EGT responsiveness

All sensor specific discussions! Temperature, pressure, steering angle, brake and throttle, etc. Post adaptations of OEM-style sensors, and also your clever DIY hacks and custom designs here too!

Moderators: JeffC, rdoherty, stieg

Post Reply
torkey
Posts: 30
Joined: Tue Sep 22, 2015 3:09 am

EGT responsiveness

Post by torkey »

My EGT setup has a couple problems. First the temps it is reading is much lower 300-400 degrees than what my stand alone gauge shows. And the response seem very sluggish compared to the stand alone gauge. The temps climb much slower. 10-15 second delay which could be deadly.
Any idea what could be going on here?

The amplifier I am using is:
http://www.adafruit.com/product/1778

The sensor is a K-type thermocouple.

stieg
Posts: 100
Joined: Fri Dec 20, 2013 1:37 am
Location: Madison, WI
Contact:

Post by stieg »

Hi Torkey,

Sorry to hear you are having issues. Unfortunately I don't know enough about how you have this setup with your Race Capture to provide useful feedback or ideas. Can you elaborate more on the setup so we can understand it better? TIA.
Andrew Stiegmann (Stieg)
Principal Engineer
Autosport Labs Inc.

torkey
Posts: 30
Joined: Tue Sep 22, 2015 3:09 am

Post by torkey »

Hi Stieg,

I'm running a RCP MK1.
The signal from the thermocouple amp is going into an analog channel that is sampling at 5Hz.
It is calibrated using these values:
0 0
350 1.7494
680 3.4625
790 4.0226
900 4.5687
Anything else you need to know?

stieg
Posts: 100
Joined: Fri Dec 20, 2013 1:37 am
Location: Madison, WI
Contact:

Post by stieg »

Hi Torkey,

A couple of questions:
And the response seem very sluggish compared to the stand alone gauge. The temps climb much slower. 10-15 second delay which could be deadly.
How are you reading these values. Do you have an RCP in car and are reading the values on a tablet that is connected to the RCP unit via Bluetooth? Or are you using some other method?
it is calibrated using these values:
0 0
350 1.7494
680 3.4625
790 4.0226
900 4.5687
How did you figure out these calibration values exactly? Did you read the gauge and measure voltage on the line? Did you use a sender spec sheet?

At what temperatures is it reading low? Is it a temperature below 680? If so have you tried removing the 0:0 measurement?
First the temps it is reading is much lower 300-400 degrees than what my stand alone gauge shows.
So this would imply that you have a EGT gauge and sender setup and you are splitting the EGT sender signal wire between the RCP unit and the gauge itself. Is this accurate?
Andrew Stiegmann (Stieg)
Principal Engineer
Autosport Labs Inc.

torkey
Posts: 30
Joined: Tue Sep 22, 2015 3:09 am

Post by torkey »

I do have a tablet in the car connected to RCP over Bluetooth.

The calibration values are based on the information provided with the thermocouple amp. (see attached image)

There are redundant EGT probes at this point. One is connected to the RCP and the other is connected to a GlowShift EGT gauge. I don't have a way to validate the readings of either one unfortunately. The GlowShift seems more believable at this point.

I've tried a couple of different thermocouples thinking that maybe I got a bad one but the results are the same. I don't think the problem is with RCP. But I have no idea how to get this to work. I know I can go in and fiddle with the calibration until the numbers are close but it's more the responsiveness that concerns me. Since this is a diesel, EGTs are a critical tuning metric and getting them wrong will kill my turbo and valves.

Are all K-type thermocouples the same? Do they make a fast response type? It's not just the delay in the tempos climbing but there is a significant delay in the temps falling too. I'm pretty sure it's the thermocouple but I've tried 2 and have had the same results.
Attachments
Screen Shot 2015-11-19 at 8.04.03 AM.png
Screen Shot 2015-11-19 at 8.04.03 AM.png (100.9 KiB) Viewed 2884 times

stieg
Posts: 100
Joined: Fri Dec 20, 2013 1:37 am
Location: Madison, WI
Contact:

Post by stieg »

