Oil Temp Sensor Location
 

Other than tapping another hole in my pan right after tapping, flushing and refilling with new oil for the turbo drain, where are all of you locating your oil temp sensor?

If the pan is the only option where did you tap it? Sensor is a 3/8 npt that came with the Maxigauges.

There are oil filter sandwich plates that have a tapped boss OR you could add a tapped boss to an existing sandwich plate IF you're using one (for an oil cooler.)

I will be tapping my pan for a GM coolant temp sensor to use for oil temp. Make sure that you tap the pan on the passengers side. It is a pita to deal with a temp sensor wire around your hot manifold.

what about the oil feed ports on the back of head??

There was some discussion about using these as an oil feed port for a turbo but don't see why they couldn't be used for oil temp.

If you guys have the pan off the engine, the way to do it is a weld bung

https://www.miataturbo.net/attachmen...ine=1396288064

The pan is so thin that I just wasn't comfortable with like 1 thread worth of engagement. I added 1/2" NPT for oil return on the drivers side and 1/8" NPT for temp on the passenger side below the baffle. I was worried with the sensor interfering with the subframe but I guess I'll figure it out shortly.

https://www.miataturbo.net/attachmen...1&d=1396302913

https://www.miataturbo.net/attachmen...1&d=1396302913

Top of the sensor is 1.5" below the baffle. I do wish this sensor was a bit shorter though.

Reminds me, I need to look, my car came with a remote oil filter. Need to see what was there for a mounting plate in lieu of the sandwich plate.

Wow that bung would have made life easier, I used an aluminum half coupler when I welded mine on. To the drivers side rear of the pan... because thats what Hustler said in this one post one time and it seemed legit.

Yeah, I just don't know if it clears the subframe yet :noes:

Got a pic of your install location? You mean yours is like, below the bellhousing?

Mine's like the mirror image of yours on the drivers side, but in the same set of ribs as the oil plug. It may have hit the subframe a little bit while dropping the engine in.

If it hits I have a spare weld bung, but its not something I look forward correcting.

https://www.miataturbo.net/attachmen...ine=1396308575

Not my pic, but its gonna be damn close at any rate.

Also, don't even think I welded that myself. Mad props to Gesso for being a TIG ninja :bigtu:

I run mine just above the oil drain plug, it is nice and out of the way.
Let me try this link
https://www.miataturbo.net/attachmen...ine=1396314712

I can't see your image, but I'm starting to think that's what I should have done. Problem is, I don't know how much oil is actually sitting in the pan at 7,000rpm and I wanted to get it as low as possible.

Image working for anyone else?

Did not want to jack the car up so it is a bad phot, but you can see that it is above the edge of the powerplant frame.
https://www.miataturbo.net/attachmen...ine=1396315083

Did this location work out for you? Is it low enough in the pan?

That position is below the low level mark and it is pretty stable on the track so I believe that it is an accurate reading, not sure how much oil is in the pan at 7k but I hope that it would have a couple inches worth or I would be seeing oil pressure drop on the straights.
That location has the hottest oil in the engine and that is what I want to keep an eye on.

Is there any data on deltaT between the sump and either a sandwich plate or an oil cooler/remote filter? I doubt very much whether there would be any appreciable difference in T between the sump and sandwich plate, although considering that the oil pump almost certainly adds heat to the oil it might be hotter than the sump :confused:. I have always used a location downstream of the pump, but I am not convinced that a few degrees one way or the other is enough to get excited about - unless you are right on the upper temperature limit for your oil, in which case sensor location is not something to dwell on, and more cooling is your focus.

I am about to install temperature and pressure sensors on the inlet side of my cooler in a engine>cooler>filter>engine setup. Next time the engine is out I might look at an alternative (or additional?) sensor location in the sump or on the sandwich plate.

