Safety Issue (no joke)
|Paul Fallert||13/03/2018 19:32:35|
|89 forum posts|
In an allegedly serious posting on Practical Machinist Forum, a VFD unit left unused for several years was placed into initial service by a professional electrician on behalf of an end user. The test run went well. Then, on the next run, the capacitors blew up and shredded the VFD interior destroying the VFD. Fortunately no-one was injured. The professional then consulted the VFD instruction manual and to his shock only on page 300+ it stated that if the VFD is not used for 1 year (or left on the shelf at the stockist), the capacitors must be reformed, or else!. Additional time in excess of one year required more extensive reformation of the capacitors.
I checked Snopes.com to see if this were "fake news", with a "not found" resultant.
Does this requirement apply to all brands of VFD? The poster said that those who restore ancient radio sets are well-aware of the need to reform their capacitors or they will blow up the entire power supply.
I was aware that VFD's do not like to be unplugged at the supply end, but rather are to be turned off at the control panel.
Does this reformation admonition also apply to the newish brushless PWM DC motor supply units which contain large power supply capacitors?
|Mike Poole||13/03/2018 19:48:04|
3306 forum posts
I think this is standard advice for inverters. The factory I worked in used SEW Eurodrive drives and motors and they recommend powering up their VFDs regularly and describe the process to follow, I think it involved bringing the voltage up gently and then a soak time. I think the manufacturers and stockists turnover stock quite quickly but in a factory immediate support is required so a drive can sit in the stores until required which can be years. I believe that ball races do not have an unlimited shelf life in one position either.
|Andrew Johnston||13/03/2018 19:54:07|
6575 forum posts
It's a well known technique - definitely not fake news. However, it only applies to electrolytic capacitors. In an electrolytic capacitor the insulating film is a very thin layer of oxide on the metal electrode. Over time, with the capacitor unpowered, this film can degrade and effectively the capacitor becomes resistive.
To narrow the scope still further I think the technique is only applicable to the larger wet electrolyte can style capacitors, such as you may find in a power supply. I've never heard of the technique being associated with tantalum capacitors, or some of the other styles of electrolytic capacitor.
A quick (not Google) search brings up a paper by ABB about reforming capacitors in their VFDs.
|Mark Rand||13/03/2018 20:05:56|
|1239 forum posts|
It's a known problem with electrolytic capacitors. Not related to the issue that many computer motherboards had in the early 2000's. The insulator in an electrolytic capacitor is a thin layer of aluminium oxide that has been formed (oddly enough) electrolytically. This layer can gradually thin or corrode away if not maintained by a voltage across the capacitor. If said capacitor is supposed to be smoothing the voltage in a high current power supply, the current going through it if full voltage is applied to the thinned/missing oxide layer can cause sufficient heat to boil the electrolyte and cause the capacitor to self destruct.
I first came across the problem in the early '80s, when an number of our 24V KDG pressure transducers, that had been in the stores for a while, stopped working as soon as they were powered up. There were many of these used permanently in power stations and process plants, but ours were used intermittently, when we did performance tests.
It took quite a lot of 'forensics' to workout how they'd died. Once I'd worked out what was happening, we rigged up a power supply in the stores, with a cable and enough plugs for all the transducers and just kept them powered up. Never had any more problems .
Edited By Mark Rand on 13/03/2018 20:09:44
|Paul Fallert||13/03/2018 20:29:51|
|89 forum posts|
Thank you for these informative replies.
VFD's may have a significant downside if not properly maintained. Or purchased "used" or old stock.
Since this seems to be important, should a Warning! be published in MEW under Health and Safety?
Still unanswered is the ubiquitous brushless motor powered by PWM power supplies which may contain electrolytic capacitors. There have been MEW articles in the last couple of years about PWM, but I don't recall any capacitor warnings. Maybe N/A ?
|Mike Poole||13/03/2018 20:52:11|
3306 forum posts
Here is the procedure for an SEW Eurodrive inverter that has not been powered for over 2 years
1675 forum posts
Any chance of a link to the Practical Machinist post please, so I can follow it there.
Particularly of interest to me as I have several used VFDs in stock awaiting the next project.
Fortunately I also have a variac, so should be able to do the "reforming" after some more research.
Edited By peak4 on 13/03/2018 21:33:34
|Neil Wyatt||13/03/2018 21:37:42|
18992 forum posts
My inverter manual says switch on for 30 minutes to 60 minutes every year (without running something) if the inverter is kept in storage.
