5 minute control power backup for soft start 2

[bentov]

We have soft starters that won't reset until 5 minutes AFTER an overload fault with control power on. The cost effective arrangement (fault relay powers shunt trip of main CB) removes power so can't reset. A UPS could of course be used to keep control power on. That seems overkill though (and would look hokey in the panel, require battery maintenance, etc.), looking instead for some kind of miniature capacitor type carryover. Haven't actually measured but I expect the consumption is only a few milliamps, only needed for the 5 minutes in that one scenario - OK if power fails after that. I see memory backup stuff but only low voltage DC so far, anyone know of something I could use here?

 

[jraef]

The condition you are describing is DELIBERATE and is a safety function built into the solid state overload protection scheme, specifically to AVOID having someone try to defeat the programmed minimum cool down time by cycling control power. When your motor overloads, you should NOT attempt to re-start it, it is ALREADY hot! The 5 minute delay is there to prevent you from re-starting it after the overload condition. The concept is to emulate the "trip-free" action of an electromechanical thermal OL relay, where it will take time for the bi-metal strips to cool down or the melting alloy to re-solidify and you cannot re-start by just holding down the reset button. But with solid state, the concept must be a function of programming, so to prevent someone from defeating it by killing power to the internal microprocessor, they build in what is called "retentive thermal memory". Some better ones change the value of the time depending on the AMOUNT of overload current and the resulting calculated motor heat.



"Will work for (the memory of) salami"

 

[bentov]

Hey Jeff, how you doing?

I'm OK with the 5 minutes, problem is my trip delay is around 2 seconds (using an external timer wired to the soft start fault relay, triggers the CB shunt). So the soft starter is stuck in fault, trips the CB again 2 seconds after you turn it on, over and over . . . if I yank the trip timer, leave power on 5 plus minutes, press the reset button, all is well, but there's no timing without power (verified with trial and error). Hours or days after the trip, we must rig power for 5 minutes to reset. Setting the trip timer to 5 plus minutes would solve that problem, but not the flaming motor thing when SCRs short!

 

[LionelHutz]

This sounds like a C-H IT starter. We even replaced one when the customer couldn't figure that out. Stupidly, it's not described in the manual at all, or at least the last time I looked.

A smaller 12V battery and replacing the power supply with a charger could work. Or, you'd need to figure out the idle current draw and then size a capacitor so it holds-up the voltage long enough with that drain. You'd also have to figure out how to re-charge the capacitor.

The simplest would be a 24VDC UPS setup. I believe Weidmueller and Phoenix both have what you need. The Phoenix is a 24VDC module you connect the 24VDC power supply into and also connect their 24VDC battery to as well as having terminals to power the load. It charges and gives a bumpless transfer to the battery. You could also likely use the external shutdown input to tell it to go off once the starter resets.

 

[jraef]

So I take it that the problem behind the problem here is that there is only one Fault relay contact on the soft starter, which trips on Shorted SCR and on Overload. So when it is a Shorted SCR trip you need to use it to open the breaker, but when it is an overload trip, not having control power interferes with the retentive thermal memory timeout.

Honestly, that's a brain fart on the part of the designer. When I worked on digital soft starters designs, we thought of that, as did most other soft starter designs I have seen. Most of them just provide separate contacts dedicated to shorted SCR detection because that is a somewhat unique situation attributable to using SCRs (although it is the same for welded contacts). Others simply use a super cap in the RTM circuit to keep the timer counting down regardless of the control power status. We did both, but the super cap was done mainly just for the convenience of the users; we thought it to be dumb to not count off time as a result of a power loss as being just as valid as any other off time. The operative issue is just not having energy going into the motor, how that happens should be irrelevant. You seem to have found one design that does neither.

So if it is the CH-IT starter described by LionelHutz, his solution might be easier than replacing it at this point. Their design requires a 24VDC power supply anyway, so there are easy battery options available for that. The other option might be to disable the overload protection in the soft starter (if possible) so that the Fault relay only works for the Shorted SCR detection, then add in a separate overload relay down stream of the soft starter.

"Will work for (the memory of) salami"

 

[bentov]

Thanks for the replies. LionelHutz, the soft starters are Siemens 3RW40s. I've just been through a 212 page manual (big improvement over the one in the box, wasn't available last time I wrestled with these). It still doesn't explain the overload reset thing though, had to discover that the hard way. Nothing new to help find a way to wire around this issue either.

We didn't catch this in the shop because we test with an undersized motor. The trip that causes (no load) clears with power cycling so works with the shunt trip breaker. It's hard to replicate the real world OL trip, lucky I was able to mess with this one onsite in a spot where I could start it a bunch of times with the OL turned down and see the trip behavior - wired it different ways (regular run command closure with control power always on, also the alternate with jumpered run terminals and run command via control power closure, etc.), also installed a (several years newer) replacement starter. It is the case that control power must be on for 5 minutes before an OL fault can be reset.

