250VAC, 4A fuse used at 700V input to SMPS?

[treez]

Hi,
Our 12kW SMPS has a 700V vin.
There are multiple SMPS “units” in parallel, each with a 4A fuse upstream of them.
We would like to just use a cheap 250VAC, 5x20mm fuse for this.
It is obviously not rated for 700VDC, but then again, its not going to flash over, so are we ok using it?

 

[betwixt]

It will still work because a 700V arc is unlikely to bridge 20mm but the time from breaking to isolating will be longer than normal while the fuse wire burns back to the length it can no longer sustain the arcing. Personally, if I had to use fuses in high voltage DC I would go for a longer length one, avoid any type with a spring construction inside it and only use an 'empty' glass one. (no sand or silica filling)

Brian.

 

[Easy peasy]

Ceramic ( sand ) filled 20mm fuse will be fine - they are designed to handle a lot more than you expect ...

 

[FvM]

a 700V arc is unlikely to bridge 20mm

Unfortunately not true, particularly for DC voltage where the arc doesn't self-extinguish at the next zero crossing. Once ignited by the blowing fuse wire, it can bridge considerably wider distances than 20 mm, depending on the actual short circuit current. That's why fuses have a breaking capability rating, different for AC and DC.

You can suppose that not all fuse series undergo a high voltage and particularly DC breaking qualification, also that a certain safety margin is involved with the specification. But it's unlikely that an arbitrary 5x20 mm fuse has a higher breaking capability than the best device on the market unless it implements a completely different technology.

A 4A fuse with > 700 VDC rating would be e.g. Littelfuse SPF series, 10x38 mm form factor. If you need only small breaking capability, a 500 VAC rated fuse in 6.3 x 32 mm fuse like Littelfuse 514 might work for you. A pure over current protection might be provided by the 5x20 fuse suggested by Easy Peasy, but it won't pass any safety qualification.

 

[Easy peasy]

We have performed a lot of DC fuse testing - at 12kW, 700V the source is likely current limited, 20A or so, sand filled ceramic fuses will handle this well, of course 38mm ( x10) would give a better margin ....

 

[KlausST]

Hi,

Ceramic ( sand ) filled 20mm fuse will be fine - they are designed to handle a lot more than you expect ..

I'd never recommend to use a device that is not retaed for this.
I agree that there will be a safety margin with every specification.
But what happens when the arc does not stop...and causes overheat, in worst case fire?
I don't want want to be responsible for this ... because I recommended it.

Klaus

 

[betwixt]

My assumption was that the distance would be sufficient as the continuous arc melted the wire back. My concern was that in a confined area such as a sand filled fuse, the evaporated wire would be confined into a narrow channel which under high temperature might create a plasma link rather than a direct connection. That is also the reason for suggesting anything but a plain wire construction should be avoided, to maximize distance across the break. Distance is the key factor so a longer fuse will always be safer.

To be honest, I've never experimented with DC arcs in confined spaces like a fuse body so I'm using intuition instead of research.

Treez, it might be worth checking with the manufacturers of high voltage fuses used in microwave ovens. They are mass produced and although expensive singly through retail outlets are probably economical to use if you buy from source in bulk.

Brian.

 

[Easy peasy]

I'd never recommend to use a device that is not retaed for this.
I agree that there will be a safety margin with every specification.
But what happens when the arc does not stop...and causes overheat, in worst case fire?
I don't want want to be responsible for this ... because I recommended it.

this is difference between knowledge gained through testing - and other noises ...

 

[FvM]

There's also a 1000 VDC rated 3A fuse in Bel Fuse 0ADBC 6.3x32mm series.

 

[KlausST]

Hi,

this is difference between knowledge gained through testing - and other noises ...

I assume the manufacturers have tested their fuses, too.
And if they are reliable to run at higher voltages - then this would be a benefit against other manufacturer's fuses.
Then they could earn more money by selling them at higher price with extended specifications.

Now why don't they do this?
I don't know. But maybe they test it on various conditions. Various series inductance, various load and source impedances, various resonances...some extreme conditions. And they can guarantee the function "at etiher case" to a limit...then they specify it to a recuced (safety margin) limit.

I personally could not recommend it, because, if I recommend something I feel responsible for malfunctions.

Klaus

 

[treez]

at 12kW, 700V the source is likely current limited, 20A or so, sand filled ceramic fuses will handle this well, of course 38mm ( x10) would give a better margin ....

This could be our permit to use an el cheapo fuse...ie, the 20A current limit means it wont be possible for an arc to form anyway i would have thought.....even in an air chambered 5 x 20mm fuse.

