1. Attachments are working again! Check out this thread for more details and to report any other bugs.

Inverter Coolant Flow Sensor Ideas

Discussion in 'Gen 2 Prius Accessories & Modifications' started by snijd, Mar 5, 2011.

  1. snijd

    snijd DIY or die

    Joined:
    Jul 5, 2008
    243
    52
    13
    Location:
    PNW
    Vehicle:
    2010 Prius
    Model:
    III
    Having seen the rather catastrophic outcomes that are possible with inverter coolant pump failure, I've thought a bit about adding a flow sensor in the coolant path. I even went so far as to buy a flow sensor, which would probably do the job. However, it's somewhat heavy, and then there's the plumbing job to take care of. Seemed somewhat challenging to pull off.

    Since this aborted endeavor, I began thinking about ways to detect surface turbulence at the top of the coolant tank, and am now thinking that might be easier to pull off, even if it requires programming a PIC to do the monitoring. My current thought is to install an LED and optical sensor in the cap, and measure light intensity received.

    Has anyone else given such a setup any thought? BTW, this is a bit like the way old slide projectors performed auto focusing.
     
  2. jdenenberg

    jdenenberg EE Professor

    Joined:
    Nov 21, 2005
    3,831
    1,824
    1
    Location:
    Trumbull, CT
    Vehicle:
    2020 Prius
    Model:
    LE AWD-e
    How about a current sensor on the pump motor? Any change in pump performance will change the current.

    JeffD
     
  3. snijd

    snijd DIY or die

    Joined:
    Jul 5, 2008
    243
    52
    13
    Location:
    PNW
    Vehicle:
    2010 Prius
    Model:
    III
    That's an interesting idea. I wonder what the typical failure mode might be? I'm thinking a bearing failure would be immediately accompanied by an increase in current, possibly followed by an open circuit, the result of overheating. In fact, impending bearing failure might be observed via uneven current load--the pump slows down momentarily, and then speeds back up. What would you expect?
     
  4. seilerts

    seilerts Battery Curmudgeon

    Joined:
    Mar 7, 2010
    3,326
    1,512
    38
    Location:
    Santa Fe, NM
    Vehicle:
    2005 Prius
    I think the thing to watch would be the coolant temperature, since this is what leads to the code and DC-DC converter shutdown. This would also capture impending shutdown due to an airlock blocking the coolant flow. There are several analog inputs to the HV ECU that relate to MG and inverter temperatures.
     
  5. snijd

    snijd DIY or die

    Joined:
    Jul 5, 2008
    243
    52
    13
    Location:
    PNW
    Vehicle:
    2010 Prius
    Model:
    III
    With an airlock, where would you measure coolant temperature, and still see high temperatures? Wouldn't you see spurious readings in most of the system?
     
  6. ccdisce

    ccdisce Active Member

    Joined:
    Oct 10, 2010
    259
    187
    0
    Location:
    Stone Mtn GA USA
    Vehicle:
    2008 Prius
    Model:
    N/A
    Monitor existing sensors and add a couple if necessary and make a decision to drive the fail light. May have to tweak it a bit based on temperature change rates.
     
  7. snijd

    snijd DIY or die

    Joined:
    Jul 5, 2008
    243
    52
    13
    Location:
    PNW
    Vehicle:
    2010 Prius
    Model:
    III
    Monitor which sensors? AFAIK no one has yet been able to monitor inverter temps using just a Scangauge, and that still wouldn't get me to a fail light. And I don't plan to use Techstream to solve this problem. A fail light is a good idea, though.
     
  8. seilerts

    seilerts Battery Curmudgeon

    Joined:
    Mar 7, 2010
    3,326
    1,512
    38
    Location:
    Santa Fe, NM
    Vehicle:
    2005 Prius
    There's two ways to do it. If you monitor the coolant temperature, such as by inserting a conventional sensor into the loop, then you would notice that something was off if it stopped working or airlocked. If you tap the HV ECU wiring harness for various MG and inverter temperature sensors, each sensor range is something 2V-5V. You could use anything from panel voltmeters to Arduino/AVR/PIC to monitor these voltage levels and configure any combination of warning lights etc.

    Here's the breakout of temp sensors going into the HV ECU from the fsm.

    Generator Inverter Temperature: H15-27 2V - 4.5V
    MG1 Temperature: H17-18 unpublished ranges
    MG2 Temperature: H17-30 unpublished ranges
    Motor Inverter Temperature: H15-19 2V - 4.5V
    Boost Converter Temperature: H16-21 2V-4.5V

    The boost converter temperature is probably the one to start with -- its acceptable range is up to only 140F.
     
  9. snijd

    snijd DIY or die

    Joined:
    Jul 5, 2008
    243
    52
    13
    Location:
    PNW
    Vehicle:
    2010 Prius
    Model:
    III
    I can see that there are a number of ways to look at this issue, but so far I find the current measurement solution the most attractive. All I really wanted to look at was the performance of the pump; not inverter temp, not MG temp, not even coolant temp.

