Why Unloaded Stator tests are BAD.

[Graham]

There seems to be way too much fussing about stator tests.
It?s simple no ground fault and factory voltage on all three stator leads equals good, if not equals bad.
Putting each other down isn?t going to salve a thing.

 

[posplayr]

There seems to be way too much fussing about stator tests.
It’s simple no ground fault and factory voltage on all three stator leads equals good, if not equals bad.
Putting each other down isn’t going to salve a thing.

There are a few instances where members here had spent a week or more (100+ posts in a debugging thread) chasing their tail due to an intermittent charging issue. From my recollection most of the reasons for those problems were eventually isolated to intermittent stator faults. Sometimes they were thermally related and sometimes just false positives (passed tests when actually bad).

The point of the thread was to be aware of these types of situations and realize that :

"Unloaded stator tests can only confirm a failure not confirm a good stator"

Guessing at the probability of occurrence of the false positive test might be 5-10% of the time. So generally (90-95% of the time) you will be able to just confirm a failure because the stator will show a short with the ohm meter.

And as to the other subject, he is on ignore.

 

[posplayr]

Most clamp-on meters are made for 60hz and may not read the amps correctly at 5000rpm. I ran into a similar situation with 3phase motor speed controls, some clamp-on meter wouldn’t read 100+hz. 5000rpm would be like 500hz wouldn’t it?

I did find that if you tested voltage from a stator lead to engine ground while running that a ground fault would show up even after an ohm test to ground passed.

I have seen stator lead-to-lead voltage test in the stator papers but I seem to have missed the stator lead-to-ground voltage test in them. Didn’t see a stator lead-to-ground test in the factory manual either.

How are you doing this test? Disconnected R/R or with the R/R connected?

As you stated this is with the engine running, but is the R/R connected ?
If it is then this is a LOADED Test. I don't recall any LOADED tests in either the manual or the stator pages.
That is the underlying motivation and point of this thread.

The UNLOADED tests are insufficient in about 5-10% of the time.


So a simple revision to the stator pages Phase B might be as simple as adding the second set of voltage tests with the engine running after first checking for an obvious failure detectable with an ohm meter.

The first part will check if there are any obvious failures which in the case of a bad stator would be conclusive 90-95% of the time (the stator insulation failed conclusively so that an ohm meter can detect it)

But if this passes there is still a need for the LOADED test:
Start the engine and measure the voltage of one lead (at a time) relative to ground and insure that it (they) increases with RPM . These would be the LOADED stator tests. I'm just not sure what you will see , although I assume the AC voltage should increase at least until regulation. For a series R/R it should keep increasing.

Measuring voltage is still an indirect measure, as a LOADED test actually requires full current which the clamp would measure. Unfortunately the type of clamp for 500 Hz AC is not in the typical toolkit. If you recall in my various stator voltage measurements, I'm using a current clamp capable of doing the measurments, but it was about $500 for the clamp alone.

 

[Graham]

There are a few instances where members here had spent a week or more (100+ posts in a debugging thread) chasing their tail due to an intermittent charging issue. From my recollection most of the reasons for those problems were eventually isolated to intermittent stator faults. Sometimes they were thermally related and sometimes just false positives (passed tests when actually bad).

The point of the thread was to be aware of these types of situations and realize that :

"Unloaded stator tests can only confirm a failure not confirm a good stator"

Guessing at the probability of occurrence of the false positive test might be 5-10% of the time. So generally (90-95% of the time) you will be able to just confirm a failure because the stator will show a short with the ohm meter.

And as to the other subject, he is on ignore.

The test I described are with the regulator unplugged and the engine running at whatever RPM the manufacture calls for. An ohmmeter is not used at it may lead to a false good or bad stator reading. An ohmmeter can only confirm a short to ground or an open coil with any confidence. The test below will confirm a good or bad stator.

All tests should be done at room temperature and again at operating temperature. This will show if there is a temperature related short or open circuits in the coils or ground.