Also, their is a 7 Ohm cable to the device...so overcurrent will mean so much voltage drop across the cable that there wont be any voltage left at the load...not enough to do damage i wouldnt think.

 

[betwixt]

I think you confuse current and voltage ratngs.

Possibly the current is limited by wiring resistance but if the fuse blows the current is severed anyway. It is the voltage across the fuse that matters and it will probably be the full voltage from the supply. As it is DC, a continuous arc will form across the gap in the fuse wire and the arc will burn until the wire has melted far enough back that the arc can no longer be sustained.

You can't treat a fuse in a DC circuit in the same way as an AC one. When carrying AC, the voltage across the broken fuse wire will pass through zero twice per cycle and the arc will extinguish itself. With DC, the voltage is sustained so only the distance across the break is relevant.

Brian.

 

[treez]

Thanks,
So our “ultimate” (most upstream) power supply is two of the below EA-PSI ones in parallel. They will provide 700V, 6.9A to the 7 ohm cable which will supply the load. These EA power supplies will be programmed to current limit at 4A each.
(There are no DC fuses on the market rated between 650V and 1Kv, only 650V and 1kV and nothing between these)
Downstream of the cable…
For this test the load is 4.5kW and is fed by three SMPS’s in parallel which each draw 6.9/3 Amps max…hence the 4A fuse just upstream of each of the paralleled converters.
Since we are programming the EA power supplys to current limit at 8A total, would you think that we can get by with the 250VAC 4A fuses for this test rig? (it wont be going out to customer).
We realise the fuses may well explode if a short occurs, so we will place fuse covers over them and wear goggles.

Cheapest Fuse for 700V, 4A (Phoenix contact 1306305)
https://www.digikey.co.uk/product-detail/en/phoenix-contact/3061305/277-15211-ND/4457997
Power supply: (EA-PSI)
https://docs.rs-online.com/e92e/0900766b816cab34.pdf

 

[Easy peasy]

DC fuses for PV strings are becoming quite common now, to 1kV rated I think ...

do you mean 2 x 4A fuses in // to give you the 8A rating ? - watch out for inrush currents

A standard 32 x 10mm HRC fuse will survive a lot of abuse - get a 16A one ( and a few spares )

- - - Updated - - -

or this:

https://uk.rs-online.com/web/p/offset-tag-fuses/0413507/

- - - Updated - - -

you can also buy these at any local electrical supplies shop ...

 

[FvM]

Since we are programming the EA power supplys to current limit at 8A total, would you think that we can get by with the 250VAC 4A fuses for this test rig?

Instead of placing an effectless fuse, why don't you simply omit it?

 

[treez]

do you mean 2 x 4A fuses in // to give you the 8A rating ? - watch out for inrush currents

Thanks, the fuses are not in parallel...there are three SMPS's in parallel...and each SMPS has a 4A 250VAC fuse just uptream of it.

Since we are programming the EA power supplys to current limit at 8A total, would you think that we can get by with the 250VAC 4A fuses for this test rig?
Instead of placing an effectless fuse, why don't you simply omit it?

Thanks yes at this stage, we could omit it..(if we trust the EA power supply current limit) but we are going to work up to eleven SMPS's in parallel...and correspondingly higher power...so then we will need a fuse, as eventually the EA power supply will have its current limit set to 28A when we have got up to eleven SMPS's in parallel.

 

[FvM]

I believe, you missed to ask some basic questions:

- Do the DC/DC modules need overload protection by a fuse, is a fuse selective enough to provide overload protection?

- Does your circuit need short circuit protection to maintain certain safety properties like isolation or prevention of board damage and fire? Or is short circuit protection already achieved by the upstream supply?

- If safety regulations require fuses, what's their rating according to DC/DC manufacturer spec and board design parameters?

 

[schmitt trigger]

Many years ago I worked with high power DC installations. At the time limited to 600 volt, up to 100 kW.

I can tell you that the answer to your question depends a lot on the DC source's impedance. I witnessed first hand what happens when there is a short under those conditions. I personally saw hockey-puck rectifiers blown to smithereens. I saw junction boxes catch fire.

As other people have already mentioned, DC will not self-extinguish like AC does, and if the source impedance is low enough, hundreds or thousands of amps will create a stupendous flash and arcover. Electric arcs are super-hot, in a few seconds a fire will start.

I know, I know, those high interrupting capacity fuses are 20 times as expensive as a plain vanilla one. But please consider, a fuse is your very last line of defense in a short circuit when everything else fails.
I am going to ask the same question FvM did: If you are going to install an useless fuse, why no simply omit it?