Following the recent thread, I bought some more isolators. I dismantled one (not really possible to reassemble, as expected) but shows the design. This is the small black Victron-style one. All the bolts etc are non-magnetic, probably plated brass.
It’s cruder than I expected, and I do not believe this is good for 275A continuous.
Indeed, a blue sea (larger) device (which has the advantage of accessible connection studs) gets warm at only 80 or so amps continuous (although it has had two or three capacitor switch on arcs -Quattro 10k- so may be a little damaged).
The input capacitance of the MPPT 450/100 is 0.15 Farads, I expect the Quattro 10k is probably closer to 0.25F, so its not surprising that there are switch on arcs even under no load all-off situations. Circa 10-15 Cb and probably charging at full battery power, several hundred amps for a few milliseconds. More than enough to cause arc damage.
The advantage of the Blue Sea isolators is that their semi-open construction allows for more ventilation and access to the terminals for pre-charging before switch-on.
I am awaiting all the components, but IMHO the basic pre-charge system might be a 25W 75ohm resistor to clip across the CLOSED terminals immediately prior to switching the isolator ON and then powering up the devices. With no heatsink these are continuously rated at 1/2 quoted power at 25Ccase to 1/10 at Tcase 250C so should be fine for a few milliseconds.
Time constant is ~5secs so 30 sec/60sec should suffice to charge the cap sufficiently.
I am adding test lead cables with medium-sized insulated croc clips and a cheap panel DVM powered by a 9V battery to check state of charge. A voltmeter would also do.
I am not a fan of using fuseholders as switches.
I think it’s fairly well known among installers that the Victron battery switches are junk. They have high internal resistance and even though I never took one apart I could tell they were crap. I appreciate you sharing these pictures.
With regard to the Blue Sea switches, the 6006 series should not be used for high-current 12V loads even though they are technically rated for it according to the documentation. They, too, have high internal resistance. I would not use the 6006 for any more than 100A continuous. We use E-series or HD series switches for 12V systems.
The 6006 switches are fine disconnect switches for 48V 5kVA inverters or MPPT RS 450/200 controllers. The 10kVA inverters are better off with the E-series or HD switches.
Ah, so you have figures for the on-impedances of the M- E- and HD- Blue Sea switches. That’s good, what are the typical figures?
The Victron and cheap clones seem to be based on the 6006 series, as you say. For info Blue sea ones now all seem to be rated at 48V/300Acont.
Interesting that you do not believe the quoted max continuous current ratings of Blue Sea isolators, peobably due to the on-impedances you have measured above. Personally, I am not enamoured of any firm that publishes specifications that claim their devices meet, but which actually cannot be met.
I presume you have inspected the internals of these switches too, I would be intrigued to know how they differ internally. I would suspect they are all similar in design.
I measured the voltage across my 9004E at 40A and got ~40mV across the contacts, suggesting a switch resistance of about 1mOhm. At 6kW (~120A) that would be about 120mV and would dissipate about 15W, so it might be expected to run warmish (!). Certainly, good heat management might be the key to longer life even for simple things like switches at 100A+ currents.
Interestingly, 400A mosfets have Rds of 1 to 2 mOhms too. Realistically, I guess switches might struggle to get much under 1mOhm in real life possibly would if they were knife switches, which I always assumed they would be, but apparently not.
PS clearly the victron/6006 switches are NOT ok given my experience with them on a MPPT RS 450/200!
I don’t recall the measurements and it was at least 16 months ago. We have a micro-ohm meter in the shop so I could take measurements and post them. I will try to do that tomorrow.
Just to be clear, that is the Victron Battery Switch? Before I again call VE attention to the inaccuracy of some of the published claims of its appropriate use I wanted to be sure that this one in the photos did in fact come out of their packaging.
I am almost 100% certain the switch was victron, as my installer is VERY risk-averse. My installer still has it but will be working away from home until next week. He has promised to return it to me and I will disassemble and examine. With PIX. I will come back when that is done but I am pretty confident it is the standard “universal” much own-labelled (in this cas Victron) switch. Probably copied from a long since out of patent switch from the 1970’s!
None of the switch specs I have come across allow disconnection/connection under load. The derating of switches for reactive loads are often quite high, and DC makes it worse.
Now here is an interesting thing …
With some nice sun the Quattro is sending over 5kW to grid which must be circa 100A via the 48V system. Voltage across the switch is a (noisy) 1 to 2 mV or a few tens of micro-ohms. I presume there are some spot welds although why today when It’s not been operated for weeks, I have no idea. It’s also running cooler (unsurprisingly, I guess). Might explain why it’s stiff turning off?
All quite interesting when looked at in detail.
Will stuff my oscilloscope across it when I have time, that might be interesting too.
I’m sure you’re right, these are YIS Marine BF441 switches (YIS theoretically holds a valid and current patent for them) but YIS sells almost exclusively as “white labeled”, so besides the one with Victron packaging we also see them with various marine store packaging and etc. I mostly just wanted to be sure, when I bring the matter up again, that it couldn’t be claimed that the photos and obvious insufficiencies are the result of some other cheaper clone!
