Thermocouples clamped on gas tubing for reasonable-cost, continuous temperature monitoring of stage outlet temps?

[ACHiPo]

If you have any specific suggestions at hand for thermal grease or epoxy available easily, affordably, and in small quantities, I'm all ears. A quick search for specific use compounds did not yield any stuff rated for my 500f minimum target. I needed the Copper anti-seize anyway, it is rated to well over 1000f, and with the copper, should be reasonably thermally conductive since it is relatively quite electrically conductive. I only need a dab at the thermocouple junction.

The thermocouples could easily find their way into an RPi / Arduino /ESP32 type control system integration eventually or even just a BT/Wifi server for remote monitoring. For now a $32 4-channel display seems pretty attractive to get me started.

No immediate suggestions—500F is pretty hot for epoxies. Your copper anti seize should work fine.

 

[Tracy]

IR cameras are great for comparing temps, not so great for absolute temp measurement—they’re sensitive to the emissivity of what they’re measuring.

True, but for diagnostic work, absolute temp measurement isn't needed.

 

[broncobowsher]

If you have an insulating cover, no need for thermally conductive goop coupling it to the tubing. Might be a degree or two off. There are bigger errors in using the basic thermocouple convertors for an Arduino. I fought that with Arduinos a few times. The dedicated thermocouple reader is a lot cleaner data.

 

[tmassey]

I wired up K type thermocouples on my compressor. I used not-adhesive thermal heat sink pad material to electrically insulate the thermocouple from the tubing: a single layer above and a single layer below. I then used a simple wire tie or two to hold the thermocouple tip up against the stainless steel tubing coming out of the head.

I actually use five of them: one at the outlet of each head, one in front of the fan/flywheel on the front of the compressor, and one immediately behind the block of the compressor, which seems like it would be the warmest place at the back, shielded from the airflow.

I then run them into a thermocouple interface, connected to a raspberry pi, which monitors them every second, and plots various graphs of temperature, which I see from a webpage on my iPad or whatever.

Given a lot of the factors you’ve already delved into, I have no idea how accurate the temperatures are. Like someone else said, the temperatures I’m seeing at that spot of the tubing are higher than I see on the top of the head. I routinely see temperatures of 145°C, which is beyond the specifications for my compressor, but the head itself is not that hot, which is where the specifications say to measure the temperature.

In the end, though, I don’t need absolute temperatures. Relative temperatures are fine. What I need are consistent temperatures. And I get that. Outside temperatures don’t really seem to make that much of a variation on what temperature the compressor is going to get to. Of course, I’m careful not to run it on hot days…. But even running it in 7°C air temperature only seems to knock a few degrees off the third stage temperature at best. Like easily less than five, or something like two or 3% difference, despite a 20°C change in outside temperature. (My compressor is in an unheated and uninsulated workshop.)

I’m surprised how relatively consistent the temperature is. You can just about determine the pressure the compressor is running at based on the temperature.

In the end, I’m not certain how valuable the information is. It really never changes, and I’ve never found a reason to change my actions based on the data from the temperature collection. But I felt very uncomfortable not being able to have any idea how my compressor is running outside of watching a pressure gauge and maybe listening to the sound. At least this gives me another way to peer into the internals and see if things are working the way I expect them to. From that perspective, it’s probably worth the effort. But if you’re not as tightly wound as I am – or as obviously tightly wound as the people participating in this thread are – I’m not sure it’s worth the effort.

 

[iain/hsm]

Instant information about exactly what you want.
View attachment 772124

Except when you're British and looking at that print of yours of a Masterline and thinking a single stage booster running at 98.4 Centigrade? that's a tad hot ...........Then reading the joined up writing again in F LOL Iain

 

[iain/hsm]

I have just been using a laser thermometer....

We do that also as a quick comparison against the fitted thermocouple RTD's but the pipework or area that you hit with the laser has to be spray painted matt black or the reading is way too low.
In addition shrouding it away from the cooling effect of the fan blast. Iain

 

[iain/hsm]

So that says to me that head temp is not really what I am looking for. Unfortunately, looking at pressure rated 316 sensors that might fit the (2nd and 3rd) stage head plugs or NPT for inline at the head exit-port in addition to the re-plumbing to accomplish it, the cost/hassle for my limited use seems impractical.

