Exhaust gas temperature / Thermocouple amplifier question....

[Doughnutter]

Hi,

I'm about to fit an aftermarket engine management system on my car and data logging system. We want to log the EGT's and the data logger is capable of doing this, however we need to source 6 amplifiers. The company supplies them for around £75 each amplifier ( times 6 needed) and someone told me that there would be a way to build these ourselfs much cheaper if we could find a plan to do this. I've had a look for several hours the net but can't find much that I understand. Is there anyone that may be able to help me please?

Cheers

Tom

 

[tcmtech]

I am assuming you are in need of a millivolts to volts type of DC amplifier so that the data logger has a more reasonable input voltage range to work with?

Odds are we have a few dozen people and designs for such a device. Good chance you may be able to build them all for under $5 - $10 as well!

 

[Doughnutter]

Hi tcmtech and thanks for taking the time to reply!

Thats exactly the sort of thing that I am looking for!

Here is a screen shot of the specifications of the sensor amplifier that they suggest. The range I am looking to record is -100 C to 1150 C although for our application it will be more in the 300 C to 1100C under normal and anti lag conditions.

**broken link removed**

 

[Boncuk]

Hi Doughnutter,

there is a cheap and precise way to amplify low voltage using an INA111 (Burr Brown) instrumentation amplifier. Gain is programmable from 1 to 10,000 via one resistor.

The INA111 is manufactured with trimmed feedback resistors in DIP-8 and SO-16 packages.

Here is the datasheet.

Boncuk

 

[MikeMl]

Hi,

I'm about to fit an aftermarket engine management system on my car and data logging system. We want to log the EGT's and the data logger is capable of doing this, however we need to source 6 amplifiers. The company supplies them for around £75 each amplifier ( times 6 needed) and someone told me that there would be a way to build these ourselfs much cheaper if we could find a plan to do this. I've had a look for several hours the net but can't find much that I understand. Is there anyone that may be able to help me please?

Cheers

Tom

I have **broken link removed** in my Cessna. It displays 6EGTs, 6CHTs, plus some other parameters, and logs it all to a csv text file readable into EXCEL. Probably a bit pricey for your application, though.

It uses ungrounded Type-K probes, with a scanner/multiplexor, all controlled by a PIC ucontroller. All probes go to a single internal reference-junction compensation circuit, which compensates for ambient temperature variation.

 

[BrownOut]

Post being edited.

 

[Doughnutter]

Hi Boncuk,

Thanks for taking the time to reply and put the download up! I'll do some reading up about that today.


Hi MikeMI,

Interesting product!

Its good to see what the aero guys are using! Your always a fair few steps ahead of us that are still on the ground, technologically speaking. I have read up on some of the history of engine management in planes, and what is new technology for us guys seems to be old news in comparison for planes!

Your right about it being a little too much for what we want. It seems impressive with its capabilities but we really only need the EGT for data logging on the dyno. We use the EGT to ensure we have correct cylinder balancing whilst doing the initial tuning. Once we have looked at the data and set the EMS (engine management system) up, we will disconect the EGT and rely on the individual cylinder knock protection to take care of any problems that may develop.

Our data logger records all the data alongside the other data such as the RPM, boost pressures, injector pulse widths and AFR's. This is very usefull for initial tuining but once we are on track the logger is mainly going to be used for monitering and recording changes in oil pressure changes during high speed drift angles and driver inputs and sector timing.

Thanks for taking the time to reply

Really enjoying this forum so far, full of useful information and clearly some very knowlegable people on here

 

[Reloadron]

Given a choice of instrumentation amplifiers to work with I would consider the Analog Devices AD595 as a good candidate. Since you are working with a thermocouple a chip designed for a thermocouple will give the best results.

MikeMl's post mentions one of the most important things to factor into things:

All probes go to a single internal reference-junction compensation circuit, which compensates for ambient temperature variation

This is known as CJC (Cold Junction Compensation) when working with thermocouples and is important for accurate results. Another nice feature of chips designed for use with thermocouples is their ability to make a non-linear thermocouple output into a linear function. Thermocouples are non-linear in nature. The chip I suggested outputs a linear 10 mV/Deg C.

