Jump to content


Photo

softstart waveforms


  • Please log in to reply
17 replies to this topic

#1 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 10 February 2006 - 09:26 AM

Hi All,
I am working on a soft start ( full bridge three phase) and wondered if you could help. I am confused by the wavefroms as they look a very ragged sine wave and not smooth(current waveforms) I am driving a big inductive load with passive pfc caps after softstart( a big nono but its run for years) pf looks to be down as 1 phase has 300A in and 140A after caps!
Does anyone know of good waveform pics i could view to see whats normal?
Many Thanks
Colin Heath

#2 jraef

jraef

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 683 posts
  • Gender:Male
  • Location:USA, California

Posted 10 February 2006 - 06:48 PM

Well, for one thing, since it is highly lookled upon as a BIG no-no to have caps downstream from a soft starter, I seriously doubt you will find any information on what the waveform would look like. That's kind of like asking what happens to a frog's skin once you stuff it in the gas tank of your brother's car (I picked that analogy because my sister actually experimented with that).

Seriously though, the rapid rise time of the SCR switching interacts with the charging current of the capaciotrs in that the cap looks like a short circuit to the SCR, and the SCR looks like an instantaeous switch to the cap. The two actions either destroy one or the other of the devices, or set up a severe resonnance in the circuit. Since as you said, this has lasted for quite a while, that means your SCRs are surviving the severe dV/dt and you therefore have a resonnant circuit, so who knows what that current waveform will look like or what it means. Also, the reasons why the caps are often destroyed is that the resonnance tends to overheat them. Your caps may not be fully functioning anylonger anyway. Did you test them?
"He's not dead, he's just pinin' for the fjords!"

#3 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 10 February 2006 - 07:37 PM

Hello Colin

Welcome to the forum.

You did not specify what the load was, is it a motor , or is it some other sort of load, i.e. and induction furnace?

Are the capacitors in circuit while the soft starter is ramping the voltage, or are they only applied after the soft starter has reached full voltage?

As jraef stated above, the connection of power factor correction capacitors after the SCRs will result in a high charging current into the capacitors when the SCRs switch ON. This wil cause a rapid rise of current and will often exceed the rated rise of current of the SCRs causing them to fail. If there is sufficient inductive reactance in the supply to the SCRs, the rate of rise of current (di/dt) may be limited sufficiently for the SCRs to survive. The other problem is the resonant circuit fromed by the capacitors and the inductive load can result in very high ringing voltages which can damage the SCRs and/or the capacitors.

You mentioned "full bridge three phase". This suggests that you may not in fact be operating the load on AC, rather on DC. If that is the case, then ignore the comments above as the resonance etc is not an issue.

Assuming that we are considering an AC load, you are concerned that the capacitors have failed because the current before the capacitors is 300A and after is 140A. Generally, when the pfc capacitors fail, there will be sort circuits developed within the capacitors. There is often a recovery mechanism built into the capacitors that protects the supply from the short circuit current. This can be in the form of fuses which disconnect the failed cell, or the cell may be made up of a metalised film such that the metalisation vapourises (self healing). In most cases, the net result is that the capacitance is slowly reduced over a period of time. This will cause a reduction in the capacitive current.
Excess current into a capacitor would usually indicate supply harmonics as the capacitor impedance reduces as the frequency increases and the capacitor effectively shorts out the higher frequency harmonics.

There are a lot of unknowns, perhaps you can help to fill in the gaps and we can comment further.

Best regards,

#4 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 11 February 2006 - 01:05 PM

Hi,
Thanks for the help so far. The load is 3 induction heating coils(in delta). This is contact induction as in the core connects with the platen to be heated and froms the secondary. It is running AC. I would have thought the capacitors would be sized so that the resonant freq was well outside the 50Hz it runs at?
I will test the caps at work for value to see if they have degraded. Im assuming a normal capacitance meter will work just for indication of difference?
The caps are connected in the circuit at all times!
The devices are well over spec which is why they are handling this ok i guess.
I have attched pics of the 3 phases(notice the one really different one)
Many Thanks
Colin Heath

Attached Files



#5 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 11 February 2006 - 01:06 PM

sorry that image is huge ill have to resize the others before posting
colin

#6 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 11 February 2006 - 01:18 PM

next pic

Attached Files



#7 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 11 February 2006 - 08:09 PM

Hello Colin

The first current waveform shows a resonant load at a frequency above the line frequency. There is an initial current peak which is the inrush into the capacitors and then the first part of the waveform shows the resonance. The second part of the waveform down not appear to have the same resonance.

The second waveform is the standard current waveform into an inductive load.

Looking at the two waveforms, I would conclude that on one coil, there is capacitance and on the other two there is not.

The current waveform comprises current flowing in the two coils that are connected to that phase, first one coil and then the other. There is an overlap in the center of the waveform depending on the conduction angle of the SCRs.

The first waveform appears to have a resonance on the first coil but not on the second. The second waveform does not show resonance on either. I suspect that the third waveform would have an initial current spike followed by a smooth curve followed by a resonant curve.

It would appear to me that there is an issue with the capacitors.

Best regards,

#8 jraef

jraef

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 683 posts
  • Gender:Male
  • Location:USA, California

Posted 11 February 2006 - 08:27 PM

Just out of curiosity, why do you have capacitors in a heater circuit at all?
"He's not dead, he's just pinin' for the fjords!"

