Semi-conductor Fuses

[theDOG]

Under what circumstances would semi-conductor (or high speed fuses) be used on a soft starter application? What purpose do they serve?

[marke]

Semiconductor fuses are an energy limiting fuse using specially shaped silver elements in a silicon sand environment. Because of their energy limiting characteristics, they are able to protect semiconductor devices from excess energy let through under fault current conditions.

A soft starter uses reverse parallel connected SCRs or thyristors in series with the supply to the motor. If a short circuit occurs between the starter and the motor, or in the motor itself, there is no limiting impedance to restrict the current flow and so the current is essentially the short circuit current of the supply. In most cases, this is enough to damage or destroy the SCRs. It is not practical to turn an SCR off once current has begun to flow, so electronic protection is not practical. Current will stop at the next zero crossing which could be half a cycle later. The semiconductor fuse is able to interupt the current flow in less than half a cycle, and the rupture time is dependant on the total energy flow.

SCRs have a maximum short term energy rating that is usually called I squared t (Current squared times time) and provided the maximum let through I2t (or total clearing I2t) of the fuse at the operating voltage of the supply, is less than the I2t of the SCRs, then the fuse should fail before the SCRs.

Semiconductor fuses can be a bit of a problem at times because their pre arcing I2t is well below their total clearing I2t and they can be stressed by the normal operation of the starter and fail prematurely. It is important to compare the overload time current curve of the fuse with the operation of the starter.

[GGOSS]

G'day 'theDOG'

 

Clearly you didn't pick up too much during your time with us. More tuition is available if you ever want to come back. All is forgiven!

 

In addition to the response from marke, semiconductor fuses must be used in all installations where 'type 2' co-ordination is required.

 

Regards,

GGOSS

[marke]

In some cases, in order to comply with standards, or approvals, you must use semiconductor fuses.

i.e. For UL508, the starters are subjected to severe short circuit tests. The normal means of passing this test is to fit semiconductor fuses. This is a condition of the test and becomes a condition of the acceptance. Also, for IEC60947-4-2, there are short circuit tests and hence in most cases, the use of semiconductor fuses is necessary for complience with IEC.

[Harry Dampers]

The problem with using semiconductor fuses on a SS application is that they will cost an arm and a leg .

 

If you rate the fuses correctly to take into account the high overload currents neccesary to start an Induction machine ....then the fuse costs more than the Thyristor it is trying to protect .

 

The only advantage is that it is quicker ti change.

 

You are also not guaranteed that the fuse will protect your thyristor under ALL circumstances anyway.

[marke]

Aggreed, but you would have difficulty getting UL approval and CE marking without the use of fuses. Both require short circuit tests at very high fault currents. IEC947 requires the unit to be fully operational after the short circuit and UL requires it to maintain is physical integrity without a flame out of any form. I've been involved in approvals testing to both standards at 600V and 60KA. If you use the devices without fuses, the device no longer complies with UL508 and you could get a shock if you need to make a claim.

[Harry Dampers]

Mark , I will do some checking on this .........I must admit to not being fully aware of the UL requirements but at least one manufacturer (I am pretty sure ) claims UL certification AND IEC compliance with no need for Semiconductor fuses ............they Certainly have CE marking....I will get back to you.

 

What exactly are the UL and IEC requirements relating to short circuits??

[marke]

UL508 is essentially the same for most motor control gear. You are required to do two tests at maximum voltage and at a high fault current. (Defined in the standard) The equipment is a) started into a short circuit and then has a short circuit applied to the output while running. The Short circuit current must be calibrated to the level proposed and from memory, for small starters, the lowest short circuit current is 10KA.

The starter must not explode, or damage a muslin cloth wrapped around it.

 

With IEC947-4-2, the starter must be able to operate after the short circuit has been applied.

 

With CE and to a lesser extent with UL, some manufacturers are able to self certify and from what I have seen, there is no way that some units could pass a genuine test / inspection.

 

An interesting area to look at is the creapage distance across printed circuit boards for uncoated boards.

[Harry Dampers]

Mark, one of the units that I have been dealing with recently complies with IEC 947-4-2 , UL508 and CE (amongst others) .

 

They are quite specific in their claim that Semiconductor fuses are NOT required.

 

Of course this is only what they claim .....but it is a highly reputable company....and it is written in their technical manual .

[marke]

Are they designed to, self assesed to or independantly tested to?

[Harry Dampers]

Mark , the questions are valid .....but you could ask the same questions regarding any manufacturers claims .......after all if Aucom for instance says their starter is rated at 500% for 30 seconds we tend to accept it ........self tested or not .

 

I will see if an email to this manufacturer produces some more info and keep you posted.

