There should be electrical and mechanical interlocks on hoist circuits to prevent both contactors energising at the same time, it is wise to at least have electrical interlock as a minimum on any reversing type application.

Emgee

Sorry inrush Current

Yup. That was a cheap 500W single phase, Chinee hoist bought from Lidl (Switch replaced by next day delivery of a better unit from cpc.farnell.com). The VFD is a 23kVA Danfoss VLT series unit.

Disconnection is no problem at all for any modern VFD. The potential reflected transient voltages seen at the VFD output in normal running are as great as those that could be caused by disconnection. This is why the IGBTs have integral freewheel diodes to cope with that problem.

The C-E voltage is limited by the DC bus voltage and the flywheel diodes.

Edited By Mark Rand on 02/11/2017 01:11:01

Allen Bradley, Mitsubishi and IMO all state that a motor should only be disconnected from the drive output when the motor is at a standstill. That is NOT an exhaustive list, just a 10 minute search. If you need to switch the drive output, read the manual or contact the manufacturer.

Agree with Michael, NDIY and others. Unless you don't mind taking risks much better to check specifications if you can. Anyone else done the course where 'Assume' is defined as 'to make an Ass of you and me'?

One risk not mentioned is the wider effect of live switching – the effect of voltage spikes on whatever else happens to be connected. A VFD with beefy modern IGBTs might be immune to spikes, but can you say the same about the insulation of your motors? 3-phase synthesised by a simple pulse device is much more likely to generate voltage spikes than real 3-phase from an alternator. You may be applying the equivalent of a severe 'Megger' test to your whole workshop every time a motor is switched off…

Joe Noci's post made a lot of sense to me; following his advice is unlikely to cause trouble. The problem with Mark's "rant" is the number of unknowns in it:

• IF the manual doesn't explicitly forbid downstream switching, AND
• IF the VFD was designed in the last 20 years, AND
• IF the VFD has IGBT output stages, AND
• IF the VFD output stages have Current Monitoring

Dave

Yes but its interesting and topical stuff as we're all faced with solving the problem of making 3 phase machinery work on single phase without breaking the bank.

My experience and confidence is growing and totally agree about the instructions but I find most manufacturers have a default or "run out of the box" page. Once its going you can start and find your way around the settings and alter to suit your requirements. I prefer larger units and "turn down" the current settings to suit, most go down a long way

I guess I've stirred everyone else and myself up more than enough with my posts, so I'll say good luck with your researches, whether handbook based or practical experiment based. Please report back what you find out about the ABB units.

As a final datum, the manual for my Danfoss VLT 5000 unit (similar ones avaiable on EBay) states under output data:-

Switching on output Unlimited.

So I know that in my case, the manufacturer has actually stated that in the specs.

Regards

Mark

.

In that case, Andrew … May I suggest contacting 'Quantum Controls' **LINK**

http://www.quantum-controls.co.uk/home-b

I found them extremely helpful.

MichaelG.

I've worked with Siemens 420 series inverters for the last fifteen years and when I've needed to make changes or fit new units I still referred to notes I kept because there are so many variables that needed tweaking.

Martin P

I stuck an oscilloscope on my lathe's built-in VFD (can't find the blurb, think it's an 'inexpensive' Delta) and the output looks like this:

It's an approximation of a 50Hz sine wave built of 5kHz pulses varied in amplitude. It's only 50Hz on average. Notice that the peak to peak voltage is 840V. Zooming in from 2.5mS to 250µS you can see the sharp internals better:

So my particular VFD is pretty grubby. At 50Hz the waveform is full of spikes and harmonics extending from the 5kHz base to VHF. When the motor windings are correctly connected, the winding inductances smooth the spikes and the motor 'sees' 50Hz. What happens when the motor windings aren't connected properly is up for grabs. The spiky waveform is liable to resonate and create very high voltages. It's not hard to imagine a situation where a mis-connected motor winding accidentally behaves like a car ignition coil. There the contact breaker pulses all the energy stored in a condenser into the coil, causing it to poke out 10kV or so at the top. Even more volts when a spark lead is disconnected. (Ouch. You can guess how I know…)

I'm sure that not all VFDs are as spiky as mine. I guess that many of the cheaper single motor units are as bad because the spikiness doesn't matter much provided they're wired up correctly. It's possible to clean up the output of a VFD and I reckon units designed to drive several motors do just that.

