Hi iGenius,

The welding did not do any harm (if you did not get metal drops inside the unit, but then it wouldn't work from the beginning). Period. There is absolutely nothing to discuss there.

The blocked ventilation should have led to a warning or a fault signal. The VLT 6000 HVAC range, for instance, would have produced alarm 29 (high heat sink temperature). Did anyone see such an alarm during operation?

Then again, inverters do fail for hundreds of reasons. But I think that the blocked air is the culprit.

I tend to see it just about opposite of skogsqurra.  Modern drives are well protected against high-temp operation and would have produced a large number of bothersome high-temp faults if the air blockage was problematical.

On the other hand, depending on the type of welding done, damaging currents could easily have been induced into some drive components causing stress damage that only appears over time, such as a month.  Many electronic components are vulnerable to this kind of failure mode.

Having said that, all of us are simply speculating.  Why not send the drive back to Danfoss and have them analyse it for cause of failure.  They would probably want to do this for warranty claim purposes anyway.

Then we can know with greater certainty.

I tend to agree with DickDV on the welding issue. There are concerns with this and they mostly depend on where the ground conductor for the welder was placed. Even so, damage could still be done if the ground was placed right next to place to be welded if the drive ground was hooked up (or anything else for that matter) when they were welding.

I would also think the drive would have tripped out on over temp or a thermal shut down, at least once if the blockage was causing problems. Then again, maybe they do not have provisions for this. Smaller semi-conductors often have a thermal limit where they shut down to avoid damage. Not sure if the drive your using (or any drive for that matter) uses any such components.

i wouldn't have thought that danfoss would be that interested in a warranty claim for a drive that had had bits welded onto it...unless you are a good customer

Hello iGenius,

DickDV & buzzp appear to have covered it well. Component failure resulting from transients that occured some time ago is not uncommon at all. In fact I could list several dozen examples of of this given the time and space.

The comments they have made in relation to nuisance trips arising from the blocked ventillation are also extremely valid.

Regards,
GGOSS

Hello all,

I think that I have to clarify a few points regarding the welding and what can happen:

First. Welding can be resistance, gas, arc etc. We do not yet know what kind of welding took place.

Second. Internal components could not have been affected by currents and voltages that occurred during welding. I think that it is safe to assume that the welding was done before the inverter was installed (brackets needed to install it) so there were no cables attached to the inverter.

Third. Induced voltages? A normal arc weld with a few hundred amperes can not induce any appreciable voltage in the traces on the PCB. There can be a few hundred millivolts induced in the cabling, but these voltages will only hit components that are designed to withstand normal mains voltage. Even a resistance weld with 10 kA (clearly overkill) does not present any problem.

Four. I have designed thyristor controlled resistance welders where the copper bars with 16 kA peak current passed close (50 mm) to the processor and control board and we never had any problems with induced voltage. And, of course, no destruction of components.

Five. The "expose now, failure later" theories are valid when you discuss ESD damage. But it is not possible to envision any such damage given the welding situation.

Six. My opinion is also that it was the thermal blocking that caused the problem. We do not know what Danfoss inverter we are discussing, but the one I mentioned does not trip on high heat sink temperature. It just outputs an alarm. I am not so sure that this alarm was taken seriously. It is possible that it was just ignored.

I said in my first posting: "There is nothing to discuss. Period" And I did say it with the purpose NOT to have this discussion. I failed, obviously.

skogsgurra....if you have a 6000 that doesn't shut down on high temp then you have a faulty one. They all shut down on high temp. The early (pre 1999) ones also used to shut down & lock (you had to cycle mains and then reset rather than just reset)

iGenius.....if the controls power up you can look at parameter 604 which will tell you the number of overtemps the unit has seen which might shed some light

I wouldn't take a welder within 10 feet of a new drive and would be amazed if danfoss even look at it as a warrenty claim

That no drive manufacturer would seriously consider warranty replacement of a drive that has been welded on speaks volumes as to the potential for damage to internal components.