Are all K-type thermocouples the same?
No. They can have different resistance values and different scales of response. But they should be similar in terms of response properties (ie they all respond linearly in the case of an ideal K type sender). How they respond is what really matters for accurate readings.
Do they make a fast response type?
Should be fast enough (read within a second or so... but that is what I have seen in practice, YMMV).
I'm pretty sure it's the thermocouple but I've tried 2 and have had the same results.
A few things that come to mind:
  • Your 0 0 mapping you previously described is wrong. It should be "0:1.25" since the output formula on the amp is T * .005 + 1.25. So if T is 0, then Vo is 1.25.
  • Check that you aren't confusing Fahrenheit and Celsius. A quick search on Google shows that the GS EGT gauge reads in units of Fahrenheit (http://www.glowshiftdirect.com/Tinted-7 ... Gauge.aspx) whereas the amp outputs its values in Celsius (previous picture). Be sure you are comparing apples to apples :)
  • Check that the wires leading to the temp probe are new and in good shape along with the probe itself. As wires and probes get used and age, their wire characteristics change, which can cause low reading values (although not nearly as bad as what you describe).
Andrew Stiegmann (Stieg)
Principal Engineer
Autosport Labs Inc.

toga94m
Posts: 127
Joined: Wed Jun 05, 2013 1:57 am
Location: Upstate NY

Post by toga94m »

torkey wrote:There are redundant EGT probes at this point. One is connected to the RCP and the other is connected to a GlowShift EGT gauge. I don't have a way to validate the readings of either one unfortunately.

Ideally you'd be able to put a voltmeter on the output of the blue board and see how quickly the voltage changes. That'd let you pinpoint sensor or RCP as the issue. The AD8495 chip used on that board is plenty fast - will update 20k times a second (20kHz response).
torkey wrote:I've tried a couple of different thermocouples thinking that maybe I got a bad one but the results are the same. I don't think the problem is with RCP. But I have no idea how to get this to work. I know I can go in and fiddle with the calibration until the numbers are close but it's more the responsiveness that concerns me. Since this is a diesel, EGTs are a critical tuning metric and getting them wrong will kill my turbo and valves.

Are all K-type thermocouples the same? Do they make a fast response type?
Mechanically, if the TC is a small exposed sensor it'll react quicker. But EGT and harsh-environment TC's in a sheath have more mass, so they'll react a little slower. EGT shouldn't be as bad because there's a large amount of heat (not temperature) available to quickly change the temperature at the sensor.

I also disagree with the scaling chart you provided. This board has an offset so it can measure negative temperatures. The formulas are right on the back of the board : Vout = (Temp * 5mV) + 1.25V or Temp = (Vout-1.25) / 5mV so a simple 2-point curve (without correcting for TC nonlinearity) would be
0C at 1.250V
600C at 4.250V
750C at 5.000V

What power supply are you feeding the blue board? It'd do best with 12V battery voltage, not the 5VREF supply from the RCP connector. Otherwise it'll fall on its face as output gets near the supply. It's only specified to 0.1V below supply so the highest valid output would be 4.900V = 730C. On the other hand, RCP won't take a voltage higher than 5V..........

I did a post somewhere else here showing how to compensate for TC nonlinearity - I'll find that and come back & edit the post with a link.

Tom
------------
Learning Race Capture Pro... on someone else's car
Learning Python/Kivy on my own PC

toga94m
Posts: 127
Joined: Wed Jun 05, 2013 1:57 am
Location: Upstate NY

Post by toga94m »

Dangit.. Steig posted 9 minutes before I did, while I was still typing. :D Same info though.

Here's that older EGT thread. http://www.autosportlabs.org/viewtopic.php?t=4518 This was for the same board, but I totally spaced on the offset feature at that time. So all my numbers need to have 1.25V added to them.

On the other hand, if you're unconcerned about EGT readings below freezing, you could just bridge across the 1.25V offset (the three-legged chip) or solderblob pins 2 & 3 of the 8-pin part together. That'll tie the REF input on the chip to common, so now 0C = 0V output. And 900C would be 4.50V, well in range of the RCP hardware, and happy to run from the RCP 5VREF output.

http://www.analog.com/en/products/ampli ... d8495.html

tom
------------
Learning Race Capture Pro... on someone else's car
Learning Python/Kivy on my own PC

torkey
Posts: 30
Joined: Tue Sep 22, 2015 3:09 am

Post by torkey »

This is really good feedback. Talk you to everyone that has responded.
The amplifier is getting 12VDC for power.
Right now the engine is torn apart. Might be a couple weeks before I have everything back together to look at this again.

torkey
Posts: 30
Joined: Tue Sep 22, 2015 3:09 am

UPDATE

Post by torkey »

I gave up on trying to come ump with my own EGT sensor. Instead I went with this.
http://www.exhaustgas.com/ProductDetail ... 1056110.58
So easy. Worth the cost easily. I have a solid and reliable EGT sensor now.

Post Reply