How to Replace the Inverter Coolant Pump

Robomod claims I have hyperlinks, blocking post.
No hyperlinks - looks like it is tripping on embedded quotes?

Will post once I am not running in limp mode

-methods

 

I just replaced my inverter water coolant pump at 292,200 after the first P0A92 code. Thought I'd share three alternative techniques that worked.

First, rather than hose clamps, I opted to use rubber plugs from Lowe's, available in the aisle with all the nuts and bolts. They're only a couple bucks. I carried the coolant control valve into the store when I replaced it (code P1121, same techniques as here) to make sure I got the right size. I didn't want to clamp my hoses since they've lasted so long—the originals—and sharp bends from crimping aren't great for the rubber.

Turns out the real benefit is that, with the system cool and the reservoir cap on, I lost virtually no coolant when I pulled the hoses off the pump and put each rubber plug in. (There was a small bit to wipe up, but with one hand pulling the hose off and the plug ready in the other, the internal pressure reduction as I pulled kept the coolant in the hose pretty well.)

20190704_143208_25 by Stephen Frost posted Jul 6, 2019 at 5:13 PM

A sped-up video from Luscious Garage covering the procedure said it took (them) about 15 min and they just lifted the inverter to get in there. Since this is only two bolts, I have skinny hands, and don't mind a little Tetris, I opted to try it this way. Removing and inserting the pumps with backing plate was through this orientation:

20190704_142934_HDR_25 by Stephen Frost posted Jul 6, 2019 at 5:13 PM

20190704_143023_HDR_25 by Stephen Frost posted Jul 6, 2019 at 5:13 PM

20190704_143053_HDR_25 by Stephen Frost posted Jul 6, 2019 at 5:13 PM

The inverter was raised 1.5 inches on wood blocks under each bracket.

I unbolted and unplugged the pump first before pulling the hoses off. (Small needle nose vice grips really helps to depress the lock lever to pull off the electrical plug. Also great for moving the hose clamps back.) I pulled and plugged the top hose, then drained the coolant in the pump into a cut-off 1-pint drinking water bottle base by tilting the pump over. I then pulled the bottom hose and plugged it. After I removed the old pump from the engine bay, I emptied a little more coolant out, then covered the water bottle with a blue paper shop towel to keep any particulates out. After the new pump was in and the bottom hose on, I used a funnel to refill the pump from the top before putting the top hose on.

20190704_143724_HDR_25 by Stephen Frost posted Jul 6, 2019 at 5:13 PM

What this did when I restarted the system was to leave almost no air in the loop. There was a little blip of a whine before going to the correct quiet hum, and the one bubble went out and into the reservoir. I did have a gallon of SLLC ready just in case, but I actually didn't need it.

 

I just replaced mine after watching this video (below) and it was very easy. If your coolant is fine, this video shows the process in a clear manner.

 

Is that code always pointing to the inverter pump or is it tripped from the temp spike?

Posted via the PriusChat mobile app.

 

Inverter pump.

 

Here's my take. The code is from a temp spike, but the system runs so much lower than the level that tripped it that, as long as you haven't dumped all your coolant on the ground or put a giant hairball in the reservoir, it is almost certainly the pump. Mine set the check engine light while in stop and go traffic on the freeway in about 75 deg F weather. I have my phone mounted while driving and a Bluetooth OBDII reader, so I was able to pull the code through Torque right then. The inverter temps for MG1 and MG2 were about 185 and 170 whereas they normally run 105 - 130 under normal driving. (The engine side always runs baseline about 175-185 where it's most efficient.)

I researched the tolerance on the inverter components and some sources said as high as 230 F, so there was still a margin. Indeed, I continued to drive it another 8 miles to a dealership which was not far off the freeway, rather than stopping and making traffic even worse, while watching the temp closely. There was some ambient cooling when stopped or coasting and I could manage the duty cycle to keep the temp from spiking. Interestingly, when the first dealership didn't have the pump, I drove to another one an added ~10 miles away, this time at freeway speeds of 70-ish. The check engine light was off when I restarted the car. I kept the inverter temps under about 180. It did not come back on.

Replacing the pump didn't take long and immediately returned the system to normal temps.

 

Neat idea - no squishing the rubber hose with clamps. What size are they? Rubber "corks" come in a huge range of sizes.

