Troubleshooting “Little Jake”: Page 4 of 5

Intermediate

Inside this Article

Troubleshooting “Little Jake”
Stepping through the troubleshooting and repair process on a vintage Jacobs wind turbine.
Little Jake needed to come down for repair.
Based on troubleshooting tests, Little Jake needed to come down for repair.
Little Jake’s tail was manually furled out of the wind
Little Jake’s tail was manually furled out of the wind so tower wiring could be checked.
The controller box
The controller box was tested and diagnosed.
Diversion controller, dump loads, and DC panel.
Diversion controller, dump loads, and DC panel.
RE systems’ balance of system (BOS) components
The RE systems’ balance of system (BOS) components, including the charge controller, battery bank, inverters, and DC and AC load power centers.
A megaohmeter (megger)
A megaohmeter (megger) was used to test the THWN conductor insulation.
Yaw slip-ring brushes.
Yaw slip-ring brushes.
Yaw slip rings.
Yaw slip rings.
A high-potential (hipot) tester
Testing for shorts from the conductors to the metal tower was accomplished with a high-potential (hipot) tester.
Testing in the shop
Testing in the shop revealed an internal short—a short from wire to wire—in the armature windings.
Testing in the shop
Testing in the shop revealed an internal short—a short from wire to wire—in the armature windings.
The front end of the generator
The front end of the generator. The blades would be attached to the long shaft.
The rewound armature.
The rewound armature.
Troubleshooting “Little Jake”
Little Jake needed to come down for repair.
Little Jake’s tail was manually furled out of the wind
The controller box
Diversion controller, dump loads, and DC panel.
RE systems’ balance of system (BOS) components
A megaohmeter (megger)
Yaw slip-ring brushes.
Yaw slip rings.
A high-potential (hipot) tester
Testing in the shop
Testing in the shop
The front end of the generator
The rewound armature.

Next, with the generator brushes reseated, we used the Sun Chaser battery bank to run Little Jake’s generator like a motor. With controller power still disabled and the tail furled, we back-fed 48 volts to the wind turbine from the junction box at the tower base. Indeed, the machine began to spin and “motor” with ease. This was both promising and unsettling—if it worked as a motor, why wasn’t it delivering power as a generator?

The final test was a simple voltage test, but had to be done with an analog voltmeter instead of a digital multimeter. I recruited friend and fellow windsmith Dave Kiedrowski to come to the MREA the following week to do this test with me. The idea was to watch the needle of the voltmeter (connected to positive and negative terminals on the generator) as we spun the rotor slowly at first, and then gradually up to a “significant” rpm, but not so out of control that we were in danger of being hit by a spinning rotor while strapped to the top of the machine in our climbing gear. According to Mick, we should have seen the voltage gradually rise, and the needle should have steadily climbed, with no spikes, jumps, or drops. But we never saw more than 2 volts register on the meter. The armature was simply not delivering voltage and, for the first time, we were pretty certain that the problem was within the armature windings.

It was too windy to unfurl the tail and work safely while we were on the tower, so we repeated this voltage test at the base of the tower, with the leads of the analog voltmeter connected to the conductors in the junction box. We manually unfurled the tail. At high wind speeds with a swiftly moving rotor, we were still only seeing a couple of volts. I relayed our results to Mick, and he confirmed that it was time to call a crane operator and take down Little Jake for repair. He agreed that the problem was in the armature.

Just to be sure the maintenance class hadn’t missed anything, Dave and I repeated the continuity checks and high-resistance megger tests from the BOS to the generator. This time, we used a digital megger with batteries instead of a hand-crank, but the results were identical.

We also checked for shorts from the conductors to the metal tower using a high potential (hipot) tester I borrowed from fellow Jacobs enthusiast Jeff Nichols. Hipot or dielectric-withstanding voltage tests look for insulation breakdown and short circuits, and are often used to check electrical appliances for faults before they leave the assembly plant. Our hipot tests showed no ground faults or shorts to the tower. This was very similar to our tests with the meggers, but was just another tool to use and verify our original findings.

MREA regional training officer Clay Sterling and I worked with a crew and a crane operator to take down Little Jake in September 2011.