I agree that the pump will increase the oil temp when pressurizing, my sandwich plate is a Moro with a 170f thermostat and did not have a temp sensor port so I wanted to stay away from measuring oil temp before the oil cooler because the cooler does not see oil flow until the thermostat opens.
I do try not to romp on it until I see the oil temp coming up and is over 100f (that can take a while in cool weather) so the oil temp in the sump or sandwich plate is not a bad thing IMO

https://cimg8.ibsrv.net/gimg/www.mia...72e31dfe55.jpg
another option is to drill and tap the oil drain bolt. Volvo did this using VDO sensors years ago. I'm using this, but I feel this location is cooler than what the rest of the motor is seeing.

I'd be wary of putting a sensor in that location, where it's so exposed. One stray rock or chunk of ice would easily snap any temperature sensor I can recall seeing.

I agree, it's an easy solution for testing. Welded bung would be better while you have the oil pan off.

My race engine builder was very forceful in telling me to keep the revs and throttle opening down until 80*c. In cold weather (~0*c) with no oil cooler that took forever, and sometimes wouldn't happen without breaking that limit. I then installed a laminova oil/water heat exchanger, which worked well to bring the oil up to temperature, and then stabilise it - without needing another oil cooler to keep the maxT under control. I haven't managed to find a way to fit one to the new car yet, so the best I can do is fit a conventional cooler behind the radiator. That will be much less effective than the laminova in stabilising the temperature, so I'll see how it goes - hence the need to monitor temperature.

No surprise I guess, but your thermostat's 170*f setting is pretty much my engine builder's 80*c.

The oil pump adds more heat to the oil than any other component in the engine. Measuring at a sandwich plate is a fool's errand. Measure at the sump or don't bother.

While I've seen no data to substantiate the claim that the oil pump adds more heat to the oil than any other part of the engine (and have doubts as to whether this is accurate), if this were true, then wouldn't it make sense to measure oil temperature after the oil pump? It is, after all, peak oil temperature which is of most interest.

It would, if the industry did it that way. They do not. They measure it at the sump, so you (the royal "you") should also measure it at the sump, since at the end of the day, the goal is to determine whether or not the measured oil temp in one particular system is acceptable as compared to generally defined industry standards.

e: As far as that claim, it comes straight from an oil engineer who cut his teeth inside JGR's NASCAR engines program, and who now works for Driven, which is JGR's retail oil sales company. Oil pumps contribute the most heat to the system, valve springs contribute the 2nd most, and pistons are third. He described testing rigs where they would spin a bare cylinder head up to speed with no bottom end attached, and hook an oiling system to it, and the oil in that system would reach 180 degrees with no combustion heat whatsoever.

I spent like 30-40min talking with the guy at PRI last year. The main takeaway I got was that oil is fucking weird.

I'd love to see some documentation on this.

Mostly because it's the kind of claim that seems to make a lot of sense if you know a little bit about fluid dynamics, but starts to make less sense as you know more about fluid dynamics. (I'll admit that it's one of those things that might start to make more sense as you get into PhD-level fluid dynamics, because physics can be weird.)

On the one hand, it seems to make sense because the oil pumps is, of course, pressurizing the oil. And we all know that when you compress a fluid, it gets got. This is how air conditioners work, it's why forced induction engines have intercoolers, it's why tire-inflation pumps get hot, etc.

Except that, like most liquids, oil is incomprehensible.

What is it about the oil pump, specifically, that causes it to put more heat into the oil than all of the bearing surfaces and piston rings?

The source is Lake Speed, Jr. You can google his resume. While I can appreciate skepticism of information that does not appear to be obvious on its face, I am not the kind of asshole who questions people like that when they tell me things that directly relate to their area of expertise. :party:

e: -4 props for sharing info about oil provided to me by the R&D director of a NASCAR team's lubrication division. Did I stumble onto Facebook?

If I had to hazard a guess, it's not the actual compression of the oil, because oil doesn't compress (yeah, yeah, I know it does, just not an amount worth caring about), but the shearing effects within the various interfaces of the engine.

