where we drive one side with a active signal from 0 to x volts and the other side with a signal from 0 to -x volts (or from +x down to 0). That way, something connected between the two sides will see a AC signal rather than a pulsing DC signal. If something was wrong with the output circuitry it could conceivably provide one-sided drive to the throttle body."
A long-shot since I expect that it would act much more broken and your Prius ran okay during testing. Anyway, I suppose you have looked at a working car for comparison???" Nope, I hadn't, and I was now regretting it. The next time a first generation Prius came in, I checked M+ & M- and found the waveforms to the left -- uh oh, they look exactly like the example in the manual (just a note - I was using 1:1 probes, but the scope was set up for 10:1 probes read 50V/div as 5V/div).
So what's going on? If the throttle servo drivers are bad, why is the car still driving so well with no warning lights. Could a bad servo cause the computer to alter it's output to the servo? I don't see how a bad servo could have caused the differences in the patterns, but then again, just because I don't understand it, doesn't prevent it from being. Good thing too -- the whole universe would likely implode if everything were contingent on my understanding :) Anyway, I've asked our customer to come back so I can do a belated post repair scope check and assuage my curiosity. Stay tuned ....