SSR current leakage problem

[C G]

I came home yesterday to finding my AC power not working.  Checked on the BMS, and found that my battery bank voltage was an astonishing 38 volts, and CFET/DFET were disabled due to overvoltage.  Yet, I saw around 80W of load and 200W of PV input.  Immediately I shut off all the PV breakers and wondered what to do next, since I don't know how I can safely draw load with that high of voltage and get things working again.  I questioned whether the BMS was reading the shunts wrong, but that wouldn't explain the high voltage.  I tried overriding the BMS control to power on the inverter, but it refused to come on.  I then realized I had programmed my Victron inverter to turn off after some lower voltage, since I never expected it to go so high.  I should probably reset that back to a high limit, so that I can use it to drain excess voltage should it ever be needed.  I then noticed that although the inverter and 12V converter circuits were shut down, the 24V circuits going through a BatteryProtect were still live.  One of these runs over to my camper trailer where a second uncontrolled 12V converter runs the DC stuff in my camper - the load was my propane furnace.  Okay, a problem, because the BMS should be turning the BatteryProtect off too, but not the main one.  In fact a little bit of a blessing because I need to run some load to drop the voltage, and I would probably be afraid to try forcing a converter designed to run on 24V to turn on with 38V input otherwise.

So I turn on all the lights and fans in the camper and get the load up to around 160W, which starts slowly dropping the battery voltage.  Then I wonder what is going on with the PV input, which should have been disconnected according to the BMS.  I try switching on one PV breaker at a time, and can never see any input coming in.  So then I start switching them all back on one at a time, and at some point, notice the PV input jumps up.  I switch that breaker back off, try the next one, and see the input jumps up there also, and for several others.  Really very strange.  I have no idea what to do and decide to just leave everything switched off until the voltage can drop down to normal levels so I can get the inverter turned on and then I can switch on a space heater if needed to draw more current from the batteries.

Less than an hour later, voltage is back to normal, the inverter is working, and the BMS is showing battery at 98% and CFET on.  Okay, let's try the PV breakers again.  I try one at a time, and I see some current coming in for each one.  The sun is getting really low so only 10-20W per SSR.  So that seems fine.  I turn them all on, and see a combined total of 188W, which seems just about right.  But then I notice when looking at the DSSR20's (all connected to the same DEXT16), that the little lights which should all be green - some are a sort of orangey-green, and one is off.  When I switch off the breaker for the one that is off, the others that were orangey-green turn solid green.  I have some flashbacks of previous unrelated problems where the lights were doing this, and I wonder if there's a little bit of insulation caught in the spring cage connector.  I reconnect the wire coming from the PV breaker where it clamps into the SSR, being careful not to push it in too far, and sure enough, then when the breaker is turned on, the light on that SSR turns green and none of the other SSR lights are affected.

So a few concerns to take care of:
* Reprogram inverter to tolerate higher voltage input, whatever the default was is probably better - as long as it won't damage the inverter.
* Add some manual override switches to the remote circuits for the inverter, converter, and BatteryProtect, in case I should ever need to use them when the battery voltage gets too high.
* Figure out why the BatteryProtect is not turning off when it should.  As far as I can tell, the wiring is correct.  I am using an optoisolator so the BMS can control all three things.  I tried switching the wires for the 12V Converter and the BatteryProtect (both Victron with the same remote plug), and still the 12V Converter turns off when it should while the BatteryProtect stays on.  Any suggestions here?
* Check all the wires better, but this had been working fine and not touched for months, then perhaps some minor arcing corroded things enough and finally it started mattering.  Unfortunately, the top connections are a little hard to inspect as they are mounted on a wall just above eye level.

My main concern is preventing this issue from happening again and the batteries getting overcharged.  Hopefully the batteries are still reasonably okay - everything does seem to be working properly since.  Why did a simple loose wire connection allow this to happen?  The loose wire caused some feedback to go from the one DSSR20 through the DEXT16 and affected multiple other (but not all other) DSSR20's.  Is this something that could be prevented via a design improvement to either of the components?  Is there anything extra I could add to prevent this from happening again?

Any suggestions welcomed.

