What's new
Van's Air Force

The definitive Van's Aircraft support community! Buying, building or flying an RV? Join our exclusive family of mentors and enthusiasts!

G3X Oil Temp probe inaccuracies

A

amaris

Well Known Member
I've been going round and round on trying to get this one solved with no luck.

I'm running the G3X/GEA system on my 7A. The oil temperate probe (2 wire UMA1B3A-2.5R with about 3' of 22 g shielded pair wire and UMA probe in GEA selected) has been reading low. The higher the temperature, the worse the variance. Testing with known temperature heated oil (and verified with 3 thermometers), it's 35 degrees too low when at 290 F and slowly gets better as the oil cools. So at 220 F it reads 207 F, at 200 F its 188 F and at 180 F it reads 169. Finally around 100 F it's nearly spot on. Garmin replaced the probe and I have the same results.
I've tested the wiring, connectors, just the probe, grounded and ungrounded the probe and can continually replicate the inaccuracy. I've tested the probe both at the leads and back at the dsub pins just to verify the wire run isn't a problem either and then ran a new run of wiring to the GEA just to take that issue out of the equation.
Garmin said to try adding an inline resistor. Which I will do just to keep the support moving forward, but I've not come across anyone who says this is a solution. I assume this will skew the entire range and while it may fix one side, the other side of the temperatures will be off. But I'm pretty remedial when it comes to electrical theory.

I've also tried all the different probe settings in the GEA and that doesn't do anything. The probe resistance tests pretty close to what it should be at various temperatures.

Has anyone come across this or have any thoughts on solutions or things to try?

Edit; Moved to engine monitor
 
Yes, certainly verify that you have the correct probe and the correct configuration.

I wouldn't simply add a resistor.

Why not post some time-correlated data? G3X indication, probe resistance as measured with a good quality digital meter, and oil temperature as measured with a verified-good digital thermometer - all readings taken as close together as possible.
 
I went through something similar early on. My oil temp on the G3x red X'd on my first flight. Troubleshooting in a pot on the stove with a multi-meter and a candy thermometer revealed that things were good and it was tracking appropriately up until about 200 degrees, then about 1 out of every 5 times if you wiggled the wires just wrong, once it hit that point it would arbitrarily shoot up to a resistance equivalent to about 500 degrees which was way higher that the range I had set for the tape so it was red x-ing.

In my case, the probe was bad right out of the box. It's unlikely but not completely out of the question to think that you may have got two bad ones in a row.

If it's not a bad probe, and you've got the correct selections in the setup menu, it seems like the wiring is the only thing left. Sounds like you've got a good handle on how resistance should change with temperature, so you know how sensitive the whole setup is. Because of this sensitivity they are uber sensitive to a bad crimp or connector, so thats another thing to check. I went with good gold d-sub pins under two layers of heat shrink for the connection on this one.
 
Hi - I had the same problem- tried everything to solve it - in the end removed the Faston connectors and soldered the wires. So no connectors between the probe and the GEA24 which appeared to create too much resistance. Perfect readings now.
 
Some general follow ups.

It's the "R" probe version. The wiring is standard mil 27500 22-2 shielded wire purchased from Stein. It is not red and yellow thermocouple wire.

Resistance readings taken from the probe leads or from the full run of wire at the d-sub are spot on with what the charts show so it lead me to start checking the GEA further. A friend said to check excite voltage, which was 1.25v and Garmin said they were expecting 5v. They're sending me a new GEA to swap out and they suspect that might be the problem.
 
amaris said:
A friend said to check excite voltage, which was 1.25v and Garmin said they were expecting 5v.
Click to expand...
Sorry, but this isn't correct, that's not how the GEA 24 works.

Post your resistance and temperature measurements.
 
mburch said:
Sorry, but this isn't correct, that's not how the GEA 24 works.

Post your resistance and temperature measurements.
Click to expand...
What should the voltage be reading? Interesting that Garmin support had it wrong?

Actual oil temp in Celsius and corresponding ohms readings
12.6c - 105.6 ohms
117c - 140 ohms
110c - 139.8
98.3c - 136.5

Then tried the probe in the heated oil connected to the GEA.
188 F oil was showing 175 F on GDU

To try to get pics for support I had to disconnect the probe take an ohms reading and then reconnect the probe to show them what the GDU was displaying.
180 F oil was 130.2 ohms
174.7 F oil (79.3c) was reading 163 F (72.8c) on GDU. And then resistance was 129.1

I then ran a new wire to the probe from the GEA just to double check that wasn't part of the issue. About 3' of new 22-2 shielded wire.

