EGT response times

[white90]

For a K type thermocouple

6mm dia T/C 3 seconds

3mm dia T/C .8 seconds

The MIMS(mineral insulated metal sheathed) type as used in the EGT guages are fine for being bent/formed into a shape to suit getting it into the ehaust gas flow.

so if you after a rapid readout a K type 3mm dia would seem to be the prefered choice.

[BogMonster]

One thing I am still unclear on is whether a "Type K" will always have the same calibration curve i.e. you could plug any K into my display unit and get a pretty accurate reading? The one I have at the moment is fairly accurate, 99 deg for a freshly boiled kettle (before I got it all sooty!)

However like you Tony my main interest is in peak temps when flogging it up hills etc or towing, fast response time isn't the main thing I need. If mine breaks like yours did I'll probably get a 3mm but until then the one I have is OK I think.

[will_warne]

You'll probably find the calibration curves aren't the same. However, my EGT gauge came with a black box that the theromcouple pluged into (mounted mine in the back of the . The gauge was then literally just an LED display. I'm not sure how the Thermoguard ones are packaged, though.

[Turbocharger]

Tony - yours broke?? What do the specified numbers mean? ie 3secs for a 20deg change in reading with a 100deg step change in true temperature etc...

[Ex Member]

A type K thermocouple will always give the same reading. They do not need to be calibrated. For a presise measurement you need a cold junction reference, but this is not a laboratory setting.

And yes, 0.8 seconds and 3 seconds for what???

Like I have said before, 6mm is too big. You will miss the important peaks. There is no reason for one either. A bare 3mm TC will last forever in this application.

Keep in mind that the TC itself is a very tiny bonding of two metals near the tip. The 3mm and 6mm stuff of simply stainless steel tubing over top of the TC and the TC wiring.

[white90]

response times for these assemblies are governed by and vary with the enviromental conditions of particular applications.

the qouted times for the above is going from 20 degrees c air into boiling water(with insulated type2I junctions)

the figures refer to the times taken for the T/C to achieve 63.2% of this instantaneous step change.

type 2G junctions the response times are 50% of the response times mentioned.

3mm will suffice 6mm is ok to for the application I need.

the above tech is from www.tc.co.uk

they supplied a catalog of all thier products/spec etc

Edited November 17, 2005 by white90

[Turbocharger]

Cheers for the info Tony. I believe the Thermoguard unit has some internal damping too, but a faster response from the thermocouple will ensure the damping 'saturates' sooner - I think I'll be upgrading to a 3mm one soon. Has anyone got the part number for the EGR blanking plate that I'll need to be drilling?

[white90]

the peaks are all I'm really intrested in as long as there aren't huge gaps in the readings 6mm is perfectly ok

the readout already changes faster than you can easily read, a 3mm would make this worse I'm guessing.

John I made the plate from a piece of 5mm cut and shaped,

the part number from L/R is:ERR4698

Edited November 18, 2005 by white90

[BogMonster]

I got my plate with the Thermoguard from Ian

[Paul]

I agree with Tony here. Most of the EGT kits you buy have 6mm thermocouples included (as my VDO kit did), and whilst a 3mm thermocouple might be better, there isn't enough in it for me to fork out the extra cash. If I stuff it up in the future, them maybe I'll use a 3mm jobbie for replacement, in the meantime, the existing one is adequate enough.

Paul

[Ex Member]

OK, so 3 seconds to change 12.6 C with a 20C step input into water. Keep in mind that water is a much better heat transfer fluid than the exhaust gas.

So let's say you start at idle, EGT is 150C you pin the throttle for a steep climb on your modified diesel. The EGT (actual) heads for 800 C very quickly. Your 6mm thermocouple won't even reach 700 within a minute. Personally, I don't think I could find a hill long enough to have a 6mm TC settle out at max throttle.

Normally how long are you using full throttle? 20 seconds, 30 seconds. You will never see the peaks. Really a bare TC, would be best for testing purposes if you are modifying the fueling or boost.

One of these days when I'm bored, I'll run a test on 6mm and 3mm and plot the response rate for you guys. I do lots of testing that involve temperature measurement. For the response time in this application, I would be testing with a bare TC and using a 3mm for monitoring. A 6mm would be completely out of the question.