#9 jraef

jraef

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 683 posts
  • Gender:Male
  • Location:USA, California

Posted 11 February 2006 - 08:29 PM

Oh, never mind, I just noticed the word INDUCTIVE. Duh.
"He's not dead, he's just pinin' for the fjords!"

#10 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 11 February 2006 - 10:17 PM

Hi Marke,
Now i see that's what i was missing. Running 3 phase you get overlap as the time is 6ms(approx) instead of 10ms.
I sort of noticed the resonant section as i build tesla coils (air cored resonant transformers) for a hobby but couldnt understand why it didnt decay like a standard damped wave oscillation. Now i do!
I have attached the other waveform but also what you say makes sense. the standard waveform reflects partially across both outputs of the caps it is attached to. The lack of capacitance would also explain the poor power factor on 2 phases and not the other. I realise now that the caps shouldn't really be there at all but the degrading of a few of them would explain these problems. just as a note there are 3 of them across each phase in parallel (3 400uF caps) meaning 1200uF each coil.
This would explain why on the other machine we have with same system has three ragged waves(not as bad as ones i posted) rather than the wave i expected which is the standard wave you mentioned. So it appears with the sytem we have i need 3 wrong waves rather than correct. i will however in the future try to get the system changed.
also any idea on a good way to check the caps in case they are too leaky to check with standard meter?
Many Thanks
Colin Heath

Attached Files



#11 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 11 February 2006 - 10:35 PM

Hi Colin

The way that I would check the capacitors, is to connect them to 240 volts or 400 volts AC and measure the current into them. If they are correct, you will have the right current. (Easy to work out!!) If they are greatly different, then you have a problem with them.
Using a capcitive meter can be a problem as the capacitors will could an internal blead resistor and also you are not checking them under full voltage. There could be an issue that does not show up at low voltage.

Best regards,

#12 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 12 February 2006 - 04:43 PM

Hi Marke,
ok this problem was really bugging me so i popped in for a couple of hours today(on a sunday!) I worked out the reactance of the cap then did as you said. I setup on 240v with variac. i went slowly upto 100v no current at all. Then ihecked with a bulb in series and still nothing. I then looked at top of cap and it's seriously swelled. Luckily they have an inbuilt over pressure disconnect and this has happened on 2 out of the 3 banks of caps so it was exactly as you suggested. I will order new caps for now while i work on the best way to do this properly. we already have active pfc back at the substation so perhaps i can just switch the caps in at full voltage as suggested?
Well thanks for the help so far.
I will update when i fit caps and get good waveform pics and it may help newcomers like me.
cheers
colin

#13 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 12 February 2006 - 05:56 PM

Hi Colin

When using a soft starter with a motor application, we always use a separate contactor to switch in the capacitors after the starter has reached full speed. This way, there is no significant harmonic current applied to the capacitors and they last a lot longer. (so do the SCRs!!)
If you application operates primarily at full voltage with a voltage ramp up, then you could do a similar thing. Otherwise, I would use higher voltage capacitors and detuning reactors to limit the harmonics created by the switching of the SCRs.

Best regards,

#14 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 12 February 2006 - 06:08 PM

I think the scr's should be fine. they are rated at 1400v and 400A continuous although i havent spent that long looking at switching spikes yet. The application heats for2 seconds then off for 2 seconds so the ramp has to happen pretty fast to make best use of time window for maximum heat. I will look at detuning reactors to prevent resonance and possibly uppiing the voltage of the capacitors. I am now concerned how the other caps are holding up. i think these machines are quite old though so they may have survived ok. well all i can say is this has been a really interesting problem with a steep learning curve. Hopefully another step in the direction of power electronics for a living.
Many Thanks
Colin Heath

#15 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 12 February 2006 - 09:08 PM

Hi Colin

I would not worry about the SCRs if they are still controlling OK. If the SCRs fail, they fail as a short circuit so are easy to test.
There are two issuses, the harmonic current into the capacitors, and the inrush current into the SCRs. The detuning reactors will help with both.
With your fast cycle time, I can see why they would not use a contactor to switch the capacitors on and of. It would not last long at all. I would expect at that rate, that detuning reactors would get very hot also. I wouls suggest that you need to put the whole process infront of the capacitor supplier and get them to supply the capacitors and detuning reactors. If they do not know how to do this, find someone who does, otherwise you will be out of the fat and into the fire.

Have a good day,
Best regards from New Zealand,

#16 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 13 February 2006 - 06:34 PM

Hi Marke,
It turns out these caps have lasted 10 years which isn't bad at all for pfc caps. So for now ill replace the caps then look at other options as you say with a good company. I think this is still worthwhile as apparently harmonics on the mains have been trouble with this sytem for some time.
Cheers
Colin

#17 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,602 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 15 February 2006 - 09:18 AM

Hi Colin,
Thanks for the update. Looks like it is all under control now,
Have a good day,
Best regards,

#18 colinheath

colinheath

    Junior Member

  • Members
  • PipPip
  • 9 posts

Posted 15 February 2006 - 06:02 PM

I have the caps on order and will let you know how it all goes.
Cheers
Colin




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users