Themocouple in noisy environments

[JMarkWolf]

K-type thermocouples are commonly used for exhaust gas (EGT) and cylinder head (CHT) temperature monitoring in automotive and aviation applications. The sensors are commonly "grounded" at the thermocouple junction, through the screw-in plug (EGT) and spark plug ring (CHT), directly to the engine block and exhaust pipe.

Intuitively, I would think that this would introduce serious ground loops, and I would expect that the ignition and spark plug noise energy would wreak havoc on the micro-volt signal at the junction.

I have not found any discussion on this topic on the web. Can anyone comment or direct me to a source for discussion of these topics?

 

[itsmoked]

I can't comment with respect to automotive. Most TC inputs can be heavily filtered. Also they are generally used slowly so the filtering doesn't pose any speed problems.

Personally I agree, I'd almost never use grounded probes.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

I have some first hand experience here.

We designed a system to measure temperature in engine cylinders. We do it with K thermocouples and doing it involves bringing the TC into the cylinder along the spark plug. The ignition system is a 30 kV high energy system, so EMI is plenty.

Grounding the TCs is death. Symmetrical filtering and tens of kohms meeting the TC leads is what we did. With less resistance, you get very nasty common mode currents that reset your micro. Using USB was impossible first but can be used now.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[itsmoked]

I think cars are electrically about a decade behind the times. The pressure on makers to engineer out every nanocent results in old thinking and old results. I've opened about 10 engine and drive train control units and I have always been amazed by the multiple fundamental violations of modern embedded control.

My T-bird had severe automatic transmission troubles that resulted in my changing out the tranny twice. It's a nasty change as you even have to drop the fuel tank to pull it off.

After putting in the car's third $2,600 transmission I went into a web frenzy. It resulted in me purchasing a 2"x2" circuit board that I bolted in the transmission's oil pan and the spicing of a wire to the board. The transmission has now run years with no problem. Archaic design circumvented by a $50 aftermarket add-on.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[NASAPeet]

@itsmoked
I'm curious: what did your board do? And how did it save your transmission?
Peet

 

[retred]

I have a lot of experience with thermocouples in aircraft. I worked on radial engines in the forties (WII), and was an a/c instrument mech in the 50's. The thermocouple ring went directly under the sparkplug as you would normally use a copper seal. The leads went directly to an analog cyl temp gauge in the cockpit. Never a noise problem. This a case of extremely low impedences on each end. Hope this helps.

 

[itsmoked]

retred; Good point with the low impedance. Can't get much lower.

NASAPeet; The trannies all consumed themselves because they would start oscillating with a partial shift. You would feel a subtle shudder at specific loadings. The board prevents that thru either lying to the Drive Train Controller or blocks the command to the hydraulic controls inside the transmission. I never tried to figure out which.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[JMarkWolf]

Update:

As Retred points out, the low impedance apparently effectively nullifies the noise.

I had acceptable results with two different circuit types: 1. a discrete differential op-amp circuit and
2. an integrated cold-junction compensated chip (MAX6675)

Aside from a gain error induced by the "operator", both seem to work equally well without noise or significant "dithering".

 

[Skogsgurra]

False conclusion warning!

Instrumentation in the fourties were d'Arsonval movements with low impedance. Right.

The inertia of the movement laughed at microsecond transients caused by ignition. Right.

BUT:

Today's instrumentation is semiconductor based and those devices do not laugh at microsecond transients on their input. As I said before (10 Apr 09 2:32 post): Use a pair of high valued resistors and then capacitors to divert transients to ground. Forget about low impedance.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[JMarkWolf]

I neglected to mention that I have series ferrite beads and caps to ground on each thermocouple conductor.

 

[itsmoked]

Thanks for the come-back JMarkWolf.

Keith Cress
kcress -

http://www.flaminsystems.com