It’s not often I get too involved with vehicles that have been to a main dealer but after a friend felt the price for a diagnosis on their 2009 Honda Insight was getting a little out of control, I offered to take a look.
Trying to remain unswayed by the dealership’s diagnosis, which was correct, I started my process.
A quick chat with the customer highlighted that there were a number of warnings present on the instrument cluster, where a warning for the IMA and a picture of the 12V battery were displayed with the ignition on. They also mentioned that when starting the engine, it sounds like it’s slower than normal, as typically it fires up instantly.
As expected, when switching on the ignition, the warnings were there. When starting the engine, it sounds much like a traditional petrol engine. Honda did something clever where a 12V starter motor is also fitted, unlike some hybrids. Typically, the motor generator would start the engine, which would be almost instantaneous.
This backs up the customer’s information on it starting slower.
Next up was retrieving any fault codes. While the scan tool was running through a global check, I noticed that the battery voltage was sitting at 11.9V with the engine running. As this is a hybrid, there is no alternator where a DC/DC converter steps down the high voltage battery voltage to around 14V in order to charge the 12V battery, another piece to the puzzle.
The fault codes retrieved were:
- P1634 – Motor Power Inverter (MPI) Signal Circuit Malfunction
- P1448 – Intelligent Power Unit (IPU) Module Fan Problem
Following the technical information and the flow chart for the above fault codes, P1448 states that if P1364 is present, then fix this one first.
Reading through the flow chart and the possible 63 steps to this fault code, I can see now why the dealership was unclear as to how long it was going to take. Also, it involves switching off the high-voltage circuit to access the connectors.
With the scan tool, I was able to apply an active test for the IPU fan, which would not complete. Using the flowchart, it advised to locate fuse 53 in the under-dash fuse box. This was indeed blown.
Now, at this point, the flow chart asks to repair a short to ground between the IPU fan module or the MPI module. Rightly or wrongly, I decided to go against the flow chart here, opting to use a resettable 5 A fuse. Should there be a dead short in either circuit, then the lower-rated fuse would immediately trip. With the fuse in place and the ignition switched on, the fuse remained set. Starting the engine saw it now start using the motor generator and not the 12 V starter motor and we could now see the battery voltage at 14 V, indicating the DC/DC converter was now doing what it needed. I understand and appreciate why there are flow charts in place, but sometimes they can send you off in a direction that isn’t technically wrong, but there is often a more efficient way.
The next step was to repeat the test with the IPU cooling fan, which failed with the fuse blown. This then tripped the fuse, causing the warning lights on the instrument cluster to appear and the DTCs being logged. Something is pointing towards the IPU fan in support of the DTC P1448.
Removing the rear compartment to access the fan, I took the opportunity to hook the scope up to see what was happening during the active test. Using the universal breakout leads, all 4 terminals were broken out, allowing us to see what was present on each terminal. After finding the ground on Pin 4, this was swapped for a current clamp around the fuse 53 extension with the resettable fuse.
Running the active test, the following was observed.
Looking at the waveform, we can see that there are some clear issues when the fan is being activated. To start with, the supply voltage on channel A in blue is being intermittently pulled down. At the same time, we can see the current jumping up to 33A! This is more than enough to blow a 10 A fuse, but what is interesting is that due to the pulsed nature, the fuse remains intact. It’s only when it starts to consistently pull excessive current do we see the supply drop with an interesting return back towards 0V. This could indicate that something is discharging at the same time, or are we seeing any effects from the MPI module, which it is also connected to.
Either way, there is enough evidence to support that the IPU cooling fan is drawing excessive current when activated, which in turn is blowing the 10A fuse for that circuit. This in turn causes the MPU to lose its ignition live and so this shuts down the DC/DC converter and all other high-voltage components.
The next step was to find a new IPU cooling fan, which was found online from a reputable second-hand source. Whilst the motor was on its way, I decided to open up the module to see what could have been causing the issue. Upon opening the back, it was clear to see that we were dealing with a brushless 3-phase motor, which made me grab the MT03A Milliohm and motor tester to see if we had a phase short.
We can see that all phases are around 80 mΩ. Should there have been a phase that was shortened, we would have expected some imbalance between the phases. Therefore, the issue must be on the board somewhere, maybe part of the drive circuit. I may look to remove the board from the motor at some point, but it does need the phases desoldering from the board before being able to remove it. Quite possibly, one of the MOSFETS has failed, causing a short.
With the new/second-hand motor acquired, I carried out a quick test on the windings to make sure all was good before putting it into the vehicle.
All phases were in the 80 mΩ giving some level of confidence when putting it back into the vehicle.
The same setup was carried out using the resettable fuse and the scan tool was connected to run an active test. This time, the active test was completed and we no longer had the peak of 33 A present, with the current only really hitting around 3 A during the high-speed test.
The scan tool actually runs through 3 different speeds, low, medium and high. Altering the timebase allowed me to see all 3 on one screen. By adding a frequency math channel on channel B, the fan speed output, we can see the different changes in speed.
With the codes cleared, the correct 10 A fuse fitted and the 12 V battery charging correctly at 14 V all was right with the vehicle.
I appreciate that some of this is going to another level and in the workshop, a new module would be fitted and then tested before handing back to the customer. I do believe though, that this particular job could have been completed within the magic golden hour, including fitting and testing of the new part. What I couldn’t quite understand is that the dealership failed to produce any evidence of their findings to the customer, leaving them with a slightly sour taste when handing over payment for the work they had carried out. Having Pico, or anything, to justify and support any diagnosis will help build trust between a workshop and had the customer in this case, been given some evidence, he would most likely have had the repair carried out by the dealership.
Going forward, I will look to see if there is anything on the board that might point towards the issue. Should this fault just be due to a failed MOSFET, then these can be purchased for around £5 which, if we were confident, could be replaced at a fraction of the cost of a new module. This type of board-level component replacement is becoming more and more popular as the cost of car parts seems to be ever-increasing with no sign of stopping.