When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
The pot of water example is not appropriate IMO. Water is absolutely gushing through the motor at that speed. The water measured at t + 1 second is not the same water measured at t. Water temperature is not homogeneous.
It is explicitly the reroute that I am proposing as the source of this being observable due to the above flow.
I wish I had that FM report, because if memory serves it clearly showed a substantial effect on measured temperature from the turbo water feed alone. I do know Mike, I may reach out to him to ask his opinion.
Couple things I would like to add to this comment from a heat transfer / thermo perspective. The pot of water, while not a direct example, is still pretty relevant. The point with that example is, the water cannot cool down instantly the moment you take away the heat source, nor does it boil the instant you apply a heat source. In a similar way, introduce flowing water to a heat source (an engine) and loop it to a cooling source (the radiator) and the water will reach an equilibrium point in any local point you measure it. You are correct that each local point will not be the same temp, but those local points still cannot change instantly due to properties of water. It takes time for any locally measured temperature in flowing fluid to change whether it be from an increase in heat addition rate (wide open throttle) or an increase in cooling (radiator fan kicking on at idle). You do not see 'step changes' in temperature in any example that I've seen. The measured coolant temperature ALWAYS follows a smoothed (slightly time delayed) curve to changes in heat rejection or heat input rates. Refer to Curly's example quoted below.
On your FM report comments, adding a turbo water feed line to the cooling system should notably increase the system temperature because you are adding a heat generator to the coolant system without improving how the system can reject heat. The increase in power (aka heat from fuel burn) that comes with a turbo further strains the cooling system whether the turbo has a water cooled housing also connected or not. This is why radiator, ducting, and hood vent upgrades are so important for our cars when boosted because that's the only proven ways we've consistently been able to improve the heat rejection side of the equation.
Originally Posted by curly
I found another log where I did a 2-5th blast in another car, and coolant peaked 20 seconds after I returned to idle, held there for 6 seconds, and dropped back down to normal operating temp within 30 seconds. That's typical.
Not to argue with this specific application, but point of perspective(s)
The technology I primarily identify with uses RF electrical energy to heat material that is mostly water in contact with metal. We can heat a small volume (say a couple cc) by 60C+ in 1-2 seconds with a time averaged power of 200W maximum. If my calculations are correct, we are talking about ~200000W produced by an engine with 300hp, so 3 orders of magnitude.
We know that the EGT can go up by many hundreds of degrees (F) and the drop by that same amount very quickly.
We also know that an A/A intercooler can remove enough heat from air to reduce the temperature in a single pass (lower thermal mass but still a fluids problem) by over 100 degrres F.
I mention the FM report and turbo because IIRC it showed that the temperature leaving the turbo while in the transient high EGT case was a step function at a macro level. Which is to say that while on boost the temp was significantly hotter than while out of boost.
Bottom line is that the data Curly has shared is undeniable and if I was really motivated to work this out (which I am not, my zeal for such things has waned) I would instrument and conduct experiments. In my current state I am good with taking the knowledgeable inputs from this forum and moving on with life. My "trust info from rando internets" filter is quite good. You and curly are good sources.
I have many thousands of WOT pulls on this setup, so I suspect it will hold together.
I think the key points in the technology you are pointing at is, heating a couple cc's of fluid by 60C+ in 1-2 seconds with 200W is a lot different than heating gallons of moving fluid with 200kW that is simultaneously being cooled by slightly less than 200kW from the radiator. What you are describing sounds oddly a lot like tankless water heaters if I'm being honest... If I'm right, there's very clear similarities AND differences between those units and powertrain systems (including A/A intercooler circuits).. I could get into what those are, but I'd rather keep thread on track.. lots of good information coming out of this thread.
@SimBa
I think you have identified a brain tickling observation (at least for me) on the temp drop behavior. As far as your temperatures go overall, I do find it interesting that the temperature spike is around 104C and it plateaus at that level in the screenshot you have in the OP. Have you seen it ever get higher than that and was it trending on a slope or a flatline? If coolant temp is climbing fast during a WOT pull and in the middle of the pull your temperature flatlines, that is NOT normal behavior. The flatline in the OP being at 104C (219F) is ironically close to the boiling point of 50/50 coolant at atmospheric pressure. Have you checked that your system is building pressure and that you don't have a faulty pressure cap on your radiator? Temp monitored by a CLT sensor will hold a peak flatline temp of 100C (pure water) to 106C (50/50 glycol) if zero system pressure exists and fluid is boiling locally in the engine or turbo. Unless you are confident that pressure cap is not an issue, I'd recommend replacing the pressure cap to remove that variable from the list of possible root causes. They are cheap and easy to swap out, and I've seen some odd behaviors like this exist on faulty caps before on other manufacturer engines.
Thanks for the feedback and discussion guys. I appreciate the attention on this as it's already a concern of mine, and will only be more of a concern as we head into summer.
@Padlock Yes, the reason I started getting worried about cooling performance is that at an early autocross this season (probably ~60 F ambient) my car was breaking up through the finish line. I pretty quickly realized that I was hitting 110 C which was where I had my rev limiter start to decrease as a protection method. That was before I added ducting and bled the system more thoroughly.