Hey all. Sorry for the delay but running a small business keeps me hopping. First, I want to retract my remark about the Victron 275A switch being ‘crap’. I still think it’s not built well, but in light of the results below I can state that from an internal resistance perspective it’s better than three Blue Sea Systems 6006 (m-series) switches we tested in the shop yesterday.
Test:
This was certainly not a carefully run test. I don’t have time for that. For each switch type we turned them off & on 15 times to ‘wipe’ the contacts. And since the 6006 tested poorly, we took three samples of the shelf for that one to make sure we weren’t testing a single bum switch. I used a Unit-T micro-ohm meter. It’s not a Fluke, but it’s more than sufficient for this testing. We used 4-wire clamps of course. We tested each switch three times to reduce errors.
All readings are in milliohms
Victron switch:
0.12 - 0.11 - 0.11
BSS 6006 (m-series):
0.54 - 0.53 - 0.55
BSS 9003E (e-series):
0.12 - 0.14 - 0.13
BSS 3000 (hd-series):
0.25 - 0.18 - 0.13
We typically only use the 6006 switches on 24v or 48v systems with 3kVA or 5kVA inverters or it’s e-series for larger systems. HD switches are only used when customers insist on using a 12V system, otherwise they aren’t that useful to us since they aren’t rated for 48V.
That’s pretty interesting. I wonder if there is any difference in the internals? I haven’t time for a few weeks but it occurred to me that I could use my little welder (semiconductor/pcm) at ~100A and do a measurement that way too. I agree with you that we only need ballpark figures here. Just out of interest what test current were you using?
Again, brand new may differ from arc-damaged as my figures are from a 9004E which has been switched on without precharge several times.
Thanks for the new numbers and it does make me feel a bit better. I’m using that Victron switch because it was one of the few that was rated for a 48 volt system. The Blue Sea 6006 is not rated for 48 volt systems. In the datasheet it specifically says up to 48 volts.
There was a post over at DIYSolarForums.com a few years back on this very subject. I believe it was when Blue Sea made a clarification about the voltage rating. At the time it was unclear if it was up to 48 volts or for 48 volt systems.
I used the Blue Sea E-Series 9003E Battery Switch in my prior 12 volt system and often pushed 200 amps through it. I tested the system once by charging my EV from the camper for a few hours and kept a close eye out for hot spots and the switch wasn’t one.
I recently performed the same test with my 48 volt system and the Victron switch did just fine. I tested for hot spots in this test with a Flir One Edge Pro infrared camera.
I have no actual testing numbers to compare with, so my results are anecdotal.
If the Victron switch is questionable I have no problem ditching it, but so far it’s OK.
What I would really like to see is a Lynx Switch. Pretty please Victron?
Blockquote What I would really like to see is a Lynx Switch. Pretty please Victron?
I’ve asked and asked and even went pretty far down the path of manufacturing a replacement to Lynx Shunt that included a SmartShunt, 500A contactor, pre-charge circuit, a tiny Arduino controller and all connected to the GX via USB with a custom device driver. I pulled the plug because making them in small quantities just cost too much and I didn’t think we’d be able to sell enough of them at a price point that would be compelling. Machining the copper for the bus bars was the highest cost.
Before I went down the path I asked my Victron rep to confirm with the product team they weren’t coming out with such a device in the next couple of years and they assured me there were no plans to do so.
It’s not cost effective for smaller systems, but you could use a Lynx Smart BMS 500A with an M8 jumper cable across the two BVT ports. This would give you pre-charge, master disconnect, and battery monitor. But that doesn’t make much sense from a cost perspective for most customers.
To be honest if I were designing a switch I would be looking to use power mosfets. No arcing problem and relatively simple to do a slowish recharge, although most take eyewatering transients.
On searching, the first 300+A one I came across was only a few $:
Electrically quite robust but a rather high Rds on ~0.7mOhm. If lower on resistance devices are not available, then some parallelling may be in order. Vds could be better, 80V+ preferred, 60V is a tad too close for comfort. I’m sure there are other devices than the first I came across.
I’ve had one of the smaller isolation switches fall apart after a year’s usage (around 100A continuous). It literally melted. Now it wasn’t a Victron branded one, so probably different in QC, but identical in physical design. I do think that these switches are prone to having their connections loosen over time with heat.
I’ve since replaced it with a Blue Sea 6006. Same spacing for the mountings although the holes are slightly smaller so had to drill out for M6 bolts to go through. It does feel a lot tighter though.
Excellent photo. This is more than capacitor charge damage and is exactly what I would expect if the switch had been used to switch off when under significant, possibly full, load. A fat welding-style melted-metal arc track. This damage did not occur whilst the switch was full-on or full-off, but during switching off.
That said, this switch could continue to function after the damage occurred, although there would be a high local resistance due to the small contact area through the welds on the contact rather than the full contact area. Heating would result. This may even be the case with the switch that failed here (due back in 7-10 days).
These switches are cheap and effective but easily abused.
As I said before, IMHO the only reliable way to switch 100’s (even many tens) of amps is via an appropriate solid state switch. That prevents arcing and can be slugged to dissipate transients.
I would recommend to always use a pre-charge circuit to ensure no inrush currents before switching on the isoltator switch, as well as to reduce the power consumption to a minimum when switching off.
Both will protect the contacts and will last you components much longer.
