DC switch selection/discussion (again..) - letter from Mr Angry

[SibertSolar]

Hi all,

Some may already know that I have been vocal over the past year or so about using isolation switches for DC that are designed for purpose rather than just achieving the requisite DC21 current/voltage ratings, ie - not simply using modified AC isolators. Anyway, I've been given the attached document from a home-owner who has had an experience of his own that he wanted to share with the Solar PV community. He wanted to remain anonymous and had found Sibert via the web when searching for related info and wondered if I might like to share his document with my contacts/customer-base. I'll be sending a copy to the BRE and others, as well as discussing with various trade publications we work with.

Any thoughts from you guys?

Andy

View attachment Mr Angry Letter 07Jun12.pdf

 

[Gavin A]

thing is though, the switches are rated by the manufacturers to be capable of switching the DC loads they're rated at, so the installers are in full compliance with the DTI guide when using them.

I understand this is an issue you bang on about a lot, but frankly I don't see the problem, although we do ensure we've got a significant margin of error on the isolators we spec anyway to be on the safe side.

These aren't industrial load situations where the isolators will be in use daily, and all customers are instructed via the stickers on the inverter and isolators to isolate the AC before the DC to remove the load and further reduce the potential for problems with the isolator. OK, so if a customer for some reason insists on switching their isolator off and on regularly at peak generation times, and doesn't follow the instuctions not to switch under load, then eventually the isolator will fail, but this is down to the customer not operating the equipment according to the instructions rather than an actual system design issue IMO.

 

[JulianC]

Can't add anything more to what Gavin says. Non-issue as far as I'm concerned.

 

[BruceB]

There is a difference between the requirements for isolation and functional or emergency switching. Some of the things a switch might have to do are:
- isolate the maximum voltage present when in the off position
- carry the maximum current specified
- be able to break the maximum current specified.

However, for the dc side of pv, 7671 only calls for a switch disconnector which is synonomous with an isolator, which has to meet those first 2 requirements but not the third. It does NOT envisage the switch being used to break the full load current.

DTI V2 gets itself a bit confused. It gives 2 ticks for the switch must be able to isolate the circuit and carry the dc current, but only one tick to it must be load break rated. Yet the note above that load break tick (the one quoted by Mr Angry) uses the word 'must' which is not consistent with a single tick.

DTI V3 changes the format a bit, but the same sentences are used so the same confusion arises.

Provided the switch manufacturers state their switches can carry the dc current then that ought to be enough as they should not be used to switch it. From memory, even SMA say their own disconnectors will only last 50-100 full current switchings before being worn out.

Regards
Bruce

 

[SibertSolar]

Fair comments Gavin, appreciate your feedback. My point is that being in compliance with the DTI guide is what's caused the problems (and there have been problems specific to installed systems and DC switches) because the guide is too ambiguous. Clarification and explicit guidance is what's needed in my (and various manufacturer's) opinion.

As for the frequency with which a unit may or may not be used, I've seen evidence of modified AC isolators being used to switch DC where just one operation of the isolator has caused excessive arcing and failure of the device, sometimes with explosive results. If this should happen to a home-owner here in the UK under these circumstances, I fail to see how far falling back on "we complied with the DTI Guide" will get anyone who installed the product in the first place but that's just my opinion of course. Also, aren't DC switch disconnectors required to be load-break rated according to the DTI Guide?

 

[SibertSolar]

Posted above but didn't see other responses before it re-loaded. Thanks for the input Bruce, invaluable as always. I've heard that the DTI v3 document will be clearer so it will be interesting to see what the final release reads like. I've a draft here that still mentions the requirement for load-break rating but I don't have an indication of the weighting. Can't see why it shouldn't be a "must" though to be honest - or does the guide also indicate that the installer must educate the end-customer as to when they can/can't switch the DC on or off? "KISS" suggests that simply using a load-break rated switch by default would avoid any potential liability issues.....

 

[JulianC]

What are these customers doing to have cause to mess about with the DC isolator in the first place?

 

[Gavin A]

I've just checked with k&n's technical guide, and they're rating their switch isolators as being compliant with the following British Standard for DC at the voltages and currents they're rated at.

When stating product compliance with BS EN 60947-3 for a given utilisation category there are a number of mandatory tests which need be completed, one of these tests in BS EN 60947 parts 1 & 3 is the ‘Operational Performance test’, below is some extracted data from section 7.2.4.2 of BS EN 60947-3.

The below tests are required as a minimum for compliance with the standard.
Verification of operational performance for rated operational current of 0 - 100A


Utilisation | Number of test operations
Category | without current | with current | Total
DC-22A 8500 1500 10,000
DC-22B 1700 300 2000
DC-21A 8500 1500 10,000
DC-21B 1700 300 2000

so even the lower rated switch isolators are supposed to be good for 2000 switching cycles, with 300 switching cycles under full load. I'm struggling to think of any circumstances in a domestic situation under normal usage conditions where this would be exceeded, particularly bearing in mind that this is at full load, whereas even on a system that was specced at the maximum rated capacity of the isolators, it's going to spend the vast majority of it's operating life operating at much lower outputs than it's peak rated output. Factor in that even the lowest rated isolator is rated for 25A at 450V, and single string systems will usually be around 8amps, or 2 string around 16amps, that's a far amount of headroom on the ratings used for those switching cycles (we actually use the 28Amp version for dual string to give even more head room).

http://www.krausnaimer.co.uk/images/pdf/pv_g_2012_27_2_lowres.pdf

 

[Earthstore]

I understand this DC switching argument, but of the 1000's of DC switches installed in the UK alone, how many have failed?

 

[Gavin A]

we've had 2 fail due to our installers dropping brick dust into them by drilling above the isolators while they were open.

think this is really an installer fault though.

 

[moggy1968]

Have just been nack to change one that burnt out. Have replaced it with one with no links.
The customer will be 'messing around' with the DC isolator because thats what it says in the shutdown procedure! Maybe I need to add a 'maintainance procedure' where just the AC isloators are switched when working on other parts of the electrical installation, cleaning the panels, or being up on the roof?

 

[Gavin A]

how often have they been doing that then?

 

[moggy1968]

dunno, but he did say that he had switched the DC off when they built the cabinet around the inverter and bits, but to be fair, this is a 6 month old installation, even if he switches it off every night to put it to bed I would expect the switch to be able to take take it!!
In reality, it's probably been off no more than 2 or 3 times. It concerned me as we used a lot of these in the early days.