    So I'm thinking I'll work up an over/under voltage comparator circuit, install a shunt resistor in the pump supply wire, and monitor the current that way. This may not tell me about an airlock, or overheating the MG if towing something, etc., but I think it'll tell me whether the pump is running, and maybe if it's about to fail.

    I'm also wondering whether the pump behaves differently with new vs old coolant. After the coolant gets old, and perhaps loses some of its lubricating properties, does the viscosity change? If so, does pump rpm change? Just brainstorming a bit.
     
  10. 2009Prius

    2009Prius A Wimpy DIYer

    Joined:
    Mar 25, 2009
    2,705
    509
    63
    Location:
    USA
    Vehicle:
    2009 Prius
    I started my PriiDash(TM) project partly because of wanting to monitor the inverter and MG temperatures. It's interesting to note that from my limited data collection the converter T tends to spike the highest.
     
  11. pEEf

    pEEf Engineer - EV nut

    Joined:
    Sep 23, 2010
    720
    576
    3
    Location:
    Berkeley, CA
    Vehicle:
    2008 Prius
    Model:
    N/A
    Guys, this is an unnecessary thing. The IPMs have thermal sensors in them and will set a code if for any reason they don't get enough cooling. The "warning" is the big Triangle of death you see in front of you. Done!

    In addition, the HV ECU will keep reducing PWM duty cycle until they become cooler.

    In fact, it's been reported on here that several people have driven their cars for many months on a failed inverter coolant pump without even knowing because of cool ambient temps and short trips!
     
    1 person likes this.
  12. snijd

    snijd DIY or die

    Joined:
    Jul 5, 2008
    243
    52
    13
    Location:
    PNW
    Vehicle:
    2010 Prius
    Model:
    III
    Not too surprised. My car ran for months with the radiator fans installed backwards. But on the first hot day of the summer, it overheated. The car is generally tolerant of abuse. I'd just rather know the first time there's a hint of a problem, and not at the point that I'm pulling over and calling AAA. Call me obsessive if you want! BTW, my car has already had one pump failure. I think this will be a relatively simple way to monitor pump activity. I'm trying to convince no one to do the same.
     
  13. pEEf

    pEEf Engineer - EV nut

    Joined:
    Sep 23, 2010
    720
    576
    3
    Location:
    Berkeley, CA
    Vehicle:
    2008 Prius
    Model:
    N/A
    The only important piece of tech you need on board is a scantool. Keep in mind that inverter pump failure almost never leaves anyone stuck on the road. You get the Red Triangle of Death, you check the code, you see it's inverter temp, then you drive sane until you get to where you are going.

    Any other Rube Goldberg device you come up with is going to do the same thing. If you still waste your time, put in the flow indicator, and then you are driving and suddenly the flow stops what are you going to do? Call AAA?

    I've concluded the most important device to have in the car (besides a good brain on the driver) is a scantool that can read all the codes. On the PRIUSCAN, It reads all the codes along with their plain-text definition. I wanted to be sure I could diagnose any problem even if I didn't have a code list or was out of cellular range.
     
  14. snijd

    snijd DIY or die

    Joined:
    Jul 5, 2008
    243
    52
    13
    Location:
    PNW
    Vehicle:
    2010 Prius
    Model:
    III
    My primary objective was to be able to predict an impending failure. A scantool is unlikely to do that. But I guess you're suggesting that pump failure is not that big of a deal. Just wait until it fails, then drive to the dealer, and have a new one installed. As long as it's not a hot day, that will probably work. But there are plenty of posts on this forum that suggest things don't always go smoothly in this event (or in this rather strange one). With the current pump recall, failures may diminish, and make this entire exercise moot. There have been failures of the redesigned pump, though.
     
  15. 2009Prius

    2009Prius A Wimpy DIYer

    Joined:
    Mar 25, 2009
    2,705
    509
    63
    Location:
    USA
    Vehicle:
    2009 Prius
    One of these days I will disable the inverter pump and look at the inverter and converter temperatures reported by one of the ECUs to see if we can get some advanced warning before the dreaded "Christmas lights". Anyone know of a connector that's easy to unplug to disable the inverter pump?
     
  16. snijd

    snijd DIY or die

    Joined:
    Jul 5, 2008
    243
    52
    13
    Location:
    PNW
    Vehicle:
    2010 Prius
    Model:
    III
    See Hobbit's photo about half way down the page. It's surprisingly easy to disable the pump, perhaps because Toyota originally intended to include some additional monitoring/control of the pump. Possibly three speeds, like the Gen III cars evidently have. I'm guessing only catastrophic pump failures are going to result in temperature increases, while impending bearing failures might show up as momentary increases in current draw. Industrial electric motors in critical applications are routinely monitored for impending bearing failures.