Looking for ideas on a DC disconnect. 1

[itsmoked]

I'm looking for some ideas from the Brain Trust.

I'm building a high energy off-grid power source. It has lithium batteries. Of course the finicky lithiums cannot be over charged or over-discharged or charged when too cold (on pain of death to lots of things including one's wallet) The system's Battery Management System (BMS) needs the ability to cut the power to and from the batteries. Commonly used for this are the typical power solenoid relays:

But those are abhorrent because they consume a ton of energy just being kept closed. They typically are too hot to touch making them scandalously power wasting. Not doing that.

Much much better is a motorized disconnect that ceases to draw any power after it's open or closed.
I use these cool valves frequently. Look at the price for a stainless steel ball valve version!!

Amazing! So now turning to switching electricity instead of water, a much more trivial effort made of some stamped metal contacts and a little bending of it and you get this:

The problem is that ridiculous price! I'm not doing that. I need about 10 of these. But damned if I'm paying that exorbitant price for it.

Here's the non-motorized price.

While OFF and ON would be nice I can live with OFF only. This means a solenoid type action where a PLC generates a brief ON could maybe pull something. Perhaps something like one of these kill switches.

Or perhaps twist something like one of these:

Or pulls open one of these:

I could run a motor too.
I have a couple of 3D printers and I'm not afraid to use them I can print most any cam, lever, shaft, housing, frame, linkage, or mechanism.

Suggestions?



Keith Cress
kcress -

http://www.flaminsystems.com

 

[Sparweb]

Violation 3: On a massively paralleled system it could be hard to tell which pulse you were on and whether you actually had turned a battery bank on or off.

Just guessing, but getting a feedback signal that a relay is open or closed is possible, even basic.

Logic: [ (relay open = FALSE) AND (battery needs to be disconnected = TRUE) ] = (pulse signal required)

An AND gate gets you the pulse only when you need it, right?
(and not inadvertently activate a bank that still needs to remain disconnected)

http://www.sparweb.ca

 

[Sparweb]

Are there latching relays that require inverted inputs?
+V to close
-V to open
0V to stay in state

http://www.sparweb.ca

 

[davidbeach]

A DC latching relay plus a couple of diodes would produce a latching relay with +V to close and -V to open.

 

[stevenal]

Link

 

[waross]

Hi Spar. Commercial/industrial latching relays or contactors often have a close coil and a trip coil.

But those are abhorrent because they consume a ton of energy just being kept closed.

I think that Keith is looking for less current.
A simple electrically held solution is quite dependable and fail safe.
In a simple system it may run without monitoring.
Adding latching relays adds additional failure modes and circuit complexity.
Additional monitoring adds cost and additional failure modes.

Keith, are those original relays suitable except for the energy usage?
Do you have one in your shop?
Have you considered some tests to determine the drop out voltage?
Once a solenoid is pulled in it takes very little current to hold it closed.
If the drop out voltage is quite low you may consider an economizer circuit.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[stevenal]

I have found latching relays to be quite reliable. No experience with devices at these current levels though.

 

[waross]

I agree Steve. One widely used type is a pump alternator.
This may be a circuit built from individual relays or an "All In One" dedicated device.
Pump alternators switch state with each input.
But, with the failure of a pump alternator you will get one pump doing all the work instead of sharing with the other pump. A few days or weeks of a failed pump alternator is not serious when compared to 10 or 20 or more years of expected service.
With a battery disconnect, a failure in the relay, or more likely a failure in the control circuitry and wiring, may lead to a battery fire.
Faced with the same choices as itsmoked, I would be weighing latched versus electrically held on the basis of cost, energy use, control circuit complexity and dependability, and the possible consequence of failure to act as expected.
Good luck Keith.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[itsmoked]

I’m Steve, not Bill.

I realize that I was thanking Bill for the lighting comment.

Requiring two separate signals OFF verse ON times 10 more channels is another PLC block module.
I tried the link. It didn't take me long to decide not to spend thousands of dollars.

Spar. Hundreds of years of relays I'm sure there's reverse polarity ones out there. LOL
Heck, the GIGAVAC one I'm using requires the DC current thru the main contacts flows in a particular direction or the lifetime is cut to 1/3 the cycles.

Good point David. My PLC doesn't offer -outputs really.