[C G]

Well, I was wrong about a loose positive wire being the problem.  It just happened again and I went and fiddled with it.  Found that the positive good but sometimes when I would tap on the SSR, the green light would go out.  Finally checked the negative wire connection on the bottom, and that was the one that was pushed too far in.  Sure enough, it had a clamp mark on the insulation when I pulled that wire out.  Put it back in, not too far, and the problem seems to be resolved now.  Still curious about ideas how to fix this more properly, as well as any ideas why the BatteryProtect remote isn't working as expected.

[Dacian Todea]

A photo of your DSSR20's and the connection to DEXT16 will be useful.

Based on your description you had a short circuit between the Batt and Heat remote signals as that is the only case where both red and green LED's will light up so check all your wiring and make sure you do not have exposed wires that touch each other.  The negative wire had nothing to do with this problem as if that will not have made contact it will have affected just that one DSSR20 and that will just not have worked at all (OFF).

So you had a short circuit between RCBatt and RCHeat either around the DEXT16 or on one of the DSSR20. Check all those wires and fix that since that is super problematic.

"that the little lights which should all be green - some are a sort of orangey-green, and one is off. "  This is the important thing that you mentioned. The one that was off was the one with the negative not making connection all the others where orange because both green and red light where ON at the same time and that only happens if the Batt and Heat remote signals are short circuit.

[C G]

Here are photos of the DSSR20's, the mark on the wire insulation where the insulation had been clamped, and a video reproducing the problem as I'm describing it.

https://youtu.be/FLhz-uDMCNQ

[C G]

In the youtube video, you can see when I make a weak connection with the loose negative wire, I can reproduce orangey-green lights on two of the DSSR20's with the light off on the one I have the poor connection to, and then the SBMS0 shows PV current coming in.  When it is properly clamped in, the lights all turn green and the SBMS0 shows 0 PV current coming in, as expected.

[C G]

In the video, when show PV current coming in when CFET is off, which should not be allowed, this is happening through the DSSR20 that I have the negative wire completely disconnected from.  I'm not sure why the light color is being affected on two other DSSR20's (and it was more than 2 yesterday, perhaps because the sun was much weaker), but those don't seem to be leaking the current - just the one with the negative disconnected or not clamped in properly.

[Dacian Todea]

Thanks for the video. That is very strange.  Where there two of the DSSR20 that where orange ? Hard to say from the video but all others looked green.

I can not imagine what can create this problem as when you remove the battery negative the Batt remote signal is connected to internal GND tough a 7kOhm resistor and that internal GND is connected to battery+ trough some ceramic capacitors thus unless those ceramic capacitors have some internal short circuit some significant leakage to lift the voltage on that internal ground I can not see how the Batt signal from non connected DSSR20 can supply some of the others.

Can you try removing the GND from any other DSSR20 to see if the same thing happens?

[C G]

In the video, when I am messing with the loose negative wire on the 4th from left DSSR20, you can see the light changing from green to almost solid orange on the 2nd from left DSSR20, and the light on the rightmost DSSR20 showing orangey-green.  Last night when the sun was weak, that is when I saw most of the other DSSR20's showing orangey-green - none of them as orange as that 2nd one from the left now.

I just went out and recorded two more short videos which are uploading to youtube now.  I will post links when they finish but I am on a rural cellular band so it will take a little while.  I cannot reproduce the problem by disconnecting the negative on any other DSSR20s, and the problem with this one happens even if I disconnect the Battery+ wire too.  Strangely, I have to enable input current to several other DSSR20's, not just one, before the problem will manifest.  But then I can disconnect the input current from most of them and the problem will remain.  Very strange that when the 4th DSSR20 is only connected to the PV input and DEXT16, it can cause the signal to other DSSR20's to be wrong and then they start leaking, and also if the positive wires are connected to the 4th DSSR20, it will feed power through to the battery directly even when it should not.

Do you think the 4th DSSR20 has some hardware problem?

Part 2:  https://www.youtube.com/watch?v=KNoiIBwlgFg
Part 3:  https://www.youtube.com/watch?v=JWO8U1erRck

[Dacian Todea]

The 4th DSSR20 has clearly a hardware problem I'm just not sure what exactly.  I will ask you to send that to me to investigate in more details.

