Thermostat call for heat. Ignitor fire and almost at the same moment the pilot flame lights. After a second or two it goes out. Where should I start? Thanks

Clean the flame sensor with sandpaper or steel wool is the first and cheapest thing to do.

I can do that; as soon as I find it.

Even if the flame sensor is a problem, wouldn't there have to be flame in the tubes for the flame sensor to even be in the 'loop'?

The ignitor would have glowed the furnace would lite for a very short time or not at all then cut off since the flame sensor would not be sensing the flame. Just clean the flame sensor and try it. There are many different looking flame sensors and if you have toi get close to the ignitor becareful since they are very sensitive and break easily.

Additional info: the pilot flame starting and going out happens in conjunction with a 'clunk' from the gas valve. I'm thinking the problem is related to 'flame rectification' ~ assuming this unit has FR. Reason I'm a bit skeptical about FR is that I don't see any wire in the vicinity of the pilot flame that would carry an electrical signal anywhere (circuit board).

Flame recification is not the only flame proving technique. Here is some info on an IR sensor that Trane used: Trane Furnace whoas - http://www.hvacmechanic.com/ Forums (http://www.hvacmechanic.com/forums/resservice/messages/13028.htm)

FWIW, In flame rectification 24 VAC flows from an insulated wire to the flame to ground. The resultant pulsating DC current is used to "prove" the flame. No flame. No current.

My igniter looks a little like the picture on the left. Except that the part on the left is attached via a gas line to the gas valve. The top left part is a diffuser which aims the diffused gas stream at the igniter.

The right side is an inverted spark plug with a cap or target about a quarter inch or so from the electrode. The spark jumps the gap to the target igniting the gas.

I'm not finding a separate flame detector. The igniter is not the two or three legged glowing grid type.

I cleaned the electrode; no effect good or bad.

The flame tube assembly is oriented L /R with the igniter / pilot light assembly in front.

Can you take a picture and post using go advanced/manage attachements.

Post exact model number for your Trane unit.

Here are two pictures (I hope): Honeywell control module and Ignitor / Pilot light flame tube.

Model # TH5 100A 94883 S/N F41320310

Two other numbers on the label that I initially thought were Model and S/N: HW820411D and 207743 C

Does any of this indicate the age of the unit?

OK, not having direct experience, the floowing makes sense:

Note the rod has a nice ceramic insulator and a heavy insulated input. The heavy insulation is reminiscent of a spark plug.

Note the metal not to far away.

HINT: It's not a hot surface ignitor, but rather a spark ignitor. Spark jumps from insulated wire like thing to ground not to far away.

There is no reason, except for being a bit odd, that the spark ignitor could not be used for the flame rectification. So, it's like it acts as double duty.

There is like two paths: Flame to ignitor/flame sensor and ignito/flame sensor to ground.

And everything looks dirty.

When the pilot is lit the rod is in the flame.

When it's igniting, the rod is arcing to the metal above.

Proving the flame and igniting don't have to happen at the same time.

Best guess. No wonder why your confused.

See page 7 for operating system of valve ignitor

See page 13 under service for testing components with meter.

http://customer.honeywell.com/techlit/pdf/PackedLit/68-0047.pdf

See this link for S8600H controler

Honeywell post date 12/1989

http://customer.honeywell.com/techlit/pdf/68-0000s/68-0085.pdf

My past experience with this system control is as follows.

Bad ground (could be under screw where the S8600H bolts/screws to the furnace.

Low readings at test points of S8600H

Follow service procedure to test.

In most cases it is the S8600H that fails(Note this control is available with many different ending letters so read the above info links to see which ones you can use to replace if necessary)

This link below for info on newest model of controlers

http://customer.honeywell.com/techlit/pdf/PackedLit/69-1954.pdf

Honeywell post date 6/2006

Here is a video: http://www.forwardthinking.honeywell.com/related_links/combustion/s8610u3009/train/intermittent_pilot.wmv

THAT is what I needed. hvac1000 and KISS; Thank you very much.