What's new
What's new

How do older Phase Perfects usually die?

SND

Diamond
Joined
Jan 12, 2003
Location
Canada
For those of you who has PP's die, did it just not turn on one day, or just let the smoke out or something while running?

I've got a 10hp blue unit, bought in 2007 I think and it has run pretty much 7 days a week since. Changed filter caps about 2.5 years ago I think, and originals sill read good on meter.
I noticed in the last few months that when a machine is running(mostly the 7.5hp lathe), the high pitch from the PP gets quite a bit louder, which isn't something I ever noticed before. Is it a sign the IGBT's are getting old and tired? I'm sure I have over 30k if not 40k hours on it. Still runs every day but I feel like maybe someday I'll flip the switch and nothing will happen?
I've been thinking of buying a new 10hp unit of the newest generation, but kinda also wonder how much longer this old one might go on?
 
I guess nobody here had one quit, or everyone's too deaf to hear the high pitch change?
Anyhow, read a bit on IGBT's failure modes and I suppose this thing probably is starting to get tired by now so I'll order a new unit soon, I'm kinda curious to see the newest generation of these. How much longer this one has might remain a mystery, but it has served me very well many years.
 
Electronic semiconductor devices in general have what is termed a "bathtub" failure curve.

Initially, there is a high failure rate, as the "good" but borderline devices get weeded out by random stresses.

Then for a long time, there is a low failure rate.

Eventually, especially with semiconductors, "wearout" mechanisms start to come into play, and things start to fail at a higher rate. Except in the most stressed devices, you can easily get between 10 and 15 years operation.

Some of the failures are:
migration of the "metallization" the forms the conductors on the chip surface, to where clearances are reduced below minimum,

cumulative effects of heat stresses gradually cracking the "solder" that holds the chip to the heat dissipating block in power devices, so that the chip overheats.

In non power devices, the difference in thermal expansion between the plastic encapsulation and the chip may eventually crack the tiny wires that connect the chip to the solder pads, because those wires may be encapsulated by the epoxy case material.

In power devices, plain thermal expansion and contraction may eventually fatigue the wires, or even solder joints.

If the parts become overheated, or even if not overheated, eventually the "doping" of the chip, the materials that are heat-diffused in to form the "works" of the chip, may diffuse further, and ruin the internal structure.

Many recent products. and PP are among them, rely on programs stored within the control unit in some form of "ROM" (read only memory). The newer types are not actual physically changed to write the program into them, they rely on stored charges which will eventually decay and "forget". (Old ROMs used "fuses" that were actually blown open, and are probably somewhat better as far as decay of the program storage)

Manufacturers work hard to make those failure mechanisms less of a problem, but 25 year old solid state electronics are getting toward the "suspect" stage in most cases. There are exceptions.
 
I would investigate the snubbers on the IGBTs.

Depending on the topology, there will be a resistor, diode, capacitor, but its unlikely they used an inductor or even a transformer to put the energy back into the supply. so anyhow any of those components can fail.

In addition to the snubber circuitry there should be a film capacitor bolted directly on the igbt across the dc bus. that film capacitor can fail as well.

I don't know what they use for dc electrolytic capacitors, but they can fail as well.
 
wheelieking had one fail. If I remember it right it fried a lot of electronics and almost burned down his house (on the same circuit as the shop), a search should find the old thread.
 
I guess nobody here had one quit, or everyone's too deaf to hear the high pitch change?
10 years ago the high-pitched whine (at idle) used to drive me crazy. When actually running the lathe, I can't hear it because the lathe drowns out the whine from the PP.

Today I have so much hearing loss (and have piled so much stuff around the PP box) that I no longer hear the whine.
 
I had a used blue one blow soon after I bought it. The board was no longer supported so I was SOL. The seller had issues ( failed to disclose that ) and sent it to PP who checked it out. The seller ran for cover but PP gave me a very good deal on a new one. I have two other blue ones that are still working. If it is critical I'd buy a new one and keep the old as a back up. I have one that I plug into a 50 amp receptacle when I need additional capacity. Dave
 
20HP PP blue box (1st gen). Purchased from original owner, he was upgrading to a bigger unit to run a new machine. It worked great for me for about a year. Shut it off one evening, went to start it back up the next day, DOA. Powers up for a few seconds (makes 3 phase) then it shuts off. PP does not support the blue boxes anymore. I have a friend that is excellent with electronics repair look at it-he also consulted with a friend that designs electronic units that share similarities to the PP. Long story made short, its an uphill battle to fix it, so I purchased one of the 3rd gen units new from PP.
 
Electronic semiconductor devices in general have what is termed a "bathtub" failure curve.

Initially, there is a high failure rate, as the "good" but borderline devices get weeded out by random stresses.

Then for a long time, there is a low failure rate.

Eventually, especially with semiconductors, "wearout" mechanisms start to come into play, and things start to fail at a higher rate. Except in the most stressed devices, you can easily get between 10 and 15 years operation.

Some of the failures are:
migration of the "metallization" the forms the conductors on the chip surface, to where clearances are reduced below minimum,

cumulative effects of heat stresses gradually cracking the "solder" that holds the chip to the heat dissipating block in power devices, so that the chip overheats.

In non power devices, the difference in thermal expansion between the plastic encapsulation and the chip may eventually crack the tiny wires that connect the chip to the solder pads, because those wires may be encapsulated by the epoxy case material.

In power devices, plain thermal expansion and contraction may eventually fatigue the wires, or even solder joints.

If the parts become overheated, or even if not overheated, eventually the "doping" of the chip, the materials that are heat-diffused in to form the "works" of the chip, may diffuse further, and ruin the internal structure.

Many recent products. and PP are among them, rely on programs stored within the control unit in some form of "ROM" (read only memory). The newer types are not actual physically changed to write the program into them, they rely on stored charges which will eventually decay and "forget". (Old ROMs used "fuses" that were actually blown open, and are probably somewhat better as far as decay of the program storage)

Manufacturers work hard to make those failure mechanisms less of a problem, but 25 year old solid state electronics are getting toward the "suspect" stage in most cases.
Concur. Very well summarized thanks.

There are exceptions.


Ditto.

NASA built as many as they could into such kit as "Voyager" DEEP space craft.

Sure did pay-off. But "cheap" it had not been. Glad they at least published, though!

Minor curio was that "at one time" a return to vacuum tubes was proposed for spacecraft. Open air ones, even.

Because one didn't have to WORRY about a tube failing over going "gassy" ...when "open air" was a damned hard vacuum .... at no extra charge.. and for a VERY long time!

No doubt some Government chair-warmer shot THAT idea down off the possible contamination off launching through a cloud of lunar green cheese flakes and mold?

Cautious, some folks can be when a gummint retirement plan is at-risk!

:D
 








 
Back
Top