233 F oil was reading 217 on GDU
I switched back to ohms and 110.5 c oil was 140.7 ohms

When I was on the phone with them yesterday I was doing ohms checks again with them and the resistance was usually within .5 ohms of the PT chart. He said he felt confident at that point that the probe and wiring were fine and that something with the GEA isn't reading the numbers correctly.
 
amaris said:
What should the voltage be reading? Interesting that Garmin support had it wrong?

Actual oil temp in Celsius and corresponding ohms readings
12.6c - 105.6 ohms
117c - 140 ohms
110c - 139.8
98.3c - 136.5

Then tried the probe in the heated oil connected to the GEA.
188 F oil was showing 175 F on GDU

To try to get pics for support I had to disconnect the probe take an ohms reading and then reconnect the probe to show them what the GDU was displaying.
180 F oil was 130.2 ohms
174.7 F oil (79.3c) was reading 163 F (72.8c) on GDU. And then resistance was 129.1

I then ran a new wire to the probe from the GEA just to double check that wasn't part of the issue. About 3' of new 22-2 shielded wire.

233 F oil was reading 217 on GDU
I switched back to ohms and 110.5 c oil was 140.7 ohms

When I was on the phone with them yesterday I was doing ohms checks again with them and the resistance was usually within .5 ohms of the PT chart. He said he felt confident at that point that the probe and wiring were fine and that something with the GEA isn't reading the numbers correctly.
Click to expand...
You could verify that by disconnecting the probe and connecting a verified known resistance. Using a known resistance, the GEA24 and EFIS should show the corresponding temperature.
 
rocketman1988 said:
You could verify that by disconnecting the probe and connecting a verified known resistance. Using a known resistance, the GEA24 and EFIS should show the corresponding temperature.
Click to expand...
Thanks, I'm planning that Monday. I have a resistor kit coming today so I can build something for the higher end temperature range of the probe since that's where it deviates the most.
 
amaris said:
What should the voltage be reading?
Click to expand...
You can't use a voltmeter to determine anything in this instance. The oil pressure input on the GEA 24 doesn't work that way.

amaris said:
174.7 F oil (79.3c) was reading 163 F (72.8c) on GDU. And then resistance was 129.1
Click to expand...
The conversion from sensor resistance to displayed temperature appears correct here for a Pt100 RTD, with any small disagreement likely attributable to differences in when the measurements were taken.

amaris said:
233 F oil was reading 217 on GDU
I switched back to ohms and 110.5 c oil was 140.7 ohms
Click to expand...
This again seems basically correct in terms of conversion from sensor resistance to displayed temperature.

These sensors have a certain degree of error and nonlinearity, but you'd have to ask to manufacturer (UMA) what those tolerances are. It's certainly possible you got a bad sensor, but I don't see anything here that would indicate a problem with your GEA 24.
 
mburch said:
You can't use a voltmeter to determine anything in this instance. The oil pressure input on the GEA 24 doesn't work that way.
Click to expand...
Shouldn't the GEA be sending out an excite voltage to the probe?

This is the 2nd sensor and a couple years apart so it seems unlikely now that it's the probe especially since the readings line up with the charts. UMA indicates a +/- .3 c error margin and I would accept is this were 5 degrees as I understand some degree of error, but I can't see 20 degrees of error as anywhere near being acceptable. And I'm not finding other people that are having this sort of issue of accuracy so just seems unlikely that this is all within an acceptable error margin.
 
amaris said:
Shouldn't the GEA be sending out an excite voltage to the probe?
Click to expand...
Yes, but it's not a fixed voltage that you can measure yourself to determine whether or not the GEA 24 is working.

Allowing for the inherent vagaries of your testing method, your G3X system seems to be displaying the correct temperature based on the probe resistance you have measured. If you aren't satisfied that your probe is behaving correctly, I would look at the probe, not the rest of the system.
 
amaris said:
Shouldn't the GEA be sending out an excite voltage to the probe?