Yeah but. Do you vacuum down your oil?

Oil pump is positive displacement with excess oil being throttled through the relief valve, at higher and higher pressure relative to RPM. Possibly that is a contributor.

Turbo is also a contributor that would not come into play on NASCAR - only testing, if that were the case.

Re: Sensor in change plug. This is done on occasion and I have not heard of damage from any that have used that method.

This discussion fits well into the other one on coolers.

If Sav's informant is correct, then yes it is more important (for accuracy, not necessarily for outcomes) to measure T in the sump. For the health of our engines, we are interested in how the oil performs inside the oil passages/bearings/rubbing surfaces. That requires the oil entering the engine to be above the minimum operating temperature, and exiting below the maximum operating temperature. Lets hypothesise that this is an 80-120*c range. So we want the sump T below 120*.

It doesn't matter (probably within certain limits) how high the oil pump pushes the T above 120, providing the oil cooler can bring it below 120, hopefully closer to 80*, to provide a margin sufficient for the oil to be heated to no more than 120* by the time it enters the sump after doing its lubrication thing. Assuming that the excess T above 120* does not do irreversible damage to the oil's lubrication qualities.

Placement of the oil temperature sensor would be less important if the post-oil-pump T never exceeded 120* (or whatever the max operating T of the oil), as that would mean the sump T was less than 120*. However that also means that regulating oil T on the basis of post-pump T would also be safe, perhaps even 'safer', given the sump T is lower than the post-pump reading.

I gave you a :likecat: to try to combat what I admit is a bit of BS behavior here.

I'm still having difficulty with this concept, though. Beyond the simple fact that the mechanism of action for the oil pump putting a large amount of heat into the oil is non-obvious, I keep coming back to the physical properties of the oil pump itself.

It's not very stout, quite frankly. Those gears are sinstered metal, and not large.


I've been trying to pin down exact numbers here, based on the known thermal efficiencies of gasoline engines, the various ways in which "lost" heat energy is transmitted, etc. The problem is that I can't find any meaningful data on the horsepower consumption of an oil pump. We're all familiar with the parasitic crankshaft loss imposed by a supercharger, and the same basic mechanism of action is at work here. For the oil pump to be dong its thing, it must be taking energy from the crankshaft and converting it into a combination of useful work (pumping oil) and waste heat. And the number I can't seem to come up with is how many HP would have to be going into that fragile little gear, and being lost within it, in order to produce a greater amount of heating than the crank bearings, the piston rings, the valvetrain, etc. But it can't be a small number.



I'm familiar with Mr. Speed, and googling "Lake Speed" "oil pump" has thus far returned a bunch of his talking about why his company's oil is the best you can buy, and not a single reference to oil pumps being the largest contributor of heat to the oiling system. I'm not saying that I doubt you heard this, merely that this is one of those concepts which is so far removed from the obvious that it begs for a citation.

I'd also like to see documentation, simply because this sounds interesting, i like to read, and previously on this forum we've hammered home "Miatas are special and not-Miata DPE doesn't matter." Hopefully that doesn't make me an asshole. :giggle:

I'm surprised that the pump puts the most heat into the oil. A pump that size cant draw more than 2 hp, and at worse must be 60% efficient, so at most its going to be putting 0.8HP worth of heat into the oil, that's like 600 watts. That doesnt seem like it would be the largest source of heat for the oil in the engine.

For that drain pul that someone drilled and tapped for the sensor, you know you can just buy a drain plug already setup like that.

Just as a data point for the earlier post about sender location; on my Lotus, which ran much too cool because of the twin stock coolers in the nose, I had first put the sender in the sandwich plate. When the temps were so cool, I doubted the accuracy of that location, so moved it to the lower part of the sump, and got exactly the same temps. At least in this instance, it didn't seem to matter, so for many, who will not have the engine out, or pan off to weld a bung, but do want to begin monitoring oil temps, the easy sandwich plate location probably makes sense.