I don't know your location but the chip can be had from Spark Fun Electronics at a pretty reasonable cost. I am sure other supply houses also distribute it.

Anyway, there you have a few things to consider.

Ron

 

[smanches]

I've been looking for thermal terminal blocks for CJC but can't seem to find any. Anyone have ideas where? They must be called something else.

 

[Reloadron]

I've been looking for thermal terminal blocks for CJC but can't seem to find any. Anyone have ideas where? They must be called something else.

This is a good read on the subject and I think what you are after is called an Isothermal Block.

<EDIT> Another very good read on the subject. This includes the 1NA126 and mentions the lack of CJC. </EDIT>

Ron

 

[Doughnutter]

Hi Ron,

Thanks for the advice! The temperature being referenced against atmospheric conditions is a good point!

I think i'll go for the suggested chips to use.

I need to build a box that is capable of amplifing 6 signals. I have very little understanding of general electronics. I assume I need 6 of the chips and then a board to mount them on and some additional components.

Could someone please give me some simple to follow advice on how to go about this. I have a friend who is willing to teach me to solder circuit boards (he worked on an assembly line doing this briefly). A circuit diagram would be great to work from.

Cheers
Tom

 

[Reloadron]

HI Tom

I think what is needed is somewhat of a plan starting in the beginning. You have a few options but before we can really move on any options we need a start.

Early on in the thread MikeMI mentioned for example the use of ungrounded type K thermocouples as he uses on his aircraft. I think we will all agree for the temperature ranges in question a type K thermocouple is the best choice. So let's start with the thermocouples. The thermocouple is no more than a bead junction of two dissimilar alloys. That junction for your application should be "sheathed" in a material that won't get eaten alive by exhaust gasses. I would suggest an alloy like stainless steel or inconel. This will make for the TC probe.

Overall the probe diameter and the actual gauge of the thermocouple wire are important as to response times. Very important is getting the probe tip (actual TC junction) directly into the hot gas flow. This is typically done by drilling and tapping a point or points in the exhaust manifold for probe placement and the probes inserted using compression fittings.

You mentioned a data acquisition device? What sort of (in detail) data acquisition device do you have? What sort of input signals does it want to see? That will determine how the engineering units like temperature are scaled for the device. The same is true of things like pressure etc.

For openers I suggest we get the thermocouples (you mention 6 channels) in order, then worry about chips etc. Got any thoughts on location and placement?

Ron

 

[Doughnutter]

Hi Ron,

Thanks for the response! Heres some more infomation on our setup.

The system we are using to log the data is a Race Technology DL1 data logger, heres a link to it...

Products » Data loggers » DL1 - Racing - Race Technology Ltd - Automotive Technical Excellence

We intend to log as high a rate as possible. The system is 100htz maximum so realistically 50 htz with hysterisis (If i'm correct).

If I am correct in my undertanding it requires an analouge signal of between 0v and 5v. The thermocouples that they supply are in the following link and the ones that we are using ...

Products » Sensors » Temperature sensors » Thermocouples - Racing - Race Technology Ltd - Automotive Technical Excellence

The suggested amplifers from the company are these....

Products » Sensors » Temperature sensors » Thermocouple amplifier - Racing - Race Technology Ltd - Automotive Technical Excellence

As you can see from the link they want 75 GBP each one. We need 6!

We also considered these thermocouples..

We are mounting the thermocouples on little bosses on the custom cast manifolds. There is one for each port.

**broken link removed**
**broken link removed**

The data logger is also tied into an ECU via CAN bus. The ECU is an Emerald K3 link to their site with more details ...

Welcome to the Emerald M3D homepage


We expect to see temperatures close to 1200 C with anti lag use.


Hope this helps

 

[Reloadron]

That was some interesting reading. One hell of an interesting data logger in a small package. I read how there are a few versions of the DL-1.