This is the 2nd sensor and a couple years apart so it seems unlikely now that it's the probe especially since the readings line up with the charts. UMA indicates a +/- .3 c error margin and I would accept is this were 5 degrees as I understand some degree of error, but I can't see 20 degrees of error as anywhere near being acceptable. And I'm not finding other people that are having this sort of issue of accuracy so just seems unlikely that this is all within an acceptable error margin.
Click to expand...
They are probably sending a current.
 
mburch said:
Yes, but it's not a fixed voltage that you can measure yourself to determine whether or not the GEA 24 is working.

Allowing for the inherent vagaries of your testing method, your G3X system seems to be displaying the correct temperature based on the probe resistance you have measured. If you aren't satisfied that your probe is behaving correctly, I would look at the probe, not the rest of the system.
Click to expand...
Sorry, I'm obviously not as versed as you are in this but how does a variance of 17 degrees between the actual temperature of oil that the probe is sitting in and what the GDU is showing me the oil temp is as a correct display? (I was using 3 thermometers, an instant read, one that takes about 3-4 seconds to stabilize and an infrared. All were within +/- 2 degrees from one extreme to the other) According to the PT100 chart, the probe provides the correct resistance for the correct temperature all the way to the female end of the D-sub pin. So somewhere from the male pin on the GEA to the display I'm seeing with my eyes, it's either introducing or removing a variable or it's not interpreting the data that it is getting correctly. And it's only doing this more and more as the temperature of the oil increases above 100 degrees Fahrenheit.
 
amaris said:
how does a variance of 17 degrees between the actual temperature of oil that the probe is sitting in and what the GDU is showing me the oil temp is as a correct display?
Click to expand...
I didn't say that your system is indicating the same value as your thermometer, I said that your system is indicating a temperature value that is correctly corresponds to the probe resistance you measured, accounting for the inherent inaccuracies of your measurement method.

amaris said:
According to the PT100 chart, the probe provides the correct resistance for the correct temperature
Click to expand...
Be careful assuming this - very small resistance changes correspond to pretty large temperature changes. For instance:

amaris said:
Actual oil temp in Celsius and corresponding ohms readings
117c - 140 ohms
110c - 139.8
Click to expand...
This doesn't look correct to me, which you can verify here if you like.
 
Allen,

On the GEA 24, where do you connect the Oil Temp leads? -- and similarly, which pins did you measure when you got a voltage; J243 Pins ???

B
 
Walt said:
Can't have current flow without voltage....
Click to expand...
You are spot on! :) The issue is what is being forced- a current or a voltage, then what is being measured: a voltage or a current, respectively. Both ways can be done however forcing a current is how these types of things are done for a number of reasons. One advantage of forcing a current is the power dissipated, along with the self heating is limited. If a voltage is forced, high current can flow if the resistance decreases too much, necessitating short circuit current protection which complicates the circuit. Again, it can be done either way. :)
 
Walt said:
Everything you always wanted to know about RTD temperature measurements.
https://www.ti.com/lit/an/sbaa275a/sbaa275a.pdf?ts=1741452926462

Looks like this is the probe spec sheet but I can’t find any ref to which class probe (class tolerance range spec):
https://umainstruments.com/wp-content/uploads/2021/10/UMAspec1BxRevH.pdf

Maybe the GEA needs a temperature compensation adjustment for those desiring higher precision?
Click to expand...
Thanks, Walt. Looks like 41 pages of fun!
Yes, that's the probe sheet I found as well. Garmin must have something different because he quoted a smaller range than what's on the sheet.
 
amaris said:
Shouldn't the GEA be sending out an excite voltage to the probe?

This is the 2nd sensor and a couple years apart so it seems unlikely now that it's the probe especially since the readings line up with the charts. UMA indicates a +/- .3 c error margin and I would accept is this were 5 degrees as I understand some degree of error, but I can't see 20 degrees of error as anywhere near being acceptable. And I'm not finding other people that are having this sort of issue of accuracy so just seems unlikely that this is all within an acceptable error margin.
Click to expand...
You are thinking of a pot, where they send reference voltage out and to ground and pick up a reduced voltage off the wiper lead. Simple resistive senders like this use a pull up resistor inside the box. They do send power on the line, but it is not a reference feed, like the pots. I am sure there is a way to test it, but I think you need to know what the box is doing inside first.

