I would avoid hi-pot as it may hurt the windings. I'd stick
to a Doble test. It also depends on your voltage rating. A megger should suffice for voltages less than 5 kV.
Thats a tough one. Given that hi-pot tests usually apply much more than rated voltage, it can be somewhat of a destructive test. However, it will definately expose any weak insulation by causing it to fail. The question is, how long would it have lasted without hi-potting given the volts during operation are usually at least half of that applied during a hi-pot test. Good question.
I lean towards a megger test in place of a hi-pot test. The hi-pot will definately verify the insulation but if it is weak at all, it will damage the equipment. In my mind, the hi-pot is just going to give you a warm fuzzy feeling if it passes, if it fails, time for a rewind. Unless its real critical, its probably better to let it fail when it wants rather than hurrying it up by performing a hi-pot. Maybe to summarize, if you have a spare motor around then hi-pot, if you don't then skip it since you might get a little more run time out of it before it smokes.
I agree with Unclebob, you do not want to destroy an insulation by HiPot every time you make a PM service. To verify the insulation condition a "Megger" test Insulation resistance according with the recommended voltages by IEEE Std 43 is good enough. For windings operating up to 1000 Volts AC test with 500 Volts DC and for 2300/4160 VAC test with 1000 Volts DC.
IEEE 43 is definitely worth a look, it will tell you all you need to know about insulation resistance measurement and assessment.
Regarding hi-pot testing: a winding that has been in service shouldn't receive a hi-pot test at the same voltage as the manufacturers test (as given in IEEE 112 etc) - 50 to 75% at most. Even the manufacturer won't (or shouldn't) apply it more than once, for fear of shortening the life of the insulation. I suspect hi-pot testing is only done on machines out of service - unlike IR testing (which at least gives some warning of deterioration), a hi-pot test can instantly push a motor over the edge.
Hi-pot testing gets discussed on this forum every so-often, you might try a search.
Where I work now, they do what they call a ramped hi-pot test where they bring the voltage up slowly. They still test it to twice rated voltage and they do this every four years. This is also a large generator and not a motor. Never really agreed with it though.
Hi Afton Chemical,
Before to perform this test made a Megger Readding and if you have a good reading perform the HIPOT Test.
The DC HIPOT is a practice to evaluate the insulation system integrity, I was used for many Years with a few problems, for a machines with 5 or more years old we use a Voltage test of 1500 V for 460 machines and 1000 V for 230 machines.For a new, rebuilt or age less than 5 Years use 2 Rated Voltage + 1000 V.
The DC HIPOT Test will help like a predictive maintenance tool, avoiding DOWN TIME, because you could detect earlier problems and take decisions. I was asked to the hipot manufacturer about the Voltage Test and they conclude all motors most be tested at 2 rated Voltage + 1000 because the most critical condition is the motor´s start and you could find high voltages like rebound voltages due to the contactor don´t have instantaneus close. Example: If you perform a DC HIPOT Test in a some 460 V motor at 1880 VOLTS and you find a failure at 1200 Volts, you most abort the test but the motor could be running, because only will fail when it reaches this 1200 Volts, but you have the time to find a spare motor or to make the preparations to send it a repair shop. This is different like the motor burnout in operation because you will have down time.
May I just add something to Petronila's post. He specified a dc hipot test, which for a particular voltage is less damaging than ac. Strictly speaking a hipot test is usually ac, the specs (both IEEE and IEC) specify a test voltage that is ac - both quote the formula:
2x(line V rms) + 1000 V ac rms.
(if a manufacturer uses dc rather than ac for a highpot test, then the above formula is usually multiplied by 1.7).
In other words, Petronila's test is a megger test plus, rather than a true hipot. Megger tests are dc of course.
The following document comments on the differences between ac and dc tests in sections 22 and 23:
I believe that a full hipot test (dielectric or flash test, in IEC-language) is more a proving test carried out to show to the customer that a new winding has been properly rated and correctly manufactured, rather than a predictive maintenance test. It should only be repeated at reduced voltage, and probably at the beginning of a maintenance strip-down. There are better methods of PM, using insulation resistance as specified in IEEE 43.