While this has been somewhat RIX-specific ...

We just need to me more specific on this IMHO.
In order for us to all sing from the same hymn sheet. If that is indeed remotely possible.
Head temperature is not quite the same as gas temperature, Not by a long shot.
And gas temperature calculated is the critical component and not gas temperature measured.
And the gas temperature measured can vary wildly depending on a number of factor including position
gas velocity, cooling effect at the measured position and a little know fly in the ointment called Von Karman sheet or vortex sheading.

Not withstanding having to manufacture your own thermowell rating it to 5000psi in oxygen under ASME with a 4:1safety WP to BP and a 20,000psi rated burst pressure under cyclic loads and high temperature
and cost less than a family of four would pay eating out on a big mac and fries.

That and ensuring the space taken up by poking a metal finger into the gas stream causing disruption to the flow and these vortex oscillations causing drag vibration and a capacity to squeal like a spiked pig at certain pressures and flows (when you get it wrong.) Iain

 

[iain/hsm]

A kick off point would be the calculated gas temperatures of the SA-6 from 0 to 5000psi
Calculated with no cooling effect, no fans, no heat sink just adiabatic gas loading using Charles Boyles Van Der Waals Henry's etc from a specific ambient air temperature loading
And all in degrees F to please our imperial masters.

SA-6 3 stage air compressor
Medium: Atmospheric Air
Ambient pressure and temperature
Inlet 70F
1st Stage 493.5 F at (100.11 psi G)
2nd Stage 447.95 F at (628.22 psi G)
3rd Stage 538.1 F at (5000 psi G)
Approach temperature each stage calculated at Ambient plus 25F

Perhaps the addition of a good 8 blade zero degree pitch angle cooling fan might be useful after all. Iain

Also note that these calculations are the same or indeed similar to any multi stage air compressor including your Bauers, Coltri's etc
Temperature is a consequence of P1 V1 over T1 over P2 V2 over T2 not a brand name
Cooling is the key and an oil free compressor is a more effective cooling medium

Compare cooling times when left idle of say a pan of hot oil at 212F by comparison to the same volume of water starting out at 212F and comparing both to the same size pan of hot air and considering why two of them cool faster than the bucket full of oil and one even faster than the bucket full of water.

 

[-JD-]

and cost less than a family of four would pay eating out on a big mac and fries.

Well, given current inflationary conditions, and allowing for world perception of US eating habits, that pretty much implies US DOD level spending capabilities. So no problem, right?

But I have a compressor to feed so the family will have to starve for Micky-D's.

A kick off point would be the calculated gas temperatures of the SA-6 from 0 to 5000psi
Calculated with no cooling effect, no fans, no heat sink just adiabatic gas loading using Charles Boyles Van Der Waals Henry's etc from a specific ambient air temperature loading
And all in degrees F to please our imperial masters.

SA-6 3 stage air compressor
Medium: Atmospheric Air
Ambient pressure and temperature
Inlet 70F
1st Stage 493.5 F at (100.11 psi G)
2nd Stage 447.95 F at (628.22 psi G)
3rd Stage 538.1 F at (5000 psi G)
Approach temperature each stage calculated at Ambient plus 25F

Perhaps the addition of a good 8 blade zero degree pitch angle cooling fan might be useful after all. Iain

Yea, that's pretty warm!

You state "Approach temperature each stage calculated at Ambient plus 25F" Should we expect interstage cooling through the coils (and seperator) be on the order of 350-400F !????? Or am I completely misunderstanding approach temperature?



Fan: check! (I'm assuming from myriad other threads that this is the "good" fan.)

So let's take a step back from trying to hit perfection of dead-nuts accurate and go with something targeted at (somewhat) cheap bastards, with decades-old pumps, that would like them to keep running for decades more. You know close enough for "grenades" (or do folks play horseshoes over there too?).


You've already indicated that you use non-contact, IR surface temp readings from the (blackened) SS tubing for cross checking against thermowell RTDs. Given that, I have some targeted questions that I feel would benefit from your experience with monitoring off of the outlet tubing.