1. Output voltage; (lead to lead) if low or high you have a short circuit in the coils and the stator is bad. If no voltage then you have an open circuit in the coils and stator is bad. If you show manufacture’s voltage range at manufacture’s RPM you have no open or shorted circuits in the coils and that will confirm a good stator if part 2 passed.

2. Ground fault test; (lead to engine) if you show any voltage to ground then you have a short circuit to ground and the stator is bad. If you show no voltage to ground then you have an open circuit and this will confirm a good stator if part 1 passed.

These tests are not for finding out what is wrong with the stator but instead are for confirming a bad or good stator witch is really all that is needed. Load testing is not needed because if it shows a short or open under load it will show up on a voltmeter under no load. No guessing or theories but proven facts, out of over 300 stators tested in a 5 years span not 1 passed the tests above and failed to work as manufactured.

 

[posplayr]

The test I described are with the regulator unplugged and the engine running at whatever RPM the manufacture calls for. An ohmmeter is not used at it may lead to a false good or bad stator reading. An ohmmeter can only confirm a short to ground or an open coil with any confidence. The test below will confirm a good or bad stator.

All tests should be done at room temperature and again at operating temperature. This will show if there is a temperature related short or open circuits in the coils or ground.

1. Output voltage; (lead to lead) if low or high you have a short circuit in the coils and the stator is bad. If no voltage then you have an open circuit in the coils and stator is bad. If you show manufacture’s voltage range at manufacture’s RPM you have no open or shorted circuits in the coils and that will confirm a good stator if part 2 passed.

2. Ground fault test; (lead to engine) if you show any voltage to ground then you have a short circuit to ground and the stator is bad. If you show no voltage to ground then you have an open circuit and this will confirm a good stator if part 1 passed.

These tests are not for finding out what is wrong with the stator but instead are for confirming a bad or good stator witch is really all that is needed. Load testing is not needed because if it shows a short or open under load it will show up on a voltmeter under no load. No guessing or theories but proven facts, out of over 300 stators tested in a 5 years span not 1 passed the tests above and failed to work as manufactured.

I now understand that your statements are in direct conflict with the premise of this thread.

Generally speaking an electrical engineer would always prefer the results of a "LOADED" test v.s. UNLOADED as the loaded test is testing the device as it is expected to be operated.

The UNLOADED test is a theoretical test as it assumes homogeneity of the conditions. It is an extrapolation of a non operating condition to what should happen in a OPERATING condition. The theory of this type of test is is easily violated by multiple factors. Here is a reference detailing the multiplicity of factors that might cause misleading conclusions from using indirect measures like your UNLOADED tests.

http://www.scribd.com/doc/24911026/Megger-Book-The-Complete-Guide-to-Electrical-Insulation-Testing

As for your 100% positive results, I can't explain that as it contradicts what has been reported here at GSR with temperature dependent and other intermittent results in stator testing.

 

[Graham]

I now understand that your statements are in direct conflict with the premise of this thread.

Generally speaking an electrical engineer would always prefer the results of a "LOADED" test v.s. UNLOADED as the loaded test is testing the device as it is expected to be operated.

The UNLOADED test is a theoretical test as it assumes homogeneity of the conditions. It is an extrapolation of a non operating condition to what should happen in a OPERATING condition. The theory of this type of test is is easily violated by multiple factors. Here is a reference detailing the multiplicity of factors that might cause misleading conclusions from using indirect measures like your UNLOADED tests.

http://www.scribd.com/doc/24911026/Megger-Book-The-Complete-Guide-to-Electrical-Insulation-Testing

As for your 100% positive results, I can't explain that as it contradicts what has been reported here at GSR with temperature dependent and other intermittent results in stator testing.

Electrical engineers don?t work on motorcycles for a living and therefore would not have the practical knowledge of motorcycle electrical repair in the field. A motorcycle mechanic would. Common sense could tell anyone this. Would anyone use a motorcycle mechanic to design a motor control center for a large factory? I think not.

