trader_4 wrote, on Wed, 31 Dec 2014 05:29:47 -0800:
I am clueless.
Maybe cracked ceramic that only opened up under red
hot heating of the metal rod?
Even then, what's the failure mode?
It would just be easier for the electrons to get to
the chassis (which is what they're trying to do
through the flame).
I really do not understand the failure mode, especially
as this thing is clean as can be.
I wish I understood HOW it failed.
1. The flame rod consists of a center electrode and an outer metallic sheath.
2. The electrode is isolated from the outer sheath by an insulating material.
3. The end of the flame rod, in the flame path, has a special high temperature
hermetic seal with a threaded adaptor which is connected to the center
4. A high temperature sensing rod is attached to the threaded adaptor.
Is the "outer sheath" the ceramic?
Or, is the outer sheath the stainless steel rod exterior?
As it says, the center electrode is isolated from the outer metallic
sheath by the ceramic. IOW, it's a center rod in the center of a
ceramic insulator all surrounded by the outer case that you see. It's
not a "rod" at all but a composite assembly that looks solid from the
On Wednesday, December 31, 2014 8:14:03 PM UTC-5, dpb wrote:
IDK what flame sensor you have there, but from the dimensions it looks
like it's a foot or more long, 1/2" in diameter and certainly not for a typical
home furnace. The one Danny has pics of does look like a small rod,
mounted on an insulator, connected to a wire. The descriptions I've
read about rod to flame conductivity also don't say anything special
about the reqts for the rod. They just say that a metal rod inserted
into a flame will conduct a current when voltage is applied. That would
seem to coincide with what he has, ie a metal rod on an insulator mount.
Those are commercial for larger boilers, etc., but the physics is the
same. There must be an isolated section for the current flow; otherwise
it would also be grounded. The _precise_ configuration is immaterial...
On Thursday, January 1, 2015 9:25:28 AM UTC-5, dpb wrote:
I agree. The isolated section is the metal rod with a wire attached.
It's housed in a ceramic insulator that holds it and keeps it isolated
from the furnace metal. The flame completes the circuit. The controller
board sends it a small current, at about 90V?. If the flame is there,
the controller detects that current is flowing, because the flame can
conduct a small current, it actually rectifies it. I had no idea how
this worked before, but that's what I learned from googling. I haven't
seen anything about any special requirements beyond that. And the pics
Danny posted of his flame sensors sure look like just a metal rod housed
in a ceramic insulator. So, from everything I've seen so far about
how it works, the physics, etc it sounds like just a metal rod.
So, if that's right, then the question remains, what failed in Danny's
detector? There have been so many posts, I'm unsure of is what exactly
he measured, what the readings were, etc.
I think he measured that there was infinite resistance between rod and
furnace metal, ie no shorting of the insulator. I think he verified
conductivity from end of wire to tip of rod, but not sure what exactly he
measured there. So, I don't understand what could be wrong with the old
I'd guess the measurements he took aren't precise enough (or weren't
actually measuring what thought he was is also a possibility of course).
Being as it's a minute current, _any_ breakdown in the insulation and/or
hermetic seal could lead to having enough leakage current to essentially
short out the actual current but still measure a high (but not infinite)
Hypothesis; we don't have enough info to be able to diagnose remotely
exactly. The key is it doesn't work; a new one does. That's good
enough for me...
In a "former life" was responsible for incore self-powered neutron
detectors for a particular commercial power reactor vendor. These are
essentially a current source also in the 100-1000 nano-amp range and are
a swaged leadwire on a Rh emitter surrounded by a ceramic in a metallic
sheath. (Sound familiar?)
Their OD is only about 1/16", the lead wire and ceramic are thus tiny.
Turned out that even though one could measure a very high bulk
resistance between the emitter and sheath that miniscule locations along
the length of the detector could have very localized thin spots and
become a shunt path for leakage currents. Also turned out that moisture
could infiltrate the end seal and gradually diffuse down the length of
the detector over time. This also reduced the overall resistance and
showed up as a systematically lower output signal but of great
variability as again the existence of the surface roughness and local
thickness made exponential changes in local conductivity.
The detectors had been tested for years at the corporate research
reactor without ever discovering the issues--turned out it required the
higher power density of the real thing for the issue to become
significant. While they're not identical, given the similarity in
construction and function I can imagine many such similar scenarios with
these devices. Mostly I imagine it is a moisture infiltration problem
with these as a first guess.
In the end, the reactor detector problems turned out to be a secondary
blessing; I got a MS thesis out of it while it kept me gainfully
employed to pay for the course work while working around the issues for
operating reactors w/o requiring warranty replacements and solving the
manufacturing problems for new product. :)
I just noticed this, so sorry for not responding sooner.
I'm still amazed that the heat sensor was bad, simply because there
was almost nothing to it.
I guess if, under heat, a crack allows electrons to "jump" to ground,
both ways, then the control board wouldn't measure any DC current
(because it would be AC). I'm guessing though.
The heater has been working (although it has been almost 70 degrees
lately) just fine with the new flame sensor.
Apparently, this is what happened:
1. Heater wasn't working for so long that it "forgot" the prior
error code, so it was just blinking incessantly.
2. Eventually, the heater got around to blinking a code for a
bad flame sensor circuit.
3. The observation was the flames would light and then shut off
within seconds, four times, and then lockout would occur until
the power to the furnace was turned off and back on.
4. All measurements to determine whether the flame sensor was
broken had failed, mainly because 5uA is hard to measure and
because the meter I had borrowed must have a blown fuse.
5. However, replacing the flame sensor worked.
How the flame sensor failed is a mystery, as it clearly was not
tarnished (it was sanded with steel wool, emery cloth, and a file,
over subsequent attempts). The ceramic "looked" intact, but may
have developed a fault such as a crack.
The assumption is that the flame sensor failed, although the
only proof of that is that the new flame sensor worked as
soon as it was installed.
Thanks for all your help and advice.
Who knew that the flames form part of an electrical circuit!
HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.