Also, you must understand that these senders do not respond very quickly. You have a 1/2” chunk of brass and they drill a 1/8” hole to insert the rtd100. That mass has to absorb the heat before the sensor sees it and that takes some time. I made me own probe and that rtd100 is tiny. If they didn’t get the epoxy all the way down the hole, you get an insulated air gap and they are even slower then.
 
Last edited:
Walt said:
Didn’t notice the missing ‘R’ on the end, I believe this is the correct spec sheet for PT-100 series.
Bottom line is without a way to adjust the ref looks like not much can be done to correct an offset.

https://irtfweb.ifa.hawaii.edu/~iqup/domeenv/PDF/pt100plat.pdf
Click to expand...
Actually I think there might be. Instead of picking the uma, choose custom. You can then build a table for volts and temp. I installed a non standard fuel pressure sender (60 psi) and did this, so i know it works for transducers and guessing it also works for resistive senders. Unfortunately, you would need to do the testing on the plane to get voltages that map to temps, but pretty sure the system will show current voltage levels in config mode.
 
Last edited:
lr172 said:
Actually I think there might be. Instead of picking the uma, choose custom. You can then build a table for volts and temp.
Click to expand...
This is not supported for temperature measurements.

These sensors are generally very accurate. From the OP's description, it just sounds like they got a bad one.
 
I hooked in 2 different resistors into the wiring instead of the probe and the GEA/GDU were reading within a degree or two of what I was expecting based on the chart. One at around 220 F and the other around 175 F. That removes the wiring and the GEA/GDU out of the equation so it's back to probe. The kind folks at Garmin are sending me a 3rd probe now. This time I'll keep the new one and test it at the same time as the old one to see if there are differences. Hopefully, 3rd time is the charm!
 
amaris said:
I hooked in 2 different resistors into the wiring instead of the probe and the GEA/GDU were reading within a degree or two of what I was expecting based on the chart. One at around 220 F and the other around 175 F. That removes the wiring and the GEA/GDU out of the equation so it's back to probe. The kind folks at Garmin are sending me a 3rd probe now. This time I'll keep the new one and test it at the same time as the old one to see if there are differences. Hopefully, 3rd time is the charm!
Click to expand...
Good technique!
 
Well, problem solved, it was another bad probe.
Received the 3rd probe from Garmin and tested it side by side with the old in hot leftover peanut oil from Thanksgiving. Yes, the house smelled like old fried turkey so I don't recommend that approach again. But the test did show a small variance in the old probe compared to new probe with the new one being more accurate. But not as much as I would have expected. I then hooked it back into the wiring as it should be and tested hot oil from 280F down to 165F. The GDU was reporting about 10 degrees cool at the high end, but started to get more accurate from 260 being 5 degrees too cool. I tested every 5 degrees, comparing an instant-read thermometer with the GDU reading. At 220 it was back to accurate and continued that way down to 165F when I abandoned the test.
Next flight will see what the engine oil temperatures are actually doing.
 
Gentleman
I’ve have a new G3X system installed and I keep getting a high oil temperatures. I can go fly for 20 min, do some aero, normal temps. Once I get in the pattern oil temps go into the yellow and twice above 250! I replaced the thermostatic oil cooler valve. It didn’t fix the problem. Help!
 
After my issues, my first test would be to pull the oil probe, heat up some clean oil in a pot to 250 then with a digitial instant read thermometer and compare the probe to actual from 250 down to 140 and see if you have any variances from the hermometer and what the GDU is reading. When I did it because I wanted to use the existing wiring (to take that out of the equation), I heated the oil to 250 put the pot of hot oil close so the probe would reach and then read temps as they started to cool. You'll have some variance because the thermometer is more instant read and the probe takes a little time to change but it won't be that different. If those are off, next step is to repeat it but disconnect the probe wiring and use a meter and check the resistance against the chart. If that lines up correct, it's wiring or a setting in the GDU. If resistance doesn't match the chart, repeat it 3 more times. Make sure to use alligator clips or something secure, imperfect connections will change the readings. If from any tests that the actual oil temp measured matches your GDU, you indeed probably have a real engine temp problem.
 
Thanks for the response. The guy that installed my system has another probe. I’m going to install that. If that doesn’t work I will follow your advise.
 
You must log in or register to reply here.
Back
Top