[marke]

This is so very true, and I think that often, big companies rely on their reputation to carry them through. I recently saw an example of a contol panel that failed CTick because of excess radiation from a temperature controller that carried the CTick mark.

The certification does not guarantee compliance, and I think that some are more honest than others.

How many end users can afford to run their own verifaction tests to check that the "certification" is valid??

There are some basic laws of physics however that do not need testing. When you apply a short circuit with a very high fault current to an SCR, it will fail. You can not turn it off, let alone turn it off before failure occurs. I would be interested to see the mechanism used to protect the SCRs under such fault currents without fusing!

Of course, out in the field, on small units, it is cheaper to replace SCRs than fuses and semiconductor fuses can be a problem with premature failure, but the standards say .....

[Harry Dampers]

Mark I addressed the query to the manufacturer in question ........I am awaiting an answer .......til then don't hold your breath .

[milliamp]

This is all inetersting stuff but can anybody tell me what type of semi-conductor is used in a semi-conductor fuse?

 

How does it work?

 

milliamp

[marke]

Semiconductor fuses are not made of semiconductore, rather they are designed to protect power semiconductor components. The semiconductor fuse uses silver elements with tight dimensional control over the elements so that they limit the energy flow under short circuit conditions. They are commonly known as fast fuses, or energy limiting fuses. A standard HRC fuse with copper elements will clear in half a cycle, but the fault energy is dependant on the fault current capacity of the supply. In the case of the enrgy limiting fuse, it will clear in less than half a cycle with the time to clear being influenced by the fault current.

Best regards,

[jraef]

I know this is an old thread and maybe nobody else is looking at it any more, but for some reason I missed it and now am interested. I wonder if Harry Dampers ever got his answer?

 

To add to the UL discussion issue, there are actually several ways to attain UL listing. One is to specifically state the fault current level that the unit can safely withstand without testing. Few people do that because the numbers are so low that it scares the average engineer away from accepting it.

 

The second method is the one pointed out by Marke, namely testing with a semiconductor fuse. The caviat is that once tested with a fuse, it must be installed with that exact same fuse, technically even the same manufacturer, which is something most people don't realize.

 

The third lesser known method is the one we did. We subjected out starters to a "3 cycle test" at UL for the fault levels available from the let through of the lowest rated circuit breaker available for any given unit frame size. Once we passed, we had UL listing WITHOUT the need for fuses when used behind any brand of circuit breaker for the fault level of the breaker itself. This is a tough spec to pass, and consequently very few manufacturers even attempt it. That may be what Harry Damper's referenced manufacturer did, and since it wasn't us (because we don't have CE) and UL told us only one other manufacturer has passed the same test, I have a good idea it was Benshaw. The CE issue raises a good point however.

 

I must note that we did not go for CE on this product, mainly because we had no demand for it as a US manufacturer. Now that we are going for it on our new products, we may need to add fuses because of the "operate again" clause mentioned earlier. That was not a criteria of UL's test. They basically pack the unit full of a cotton-like substance, apply the fault, and if any of the cotton turns black you lose. When we tested our 1250A frame at 85,000A, the wires feeding it at the test lab failed and the cotton substance all turned black in a flash. I hear it was quite spectacular!

[marke]

Hi Jraef

 

Never heard how Harry got on with his answer, I assume it was not forthcoming.

 

It is a problem with these standards etc, there are some glaring holes and many of the requirements are very open to interpretation. At the end of the day, I consider the standards are only part of the equation, I believe that it is more important that the equipment is designed for the environment and purpose, than it is designed to pass the standards. Correctly designed, it will pass both.

 

I recently experienced an EMC test lab in asia that failed some equipment that had passed at another lab, then asked for a fee to correct the problem. I also heard of another accredited lab that had a fee to test to EMC standards, and another fee to certify without testing!! (second fee was higher than the first but did not involve any testing)

 

I have always gone down the approach of UL testing with semiconductor fuses as this is the safest way if you wish to retain the integrity of the equipment. For CE it must continue to operate and this is more difficult with circuit breakers, especially where the fault current potential is very high.

 

There is equipment out there that is "self assesed" or claimed to be certified, and I can not see any way that it would pass the testing regimes that I am used to.

 

There is also a lot of "marketing license" applied to claims statements etc.

 

An interesting subject!!

Best regards,

[Guest]

Originally posted by marke

Semiconductor fuses are not made of semiconductore, rather they are designed to protect power semiconductor components. The semiconductor fuse uses silver elements with tight dimensional control over the elements so that they limit the energy flow under short circuit conditions. They are commonly known as fast fuses, or energy limiting fuses. A standard HRC fuse with copper elements will clear in half a cycle, but the fault energy is dependant on the fault current capacity of the supply. In the case of the enrgy limiting fuse, it will clear in less than half a cycle with the time to clear being influenced by the fault current.

Best regards,