The cap is from Mishimoto which is about a year old at this point. I don't have a reason to think it is faulty but I believe I have an old cap from my previous rad laying around. I'll see if that will swap on and see if anything changes.
I did a few more pulls on the way home from a rigorous DDR session at the arcade last night. Ambient temp was 8 C ( 45 F). This was taken via Bluetooth to MSDroid. I still don't know if that plays a factor compared to USB so I'll keep mentioning it.
I also turned off any filtering of the coolant sensor when taking this data, which is why it looks more jagged that the original photo I posted.
This was 1st-3rd flat out. Temp differences during shifts were
1-2 90->81 C
2-3 90->79 C
3-4 92->85 C
Average temp over the 3 seconds after the pull was 85 C. 10 seconds later temps had dropped to thermostat temp, right around 80 C.
If all of this was heavily smoothed then it would seem reasonable to me. IE if the temps during that log went from 80->85 C throughout the pull without the major spikes and drops.
I'm still leaning towards this being a grounding issue like Curly said. Something seems to be introducing noise as RPM increases. I will be out of town this weekend but will likely pick this back up next week.
Datalog is attached if anyone wants to poke through it. I also noticed that there is a pretty noisy bit at the beginning of the log during cruise as well, so I'll poke into that a bit more as well.
Swapped out to my old cap today and drove around for a bit. No apparent change in cooling performance, but I wasn't able to push the car much since there was traffic everywhere. I'm going to leave the cap on for a bit longer to give it the benefit of the doubt, but I don't think that's going to be the fix.
I was looking at some datalogs and graphed voltage vs coolant. I thought I was seeing a correlation so I stitched together all the datalogs I had from a PCA event a few weeks back and made a histogram. I always forget about that thing but it's an amazing tool when you need it. Most of this data was after I had turned the boost down a bit.
My theory was that my battery voltage was dropping and causing the coolant sensor to read high. Based on this data it doesn't look like that's the case. In the cells/rows where there are a larger number of data points the coolant temps look pretty close across all voltages.
I'm still suspecting this to be related to an electrical issue, such as grounds. I still intend to try to improve those. Any tips on how best to do that are appreciated.
I'm so happy this thread exists. I've also done a lot of what OP stated earlier this year when I pulled the engine to reseal it, timing belt, coolant reroute etc and I would say the resultant cooling data is anything but predictable. I'll give some for instances.
'02 VVT
FM Coolant Reroute
OEM sensor is in the reroute
Aftermarket thicccc radiator (cobalt)
Driver and Passenger side fans, drivers kicks on at 195, passenger at 200.
180F thermostat
FM 20psi radiator cap
Almost completely water and water wetter, any coolant is just by accident
I've bled, rebled, and re-rebled with the nose way up in the air
Still have AC condenser and use it (I'm in florida y0)
I'll go pick up my kid from school in 85F weather. Its a 10+ minute drive through rural roads at 65mph with no stop lights. On the way there, with AC running, I'll be at 190F coolant temps with 2 to 5 degree variation. After stopping for 5 minutes to pick him up with the car running (temps will generally rise to 203 to 205 while stopped). Most times driving back now the car will be 205 to 215F and sometimes the car will cool itself back down under 200F. The point is, its wildly inconsistent and I have no reasons as to why that would be. I have noticed that again, sometimes, if I turn off the ac the temps will drop, but sometimes it wont. I've noticed if I repeatedly go off throttle, think of it like the ultramiling that prius owner's do, it will get it to drop. I'm left VERY confused by the lack of consistency.
I think our situations are actually quite different. It seems like your coolant temps are inconsistent over a longer period of time or situation.
Mine is more immediate, more like a noisy signal to the ECU. If I average my coolant I usually don't see more than 5-10C over thermostat temps while cruising. Keep in mind we've barely gotten to 80F this year so far. However, I might see 90 C one second and 80 C the next second, which is what I'm concerned about.
I think our situations are actually quite different. It seems like your coolant temps are inconsistent over a longer period of time or situation.
Mine is more immediate, more like a noisy signal to the ECU. If I average my coolant I usually don't see more than 5-10C over thermostat temps while cruising. Keep in mind we've barely gotten to 80F this year so far. However, I might see 90 C one second and 80 C the next second, which is what I'm concerned about.
Definitely - I was just glad someone else opened a new thread so I don't get another -1 cat!
Not saying you have to do this, but to remove the OEM wiring/sensor from the equation, have you considered adding some sort of secondary coolant temp sensor to validate the measurement you are getting is "right"
Quick example of something I found in 0.002 seconds from Google...
@Padlock , funny you mention that. I asked on the SpeedyEFI troubleshooting group and another user who is very competent ( @Alejo_NIN ) said that they see the same phenomena that I am. They said it's been this way since day 1. The guys who make the boards said the same thing as you, to try installing another sensor and see how it reads.