Thank you!

This thread had gotten me to trying to remember where I mounted the temp sensor back in 2006 when I was just starting into the '92 build. That was it. I found an adapter that was M14-1.5 male x 1/* NPT female. Just do a Google search. Beware that many (most?) of these fittings don't have the flange around the large end, and will thus leak. I thus wound up drilling and tapping an OEM plug.

It was slightly annoying, as you had to unscrew the sensor wire before pulling the plug to drain the oil. After after about two years, something hit it while I was driving on a rough road and broke the stud off of the sensor. Fortunately, the oil stayed in the engine.

I've been searching back through historical photos trying to find a picture of that install. This is the first photo I found showing the gauge in place, but I can't find a picture anywhere near it showing the sensor itself.

https://cimg1.ibsrv.net/gimg/www.mia...013928f98f.png

That photo (and those around it) brought back some good memories. I really loved that car. And also living in Carlsbad, CA, where I really got to stretch its legs.

No arguments there. I didn't spend a whole lot of time doubting him, mostly for the aforementioned "don't look a gift horse in the mouth", but also because right after he said that, he told me the bit about the valve spring testing, and I spent most of our chat trying to wrap my head around that.

In a Fermi-esque way, I would start by figuring out how much oil volume a normal Miata pump is designed to move (or really any gerotor pump, or in absence of that data, any engine oil pump at all), and then look to either the beverage industry (soda/beer/cider) or maybe the petroleum industry for a similar pump with a raw horsepower rating.

Joe, so you’re comment about the sensor being damaged WAS anecdotal, NOT conjecture.

Nice pic.

I'm imagining a thicker oil creating more heat. There's got to be a trade-off between film strength for protection and oil thickness causing drag and heat.

What's wrong with measuring something at its hottest point? If your gauge never goes above 250 or so in the sandwich plate, would it not be safe to say the pan will be the same or cooler? Seems like the same idea as measuring the coolant temp near/in the cylinder head vs in the radiator or something along those lines. Just thinking out loud a bit.

Correct. Thicker oil will run hotter than thinner oil. Too thick and it's entirely feasible to push the oil to a point where a hot 20w50 may not work as well as a cooler-running 10w30, for instance. Ergo, it's best to run the thinnest oil you can get away with given your bearing clearances. I run 10w30 in stock engines and 15w50 in built engines (specifically, in any engine with a non-stock bearing package).

Has someone said that measuring the oil at the (suspected) hottest point is not theoretically optimal?

For most of us, I think that practicality (in terms of ease of installation) is probably the key factor, mostly because we're interested in trends (is the engine running hotter / colder today than usual?) rather than in using highly accurate instruments to establish precise, exact measurements.

That being said, it would be interesting to see someone instrument an engine with oil temp sensors in multiple locations, and compare them (using sensors and gauges of higher-than-Autometer precision, of course.) I, for one, would love to see comparative, simultaneous measurements of the oil in the pan, the oil at the filter (before the cooler), the oil after the cooler, and the oil in some location after it's passed through the majority of the engine I'm not sure if there's a passage in the head which is easily instrumented in a location that isn't a dead-end. Maybe drill a tiny hole into one of the galley plugs and pass a K-type thermocouple through it.

But I'm diverging from the practical into the academic here...


You can has plastigage?

I agree with you Joe. Most of use are just looking to for changes in setups or watching for issues. Trying to compare values to anyone is the internet is always a crap shoot because there are so many factors. Accuracy of the sender, accuracy of the gauge, location of the sender, etc.