Dave

Neil has just reminded me about how the company I worked for used their inverters, one piece of equipment had two three phase semi submerged pumps one running and one on standby i.e. after 24 hrs the running pump is stopped and the stopped pump is started up which then runs for 24 hrs when the process happens again, both pumps have contactors between them and the inverter and when the pumps are stopped the inverter is shut down to change over and the inverter is switched back on to ramp the pump back up to running speed.

Martin P

There seems to be some confusion about what the output of a VFD looks like. Here is a sketch of a typical VFD output stage, essentiallly a 3-phase bridge:

The 0V line is the same as the mains neutral and the DC_BUS is the rectified mains, around 320V for a single phase input. A motor is connected the the three terminals U, V and W.

Each pair of IGBTs can be in one of four states:

Both IGBTs off – the output will float and do it's own thing

Upper IGBT on, lower IGBT off – The output will be DC_BUS, minus the Vce(sat) of the IGBT

Lower IGBT on, upper IGBT off – The output will be 0V, plus the Vce(sat) of the IGBT

Both IGBTs on – exciting stuff, as the DC_BUS will effectively be shorted. Been there, done that as a result of a software boo-boo, short circuit current was on the order of 3000A.

Let's assume the the IGBTs are fed with 1:1 mark:space square waves, so the upper IGBT is on for t seconds while the lower IGBT is off, then vice versa. The output will be a 1:1 square wave. If this is filtered it will be a DC voltage at half the DC_BUS voltage, ie, mid-rail. This is essentially the "neutral" point of the 3-phase output. If all three arms are driven in this fashion then any motor connected will just sit there, as the same voltage will appear on each terminal of the motor, at zero frequency, and no current will flow.

Now suppose we ensure that the upper IGBT is on for longer than the lower IGBT; filtering will result in a voltage above DC_BUS/2. Similarly if the lower IGBT is on for longer than the upper IGBT the filtered voltage will be lower than DC_BUS/2.

If we're clever with the upper and lower timings we can get an output that, when filtered, looks like a sine wave. However, the output of the VFD (on V, U, W) is a PWM signal swinging from 0V to DC_BUS.. It is not a sine wave or anything like, as there is no filter on the output. Since the IGBTs are either full on or full off the dissipation in them is relatively low, compared to them being driven linearly to produce a sine wave output – essentially an audio amplifier.

So how does that explain the SoD waveform? The 0V line is not usually brought out, so I think SoD measured phase to phase. This will give a sinusiodal like waveform, albeit with the switching waveform imposed on it. The spikes and noise are a combination of two effects. One, there will be parasitic L and C which will give spikes during switching due to high di/dt and dv/dt. Second, unless one is using a high quality probe with proper "earthing", the probe itself will generate "spikes" and "noise". The bog standard ground lead you see on 99.9% of scope probes doesn't cut the mustard for high speed measurements. It has too much inductance.

Right, still got a long list of things to do this evening. The first of which is check on the medlars simmering on the hob. The plan is to make medlar jelly, but at the moment the liquid is looking like brown goo, not translucent red.

Andrew

Oooo! Andrew's post has given me much to think about. Did I measure phase to phase or not? I can't remember and, although I took notes, they're not in my usual book. So it's possible I didn't measure what I thought I was measuring; the point about parasitic LC rings true. It sort of aligns with what I was saying about the possibility of a mis-connected motor winding producing high voltages though.

Never mind, it's all educational. Back to the drawing board!

Cheers,

Dave

PS Mentioning Medlar Jelly flashed back me back 50 years to Sunday dinner. Mum used to make it. One year it turned out horrible and she gave up. Pity.