My advice would be to read and follow your VFD manual instructions.
|Mike Poole||13/03/2018 21:47:58|
3306 forum posts
I visited a business that repaired the power module for KUKA robots, the power module is basically an inverter for 6 axes. The first thing they did was replace all the large electrolytics as a matter of course. The cost of the caps is cheap compared to the trouble they cause.
|Mark Rand||13/03/2018 22:04:08|
|1239 forum posts|
I think it was actually in the other "other place"
8492 forum posts
Very much so - not all electronics respond well to being brought up slowly on a variac. The maker is the best source of information on what suits his kit. Horrible job I know but best read the manual for the particular model you have.
Having been close to two capacitor explosions I can say they're not particularly violent. Bang, smell and mess - yes. Shrapnel, perforated ear-drums, blast damage, emergency services - no. Both occasions were funny rather than scary. Of course nearby delicate electronics can damaged. In my experience neither explosion broke glass valves about two inches away on the same chassis. It was the scattered goo and shredded foil that caused most trouble.
Older capacitors are more exciting than new ones. WW2 condensers come in heavy aluminium cans that can take more pressure before popping. Modern capacitors are thin walled and fitted with vents and/or weakened tear lines. They rupture before the pressure rises much.
Although it can and does happen failures aren't all that common. I expect most of us spent years watching valve TV sets full of high voltage electrolytics without being injured.
Edited By SillyOldDuffer on 13/03/2018 22:23:00
|1504 forum posts|
Standard practice for Siemens Micromaster VFDs was to reform the capacitors if they had not been used for specified periods, using a Variac.
The time taken at each step varied with the out of use time, if 1 year or less then no reforming, if 1 - 2 years then full power applied for 1 hour before use, 2 - 3 years then the Variac is used to apply : 25% for 30 mins, 50% for 30 mins, 75% for 30 mins, 100% for 30 mins. Total time 2 hours
If 3 years or more then the Variac is used for the same percentage times, but the intervals are 2 hours. Total time 8 hours.
This advice was contained early on in the user manual, under "Electrical Installation"
This doesn't apply to 400 volt Static Phase Converters, (Transwave type) which use a different capacitor.
Edited By Robbo on 13/03/2018 23:43:33
|Paul Fallert||14/03/2018 01:07:48|
|89 forum posts|
"I think it was actually in the other "other place"."
Your link is the correct one.
|Paul Fallert||14/03/2018 01:12:00|
|89 forum posts|
I was reading on PM where the OP was having trouble repairing old oil-filled capacitors in an old Monarch lathe repair and it was recommended that he substitute a "non-reforming-required" type of capacitor. Orange drop or film type. Any possibility going in that direction?
|jimmy b||14/03/2018 06:59:16|
780 forum posts
Kind of makes you realise how important it is to READ the instructions......
|Norfolk Boy||14/03/2018 07:56:28|
|66 forum posts|
When I was at college studying electronics some bright spark put a small electrolytic across the bench power supply. Nothing happened for about 3 minutes...then Crack and the room was full of snow. Amazing how much fluff could come out of something so small.
I have been looking at VFD's recently having just bought a couple for the lathes to convert and I read about the reforming. One instruction said after reforming let it sit as that is the time the oxide layer is then rebuilt.
Also advised by a very helpful supplier that they just put it on a 110Volt supply for a while first and that was broadly across all makes.
Type "abb capacitor reforming guide" in google for advice on their processes.
|Joseph Noci 1||14/03/2018 09:16:11|
|1070 forum posts|
Having been in electronics for 40 years or so, mainly military, but quite a bit of industrial electronics as well, power supplies in the 10's of KW range, dc-dc and AC invertors, etc, I have in all that time NEVER had an electrolytic explode due to reforming requirements. Military stock often sits in stores for years, esp large power supplies that would run off 400Hz 3 phase 115volts, big electrolytics after the recifiers, etc. I have bought up surplus VFD from factory closedowns and stored many of them ( a few dozen in fact) for 6 years if not longer, and they just work on power on. Not to say what has been said in posts here is not valid, but I am not sure this should be a rush to the Health and Safety Gods...These components are generally in an enclosure, the VFD case, etc, so not really a bomb in your briefcase..If however you are working on the bench with the unit open, then common sense applies - you do not need a Health and Safety pamphlet/officer/MEW reminder, etc, when machining on your lathe, do you? And it is no more safe than Electricity on your bench...
That said, I have had MANY electrolytic explosions and in every case this has been due to poor capacitor rating for the job. Savings in production ensure dumb choices as to capacitor working voltage and ripple current capabilities, the two main killers of electrolytics. The third killer is temperature of working environment. Many of the larger electrolytics have a small rubber valve which allows excess pressure to escape. Most electrolytics have a wet or gel electrolyte, and with the escaping pressure goes some of this 'wetness' - excess temp or high ripple currents causing high temps will help the capacitor dry out, resulting in its Farad value decreasing drastically. High ripple currents also generate heat which does not help.