This is an interesting paper I ran across:

The roller bearing crankshaft | Schaeffler Symposium 2018

The purpose of the paper is to present the friction benefits of adopting roller bearings instead of flooded bearings for the crankshaft (i.e., Porsche 356 Carrera). The test engine was a Euro Ford 1L Ecoboost engine. Section V contains some interesting measurements where the stock engine was powered externally to determine torque consumption of various subsystems. The graphs are a bit hard to interpret because they are scaled in %Torque, which would necessarily be an increasing absolute torque value (somewhat linear?) with increased RPM. In any case, here's the graph:

https://cimg6.ibsrv.net/gimg/www.mia...b168162bea.png

There aren't any real surprises here. The main %Torque consumer at all tested RPMs is the pistons/conrods. Next in line depends upon RPM. The %Torque consumed by the crankshaft steadily increases with RPM and, if the engine had been spun to a higher RPM, would have become the main consumer (not a surprise to anyone familiar with ARPs "Reciprocating Weight vs. RPM" graph -- https://arp-bolts.mobi/p/tech.php?page=2). At 6000RPM, it was second behind the pistons/conrods.

To Andrew's point, the oil pump was a consistently heavy %Torque consumer with a more-or-less stable percentage across the RPM range. At low RPM, the oil pump consumed more %Torque than the crankshaft. This relationship reversed around 4000RPM.

The valvetrain was interesting, as it represented a decreasing share with RPM. The Ecoboost uses roller lifters, so I'm not sure how well this relates to the Mazda B.

Incidentally, "FEAD" means "Front End Accessory Drive." I had to look that up.

I suspect most people are looking for validation that their oil temps are in line with what the larger automotive industry as a whole deems to be acceptable. To that end, measuring the oil temp in the same place that the industry does becomes quite important. If you measure at the sandwich plate pre-cooler and get a temp of 280*F, but the sump temp is 240*F, and the industry says that a sump temp of 240*F is perfectly acceptable given the oil you are using, then you might spend an awful lot of money to solve a problem which doesn't actually exist.

Nice find. I wonder if the rings and pistons contribute a higher share of the friction torque, but less of that friction (heat) is shed into the oil since the oil doesn't come into contact with the pistons/rings nearly as much as it comes into contact with the oil pump?

I would suspect so. In particular, there is a heat path to the cylinder block water passages for the pistons and rings. So the heat from that friction would be absorbed by both oil (in the bearings) and water (cylinder walls). OTOH, the crankshaft and oil pump would reject their friction heat almost exclusively to the oil.

Now that is an excellent second-source which validates Sav's claims regarding the oil pump.

I truly would not have expected that.

To explain, my doubt was fueled largely by this ASME paper, which specifically explores the use of variable-displacement oil pumps: https://www.asme.org/engineering-top...etter-oil-pump

They are seeing overall efficiency gains in the 3-6% range, but yes, if you decouple the dominant source of friction in the engine (pistons) from the heat-to-oil calculation (which is obvious, but I hadn't realized the magnitude of this contribution to overall friction), then 3-6% overall starts to become meaningful in the context of heat-to-oil.


I stand corrected.

I wonder how they are breaking down "crankshaft and seals" and "pistons and conrods". I assume that the rings are included in the pistons/conrods section, and I also assume that the *vast* majority of the friction created is by the rings, as the pistons should technically never actually contact the cylinder block during normal operation (or at least no more so than the bearings contact the crank). I do wonder if the rod bearings are included in the "crank" or the "conrods" category, and how much of that friction is due to windage losses from the physical crankshaft itself and how much is due to the friction in the bearings themselves (I know the frictional losses to the seals are extremely minimal).

Given that the analysis was subtractive (measure the load, remove a part, and then re-measure the load, then remove another part, etc), I think we can assume that windage losses were accounted for. As these include both the air contact with the rotating assembly as well as pumping losses on the underside of the piston, it seems safe to assume that the rings themselves constitute a vast majority of the combined pistons / conrods / crankshaft friction figure.

On the other hand, these losses are still friction, they still generate heat, and it's reasonable to postulate that at least some of that heat is going to be absorbed by the oil. (Much of it will, of course, be vented by the PCV system, and absorbed into the structure of the block and, therefore, the water jacket.)