# 8 Channel Thermocouple Based Temperature data logger

#### AtApi

Joined Nov 14, 2007
1
Hi
First of all sorry for my English, not my native language.
Im searching for a 8 Channel Thermocouple Based Temperature Data Logger kit.
Im gonna use it to log the temperature that coming out from my dragster mufflers

Heres some specs:

The Data Logger should be battery power so i can put it on-board.
Im gonna use a Thermocople probe type K, that i already have, off course in this case i have 8 of them.
Should have a temperature acquisition range between 0F and 2000F for every channel.
Accuracy is very important.. it should be i think at least 8 bits
Sampling rate per second . this cars use nitro-methane where there s detonations and temperature shock (extremely rapid changes in temperature), i guess has to be at least 3 samples/sec.
It would be nice to have a LCD display but not necessary.
Storing the readings - obviously EEPROM is an easy option but if we can implement a SD-card would be great!
Off course if you choose EEPROM a USB or SERIAL connection to the PC is required.
Look at this Data Logger http://ifamilysoftware.com/8307.html is has pretty much same specs i need but single channel.
And has to be strong enough so G forces will not tear this unit apart.
My Type-K thermocouple assemblies (Nickel-Chromium/Nickel-Aluminum) are terminated with a two pin male miniature-connector.

Ok i think i give you the idea, Theres someone that can help me?
Thank`s for your attention and if you have some question please write me back.

Best regards

AtApi

#### beenthere

Joined Apr 20, 2004
15,819
Here is link to a reputable equipment supplier that might suit your needs - http://www.dataq.com/data-logger/features/index.html. I chose the page that talks of the stand-alone operation with data stored to a removable card.

You will want to select a better A to D conversion if accuracy is important. 8 bits only gives 256 measurements levels from 0 to full scale. 16 bits gives you 32,768 levels on the same span, which is a lot more detailed.

If I may be critical, three samples/sec will not catch those rapid temperature changes. The worthwhile sampling rate will be determined by the thermal mass of the chromel/alumel thermocouples. If they have to go directly in the exhaust stream, they will have to be fairly large to stand up to the gas erosion. That will slow their response to temperature change. Omega Corp. will have some tables that will let you figure the thermocouple latency.

Looking for detonation might be easier with an acoustic pickup. The waveform is quite different from normal combustion. A web search can probably bring up something relevant. If you need to conserve sampling space in the recording, don't enable the acoustic sampling until the N02 valve opens. But you only need 5 - 8 seconds of data for the run, I guess.

#### scubasteve_911

Joined Dec 27, 2007
1,203
Beenthere,

Are you sure about the response time problem? I thought that thermocouples that size are quite slow, maybe 500mS response time.

Do these gadgets come with linearizing circuitry? Usually people get their actual data and use a lookup table or a polynomial fit. I guess the question is, how much accuracy do you really need?

Steve

#### beenthere

Joined Apr 20, 2004
15,819
Could be valid points, scubasteve, but the OP never came back.

#### SgtWookie

Joined Jul 17, 2007
22,227
Yeah, stuff happens really quick in a dragster engine. I doubt that a thermocouple would be of much use in preventing severe engine damage due to the inherent delays involved.

Oxygen sensors coupled with detonation sensors is how I'd go. Of course, you'd have to pre-heat the 02 sensors before the run, but you'd get much more timely info throughout the RPM range, and the detonation sensors would enable one to build an automatic shutdown circuit of the injectors to cut fuel to lean/detonating cylinders.

#### Dave

Joined Nov 17, 2003
6,969
Do these gadgets come with linearizing circuitry? Usually people get their actual data and use a lookup table or a polynomial fit. I guess the question is, how much accuracy do you really need?
They do come with correction circuitry to map the EMF to temperature, although for a K-type thermocouple the EMF to temperature profile would be linear over the region of interest anyway. As for accuracy, you can only typically expect +/- 2.5C from a K-type.

More of a concern of all that is that K-type thermocouples have been known to suffer a hysteresis effect at higher temperatures where they under read when the temperature is rising and over read when the temperature is falling - this phenomenon, as far as I am aware, is not understood, although it is probably down to some sort of short range ordering. There is a chance that although the thermocouples you intend to use are temperature-rated for your application, they may not be suitable for any type of accurate temperature measurement.

You may be better advised to look at the more expensive but robust B/R/S-type thermocouples.

Dave