After reading several threads on this subject it still isn't clear to me what the geometry of the coolant path is. My best guess is that there is a hose to the inverter case from the tank (or maybe the tank is bolted directly to the inverter case, the connection is never visible in the pictures), through the inverter case somehow to pick up heat, another hose from the inverter case to the top of the pump (seen in pictures), a hose from the pump to a radiator to dump the heat (pump end of hose is seen in pictures), and a last hose from the radiator back to the tank. From the shape of the pump it looks like the fluid flow direction is into the pump from the top and out the side of the pump.

(EDIT: WAY off on that path, see the image here:
Inverter mechanical
)

What exactly is going on with the bleeder screw? Is that at the top of a void within the inverter case which should be entirely fluid filled for optimum function, so that the screw is the only way to get (all) the air out? Or is it just the high point in the entire fluid path, so that is where air accumulates?

Would it be a good idea to prefill the pump on reassembly - put the side hose on, pour coolant slowly into the top of the pump until it is full, then put the top hose back on? Because some water pumps never move fluid if there is a large bubble of air around the impeller.

 

(Not sure what I did to cause the pictures to top post like that, sorry!!!)
Changed my pump yesterday by: remove black "tray" under hood on top of bumper, loosen bumper driver's side, remove headlight driver's side, remove 3 screws holding old pump, release both hoses from nearest hose clamps, pull it forward and into a plastic bowl (to catch leaks), remove electric plug, test that power gets to electric plug with voltmeter in Ready state (it did), pull hoses and plug them with rubber "cork", attach hoses to replacement pump, temporarily attach electric plug and test (it worked), remove electric plug, put pump back onto plate and fasten with screws, place electric plug permanently, replace headlight, replace bumper, replace tray, remove air bubbles using plastic hose and bleeder screw, top off fluid (added about 1/2 cup).

A few notes...

1. I used 15/16 x 11/16 rubber stoppers from Home Depot. These were one size too large. They worked but were hard to jam in and only went in a few millimeters. Not very much fluid came out, in total about 1/3 to 1/2 a cup was collected.

2. The electric plug simply would not release with any amount of squeezing on the lever, including with pliers. I had to put a small screwdriver in the bottom and twist, and of course that was impossible while it was in place.

3. The replacement unit was a Dorman from rockauto.com. It was about an inch taller than the original unit (first picture). We will see how long it lasts...

4. With the hoses on and the replacement in place it was very hard to reach the back screw (the one closest to the inverter) to put it back in. Perhaps the Dorman is a little wider than the original? Anyway, I have a tiny ratchet for tight spaces (second picture) which worked.

5. Rodents at some point must have lived in this car, see the mouse poop which was found under the headlight (third picture).

6. Each of the rubber hoses is held about 6 inches from the pump by a plastic clamp. A "C" shaped section holds the hose. Until these are released it is hard to pull the pump out very far. For the lower one I just pushed the hose out the open part of the "C". For the top one the push connector was pushed in with a screwdriver and the whole clamp released from the metal. The screwdriver in the fourth picture points to this part of the clamp.

It would have taken about 2 hours in total except I stopped for four hours because it was too hot in the direct sun to keep working. Also pay more attention than I did to how the bottom plastic pieces of the bumper go together. I wasted half an hour not getting it back together when one flap went over when it should have gone under!

 

Dorman 601015 (601-015) was $78.82 with CA tax and shipping. It came in 2 days - rockauto has a warehouse near me so those arrive as if it was 2 day shipping, but with no extra cost. The description says that the design was modified to be more reliable than the original (ie original design) inverter pump.

https://www.rockauto.com/en/moreinfo.php?pk=8457452&cc=1434449&jsn=8

My hearing sucks so I'm the last person to ask about noise. That said, I could only just barely hear it. If my finger was on the outlet hose there was a little vibration while it was running, don't know if that is also the case for the stock one.

 

This weekend I had a little time to examine the failed OEM pump which was removed from the car. When 12V was applied to it it would not turn. Resistance was high in one direction and lower in the other across the two contacts. A long skinny screwdriver could be inserted through the outlet and used to turn the internal rotor. This confirms that the pump was dead - as opposed to something like an intermittent failure of the inverter pump relay. Probably the DC motor inside failed. The electronics may be OK, since the asymmetric resistance has been reported by others. In any case, it wasn't electrically open or shorted.

 

Is Aisin the OEM manufacturer of the inverter pump?

 

No kidding! I will be replacing failed pump #2 this weekend. First pump replaced prior to failure under the recall. It failed after about 65k miles. That pump just failed at 44k miles. Love that the new pumps are so much more reliable! (Sarc).