Steve that Hersee is a nice one. Still $1,500. Groan.

Bill; the three I just bought have it built in - part of the "control circuit". The energy penalty for 10 would be 1.4 x 24 x 10 = 336WH per day. Way better than most others that land in the kWH per day realm. For the GIGAVAC T model; $0.10 on grid or 1/2 a solar panel off-grid. The losses across the main contacts are frequently hideous. Except for these GIGAVACs I've seen 10's of watts in some cases. They'd be smoking hot, all wasted energy.

I considered using some MOSFETs on a board. A lot of them have 300A ratings and 2mOhm drops. You need reversed series ones to block flow due to the body diodes. That became a mini-engineering project requiring heat sinks and how to connect 2AWG cable to a surface mount part. Nasty!

I really liked the idea of using one of these:

And a small chain to one of these:

20KGF is probably way more than needed so I could buy a ten pack of 1KGF servos for ??$30?

The chain connection would allow one to walk up and turn off a battery regardless of the PLC not turning it off. Then manually turning it back on all without digital intervention.

But! Servos require that 2ms pulse train to keep them in position. That required adding something like this "servo tester":

That servo test board that has position memory (the pot). I'd then need the PLC to power-ON the servo test board that would wake up and drive the servo that would pull the chain disconnecting the battery.

Started seeming pretty cumbersome. Then you read some negative (gets hot!) reviews on the switch and you start wondering if you're building your house on a low tide sand bar.



Keith Cress
kcress -

http://www.flaminsystems.com

 

[Sparweb]

https://www.mrs-electronics.com/products/detail/toggle-relay-mechanical-24-v

Quiescent current = 1 mA

Sorry, can't find anything bigger.

http://www.sparweb.ca

 

[itsmoked]

Bad link Spar?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

Can you use IGBTs?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[itsmoked]

Can you use IGBTs?

Nope. Same problem I had with MOSFETs but worse.

A 160A rated IGBT module at 100A drops 1.5V!!! That's 150W I'd have to do something with and wasted power out of expensive batteries.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[BrianE22]

Can you use a bi-stable relay? By itself, it's not fail-safe but maybe you build a fail safe circuit that discharges a large capacitor into the "open-it" coil.

 

[itsmoked]

Hi Brian. About half the posts above are about bi-stable relays or in this case contactors (big muther relays). They were all proving to be hecka expensive at the needed current.

It should be noted that all the batteries involved are capable of probably delivering more than 3kA into a fault and have fuses for current limiting protection. These contactors are to simply isolate them due to temperature, or out-of-spec cell voltages.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MikeHalloran]

One shot idea: just run a battery cable through a steel ball valve as per your example above.





Mike Halloran
Corinth, NY, USA

 

[waross]

Keith; On at least one relay, I strongly suspect that bi-stable means that it will be either open or closed, depending on current through the coil, not latched. (Like you can consider a motor starting contactor to be bi-stable. On or off but not halfway.)
Have you considered a solution to disable charging upstream rather than disconnecting the batteries?
Can you use lower rated multiple pole relays or contactors to disconnect the solar panels?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[itsmoked]

Mike...

I like it!



Bill; Since I have 10 of these batteries in parallel I can't use the charger as the ultimate disconnect. The most likely issue will be one cell out of 80 being low capacity. This, of course, will cause its battery to be its lousy capacity. This means that battery will 'fully' charge early AND 'fully' discharge early and then needs to be disconnected from the rest of the bank. Meanwhile the inverter can continue drawing and/or the charger can continue charging the bank.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[itsmoked]

News Flash!!

After 4 days of waiting for the 3 GIGAVAC I got an email from the seller saying, "Since I don't get out to the warehouse very often I didn't realize we'd already sold them. So I canceled your purchase. Sorry (sucker)."

That must've been an omen to go back to the drawing board to come up with something that draws [highlight #FCE94F]ZERO[/highlight] static power like I wanted in the first place. I'll be making something. Stay tuned.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[BrianE22]

Pretty straight forward to make an over-center (bi-stable) mechanism using a spring and a couple of solenoids. You'd have to find some contacts. It would probably be more expensive then buying something though.

 

[itsmoked]

I think I've got a clean solution. Cheap and what I wanted. I'll fill you all in with it when I get a moment.

Keith Cress
kcress -

http://www.flaminsystems.com