And, as far as drives that don't trip on high heat sink temp, I doubt that any responsible supplier would design a unit that way, but, if you happen to have one, note the brand and choose something else next time.

If your drive only had the capacity for three faults, I would think high temp would be one of them.

DickDV....you are right, no-one would make a drive that doesn't trip on high temp, as a clue, Danfoss drives have 25 or more faults that will always trip it and maybe twice as many that will give a warning and can be programed to trip it.

Re the warrenty, I don't think danfoss have commented as yet but i would be surprised if they even considered itas a warrenty claim......welding bits on is abuse even if it didn't cause the problem

My take on Skogs replies:
1. You are right we do not know what kind of welding occurred.
2. I agree that cables or grounds may not have been attached during the welding for the sake of the drive, however, a hot and return have to be attached to weld (assuming electric welding). If the ground clamp was placed on the bottom of the drive then, of course, all of the current has to make its way to the ground clamp. The paths it takes to get there are all over the drive. These multiple paths could present problems with components on the board. They can also induce voltages on parts close by to a level they are not designed for (as little as 5-10volts or maybe smaller depending on the processor). This could have weakened components that are normally there to protect in such a case to where when the next incident the component failed. Very easy, to me, to see happening.
4. 16kA flowing within 50mm of a processor is not a good idea. Typical in these types of cases, things appear to work ok but occasionally, something bizarre happens that is not explainable. Attribute this to 16000 amps flowing next to the processor.
I am not saying your statements are entirely wrong but welding can and will damage components. Any welder at an electrical shop can probably tell you cases of welders damaging components.
In any case, I still stand firm that welding could have caused the drive to fail later on.
  
 

Hello again,

Sorry that I seem to have upset you. But the fact is that you cannot hurt a modern frequency inverter by welding on its case (if you do not let molten metal into it or overheat it through prolonged welding). The reason is that there is isolation between the power electronics and the chassis. There is also isolation between the electronics and the power electronics. The electronic parts are connected to ground in one point or (if grounding is done around the edges) to a solid piece of plate or aluminium which does not develop voltage gradients as a result of the welding current. That is part of the EMC strategy and no inverter will survive its own EMI if you fail to do that part right. "Connected in one point" means that there is no way for the welding current to enter and leave the active parts - hence no current.

Regarding thermal tripping. My personal experience is only from the 6000 HVAC and in it, parameter 411 decides if the inverter shall tripp on overtemperature or not. If this has been programmed not to trip (and I can very well think of a situation where it was decided to program it not to trip instead of cleaning the air intake) then it simply will not trip.

I am afraid that this question has turned into a "conversation piece" and I do not mind at all. But the main issue was/is to help iGenius with his problem Are we doing that?

skogsgurra.....411 does not stop it tripping if it get too hot

Setting 411 to 1 will make the drive reduce frequecy & output power as the temp approches the trip level in an attempt to stop the temp rise. If this fails and the drive hits the trip temp (80 or 90 deg) then it trips and can't be reset until it cools down.

iGenius...got an update for us?

Yes panelman. You are probably right.

I never had that situation. Never ran it that hard. Does it really trip if the frequency and current reduction doesn't help. Manual incomplete?

We need feedback from iGenius.

skogsgurra

If reducing the switching frequency and current don't stop the temp hitting 90 deg (maybe 80) then it trips and stays tripped until it cools to i think 60

You are right, entertaining as this is we do need an update...over to you igenius

Let's go back to basics fellows.  What components failed?  Then we may be able to determine the cause.

Lots of interesting replies here.

We used to install a lot of Danfoss drives, Danfoss strongly recommend the fitting of a back plate if the drive is not mounted directly, to ensure all cooling air is dragged passed the cooling fins.