Electrolytics without the small rubber bung still release gases through the base of a radial connection type, and through the positive connection end of a axial connection type - both constructions have rubber insulation seals which vent.
Capacitor re-forming is normally applied to regenerate the capacitor back close to its nominal capacitance value, and this by default will permit the specified ripple current to be handled.
I mentioned the importance of capacitor working voltage - Generally the capacitor specified voltage should be about 20 to 30% higher than the max applied voltage. This ensures suitable inner plate arcing prevention, and appropriate plate oxide maintenance. It is folly to use an electrolytic of greater than 100% more than the applied voltage in an industrial application ( such as in a VFD) - Oxide maintenance is reduced over time, and the capacitor value reduces, as does its ripple current capability. It will eventually fail as will one with too low a voltage rating. You will note that in 230V mains rectifiers, the smoothing capacitor is normally rated at 400VDC - below the 100% limit.
As to SillyOldDuffers comments on the explosions not being dangerous - everything is or isn't dangerous, just be careful, and use common sense. Legislating it dangerous won't stop it from killing you, nor will pages of H&S gumph all over the place. That's why folk don't experiment and toy anymore - they are not allowed to or are too scared...
By the way, the small electrolytics (say up to 15mm diameter or so) are quite dangerous when they explode un-restrained. The case is small, but the aluminium wall thickness cannot easily be scaled down per the case size, so the case rarely splits - for a radial lead device, it just pulls off the rubber base bung and shoots out behaving like a bullet. Watch out for those...
Abused Tantalum capacitors, on the other hand, rarely lose capacitance or dry out or behave normally - they just become a low impedance or short circuit and burn up everything around them..This happens very quickly by exceeding its working voltage by even 5%....
My diatribe gives no advice as to reforming capacitors - have just never needed to do it, not even on an Old Phantom 3phase 115 VAC to DC supply to the cannons after 20 years in the store.....
edits to fix damn typo's
Edited By Joseph Noci 1 on 14/03/2018 09:19:45
8492 forum posts
As always Joe's post is good stuff. Just a slight point - I didn't say capacitor explosions weren't dangerous, I said "they're not particularly violent". You would likely be hurt if one went off in your face.
Trying to put scale on an explosion is quite difficult because most of us lack experience. None of them are 'safe'. What's a capacitor explosion like? Imagine someone setting off an old-fashioned British banger firework in your living room: not as bad as that. Much less dangerous than a model boiler explosion, and orders of magnitude less dangerous than a hand-grenade.
The first time I saw one go bang was funny because of the circumstances. A newly married friend had bought an old radio. Under sufferance he was allowed to keep it in a recently redecorated spare bedroom. This featured a new carpet, cushions, and all the other tidy appurtenances of an exceptionally house-proud female. We took the case off, put it on the floor and switched on. All seemed well - bargain - until one of the electrolytics let go. It made quite a mess. When his SWMBO arrived I felt it best to leave immediately. She was quite cross...
|Norfolk Boy||14/03/2018 10:31:28|
|66 forum posts|
The company I work for took it seriously enough to routinely (6 months) cycle all the power supply cards that were in the spares cupboards with the working units and swap to reform the capacitors.
When the accountants started running things rather than engineers they decided sending people out to do this routine was a waste of time. The result was not wholescale exploding capacitors (certainly never heard of it) but given the scale there was an increase in dead on the shelf power units when needed. Enough to reinstate the cylcle routine, If nothing else to identify duff spares early. This equipment was put in in the 1980's and is still working and doesn't have a huge failure rate in service. The degradation of the plastic runners holding the cards has been more of a problem.
|Andrew Johnston||14/03/2018 11:07:17|
6575 forum posts
Tantalum capacitors are a PITA as they tend to fail short. So much so that you can buy "special" ones for power supplies that are designed not to fail short. In reality they have an internal fuse that opens when they do fail short. When tantalum capacitors fail they often release a rather pretty purple smoke. At one company, after a significant number of failures, I was asked to take my experiments to another part of the building where purple smoke was less of an issue.
Personally I try and avoid electrolytics and tantalums if at all possible. Neither of them, electrolytics especially, are particularly reliable.
Please login to post a reply.
Want the latest issue of Model Engineer or Model Engineers' Workshop? Use our magazine locator links to find your nearest stockist!
You can contact us by phone, mail or email about the magazines including becoming a contributor, submitting reader's letters or making queries about articles. You can also get in touch about this website, advertising or other general issues.
Click THIS LINK for full contact details.
For subscription issues please see THIS LINK.