Some of it left me a little confused. Sometimes they seem to refer to the 8 analog input channels as 12 bit (4096 quantization levels) and other references were to 16 bit. Still wondering about that. They also cover how of the 8 analog inputs that one is reserved for battery voltage leaving seven channels. Of the seven channels three are 0 to 12 volts and four are 0 to 5 volts. Then there are two frequency inputs for tachometer signals and four frequency for wheel speed.

Then under sensors I found this:

1.2 External Sensors

The DL1 allows you to attach several external sensors for logging. These can be frequency or analogue inputs. There are 4 frequency inputs, generally used for shaft or wheel speeds, and 8 analogue channels (12v input). The DL1 is supplied with connectors for the external inputs. For each input, the positive wire connects to the input, and the ground to one of the ground connectors. If the sensor needs a 5v reference supply, one of the pins on the analogue connector side supplies this. You can have more than one sensor attached to the ground and reference output.

They refer to all 8 analog channels as being 0 to 12 volts and make no mention of the first being reserved for battery? Somewhere in all of this I am missing something.

I looked at the temperature transmitters or as they call them temperature amplifiers. They list two flavors which are basically low temperatures -100 to 210 degrees C for things like coolant and oil and then -150 to 1150 degrees C for high temperatures like the EGT. Both output 0 to 5 volts full scale. The sensitivities for the low and high range units differ as expected at 3.6 mV/deg for the high and 12.3 mV/deg for the low range. Then they amplify for a linear (close to linear) span of 0 to 5 volts out for the given input range. That is how I read it. Looking at the actual thermocouples I didn't see compression fittings to mount them to the exhaust manifolds? Again, I am missing something.

If you run six EGT temperatures will you have enough analog channels for other sensors?

I'll post more tomorrow.

Ron

 

[Doughnutter]

Hi Ron,

Thanks for taking the time to reply again and read through all the information.

We only intend to use the EGT monitoring for initial cylinder set up and balancing. Once we have done the mapping and sorted individual cyl fuel and spark trims we will remove the EGT from the car. We have considered mounting just one sensor either side in the collector before the turbo for regular use, but this is unlikely to happen.

We want to monitor other data channels in normal running. Oil pressure in relation to lateral G is one we'd like to look at to see if there are any adverse effects to oil pressures when cornering at high speeds. As we do a lot of drifting (the car has been built for the British Championship) we would like to look at our wheel speeds in comparison to the GPS speed trace. Then the usual RPM, MAP, AFR and steering input will be monitored too along with coolant temperature and oil pressure.

Whist setting the car up we may also look at data to include the exhuast gas back pressures, intercooler before and after temeratures.

We are aware that we are limited to 8 channels and the four digital channels for the wheel speed and this did concern us to start with that we may not have enough inputs to cover every situation. Unfortunately our budget only stretched to the DL1 and not the DL2. This is why we intend to log the EGT's, exhaust back pressures etc for initial balancing and then the more 'regular' vital signs in regualr use with the capability to swap as nessasary.

Cheers
Tom

 

[Reloadron]

Hi Tom

Thank you for the explanation as I was sure confused as to channel assignments. I have always found today's automotive systems to be really interesting as compared to what I grew up with.

You guys have several options including building your own temperature transmitters. There are plenty of chips out there like the one I mentioned (the AD595) and another one I was looking at today. This chip is an option also. I was trying to come up with something simple and requiring as few external parts as possible. The downside to most of these approaches is the best way to do it would be to use 6 chips for 6 channels and make a printed circuit board. Board making and layout isn't my forte.

I also looked at another option. What the company that makes your data acquisition unit is selling is pretty high priced for thermocouples and thermocouple transmitters (what they call the thermocouple amplifier). I looked at some reasonable priced type K thermocouple transmitters like this one distributed by Cole Parmer and should be available in Europe at about 36 GBP. I use advanced versions of these that are fully programmable loaded with features you don't need. They output 4 to 20 mA as shown but a simple 500 Ohm resistor on the output provides 2 to 10 Volts. They are easily scaled for your application and data logger.

Tomorrow at work I'll look deeper into things.

Ron