Anyway, why on earth would anyone weld brackets onto a sophisticated piece of equipment???   Channel bar (unistrut) is a wonderful thing which only requires nuts & bolts!!

Alan

Hi to everyone! Sorry for only getting back to you now but I was out of the office for a couple of days.

I appreciate the enormous response I got and must admit that I am somewhat overwhelmed but the varying opinions. We take note of them all and this really helps us to view this problem from all angles.

I regret to inform you that I cannot give you much final feedback at this point in time. You need to keep in mind that I work in South Africa, and in this wonderful country everything happens a little slower than in the rest of the world (we call it "Africa Time"). We are still in the process of conducting our own investigation and Danfoss will assist with this. All of your comments will be used to assist with achieving an accurate outcome.

Some people enquired about the type of welding - we used electric welding. And no cables were connected to the VSD during the welding process.

As soon as we have a final verdict, I will inform you and also give you the reasons for our final conclusions. But as I noted, this might still take some time.

Kind Regards

Thanks iGenius. We are waiting "African way". So take your time.

V.F.D's just aren't reliable in any working conditions. They are very complicated electronic devices, and unless you are an electronic wiz they are not servicable. I would suggest switching to a V.S.D. I did. You should check out payback.com. I must sound a little one sided, but this mystery problem happens all the time with v.f.d's.

ejnjkj

"V.F.D's just aren't reliable in any working conditions"

We've been using them with for many years with few problems, particularly modern types.

“They are very complicated electronic devices, and unless you are an electronic wiz they are not serviceable”

The old versions were easy to test & replace individual parts, although I agree modern types do require specialist attention when they fail, although they are very reliable.

In my experience modern electronic devices are much more reliable than older electro mechanical / mechanical devices.

Alan

ejnjkj
 
In Europe VFD and VSD is pretty much used interchangeably. What's the difference in your part of the world?

------------------------------

If we learn from our mistakes,
I'm getting a great education!

looking at payback.com it appears that his idea of a vsd is some sort of electro-mechanical clutch used with a full speed motor

what happened to the original poster? some feedback would be nice

The Payback product appears to be an eddy current coupling rather than a mechanical clutch. It must be very lossy in the coupling. Although torque transmission doesn't seem to be a quoted parameter, from the principle of operation I doubt that it will have very good torque characteristics, which would limit its application to fans and similar loads.

It is an interesting idea, but the addition of more bearings into the system plus the control electronics doesn't seem to eliminate the perceived reliability issue, it just moves it from one place to another.

------------------------------

If we learn from our mistakes,
I'm getting a great education!

scottyuk
Yes eddie current drives are for variable torque applications(and in my opinion the only way to do it), and are not used on all constant torque applications because of slip losses. So there are lots of places I would install a VFD and a heavy duty inverter rated motor.But only where a VSD will not perform. This is all way off topic from igenius(sorry). And my first statement might have been little harsh towards the VFD.

Scottyuk must have stock in a network of electric motor rewind shops.  Either that or the electricity in his area must be free!

Only use an eddy current clutch on variable torque loads???  Mercy!  I haven't heard a comment like that in at least 15 years and I hear a lot of comments!

Scottyuk, I'd be genuinely interested in a rational explanation of why you hold that opinion.  I'm not trying to be offensive but I am a bit stunned by your statement and am interested in why you have this view.

Sounds like fighting talk to me says he standing well back

I certainly don't want a fight here.  I'll just withdraw my comments, scottyuk.  Sorry.

I was so surprised at your response that I guess I just went over the edge!

So, again, my apologies.

I'm sorry.....I'll have to get one of those smiley things to show what is not meant to be taken seriously

Engineering debate, differences of opinion and exchange of ideas is what this site is all about so carry on

I'd still like some feedback from the original poster...iGenius where are you?

got one

Hi DickDV,

Apology declined - none needed!

