My home has a natural gas gravity floor furnace from about 1930 which had been retrofitted at some point with a standing pilot. Two years ago, I replaced the standing pilot with a cycling pilot control, consisting of: White-Rodgers 36C84-426 Gas Valve, 3098-134 Mercury Flame Sensor, 5059-23 Pilot Relight/Safety Timer Control, 760-56 Electrode Assembly and F115-0064 Harness Assembly. The Pilot Burner is a Honeywell Q303.
When the thermostat calls for heat, the pilot lights, the spark ceases, and about 45 seconds later the flame sensor energizes the main gas valve. All correct, so far.
However, when the burner is cold, the main valve cycles open 1 second/closed
2 seconds perhaps 7-10 times before it stays open. When the burner is hot, it stays open without cycling, or at least cycles fewer times.It appears to me that when the flame sensor has been heated just enough to open the main valve, the inrush of cool air as the burner ignites cools it back below its shutoff point. I suspect that while the burner cycles a few times, the flame sensor gets heated further by the pilot so that its temperature no longer dips far enough to shut the valve.
The pilot must be set high enough so that the flame engulfs the sensor bulb, but not so high that the flame separates from the pilot orifice or sputters. Within that range, I can't find a setting that eliminates the main burner cycling problem. The manifold pressure is set to 3.5" WC, which is what it was with the previous gas valve. There are no specs on the furnace, so I maintained the existing pressure and damper settings. The main burner has no floating, lifting or flashback.
I have had a tech from the gas company inspect the furnace, and also an HVAC professional. Neither was able to correct the problem. They both did verify that the installation was correct and that the furnace was functioning properly (aside from the cycling problem) and vented safely.
Any suggestions would be welcome. (Replacing the furnace is not an option right now.)