1. Would you have reason to expect a functionally-significant difference between non-contact IR measurement and a reasonable well-mounted thermocouple (i.e. using a thermally conductive compound like copper anti-seize and firm compressive and externally-insulating mounting like silicone tape and clamp)?
2. What distance (tubing path) from the head (outlet) connection should the thermocouple mounting point be for each stage for optimal monitoring, balancing the heat-sinking of the head vs. the radiation and convective losses from the tubing?
3. What would measured do-no-exceed temps be?
4. What would nominal target temperature ranges be (assuming proper cooling - fan, shroud, etc.) at 3000, 4000, and 5000 psi be, when pumping air, for a "healthy" unit? (5000 psi is aspirational for banks someday, but since I dive 3500psi tanks, aproaching 4K hot is reasonable)

I guess I'm seeking more specificity than poetry as I make no claim to be versed in compressor engineering.

With all gratitude and candor @iain/hsm , your insights into these brilliant units is highly valued.

 

[Jai Bar]

Well, given current inflationary conditions, and allowing for world perception of US eating habits, that pretty much implies US DOD level spending capabilities. So no problem, right?

But I have a compressor to feed so the family will have to starve for Micky-D's.



Yea, that's pretty warm!

You state "Approach temperature each stage calculated at Ambient plus 25F" Should we expect interstage cooling through the coils (and seperator) be on the order of 350-400F !????? Or am I completely misunderstanding approach temperature?



Fan: check! (I'm assuming from myriad other threads that this is the "good" fan.)
View attachment 832952




So let's take a step back from trying to hit perfection of dead-nuts accurate and go with something targeted at (somewhat) cheap bastards, with decades-old pumps, that would like them to keep running for decades more. You know close enough for "grenades" (or do folks play horseshoes over there too?).


You've already indicated that you use non-contact, IR surface temp readings from the (blackened) SS tubing for cross checking against thermowell RTDs. Given that, I have some targeted questions that I feel would benefit from your experience with monitoring off of the outlet tubing.

1. Would you have reason to expect a functionally-significant difference between non-contact IR measurement and a reasonable well-mounted thermocouple (i.e. using a thermally conductive compound like copper anti-seize and firm compressive and externally-insulating mounting like silicone tape and clamp)?
2. What distance (tubing path) from the head (outlet) connection should the thermocouple mounting point be for each stage for optimal monitoring, balancing the heat-sinking of the head vs. the radiation and convective losses from the tubing?
3. What would measured do-no-exceed temps be?
4. What would nominal target temperature ranges be (assuming proper cooling - fan, shroud, etc.) at 3000, 4000, and 5000 psi be, when pumping air, for a "healthy" unit? (5000 psi is aspirational for banks someday, but since I dive 3500psi tanks, aproaching 4K hot is reasonable)

I guess I'm seeking more specificity than poetry as I make no claim to be versed in compressor engineering.

With all gratitude and candor @iain/hsm , your insights into these brilliant units is highly valued.

At some point you will have to compromise accuracy for practicality. Determine to yourself what is an acceptable temperature measurement accuracy and set up your sensors accordingly.

IMHO you can use the thermal image posted earlier as a reference to see the hottest spots, or you can sample a few points with your Thermo Couple set-up (you don't need to glue them or insulate them for this, just gently press them against the hot surface maybe with a small wooden toothpick etc.). This way you will ne able to fast check several locations and find your sweet spot for placing the TC in a more permanent manner.

For testing, as suggested, you can place the TC in a cardboard cup with water and ice, then boiling water. Probably will get a reading of 0-4 Celsius and 92-96 celsius, respectively, good enough.

Many years ago (a couple of decades?) when my first daughter was a baby, she had high fever. We tried several types of thermometers but she was crying and moving so much we couldn't hold any type of thermometer long enough to fix a reading. We got one of those non contact fast reading thermometers that you place near the ear and in 1 second get a reading. It was repeating 42C, so to the hospital we went. There, the nurse measured only 39.5C but it was after some medication and bath etc. So we spent the weekend at hospital "for monitoring and doing tests". Fever never peaked more than 38C for the whole time.

Back at home I was annoyed, having about a dozen different thermometers I prepared a mixture of water about ~38C and placed all thermometers comparing with an accurate one from the lab and all were "accurate" to within 2-4 degrees... For determining whether a fever is dangerously high a difference of 2-4 degrees can be 39 or 43 it is important. For a compressor if it reaches 75 or 79 degrees, maybe not so important?

Cheers!