Buzzp:
A “Controlled Overvoltage Test” per IEEE STD 95 is a very good test to evaluate the winding insulation condition. For that test the maximum DC voltage is selected around 2.1 to 2.55 times the rated AC machine voltage. The initial value is 30% of the maximum and the steps not to exceed 3% of the final value (around 7.5% of V nominal). This test is recommended for large machines since it takes long time to test and discharge every winding phase and the operator should be well trained to stop the test before destructive conditions are imposed.
Aolade,
I would like to see any test tech stop such a test before damage occurs. This is basically a go-nogo type of test in that when it breaks down, it goes quick; there is no lag time where a human could ever possibly stop the test. What I don't know is if this breaks down, how much damage is done? Will the machine continue to run? Its anyones guess and depends on the voltage level where it broke down.
Buzzp;
That is not a go-nogo test. The voltage is kept for 60 seconds on each step, at that time the leakage current is evaluated aginst the test trend, it could take around 50 minutes to test one phase and 2 hrs to discharge it. Please read IEEE Std 75.
Your talking about something different than I am. I am only referencing our standards (government). Wherever their standard came from that we use here goes like this: the voltage starts at zero and is ramped with a slope of 1000V per minute up to the test voltage (2X + 1000 basically). We compare that with previous tests (ramp time and amplitudes)which will only deviate much if the windings are damp or something. But I have seen a bad winding. When they go, they go quick and the current is off the scale on the test equipment. There would be no way to stop the test prior to this occurring. The current limited output of the test set would trip before an operator had a chance to stop it, which would be your only saving grace of maintaining any integrity in the windings.
Whats interesting about what your saying about the discharge portion of this test. Do they want this charge to bleed off on its own or do you put some high impedance load on the winding to discharge it slowly? Our procedures here (again I dont agree with) allow us to ground the winding immediately after the test (then a ground strap is placed on the winding). Which, causes problems of its own.
I have never read IEEE 95 but it sounds like I better. I am not debating that standard (as I never even read it) just how the ramp tests are done here and the value of the test, as it is done now.
Say we are talking about the IEEE 95 standard, why would not the winding be damaged when you had an insulation breakdown in doing a hi-pot test?? I am not talking about damage being done below the rated voltage but damage that may occur by applying a higher voltage during the test. The equipment may not ever see this high voltage in operation, which means it would never break down until the insulation was weakened further, which would occur if the winding broke down during the test at whatever voltage.
Alright, Ill go read the standard and see where the tests we are doing came from.
I saw cables tested years ago and am wondering if the same method would be productive with motors. The test set was DC and the current was limited to a very small value. The voltage was ramped up and the current was watched carefully. When the knee of the breakdown curve was reached, the current started to increase disproportionately to the voltage. The test was terminated at that point before the cable actually failed.
Is this type of testing still in use, and would it be appropriate for motors?
yours
I am of the opinion that AC hipots should be utilized for proof testing recently rewound motors and for testing inservice motors when there is a sufficiently large window to permit the rewind without creating a production problem. I do not use this method for periodic testing on short shutdowns and the large size of AC hipots for large motors limit their availability except for reduced frequency models which are still very large and very heavy. A DC hipot utilizing graphing techniques or utilizing Schlief methods to stop the increase of voltage at the knee of the graph is useful as it limits the effects of flashover and allows for the continued service of the motor or generator while contingency plans to repair or replace are made.
Hello oftenlost, what is the Schlief method? As you say, a carefully done DC hipot can prevent immediate damage to the machine but practically speaking it would need to be an automated procedure.
Pete, in theory, current in a purely resistive circuit is linear with an increase in voltage. In large motors, the increase in voltage is not linear as we have inductance, capacitance and insulation polarization occurring. Schlief developed a formula which calculates this deviation during the first step of the process and allowing the operator to vary the timing of the steps so that a straight chart line is possible. Any curvature is indicative of nonlinearity and allows the timely interruption of the test to protect the winding. A little tedious but useful in very large and expensive machines.