I was only observing how the Payback unit appeared to operate from observation of the sectional drawing on the vendor's website. I wasn't advocating use of it on a variable torque load - my post actually said "It must be very lossy in the coupling". But that appears to be what Payback are using it for in order to get variable speed to the load. The high losses presumably account for the copper lining material that they highlight as a feature of their product.

It sounds like I might have to withdraw my comments and allow you to post a more knowledgable explanation of use of eddy current clutches. They are not somthing I've worked on to any great degree - perhaps half a dozen times - and I'll hopefully learn something new from someone who knows better.

Please keep posting - I enjoy reading your posts. I have no problem at all when someone tells me I'm wrong: it proves they are reading my posts if nothing else!

------------------------------

If we learn from our mistakes,
I'm getting a great education!

As I look back over the thread, I see that, in addition to being a bit blustery in my disdain for eddy current clutches, I also mis-identified the source of the comments I was responding to.  Turns out it wasn't Scottyuk but rather ejnjkj.  Guess I was having a really bad day.

Regarding the eddy current clutch, while it is functionally acceptable, the efficiencies are terrible.  Even on variable torque loads, the motor is fully magnetized at all loads which, compared to a VFD is wasteful.  The clutch itself functions to pass motor torque directly to the load while reducing the speed based on the exciter current level.  At maximum speed, the clutch still slips around 3% usually, and as speed is reduced, the losses go up accordingly.

The speed-reduction losses can be calculated bases on the familiar equation     hp = torque x rpm/5252 where torque is the load torque in ft-lbs and rpm is the difference between the motor rpm and the output shaft rpm.

It is easy to see that at 50% speed, the losses in the clutch alone are equal to about half the motor hp.  Is it any wonder that larger eddy current clutches are water cooled?

By contrast, an electronic inverter will consume only the required load hp plus about 3% for inverter losses plus a little for motor magnetization.

I do a lot of stamping press conversions from eddy current clutch drives to straight AC drives and find that the energy payback is often 12-15 months where energy costs are in the vicinity of $.08/kwhr.  If the building happens to be air-conditioned (not common), the energy cost recovery is much faster due to reduced cooling load.

   If you were to draw out a fan curve comparing VFD to a VSD( I am using specs from the payback drive) with Kilowatts vs load. You can see at 50% the VFD is under 5% more efficient. That’s pretty close.   
   Having the motor energized and spinning full speed is not as wasteful as you think, because without a load it pulls very little kilowatts. The power consumption throughout a fan curve(variable torque) are really close to a VFD, although 3%-5% less efficient.
   
   On a constant torque apl.(stamp press) I can not argue with you, because there are kilowatts being abused.

   Dickdv-This comment was supposed to be for me?
“Scottyuk must have stock in a network of electric motor rewind shops.  Either that or the electricity in his area must be free!”
I must disagree. I would love to see a motor harmed in any way by an eddie current drive!! I thought everyone knew about VFD’s causing premature motor failure due to harmonic distortion and heat. And with 4% more energy being used, no one will be going hungry.
  One more point that is being overlooked is maintenance. Sure you save a little energy with a VFD. But what if you’re the guy they point to when it burns up or keeps dropping out of service for no reason. It’s nice to know that I can buy the entire electronic controller for a 200 H.P drive for 195 dollars!

hello ejnjkj

I am interested in your comments that the eddie current coupling is almost as efficient as the inverter drive system. Is this what you mean, or am I missing something here?
In my understanding, if we look at the inverter drive system, we have the motor losses, plus a small amount due to harmonics etc, and minus an amount due to the reduced fluxing, (assuming a modern drive). We have perhaps 5% losses in the inverter itself, but we have no additional slip losses.
So if we take an application where the load is running at 75% speed and 50% load, then the motor losses would concievable have reduced by a small amount, the slip losses (in the motor only) would be in the order of 2% and the inverter losses would be in the order of 2.5%

With the eddie current coupling, the motor losses are not altered. The eddie current coupling and control has losses that can be as high as the inverter losses (or higher) plus there is the addition of the eddie current slip losses.
In the example above, we have an additional slip loss of 25% of the shaft power which is an additional 12.5% (assuming that the basic losses of the eddie current coupling are of the same order of magnitude to the inverter losses.)