 

Replace pump only, left bracket in place, in about 45 minutes total. Removed headlight and used rubber plugs for the hoses. Also utilized some left over puppy pads to catch any drips and minimize the mess when pulling the hoses. Literally had maybe one small bubble and that was it. The pump is super quiet.

 

Update on the Dorman.

Yesterday the car sat in a toxic waste drop off line for 15 minutes with the A/C on. Something like 90F ambient, but there was no shade. After leaving there we got on the highway and a few seconds later the red triangle of death appeared. This is with the Dorman unit pumping the inverter coolant. Dropped speed to 55mph and took the next exit. Turned the car off. Popped hood and took off the cap on the inverter tank. Pushed power button two times. Couldn't hear the pump (too close to the highway, also my hearing isn't so great), so could not tell if it was spinning or not. The level was much higher than normal and rather than swirling it was pulsing a bit up and down. The fluid was above the round hole though, so I had no idea what it should be doing at that fill level. Turned the car off, then on, then drove the rest of the way home. Plugged in the miniVCI and ran Techstream to find P0A93 with subcode 346. At that time the car was also making a sound identified by dolj in another thread as a failed 3-way coolant control valve shortly after it was turned on. (Sound file attached.) Have not heard that sound since that recording was made and it sounds like it is pumping coolant into and out of the tank as it always has. There was not a P1121 code, all ECUs were checked and only the P0A93 was set. The inverter temps were quite mild, 135F and 142F for MG1 and MG2, respectively.

Assumed the Dorman was having a problem called the local dealer's parts department, and this time they had an OEM pump in stock, so went out and bought it.

This morning installed the OEM. Loosened bumper and pulled driver's side headlight. With better access, tested the Dorman. On the first test, cold car, press power button twice, no foot on brake. Car goes to normal state except the pump didn't start - no vibration, no swirl. Tested it 4 more times the same way at 30s 60s 120s and 600s intervals. The pump started in every successive test. Put a vinyl hose on the bleed valve and opened it while the pump was running, one really small bubble was seen and that was it. So it does not look like the system filled with air and failed because of that. Also overnight the fluid level had dropped back to where it was supposed to be. Well, slightly less than full, but some coolant splashed out when I took the cap off to look at the tank when it was too hot. After installing the OEM pump, bled the system (oddly, again, only a few tiny bubbles, even though both hoses were off the pump) and filled to the "full" line, which required only a tiny amount of coolant. The amount of churning seen in the tank is less with the OEM pump than the Dorman.

So what happened? My best guess at this point is that the Dorman pump sometimes does not start. Since the car can run for short trips without overheating the inverter it would be really hard to tell that this was happening. I had checked the tank for motion at least a dozen times in the last month and a half, and it was moving every time. So yesterday I suspect that a nonstart of the pump coincided with a hot day and a trip on the highway and the inverter got hot enough to trigger the code. When the car was stopped to look at the tank the pump started normally, because it does the vast majority of the time, and in the short time it took to drive home it had cooled back to normal temperatures. (Conversely, when the previous pump failed the inverter coolant was 40F degrees hotter when we got home.)

Other variables...

A couple of weeks ago the low side of the A/C system was replaced under warranty at the dealer where we bought the car. That shiny new metal pipe runs right under the inverter tank and who knows what they took out and put back together to do that work. Some pink crystals were observed on the bleed valve. This time after bleeding the valve was washed with water and a toothbrush and dried. If more crystals appear that would indicate a slight leak there.

The car is at 117K miles and we have only had it the last 2K. The Carfax notes list when some standard service interval work was performed. When would the inverter coolant and engine coolant normally be changed? Perhaps the former is old and has some gunk in it which the pumps don't like?

 

Many thanks to all the advice in this thread. Here's what I can offer:

I used the video from post #203 above and read many other posts, too. Bought a set of cheap hose clamps from Harbor Freight for about $6 and the hoses held up fine.

I will throw out there that not every set of directions includes bleeding the system. There was quite a bit of air in mine. I hit the bleeder screw with some PB blaster and let it sit for a while. The screw wasn't happy about loosening but it eventually did. I ended up needing about .5 cup of the red coolant to top it off.

After reading stories about aftermarket parts failing (mostly Amazon reviews) I decided to pony up for the OEM part from a dealer.

The car is a 2007 with about 154k. It appears that this had never been done before. If it lasts another 154k I'll be happy!

 

Put me in the OEM club. The Dorman did not work for me.