What scares me is that many people are out there playing with there injection pumps and turbos and have know real idea what is happening to there EGTs.

[BogMonster]

My observations seem to indicate that while there is a lag time it does level off fairly quickly with the t/c supplied. Peak change rate seems to be about 20 degrees per second with the 6mm, maybe a little more.

I have managed to hold absolutely full throttle up a steep hill in 4th gear for slightly over a minute on two occasions while testing, and in the last 30 seconds of that the temp only increased by about 50 degrees or so IIRC, and the last 15 seconds probably only changed by 3 or 4 degrees, peaking at 609 (standard unmodified 300Tdi engine) on one run, the second run I think got to about 585 but again hardly moving towards the end, if I had held it for another 30 seconds I doubt I would have got more than +10 degrees as a peak.

I guess what you mean though, is that a quick "blast past" somebody might have a short sharp and very high EGT which you never see?

One thing I had not thought of and I don't know how relevant it is, would the main components not also have a "thermal lag" i.e. the time the pistons/valves take to actually heat up to the EGT will also be delayed, as the thermocouple is? I guess this wouldn't really apply to the turbo impeller though, as it is so lightweight, so would change temp quickly? What I am thinking is that if a melted piston melts at 800 degrees (say) then if it takes 2 minutes for the temp of the piston's metal to get past 700 even with an EGT of 800 then the thermocouple is probably telling you the "dangerous temperature" which is the temp of the components? Just a few thoughts - interested in comments!

I am not going to push the limits too much with mine anyway - it's just to have some idea of what is going on instead of none at all which is what I had last time.

[white90]

From Ian(thermogaurd)I was discussing these details with him

his response which I appreciate as he knows his stuff:

Hi Tony,

For the record, I did some 'seat of the pants' time-constant measurements on one of my T/Cs some time ago. I used an ice/water slush (0C) and vigorously boiling water (100C) and timed the changes from 0 to 63-ish degrees and 100 to 37-ish degrees. The average was about 8 seconds. I find the 6mm: 3 sec and 3mm: 0.8 sec times a bit difficult to reconcile. Are you sure these aren't for bonded junction or exposed junction T/Cs, rather than insulated MIMS junctions?

they are from the TC catalog I mention above MIMS junction T/C

Now about response times and 6mm Vs 3mm probes: I honestly don't think the response time matters that much. I've had a multi-channel thermocouple data logger for a while now and a couple of months ago I made some records with a thermocouple in the No.1 cylinder exhaust branch and my original probe in the EGR 'pocket' The No.1 cylinder probe did react a little more quickly, both on rising and falling temperature but, strangely enough, it read lower (by about 30C) during steady load driving. [i've not yet gotten to the bottom of this; could be due to No.1 actually running cooler than No.4 (due to worn/mismatched injectors?). I've also yet to confirm the two thermocouples are reading accurately and consistently (by sticking both into boiling water), but it's unlikely that would be the problem.]

As I see it, a fast response time only matters if it is critical to pick-up short-term peaks and troughs rather than look at the 'moving average' over a relatively short time span. I'd argue that the main objective of EGT monitoring is to watch for temperatures above which damage is likely to occur to the components of the turbo-charger, specifically the turbine housing, the wastegate valve & seat and the turbine wheel (particularly the blade tips). Now, I'd propose that these metal components don't heat up and cool down anywhere near as fast as the actual gas temperature does (even the blade tips, which would react quickest). So, having the sensor in the 300Tdi EGR 'pocket', with it's inherent small lag and signal filtering action, may in fact be giving a more realistic picture of the temperature the metal surfaces are reaching?

OK, let's say we have a sensor with negligible lag (very short time constant) and it picks up a high but short duration peak temperature - are you going to drop your foot off the throttle for a moment and then plant it again, trying to respond as fast as the display does? I'd suggest this is not how things work in the real world. I look (in my mind at least) at the temperature trend as well as the instantaneous value. If it gets to 700C and is still rising, I begin to ease off the throttle. I then make a judgement of the further load to come (for example, if towing up a long hill, I look at the remaining gradient and distance to the top of the hill) and decide whether to stay in the current gear at reduced throttle or change down and do the rest of the climb at a lower speed. In another way what I'm trying to say is, the 'control loop' consists of the measurement side (T/C and indicator), the 'process' (engine and turbo components) and the 'control mechanism' (you, your right foot and the throttle/injector pump). The behaviour of the whole 'loop' depends on the response times of the whole loop. Having one part with very fast response is of little use if the remainder of the loop does not/cannot react as quickly.

