Maker Pro
Maker Pro

100VDC bus & arcing connectors?

eem2am

Aug 3, 2009
429
Joined
Aug 3, 2009
Messages
429
Hello,

We are supplying twenty 10W LED lighting fixtures from a 100VDC bus.........is it true that when we connect a fixture to the 100VDC bus, there will be such severe arcing between the metal contacts of the connector that the connector will not last more than a few months?

Can we alleviate the problem by using some kind of inrush protection , whereby the arcing current is limited by a resistor which gets switched out after the initial arc?

Does the attached schematic offer a solution?.....ie it gives us a chance to make the contacts when there is no current flowing because the PFET is OFF........then there can be no contact arcing because the contacts are already closed before current starts flowing?

Non arcing HVDC switch?
http://i44.tinypic.com/23wryba.jpg
 

Attachments

  • 100VDC contact switch.pdf
    8.7 KB · Views: 97
Last edited:

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
Making a DC contact should not be much of a problem. The contacts will be cold and will be approaching rapidly so the short arc will be quenched quickly.
Breaking a DC contact is difficult since the arc will start as soon as the contacts separate by the smallest amount and will continue as the contacts separate. If the supply is AC, then the arc will die when the current drops to zero so it is much easier on the contacts.

A soft start circuit can be made with a relay, two resistors and a capacitor, you may be able to modify the circuit you have given to soft start with the addition of a capacitor.
 

eem2am

Aug 3, 2009
429
Joined
Aug 3, 2009
Messages
429
We will add another switch which turns the bjt off before we wish to open the contacts...that way there will be no contact arcing at disconnection, because the blocking fet will be off.
Does that sound right?
 

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
There will be inductance in the cicuit, maybe a little, maybe a lot. This will have energy which will need to be dissipated by the power switch. A zener diode across the fet with a voltage lower than the fet's limiting voltage may do. It will need to be fat enough to take the energy without destruction.

Perhaps a zener across the load could limit the switch off voltage to a level that the switch contacts can stand.

Of course, you can go to a big contactor that will handle the load. On big installations they use air blast breakers which blow the arc into a loop thus lengthening it and keeping the heat away from the contact.
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
I don't see why there should be any significant arcing. LEDs are not inductive. The current into each fixture is only 100 mA.

If there is any arcing, it will only occur on connection and disconnection (mainly on disconnection). How often are you planning on connecting and disconnecting each fixture?

Can you show us a schematic of the lighting fixture?
 

eem2am

Aug 3, 2009
429
Joined
Aug 3, 2009
Messages
429
led light circuit is as pag 29 of this

http://www.ti.com/lit/ds/symlink/lm3421.pdf

..(vin pin protected against 100v in our cct, and ours is 10W...and its 23 leds in series)

from what you say all is well, but i cant find 100VDC rated switchs anywhere at a decent price
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
I see. So the fixture includes a switching supply that powers the LEDs.

Just as an aside, if you're always running the fixture from 100V DC, and there are 23 LEDs in series in the fixture, wouldn't it be possible to omit the switching supply and just use a dropper resistor? I guess there's some reason why not...power dissipation, maybe?

So that circuit has a capacitor, CIN, across the incoming power supply rail. That's going to cause a spark when you connect it to the 100V DC rail, as CIN charges up rapidly to 100V DC, and probably nothing when it's disconnected.

So you want to add a soft-start circuit in the fixture? That shouldn't be too difficult. The circuit in the PDF in your initial post won't do it. I prefer your idea of a current limiting resistor that gets bypassed by a MOSFET. You should also disable the LM3421 until the MOSFET is turned ON, otherwise it will interfere with the CIN charging.

You may need to worry about power and energy dissipation in the resistor though, and that depends on the CIN value. It looks like a typical CIN value is 4.7 uF. Assuming you want to limit the initial current to 100 mA, the same as the operating current, the resistor needs to be 1k and the initial power dissipation will be 10W. The voltage across CIN will reach 95V after about 3RC which is about 14 ms. It seems to me that you should be able to use a resistor rated for less than 10W considering that these inrush events will not be repeated often enough to cause a cumulative heating effect.

I've had a look on Digikey for resistors rated between 3W and 10W. Some of them are described as "pulse withstanding" and their data sheets say things like "superior surge performance", but this doesn't seem to be quantified anywhere, and that subject is outside my experience.

If you can tolerate a startup delay longer than 14 ms, which seems likely to me, you can increase the resistance of the surge resistor. That will decrease its power dissipation, but increase the duration of the dissipation. For example a 10k resistor will dissipate 1W initially and will charge a 4.7 uF CIN capacitor to 95V in about 141 ms, so a 1W component should be fine if you can accept a 141 ms startup delay.

Then you will probably want to use a MOSFET to bypass the resistor when the voltage across it has dropped to, say, 5V. An N-channel MOSFET switching the negative rail would give better performance per dollar than a P-channel MOSFET.

Have a think about all of this. If you decide that's the way you want to go, I'll have a think about the design of the MOSFET circuit. If you or anyone else has suggestions, please post them on the thread.
 

eem2am

Aug 3, 2009
429
Joined
Aug 3, 2009
Messages
429
yes, thanks, Also, i will consider the disconnection situation, which could be more problematic and harder to cater for.

by the way, does anyone know why high voltage DC bus's for lighting don't exist?...i mean it should make sense.....do the PFC in one central place, then bus high voltage DC from there on....what's wrong with this?
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Why do you think disconnection could be problematic?

I saw your other thread regarding DC voltage distribution for lighting applications. I don't personally know of any reason why it wouldn't work; I think the reason that it isn't often done relates to the fact that there's no standard for interchangeability, which would lead to monopoly supplier issues including lack of competition, and the lack of benefits to outweigh the logistical issues involved.
 
Top