There is some sort of leakage current at least around 1mA from the Batt remote signal on that defective 4th DSSR20 and that is what enables turns ON the 2 and the last DSSR20. The light looks orange as both the green LED is ON and the red thus the combination will show up as orange but they will not turn the battery charging (red LED) ON unless they have at least a 3 to 4V on the Batt+ connection and that clearly comes from the 4th DSSR20 tho is a bit strange it comes from PV likely something got damaged maybe related to missing GND (battery negative) but maybe I can narrow down better what happened when I get that 4th DSSR20 to see what is damaged.

[Dacian Todea]

Just made some tests and it seems there will be some small leakage current if the GND (battery negative is not connected). This may actually be enough to turn ON other DSSR20's that are parallel connected.

So there may be nothing wrong with that 4th DSSR20 other than when it was missing the battery negative wire.

What happens is that the disconnected DSSR20 tries to self power from the PV+ and close the circuit trough the Batt remote wire witch is connected to GND trough 7kOhm equivalent on all DSSR20 so it will go trough the 7kOhm on the non connected DSSR20 and push current to battery negative trough some other DSSR20 that has the same Batt remote connection.

So if you expect that a bad battery- connection may happen again the way to protect against other DSSR20 being turned ON will be to add a 100V+ rated diode on each of the Batt remote signal wires so current can not flow from the floating DSSR20 to the other DSSR20's. 

[C G]

Well, I don't expect it to happen, but it is difficult to be sure with the spring cage connectors.  Also maybe some wire could get severed unintentionally.  I think it wisest to add those diodes to be on the safe side.  Is this needed only on the RCBatt+ connections or also the RCHeat+ connections?  Could you please provide a link to an appropriate diode?

In the past, I had a problem due to one of the positive connections being poor.  It seemed to be working okay but I was testing to ensure I was getting the correct wattage from each pair of panels and noticed it was lower for one, then noticed it felt warm to the touch, presumably from arcing.  That may have been this same one, I'm not sure as that was months ago when I added more panels to the system.

[Dacian Todea]

The wire needs to be striped sufficiently so that there is no possibility to clamp on the insulation instead of conductor. Positive connection missing should not result in anything similar.

The leakage current can only happen on Batt not on the Heat remote connection.

Those connectors are very low resistance and rated for fairly high current much higher than they are used for DSSR20. If you had any connector heating it means you just did not clamped the conductor so maybe inserted the conductor while lever was low thus wire was just pushed there not clamped by the spring.

Connectors work great with solid copper wire or just a few thicker strand as you usually have for PV cables. It may look that you use silicone cables with very fine strands and in that case ferules will be a good option for good contact.

The insulation needs to be removed 12mm (half inch) that way you can not pinch the electrical insulation when closing the lever as wire can not go deeper and the conductor will always be the one to make contact.

You need to also test the wires on the circuit breakers as over time due to thermal cycling the connections can get loose as those are not spring loaded. 

[C G]

The heating happened because insulation was in the clamp rather than the wire.  Some wire was touching the metal of the clamp but not enough for a strong connection.  So only half of the expected PV input wattage was making it through, and the rest was getting turned into heat.  This issue with the negative wire didn't result in any heat, but is the same core problem of a few millimeters too little insulation having been stripped from the wire.  I had not considered using ferrules for this large of wire, but I certainly could do that.  Yes, I am using silicone insulated cabling with fine strands which is very flexible.  I think the breakers should be okay as they are clamp connectors which I screwed very tight, but I will check them periodically.

As for the diodes, I am really not sure what would need used for this.  Am I correct in assuming that the amp rating could be quite small?  Do these need to be wired in a particular direction?

[Dacian Todea]

Yes any diode will work as current is super small 2 or 3mA only so 1N4002 to 1N4007 are super common and any of them will work or anything similar with a voltage rating over 100V.

Yes the direction is important the white/silver band (Chatode) should be connected on the Batt grey connector while the other side will connect to RCBatt+ on the DEXT16

The diode will prevent current flowing trough any DSSR20 that may happen to have the GND (battery negative disconnected) while PV+ is still connected.