Currently, I believe this is an issue with the grounding of the sensor. I was talking about this with someone at autocross this weekend and without being prompted that's immediately what he called out. I don't know the guy, but he seemed to be pretty familiar with ECUs and electronics. I do not believe that I am actually hitting the temperature spikes that I originally reported, and view them as an anomaly more than an issue.
This weekend I ran at a fairly large autocross course, getting into 3rd gear in places, ~50 seconds per run. I cranked the filtering on the coolant sensor up to 220 out of 240. In the original post I believe it was at 120, and after that I dropped it to 0. The tool tip in tuner studio says a value of 180 is recommended. I believe the SpeedyEFI board is laid out according to the official Speeduino designs, so I plan to ask on the Speeduino discord if this is a known issue. With the filtering cranked I saw peak temps around 100 C (212 F) and ~2-3C drops when getting off throttle. I will try to get some screenshots and/or logs uploaded later.
I am going to lower the filtering a bit and see what it looks like. This weekend we driving in 28 C (82 F) ambient temps and the coolant was at 87 C (189 F) by the time I returned to grid, so I feel comfortable that if my coolant temps are getting over 100 C, they are not staying there for long, and very likely not getting close to 110 C like I originally saw. While I would like to figure out what's actually going on here (and plan to continue pursuing this), I also have a lot of other odds and ends I want to tidy up on the car.
Now that I have coolant temp from the Megasquirt on my digital dash, I've noticed that my ECU's coolant temp readings are much spikier than my separate coolant temp gauge that I've always gone off of.
When I set up my MS, I already had the coolant gauge installed, and assumed it to be accurate since it perfectly correlated with my 195*F thermostat opening temperature. I played with my coolant temp sensor calibration in TS until the CLT in TS read within 2-3* of my standalone gauge at all temperatures. This was done with the car idling in my garage, though. I didn't look at coolant temps much while driving with TS open because I had my standalone gauge.
Doing some pulls on the way home yesterday, I saw the CLT in TS on my dash spike quickly from 190* up to 196*, then down to 187* and back to 192* during the end of a pull and return to 6th gear cruise. Did another and watched my standalone gauge slowly creep up a few degrees, then back down to 190*, with no spikes. I haven't delved into it further because I haven't seen any evidence of deficiencies in my cooling system, but figured this information might be pertinent.
Edit: both sensors in question are mounted in my coolant reroute/T-stat housing, so no discrepancies in measurement location.
I also prefer only one sensor, as long as it's a known good one. Just like a compression gauge, it's mostly used as a comparative sensor. As long as it's not spiking to 230 when it usually never goes over 210 for instance, I'm ok with being a few degrees off from some laboratory measured coolant temp.
NAs do seem to be more accurate using the "rx7 S5" calibration vs. Mazda, fyi. While at room temp, coolant matches GM's AIT calibration, and thermostat opens/closes at expected values. With the "Mazda" calibration, they'd ready ~10 degrees higher than MAT when cold, and also be off by 10ish degrees at thermostat temps.
Temp sensor refresh rates should be in the .5-1hz range, where as RPM/oil P/fuel P I would prefer in the 10-20hz range. I believe that's kinda what you're adjusting when you play with filtering,
I also don't know exactly how the filtering works. I'm assuming it is averaging while maintaining the same sampling rate. Still a workaround, but it'll work for now.
NAs do seem to be more accurate using the "rx7 S5" calibration vs. Mazda, fyi. While at room temp, coolant matches GM's AIT calibration, and thermostat opens/closes at expected values. With the "Mazda" calibration, they'd ready ~10 degrees higher than MAT when cold, and also be off by 10ish degrees at thermostat temps.
Funny you’d mention that. I started out running the stated RX7 calibration after making a thread about CLT sensor calibrations, which I believe you answered. One of my first posts on this site lol. Ran it no problem but ended up tweaking the values a bit after seeing coolant temps in TS reading almost 20* lower than my standalone gauge at higher temperatures only. I’m running an Autozone CLT sensor, though, so it’s very possible the resistance values from my sensor are just hella off. It’s stayed consistent for two years, though. I’ll check IAT vs CLT next time I go to cold start the car.
one thing i was able to do was fix my "grounding issue" which was also affecting my TPS and voltage readings.
the connections to the ALT were corroded...i cleaned this and it fixed the issue on those two.
however, I even swapped to a GM clt sensor and same results.
i have not tried another sensor to check but to go from (Fahrenheit) 190 to 220 in a second then down to 198 the up to 230 then down to 190 then up to 240 and then down to 200 its not normal behavior.
this would call for a faulty temp sensor. and I've tried many.
hell, I even tried it on the bench, and guess what, it works perfectly.
it has to be a ground issue. i cannot replicate it on the bench
Interesting that you call out the TPS. After reinstalling the engine I have struggled to get AE working the way I want due to noise on the TPS. Previously I had no issues.
I am curious if this could have to do with the larger injectors I installed, as I assume they're drawing more current than the stock ones.
So cleaning the alternator connections helped, but did not solve your issue if I'm reading that correctly? Thanks for the data point, glad to see this isn't an isolated case.
05-07-2024, 03:58 PM
1