Unloaded test are not theory and do not assume anything. They are common practice and are manufacture approved. There is a load test for motorcycles but it is for the hole charging system and not the stator only. If there is a problem with a stator under load after testing to be good then the problem is in the load not the stator.

There is nothing in the Megger book about motorcycles; it seams to be about industrial motors, transformers and generators.

The GSR is a form of motorcycle mechanics and non-mechanics. How can someone in their right mind use it to base stator test results. Don?t most people on the GSR use the stator papers for help anyway. Maybe the 5-10% is from the stator papers being insufficient and not if it's under load or unloaded.

 

[duaneage]

Isn't this just a simple matter of replacing the OEM RR with a Honda unit, or just about any other modern RR, bucking up a new stator if needed and calling it a day?

I think the only test that really matters is do you get around 14 volts across the battery when the engine is running at various speeds. At that point you can forget how it all works and focus on the riding aspect of owning a GS.

 

[posplayr]

Here's a question...

What about loading it up with high wattage resistors of a suitable value and measuring the voltage across the resistors with a normal multimeter?

No idea what value they should be, but just a thought...

I mention this because I intend to check mine tomorrow during its first start since the rebuild, and all I have available is a multimeter...

Pete,
I think you might be on to something here. A three phase dummy load would be an excellent idea. The question is what value. I found some audio resistor like the one listed below which are 8 ohms.

The stator puts out nominally 70-80 volts at 5KRPM; with 8 ohms resistors that would be about 80/16=5 amps and a nearly full voltage test. P = 5*8=40/.707=28 watts at 5K RPM.

http://cgi.ebay.com/50W-cement-powe...405?pt=LH_DefaultDomain_0&hash=item2a0e80a565

I would put them in a Y configuration and not spend alot of time at 5K RPM as they will get pretty hot.

Too many glasses of wine at the moment, I'll do a little spreadsheet tomorrow to see what the power per resistor works out to.

Nice; I think I missed/dismissed this the first time.

Jim

 

[Guest]

Cheers for that, I sorta figured it wasn't a real flash idea, but maybe it isn't so bad after all... see what you think after the few wines have worn off

I popped out of the thread for a bit because I don't know enough about this stuff to offer an opinion one way or the other and it was getting a little too heated for my liking... both you guys have given me very useful advice along the way so I don't want to take sides on something I know nothing about...

 

[bakalorz]

Pete,
I think you might be on to something here. A three phase dummy load would be an excellent idea. The question is what value. I found some audio resistor like the one listed below which are 8 ohms.

The stator puts out nominally 70-80 volts at 5KRPM; with 8 ohms resistors that would be about 80/16=5 amps and a nearly full voltage test. P = 5*8=40/.707=28 watts at 5K RPM.

http://cgi.ebay.com/50W-cement-powe...405?pt=LH_DefaultDomain_0&hash=item2a0e80a565

I would put them in a Y configuration and not spend alot of time at 5K RPM as they will get pretty hot.

Too many glasses of wine at the moment, I'll do a little spreadsheet tomorrow to see what the power per resistor works out to.

Nice; I think I missed/dismissed this the first time.

Jim

-- WARNING --

POSPLAYER IS PROVIDING DANGEROUSLY INCOMPETENT ADVICE IN THIS POST

DO NOT UNDER ANY CIRCUMSTANCES ATTEMPT THE TEST DESCRIBED BY HIM

This is DANGEROUS overloading of the resistors.

You're too incompetent to even do a simple power calculation.

I have no idea which orifice you pulled your power calculation from, but

P=V*I
or
P=I*I*R
or
P=V*V/R

take your pick.

If you were to assume that the phase voltage were 40 volts (1/2 of 80 volts) the power would = 200 watts.

But even your assumption of 40 volts is incorrect. The phase voltage of each leg of a y connected stator is 1/sin(60) that value.

