Author Topic: Datalog from a dying transformer  (Read 4974 times)

0 Members and 1 Guest are viewing this topic.

SparWeb

  • SuperHero Member
  • ******
  • Posts: 3225
  • Country: ca
    • Wind Turbine Project Field Notes
Datalog from a dying transformer
« on: March 25, 2013, 11:36:24 PM »
I have been using a datalogger to snoop on my furnace for a little while.  I'm doing this for several reasons:
   - Set the most efficient heat cycle on the programmable thermostat
   - Watch for unusual behavior (sometimes it starts heating, briefly stops, then resumes again, for no reason)
   - Have comparison data to use against the eventual replacement.

The natural gas central heating furnace of my house is almost 20 years old, so it's due to be replaced soon anyway.  The logger just happened to be hooked up when the transformer inside it burned out, giving some interesting results.

I woke up last Wednesday morning to a very cold house.  Uh-oh.  I live in Canada where this can be a big deal, even in March.
At the thermostat (a programmable type) I could see that it was still set to 19C, but the actual temp was at 15C.  The thermostat's own circuit was on, but the furnace wasn't turning on.

Circuit breaker not tripped, fan switch still on, so the problem is probably inside the furnace.  Manually threw the breaker and started to open up the panels.

Inside the control box was a definite smell of burned wire, but no damage was immediately visible.  Since this furnace uses a small transformer to step down the 120V line to 24VAC to run the natural gas solenoid valve, I would normally hear the transformer humming.  I couldn't hear that sound so that was my first target.

Here's what it looked like when I removed it from its mount...



And here's what it looks like on closer examination....



The 120VAC winding is burned out, while the 24VAC winding looks fine.  Replacing the transformer with an identical one got the furnace started again.  It's been working normally since.

I don't know why the transformer blew out, though.  In a sense, I was monitoring the wrong circuits of the furnace to positively find out.  It would be nice to know, because it means I don't know what could cause the same thing to fail again.

Now for the interesting part with the datalogger.  As I said, I was logging data from the furnace.  I was only watching the 24VAC solenoid valve at the time.  Despite that, the datalog clearly recorded a series of voltage fluctuations on that side of the transformer, while the 120VAC side was in the process of burning out.

Here's what the recording looks like, for the 3 1/2 minutes it took before everything quit:



For comparison, here's what a day of readings normally looks like:



Typically, long periods of time Off, separated by periods about 20 mintes long of On, constant voltage.  It runs until it's done, then stops.
During the failure, the low-voltage side saw repeated bursts of 10 volts AC or so, flickering dozens of times before the damage was permanent.  It's kind of like the high-voltage winding suffered an arc through part of the winding, involving only 1/2 of the wire.  Since the 1/2 of the winding has 1/2 the resistance, the current was too high and slowly cooked the insulation, rather than burning it up all at once.  It's also interesting that the arc was able to burn its way through a wire and render the winding open-circuit, rather than trip the circuit breaker on the main panel.  Since the normal current demand is 7 Amps, but the circuit breaker is rated at 15A, maybe there was no cause for the breaker to trip under these circumstances.

I can't see any actual burning from an arc, though.

Despite the mystery, it's a fun insight into something in the process of failing.
Just one of the many uses for data loggers.
No one believes the theory except the one who developed it.  Everyone believes the experiment except the one who ran it.

System spec: 135w BP multicrystalline panels, regulated by Xantrex C40, DIY 8ft diameter wind turbine, regulated by Tri-Star TS60, 800AH x 24V AGM Battery, Xantrex SW4024

bart

  • Full Member
  • ***
  • Posts: 201
  • Country: us
Re: Datalog from a dying transformer
« Reply #1 on: March 26, 2013, 06:36:16 AM »
   If your data logger voltage is reading correctly, the control voltage is to low for my liking. For most AHU's with a condensing unit, 40va transformers are used. Can't tell from photo what it is. If it is a 40va, I would look through the control circuit to see if some thing is dropping the voltage. Or the primary voltage is to low.
   As far as the furnace short cycling, most of the time it is caused by skinned up T-stat wiring, right behind the T-stat or where it goes through the cabinet of the AHU.

SparWeb

  • SuperHero Member
  • ******
  • Posts: 3225
  • Country: ca
    • Wind Turbine Project Field Notes
Re: Datalog from a dying transformer
« Reply #2 on: March 26, 2013, 12:49:50 PM »
Thanks for the tip on the short-cycling.  I'll look at the thermostat wiring too.

I should double-check the scale factor on the control voltage used in the logger's output.  The sensor reads 5V, which has been dropped from the 24VAC source with a voltage divider, a diode, and a capacitor to soak up most of the ripple.  Now that you point it out, I may have forgotten to scale the voltage from 5V + Vf from the diode.  That would account for the incorrect voltage.  If it is measured correctly, then something else needs looking at.  You are right to point out that low voltage is a sign of an overloaded transformer (it's 30VA, by the way).

I haven't reconnected the logger to the furnace again, but doing so might reveal some other differences.
No one believes the theory except the one who developed it.  Everyone believes the experiment except the one who ran it.

System spec: 135w BP multicrystalline panels, regulated by Xantrex C40, DIY 8ft diameter wind turbine, regulated by Tri-Star TS60, 800AH x 24V AGM Battery, Xantrex SW4024