In a constant torque application, operating at half speed, you would have the motor producing full torque at full speed, (full load) the eddie current coupling would be dissipating half the motor rating and the load absorbing the other half.

Best regards,

Mark Empson
http://www.lmphotonics.com

ejnjkj, I agree with you that the load kw on a variable torque load as a pump or fan drops off dramatically as the speed slows so the losses in the eddy current clutch drop dramatically too.

However, your statement that no-load kw in an induction motor is very low is not really true.  Typical no-load currents in induction motors run between 22 and 35% of full-load current.  This represents considerable kw losses simply to magnify the motor before any useful torque is produced.

Also, modern AC drive losses are running about 2% of load kw not 5% as stated.  In addition, properly designed, an AC drive and motor need not be a high failure rate item even over 10+ years.  After all, even tho you can replace an eddy current clutch exciter circuit for $195, the clutch does occasionally need bearings and, if memory serves me right from my motor shop days, it will take a lot more that $195 to replace bearings even on the smallest clutch.  And, in reasonably normal industrial use, the bearings will probably need replacing in 7-8 years if not sooner.

But, anyway, if clutches are what you are comfortable with and you have good sources for them, a variable torque load would be the best possible application for them.  Also, applications that don't normally run with much speed reduction are a better choice too.  Many pumps run like that.

Back to the original problem of a failed VFD....
There are as noted above many reason for a drive (or any device to fail). It may very well turn out that neither the welding or the blocked fan was responsible.

Having said that though , I am in the camp of the blocked fan.  Although most electronic starters have thermal switches designed to trip at a given temperature there are usually (in my experience mounted on the heat sink or the actual switching device (IGBT, SCR...)

My practical experience  is that thermal damage usually occurs over time, at lower temperatures and effects the control board rather than the power components. This means that a unit can be damaged due to high temperatures but never trip on "over-temp"

Hi to everyone!

We have finally finished our battle with Danfoss in South Africa and unfortunately we have not received the answer we had hoped for.

First of all, my apologies for the long delay in this process. This is not a common problem that occurs with VSDs, so it took both ourselves and Danfoss a long time to sort it all out.

Below is the final e-mail received from Danfoss SA based on their investigation of the problem:

<Quote>

Hi,
 
With regards to our investigation about the welding of brackets onto the Danfoss Drive.
 
This welding onto the chasiss of the drive can be detrimental to the components of the unit.
The first damaging effect can be caused by excessive heat transfer onto the components.
The second problem is the stray currents that flow through the unit while welding.
The effects may not be physically noticable, however some of the components are weakened, and may fail at any time.
 
I would therefore recommend not welding any material onto the drive at any time, unless however all components are removed prior to the welding being done.
  
 
Dave Dyce
Product Manager
 
Danfoss (Pty) Ltd. - member of the Danfoss Group
Motion Controls
Johannesburg, South Africa

<Unquote>

In a case such as this, it is extremely difficult to determine exactly what caused the failure. Apart from the technical difficulties, there are such a variety of different opinions about this that it cannot be said with authority who is right and who is wrong.

We have learned a very important lesson from this - do not bring welders close to any sensitive electronic equipment.

Thank you very much to everyone who responded and gave their opinion. Personally, I have learned a great deal from each of the responses and I appreciate the assistance that you all gave in this regard.

iGenius,

Given the circumstances. What else had you hoped for? Of course, Danfoss takes the bite that you dangled right in front of them.

I can accept that the heat that resulted from the welding can have done something to the thyristors, even if that seems rather far fetched. But the stray current theme is utter bullshit.

I agree that you shouldn't do such things to inverters. It is not good practice.