So the actual power is 267 watts per resistor ... almost 10 times more than you erroneously calculated and more than 5 times the allowable power.

WAY TOO MUCH !

And if you happen to go over 5000 rpm while trying to measure the currents the power goes up exponentially, if you touch 7000 rpm for a second, you are at over 500 watts per resistor ... more than 10 times the allowable limit.

To anyone following allong ... Do not attempt this test unless you want to take a good chance of setting your bike on fire.

And Posplayer, don't drunk post dangerous advice with the caveat that "I'll check to see if its safe in the morning"

What you just advocated is dangerous and criminally negligent ... period.

If you're to incapacitated to figure it out, wait till you've sobered up to post it at all.

 

[tkent02]

How about strapping an AC voltmeter to the handlebars and going for a long ride? Hook the meter between any two stator leads for a while, then another pair, then the other pair.
Seems like you would see any problems this way, whether temperature related, or load related or anything else. Maybe having a DC meter hooked across the battery at the same time would be helpful as well.
I dunno, in all the GSes that have been through here, I've never had a stator failure other than one that was clearly bad when I bought the bike.

 

[posplayr]

How about strapping an AC voltmeter to the handlebars and going for a long ride? Hook the meter between any two stator leads for a while, then another pair, then the other pair.
Seems like you would see any problems this way, whether temperature related, or load related or anything else. Maybe having a DC meter hooked across the battery at the same time would be helpful as well.
I dunno, in all the GSes that have been through here, I've never had a stator failure other than one that was clearly bad when I bought the bike.

Kent,
Other than the difficulty of the logistics, the main issue is that you still can not isolate absolutely between a R/R and the stator. When the R/R is disconnected then everything is unloaded. So the solution is to have an external dummy load in order to perform loaded tests.
If there is a hard failure then, there is no need as the unloaded tests will confirm failure, just not a good stator.
Jim

 

[salty_monk]

I'm all for a good debate about anything but there is no need for this to get personal.
Calls of incompetence etc. It's nasty.

At the end of the day this is a forum where friendly people give friendly advice. You have to do your own research & make up your own mind if you follow it.

Bakolorz it's great that you have technical knowledge to help others but you can disagree & phrase your posts a little better without getting personal.

Stators have not been a real issue for me yet either... touch wood!

Dan

 

[Graham]

How are loaded tests going to confirm a good stator? How much current is a stator going to put out under load? Does the manufacture provide that information? If not where does someone get that information?

Wouldn?t a test under load show the current used by the load and not the maximum safe output current of the stator? When someone tests the full load amps (current) on a three-phase motor (load) it doesn?t show maximum current output from the power plant. An electrical engineer would know this.

Why is it that unloaded stator test don?t confirm a good stator and where did that idea come from? Motorcycle repair shops, manufactures, posts on forms, magazines, or a panel of electrical engineers?

It still hasn't been explained why unloaded tests are bad and how that conclusion came about with sources?

 

[posplayr]

How are loaded tests going to confirm a good stator? How much current is a stator going to put out under load? Does the manufacture provide that information? If not where does someone get that information?

Wouldn?t a test under load show the current used by the load and not the maximum safe output current of the stator? When someone tests the full load amps (current) on a three-phase motor (load) it doesn?t show maximum current output from the power plant. An electrical engineer would know this.

Why is it that unloaded stator test don?t confirm a good stator and where did that idea come from? Motorcycle repair shops, manufactures, posts on forms, magazines, or a panel of electrical engineers?

It still hasn't been explained why unloaded tests are bad and how that conclusion came about with sources?

Unless you stress the insulation in excess of the operating condition you don't know whether it will break down and is therefore inconclusive. If you measure it and it did break down then that is conclusive. So the question is what does it take to stress the insulation?