One last consideration is the mechanical robustness of the system. I fully admit a 3mm T/C is still a fairly robust bit of gear but it simply cannot have the strength of a 6mm one, especially with a longer unsupported length, such as you have with a sensor extended to reach beyond the EGR 'pocket'. There could be the possibility of getting into the region of the probe's resonant frequency, which would lead to premature failure of the internal connections and possible physical breakage. I'm not trying to be alarmist and I have no data to support this (mainly because I don't have the resources to get any). But if I were doing this measurement in a important industrial application, I'd look into the resonant frequency characteristics of different length/diameter probes and try to get data on the gas velocities of the 'process', etc.

One more bit of trivia which you may find interesting. Earlier this year while examining my 'spare' 300Tdi manifold/turbo assembly (long story), I noticed that it is a split-pulse manifold design. Cylinders 1 and 4 are joined at the 'front' of the manifold, under the EGR port. Cylinders 2 and 3 are joined in a separate chamber behind the 1 & 4 chamber. And, as far as can be seen in this 'all one casting' design, the two chambers remain separate until they impinge onto the turbine wheel. (It might be possible to see this better by removing the turbine from the exhaust turbine housing but I've not had time to do this.) Therefore, the best we can get on a 300Tdi is the EGT of cylinders 1 and 4 combined.

Split-pulse designs give better low rpm turbo response, as the exhaust pulse from any one cylinder does not get as 'lost' in the large exhaust manifold chamber as with a single chamber design. I'd previously seen them only on six-cylinder turbo engines (with inherently larger volume exhaust manifolds).

[Ex Member]

Yes, well, Ian and I disagree on this and have in the past.

I told him that by the time you slow down and pull over to fill up the EGT is safely low (say 30 seconds). He says it isn't from his experience. I say his TC is too big..........

I'll have to run a test to convince you guys. Really I do run non steady state heat transfer and temperature testing all the time. A 6 mm is too slow to respond for this application.

Trust me that a 3mm is more than strong enough to last forever in this application. I run that size in much much more severe conditions ALL of the time.

Hey it is your turbos, you do what you want with them. Personally, if I was going to play with fueling or boost, I would want to know what the real peak temperature is. This means full throttle AT PEAK POWER for sustained periods. Personally, I would run a bare ended TC during testing.

[white90]

any testing etc would be intresting to see.

this thread was started to debate the pros/cons

the 6mm I have takes approx a minute to come down from say 550 to below 200 then I switch the engine off

this gives a period to allow the turbo to cool to and lessen the heat soak effect rather than starving the bearing of oil after a long run.

sustained high temps have resulted in cracked manifolds to my knowledge.

bogbusters set up is ideal as being untampered with before he starts mine had been adjusted so I have to go with the readings I have.

as for hills I have a local one starting at 200ft above sea level rising to 1200ft straight climb so max power/full throttle is achieved for a sustained period that is where my gauge hit 735 towing a trailer loaded.

wouldn't the 3mm one be rather erratic? the 6mm readout changes very quickly and holding any temp isn't easy.

Tony

Edited November 19, 2005 by white90

[BogMonster]

I'd be interested to see some test results too

I would think a faster acting TC would be better suited to an analogue temp gauge with a needle on it - with the digital one it would be a bit like a digital rev counter on a race engine - above a certain rate of change it just ends up looking like garbage. I just like the digital one because it has the max/min function which gives peace of mind that you haven't missed anything!

That was my only concern with a smaller TC was the mechanical strength and possible effects of half a TC going through a turbine doing 100,000rpm - but I have no experience of what the failure rate of these things are, just the general feeling that twice the size probably meant twice as strong

[white90]

Well in true change your mind fashion I just ordered a 3mm one to try.

I'll report back once fitted.

[Turbocharger]

A picture tells a thousand words; I think this is what we're all talking about:

What worries me is that it's likely the figures we see occasionally (735°C in my case) are actually occurring all the time, but it's only when I allow the thermocouple to reach that temperature that they're evident. In the example above, the turbo is above the 'danger' line for 12 seconds. Ok, all the numbers are complete guesswork but the principle's sound. In my oh-so humble opinion, the t/c can't react too fast, even if the display's unreadable. If it starts with a 7, back off. If you can't read the 'hundreds' column, back off quick!.

