General STS thread

[kevm14]

I used my coolant tester. It is fun to use.

I pressurized the pressure tank to 15 psi and shut the valve. Pressure did drop slowly but I also left it for a while and it didn't really get below 13 psi.

I looked for leaks for over an hour, sometimes with the engine running, sometimes not. I saw NOTHING. I checked lots of stuff including the thermostat housing. I used a mirror and checked behind/under stuff, including the water pump. I also looked behind the lower cover. I also did not smell even a whiff.

Recall that I said the cooling system doesn't seem to hold any pressure. I figured that was due to a leak. Well, with the leak tester threaded on, the system had no issue building pressure, and it would actually go above 15 psi after starting it around 12 and I would have to bleed off pressure, so...I think one of my problems is actually a faulty pressure tank cap. This may be where the smell comes from also.

Now I am not sure how this would lead to an actual coolant leak but it seems like I do need a new cap so I will order that. It would be quite funny if that was somehow also my leak but I don't really understand how. Easiest thing is order a cap, install it, and see if the coolant loss changes at all and go from there.

I wonder if any of my other vehicles need a new cap.

[kevm14]

$10 shipped for GM OEM according to Rockauto.

[bill25]

Hopefully it wasn't building pressure and leaking slowly through a gasket (head?). You would smell/see the white smoke though.

[kevm14]

Got my new cap. I "installed" it and did an errand. Upper radiator hose was firm under pressure when I got home. Finally. I wonder how long the old cap wasn't holding pressure. Well now I am very interested to see if it loses any coolant.

Oh the old cap was definitely failed. I tried to blow through the new cap and no air went through it. The old cap had no restriction at all (I.e. a 0 psi cap). So that is extra confirmed.

[bill25]

Easiest fix so far!!!

[kevm14]

Bill was asking about the rate of coolant loss. I don't know the quantity but maybe about an inch in the tank every 3-4 weeks or something.

[kevm14]

Wow, now this is a strange one!

https://www.cadillacforums.com/forums/c ... ts-wi.html

Dozens of people chimed in as having the issue at the same time.

Call it a Y2K-type bug. But it would appear that the next time I have to disconnect my battery, my clock will stop working. Because GPS week rollover something something. I wonder if the CTS-V and SRX will have the same issue.

[kevm14]

I no longer smell coolant after shutting off the car. It is too early to call whether it fixed the leak. I will need to get a couple weeks of commuting on it and then check the level. Current mood: Cautious optimism.

[kevm14]

Wow, now this is a strange one!

https://www.cadillacforums.com/forums/c ... ts-wi.html

Dozens of people chimed in as having the issue at the same time.

Call it a Y2K-type bug. But it would appear that the next time I have to disconnect my battery, my clock will stop working. Because GPS week rollover something something. I wonder if the CTS-V and SRX will have the same issue.

What I think is happening is that if you remove power (and reconnect it) you are doing a full reboot of the Nav. The Nav then looks for a GPS signal and first decodes the week count, then the second count. From that it knows the date and time.

I think what's failing is the interpretation of the week count because we are beyond a certain number of weeks that started in 1999. If that fails then the Nav time and date are hosed. We really don't have a way to set it from scratch, only adjust it after it gets time from GPS. That's why the buttons are greyed out. If you think about it we've always had the date, but you can't set the date, just the hour and minute. The only reason for that is if you're not in one of the selectable time zones.

The reason I think you only have the problem after a full reboot is, I don't think the Nav needs to decode the week count after it does initially. It only needs the second count to keep time. It would never drift off by a week when it's free running. That's why those of us who haven't removed power don't have the problem (yet).

Crazy. I blame Denso.

[kevm14]

Ah, I missed this the first time!

So check this out. The last GPS week count rollover was Aug 21 1999. What if the Denso programers messed up and only used 3 digits for the week count instead of 4 (decimal digits not bits).

Receivers should go to 1024 weeks, but what if ours only go to 999. That point in time would have been around the first week of November.

But then it takes a full power cycle by removing power to the unit, so there would be a delay before the first person would see the problem.

I know I'm way into the guessing game now, but it's a possibility. If that's it something will happen in April of 2019 with the next rollover. Either it will be fixed or we'll get the wrong date displayed.

https://tycho.usno.navy.mil/gps_week.html

So I did a little Excel math. Aug 22, 1999 + 999 weeks = 10/14/18. That is right around the time of people first reporting this issue.

So like he said above, when it actually rolls over next April, MAYBE it will start working again since Denso threw away 24 weeks (from 999 to 1023). Curious if it will show a date with a year of 1999 or 2019. Again, mine is fine now, though I am quite sure if I disconnected my battery I'd lose the time and date (supposedly nav still works fine).

I think one of these three options is likely:
- Even after next April, nothing changes and the clock never works again after disconnecting the battery
- It resets to week 0 which is 22-Aug 1999, but with the correct time so at least the clock can be set (I don't have much use for the month/day/year and never display it anyway)
- Everything magically is fixed and it has the correct 2019 date.

I vote #2 happens. But #1 could happen. #3 is highly unlikely, unless there is another field in the GPS data that identifies the active epoch. It would then know that the next epoch begins in 2019 and #3 is actually feasible in that case.