Insulation testing was covered in detail in that reference I posted.
Insulation breakdown is a function of many factors but looking at the Megger tests, a high +250V is used to exceed a nominal rating (e.g. 125V). At 5K RPM there is 80 volts leg to leg so that is much better than a ohm meter. The other factor is temperature, where apparently the insulation resistance drops by 1/2 for each increase in 10 degC. These are probably the two most dominant effects other than having the stator be smoked/cooked.

The two tests that you described are probably about as good as you can do as is indicated by your results.( the voltage tests to ground need to be added to the stator pages). The main shortcoming is getting the stator hot, but riding the bike around could do that. These no load tests still don't have any current flowing, which is typically what you would need to get electrical stress, But because the only issue is insulation breakdown when testing the stator alone, the 5K RPM voltage tests will probably capture most all leg to leg and leg to ground faults that the 80V leg to leg voltage can cause to break down.

The dummy load could probably get about 5 amps to flow through each leg without dropping the voltage too much. This would be more thorough if for no other reason to get the stator hotter.The dummy load is pretty easy to configure as the only thing that really matters is the leg to leg voltage is is know. A dummy Y load with 10 ohm resistors

 

[tkent02]

Kent,
Other than the difficulty of the logistics, the main issue is that you still can not isolate absolutely between a R/R and the stator.

Pop a known good R/R on there, you would know.

 

[posplayr]

Pop a known good R/R on there, you would know.

yes has been mentioned. Got a good one?

 

[tkent02]

yes has been mentioned. Got a good one?

A whole bag of them.

 

[duaneage]

yes has been mentioned. Got a good one?

Yes. Quite a few actually:dancing:

 

[Graham]

Unless you stress the insulation in excess of the operating condition you don't know whether it will break down and is therefore inconclusive. If you measure it and it did break down then that is conclusive. So the question is what does it take to stress the insulation?

Insulation testing was covered in detail in that reference I posted.
Insulation breakdown is a function of many factors but looking at the Megger tests, a high +250V is used to exceed a nominal rating (e.g. 125V). At 5K RPM there is 80 volts leg to leg so that is much better than a ohm meter. The other factor is temperature, where apparently the insulation resistance drops by 1/2 for each increase in 10 degC. These are probably the two most dominant effects other than having the stator be smoked/cooked.

The two tests that you described are probably about as good as you can do as is indicated by your results.( the voltage tests to ground need to be added to the stator pages). The main shortcoming is getting the stator hot, but riding the bike around could do that. These no load tests still don't have any current flowing, which is typically what you would need to get electrical stress, But because the only issue is insulation breakdown when testing the stator alone, the 5K RPM voltage tests will probably capture most all leg to leg and leg to ground faults that the 80V leg to leg voltage can cause to break down.

The dummy load could probably get about 5 amps to flow through each leg without dropping the voltage too much. This would be more thorough if for no other reason to get the stator hotter.The dummy load is pretty easy to configure as the only thing that really matters is the leg to leg voltage is is know. A dummy Y load with 10 ohm resistors

If you want to heat the stator to operating temperature then just idle the bike about 25 minutes with a fan blowing on the engine.

Using resistors as a dummy load could pop the resistors or short the stator out. Wouldn't you would need like 400watt resisters at whatever ohms to handle a 5amp load at 80volts. I won?t use a resistor dummy load to test a possible good stator and take a chance of burning it out.

I understood from earlier post a load test was for checking amps (current). And now the load test is for operating temperature and voltage testing.

All a Megger would be good for is checking a ground fault and that would be at very low amps (current). The Megger thing is out of the question anyway because of price alone, Megger $400 vs. stator $100. Motorcycle manufactures do not give the breakdown point for the insulation of stators like the manufactures of industrial motors do, so again the Megger is out.

If a load is really need then just hook it up to a good regulator and turn on all lights, flashers and whatever and test the amps or volts on the stator leads. You still won?t know what the manufacture specs are for this test on a good stator. With out that information the test would be useless.

It seems loaded stator test are bad because of complexity, guessing, theory, test equipment, test procedures and lack of manufacture specifications. How could they be conclusive?