# Success: How to troubleshoot your house furnace

#### MrChips

Joined Oct 2, 2009
27,699
We've been without heat for the past two days with nighttime temperatures hovering around 0°C.
So I learned a few lessons on how to fix a broken furnace.

1) I measured the resistance across the gas valve and measured 4MΩ. Actually, there are two coils in the valve. They both measured 4MΩ to common while they measured 11MΩ across the two. I thought to myself that that is odd. I was expecting a low resistance across a solenoid valve and open circuit if the coil was blown. Why did two 4MΩ together add up to 11MΩ? Had I gone to the internet in the first place I would have been given the correct answer but of course I didn't. More on this later. Lesson #1.

2) I called in HVAC service and the young technician said he thinks the gas valve is ok because he can hear hissing and can smell gas. I can assure you that his hearing and sense of smell is a lot better than mine. So he checks the igniter and reports that there is only 80V going to the igniter and it should be 120V. He thinks that there is a problem on the controller board and needs to be replaced at a tune in excess of $800 and 3-5 days to get a new board. Checking the internet was Lesson #2. 3) So I said, if there is a problem on the control board I am willing to have a look at it instead of waiting 3-5 days for heat to be restored. Result: Success today and heat is restored. Here are the lessons learned (after going to the internet). Lesson #1. While I was expecting to read low resistance across a solenoid, apparently this is not so. There is a bridge rectifier in the solenoid circuit that shows 4MΩ on an ohmmeter. This part I have not figured out as yet. In any case you cannot do a resistance test to check a newer gas valve. Lesson #2. In our previous furnace the igniter was a HV spark gap. Not anymore in newer furnaces. They use a hot surface igniter (HSI) which is a silicon nitride high temperature resistor. This should measure about 10-20Ω on an ohmmeter. I can only assume that the technician verified this and placed the fault on the controller board quoted in excess of$800. The knowledge I gained was that there are 80V HSI and 120V HSI. My device was the 80V kind.

The resistance of the HSI was about 4kΩ, yet when I applied 120VAC directly to it, it did glow, but only one time. I assume that applying 120V on the 80V device killed it but this is purely speculation. I can only assume that the HSI was faulty and by applying 120V it gave its last breath of life and then died.

The old HSI part is White Rodgers 768A-15 which has been replaced with 768A-815. Delivery was quoted as 1-4 weeks.
I ordered a 768A-843 which arrived in two days. I figured that the only difference was the style of the connector that has to mate with the furnace. The only problem is one needs the proper tool to extract the molex-type contacts. Fortunately for me I do have that tool. So I just swapped out the new contact housing and replaced it with the old one.

The bottom line here would be to check the HSI first and it should read 10-20Ω. Comments on the internet state that the HSI has a finite lifetime and should be replaced every 5 years. Our furnace is 13 years old.

All in all, I paid for a service call and the cost of a new HSI from the internet. The cost of the service call was my lesson on how a house furnace works. I saved over \$1000 by doing the repair myself. My fingers and toes are now warm and cozy as I wrap up this post.

Edit: Here is a photo of the HSI.

Last edited:

#### michael8

Joined Jan 11, 2015
333

Perhaps, perhaps not... Gas dryer, uses HSI, 50+ years old, don't remember a HSI replacement...

Gas dryer would start normally but was taking a long time to dry clothes. Checking
the failing case, could see the HSI light, hear the optical(?) sensor click indicating
it thought the HSI was hot but no gas valve noise and no gas, no flame.

This was due to an intermittent 120V AC gas solenoid coil which would open when hot. After
enough failed relights, the coil would have cooled down and the dryer would lite until the coil
warmed up again.

A new replacement coil, wired correctly by color code, resulted in no gas ever.

It turned out that the replacement coil color code was backwards. This AC solenoid coil
was one of two on that solenoid and as a safety interlock both had to aid each to open the gas valve.

A sci-fi moment: "Reverse the polarity" And everything was fine...

PS: The response from the appliance parts support people on explaining this problem was
a suggestion to check the outlet for correct wiring...

#### geekoftheweek

Joined Oct 6, 2013
916
Good information and glad you found it!!

Luckily mine was a bit more obvious when it went bad... (from around this time last year)

#### MrChips

Joined Oct 2, 2009
27,699
The HSI received measured 14Ω when cold.
Hence the inrush current on power on is 5.7A, power is 450W.
The box label says 92W. Hence the hot resistance is about 70Ω.
The inrush current would have to drop from 5.7A to 1A.

I suspect that the MCU driven controller board drives the HSI with a PWM signal to gradually bring the HSI to working temperature of about 1000°F or 550°C. The warm up delay is about 10 seconds before the gas valve is opened.

#### MrChips

Joined Oct 2, 2009
27,699
To answer my own question as to why you cannot do a resistance check on a gas valve, I tested one 1N4001 diode and then two 1N4001 diodes.

A DMM does not supply a high enough excitation voltage to exceed the diode's knee voltage when measuring resistance. Hence the diode registers as a high resistance.

My analog test meter on the other hand does show low resistance. Be aware that the polarity of the excitation voltage on an analog meter is reversed because of the way the battery has to be applied in order to measure positive current on the analog meter movement.

#### killivolt

Joined Jan 10, 2010
834
To answer my own question as to why you cannot do a resistance check on a gas valve, I tested one 1N4001 diode and then two 1N4001 diodes.

A DMM does not supply a high enough excitation voltage to exceed the diode's knee voltage when measuring resistance. Hence the diode registers as a high resistance.

My analog test meter on the other hand does show low resistance. Be aware that the polarity of the excitation voltage on an analog meter is reversed because of the way the battery has to be applied in order to measure positive current on the analog meter movement.
I’ll start with strange, I’m working on a Toaster while your here, didn’t see this till now lol

My furnace went out last week Monday, I took the day off to discover the many things going on with it. I fortunately did furnace repair work Service Tech back when I was 24, now 65 on the 27th, promote lol

Anyway, HSI (Hot Surface Ignitor?) I’m guessing by the acronym, the circuit and coils on the gas valve explained well by your investigations, conclude a resistance as well as voltage amperage operation through the ignitor, there is a flame sensor involved but for the most part you discovered what I would have assumed and pulled from my truck and installed to find the solution. On mine it’s 120vac to the HSI, then the flame sensor controls or in this sense is a safety, no flame no valve operation.

The flame sensor is resistance driven, sometimes it can get dirty, cleaning will restore temporary operation until you get a new one, just saying if it were a person who needed to know a little about their furnace similar type and design of course.

I planned to post my experience on my furnace as well, but the Toaster made it first lol

kv

Edit: Later headed up to SLC Utah to help a good friend move the last of his stuff from his old shop. Have a good one.