Bench Tests & Repairs

Once Little Jake was safely on the ground, MREA site manager Mike White disassembled the machine and packed up the generator for a trip to Mick’s wind turbine repair shop, just a few hours away. On a frigid December day, the three of us cracked open Little Jake and hoisted the armature onto the test bench, a factory original 1940s Jacobs wind generator testing system. A variable-speed DC motor arced and sparked as it drove the shaft of the armature from a slow crawl up to full output, and the needles on the hand-drawn calligraphic meter scales showed us voltage, current, and rpm.

With the generator on the test bench, the problem became quickly obvious. The voltage needle on the analog voltmeter was volatile—but only at an extremely low rpm; it spiked and collapsed with each slow rotation, indicating that there was indeed a short somewhere in the armature windings, likely due to insulation breakdown. Over time, wire insulation can break down with heat, movement, or extreme environmental conditions—three things that are unavoidable in any generator. The Jake is a 1946 machine, and the insulation was 64 years old, having seen some 50 years of service.

Comments (1)

Lindsey Roke's picture

I was interested in the trouble-shooting approach.
It appears to have followed what we might call “the aardvark method”.
(“I am thinking of a word in the dictionary, what is it?”
“Is it aardvark?”
“No”
“Is it aardwolf?”
“No”
Etc.)
This method will eventually find a fault but it but it is not the fastest – and without due care has the potential to introduce more problems if (for example) connections are not done up again correctly.
Here in New Zealand I work for an appliance manufacturer (Fisher & Paykel Appliances Ltd.). In managing production engineering I have always been intrigued by various approaches to trouble shooting. We learned a great deal from the late Dorian Shainin when we had him here running some training sessions for us. (Look him up on the web is you are not familiar with him.) The technique he would propose for the dictionary would find any word with about 17 questions requiring just yes/no answers. (Find the middle page and ask if the word is before that page. Then go to the quarter or three-quarter point as appropriate. 11 questions will get you to the right page in a dictionary with up to 2048 pages. The next question gets you to the right column. Assuming there are no more than 32 words in a column, 5 more questions will get you to the right word.)
I recall one factory break-down where the photocell sensor was supposed to detect a cabinet on the assembly line, send a signal to a microprocessor controller, which in turn was supposed to drive a solid state relay which switched a hydraulic valve that operated a piston that did something useful. The problem was that when the cabinet went in front of the photocell the useful action didn’t happen. While I could have started by checking that the photocell was working, it was much quicker to repeatedly “cut the problem in half”. Because we had indicator lights across the electrical connections on hydraulic solenoids, (put there when the equipment was installed for just such trouble-shooting) the first check was to see if there was power getting to the solenoid when the photocell was triggered, etc. etc. (Without the indicator lights, the same thing could have been done with a meter – but when an assembly line is stopped, time is of the essence.).
You can bend the rules a bit if there is a very easy point to check that is not mid-way through the problem – or if there are other clues as to where to start looking.
I have learned that, if a piece of equipment has been operating OK and it stops, unless somebody has already had a go at trouble-shooting it by swapping components in and out before you arrive, it will very very rarely have more than one fault.
My objective, when called to an intractable malfunction, is to determine which component is faulty, get the tradesmen to replace it and walk away with confidence that the machine would work when they have done so.
While the mid-point of the Jacobs system in the article might be the input to the slip-rings, to save climbing the tower, checking first whether or not there was a voltage at the bottom would have isolated the problem to either the Balance of system and wiring to it from the tower or – to the generator, slip-rings and tower wiring. The next check at the generator terminals would have put the fault in the generator or in the slip rings and tower wiring etc.
(I’m not decrying checking such things as the brush and slip-ring condition while up the tower – but as maintenance tasks not as trouble-shooting in this case.)

Shainin, in my opinion was better (at least in this area) than either Joe Juran or Edward Deming – but not so well known, probably because he was better at solving problems than writing books.
I trust this might be useful.

Show or Hide All Comments

Advertisement

X
You may login with either your assigned username or your e-mail address.
The password field is case sensitive.
Loading