JB

All graphs copyright TeamNutter. No diesel was burned in the production of this fiction.

[Bush65]

A picture tells a thousand words; I think this is what we're all talking about:

What worries me is that it's likely the figures we see occasionally (735°C in my case) are actually occurring all the time, but it's only when I allow the thermocouple to reach that temperature that they're evident. In the example above, the turbo is above the 'danger' line for 12 seconds. Ok, all the numbers are complete guesswork but the principle's sound. In my oh-so humble opinion, the t/c can't react too fast, even if the display's unreadable. If it starts with a 7, back off. If you can't read the 'hundreds' column, back off quick!.

JB

All graphs copyright TeamNutter. No diesel was burned in the production of this fiction.

Unless i'm missing something, either the chart legend is wrong or the temps in the note (on the chart) and your comments are incorrect.

When you say, "the numbers are complete quesswork", what are you referring to? - those for the chart?

[Ex Member]

Yes, it is not real. It was just a fabricated example of what it may look like. To produce that would be very expensive if even possible. I'm not sure what the point was either than to confuse everyone.

The best reasonable test is to compare response times inserting TCs of different sizes into a stable gas stream similar to exhaust. My plan if I get time would be to use a propane torpedo heater and data log the measured TC whe putting it into the exhaust stream.

No diesel was burned in the production of this fiction.

[Turbocharger]

Err... yes. The legends are wrong. Blue is gas stream, yellow is (for example) the temperature of a tip of a turbine blade and purple is what the EGT meter shows. What I'm trying to show is that during transients parts of the turbo may be hotter than the gauge shows, and if you don't allow the system to reach steady state those readings may misrepresent what the exhaust system is seeing. I've guessed coefficients for the two components to illustrate what I was showing. I'm sorry if you found that confusing.

[Bush65]

Err... yes. The legends are wrong. Blue is gas stream, yellow is (for example) the temperature of a tip of a turbine blade and purple is what the EGT meter shows. What I'm trying to show is that during transients parts of the turbo may be hotter than the gauge shows, and if you don't allow the system to reach steady state those readings may misrepresent what the exhaust system is seeing. I've guessed coefficients for the two components to illustrate what I was showing. I'm sorry if you found that confusing.

That is more or less what I assumed, but I thought it may confuse/mislead others.

[white90]

Well I've ordered a 3mm and 6mm T/C and am having a good mate make a boss to fit it to the EGT plate so the T/C goes in straight rather than being bent to suit.

So I'll be able to compare T/C s against readouts soon, for a climb up the same step hill then allowing it to cool to measure response times etc

Edited November 28, 2005 by white90

[Bush65]

I still think Ian has made some good and valid points, which Tony posted.

I would be interested to know how much real world evidence there is for turbo failure, in operating conditions like we experience, because the driver relied on egt readings from 6mm thermocouples.

I have Ian's thermoguard systems for both my 300Tdi and Isuzu 4BD1-T engines. Unless I see better information, I would rather do as Ian suggests:

If it gets to 700C and is still rising, I begin to ease off the throttle. I then make a judgement of the further load to come (for example, if towing up a long hill, I look at the remaining gradient and distance to the top of the hill) and decide whether to stay in the current gear at reduced throttle or change down and do the rest of the climb at a lower speed.

Than risk a 3mm thermocouple failing and taking out the turbo when I was in a remote area.

[white90]

the 3mm-6mm are both rated to 1100degrees Celcius so I expect them to be as strong enough

the one Ian supplies doesn't get into the gas stream though and sits in the EGT pocket, try removing it and have a look/feel inside the manifold, it is approx 55mm from the EGT plate to get into the flow from number 1-4 cylinders.

the temp difference on min when I modded the T/C was approx 50degrees C

so when I saw a 735 max I would have only seen high 600's and not eased up.

Turbo failure is lower on the expected failure list than a cracked exhaust manifold which has occurred twice pre EGT gauge on a friends 300TDI, same Turbo reused with no damage.

Edited November 28, 2005 by white90