Though we built this setup, I know we installed a factory built pump panel a few years back that had the same components (i.e. shunt trip breaker, external trip timer, NO isolation contactor). I expect that one has either never tripped on OL, or somebody figured out pulling the trip timer to accomplish OL reset - timer's probably lying in the bottom of the panel (or at the bottom of a nearby canal!). If so, now if a thyristor shorts the motor winding will likely fail . . .

I also know from experience that a UPS for one of these has to be way oversized (like 1000VA on a bigger unit, where a 250VA CPT is adequate) to be able to pull in the bypass contactor when the time comes.

So the magic bullet would be a little component array that can be stuck into the A1 A2 terminals. It must be OK to be always powered, and not cause trouble while in parallel with the load (control board power supply plus bypass contactor coil, inrush and holding). Now I've described it, it seems pretty unlikely (sort of a reverse snubber, probably zap the board).

I guess the options are the fat UPS (with guaranteed battery failure in a couple years), or some miniature capacitor version to be swapped into A1 A2 via relays at power off. When power is removed (via shunt trip)it needs to supply the 120VAC circuit board power only for the 5 plus minute thermal timeout, so when main power is restored the thing can be reset. It's OK if it peters out, probably cause a low volts fault but that's cleared via power cycle anyway. Then we just tell the user to always wait the 5 minutes, then push reset immediately after resetting the breaker. Then if it WAS an OL fault they can run again, but if there's a (non-resettable) device fault like the failed thyristors the short timed shunt trip will protect the motor as intended.

I know an isolation contactor would take care of this, of course, but hate to give up on this arrangement - seems tantalizingly close to everything we need without the extra expense and panel space . . .

 

[bentov]

jraef, guess we were writing simultaneously . . .

Brain fart is a kind way of putting it, couple other epithets occurred to me while figuring out the sequence.

You have described the problem behind the problem exactly. I was going to list the option of having the manufacturer redesign the product with the separate fault contacts and/or internal timer for power off but I know how realistic that is (not!).

As usual, trying to work around this in a complicated way I overlooked the obvious solution you suggest, the separate OL relay. It is indeed possible to turn off the onboard OL (via the Class dial, OFF being an available position - only way I could help the user reset the dang thing over the phone a couple time till I finally had time to visit).

So that's probably the most cost effective thing I can do, though it feels a little dumb to add an OL to a starter that already has one.

Thanks again, you are the man!

 

[bentov]

Still, the capacitor thing would be cool (and cheap!).

I just read a bit on AC voltage capacitor charge discharge behavior, looks like all I have to do is some Fourier analysis, solve some simple equations for (presumably varying over a wide range in this case) resistance and reactance (where "j is the imaginary component, and w is the angular frequency"), apply some derivative functions and a Laplace transform or two . . . OK, I'm back, results were, uh, inconclusive.

Is there a chance of success, and a safe way to experiment with an always-connected super cap on the 120VAC side here?

 

[LionelHutz]

There is no chance of success. A capacitor won't work as a storage device on the AC side.

 

[waross]

How about a relay that both disables the shunt trip and inhibits the soft start from starting until the timer has timed out. This could be on a six minute timer to ride out the five minute cool-down timer.

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

 

[jraef]

How about a relay that both disables the shunt trip and inhibits the soft start from starting until the timer has timed out. This could be on a six minute timer to ride out the five minute cool-down timer.

Because if it IS faulting on a shorted SCR (or more specifically 2 shorted SCRs in opposing phases), you will have unrestricted current flow through the motor winding for that 6 minutes. In fact, the Shorted SCR scenario can only be detected when you turn the soft starter Off, because it is the LACK of a voltage drop across the SCR stack that tells you that you have a short, and while it is being fired, you have the same condition. That's the reason for the shunt trip, and thereby the problem with it being a common fault relay with an overload trip. They are very different issues and should be dealt with differently, having only the one fault contact in common is the brain fart.

bentov,
I have a low cost solution for you, based on how we discriminated the really bad condition I described above (unrestricted current flow) from a simpler problem of having only one SCR shorted, which is bad, but not worth shut tripping over. Get 2 cheap little current relays, a simple Go-No Go type, capable of fitting over your conductor, but with a very low current value, say 1A if possible. Put them on any two of your output conductors. Then wire the Fault Relay to another relay, and one contact off of that slave relay goes in SERIES with the Current Relays (in parallel), to the Shut Trip. So the logic is, if there is a fault, AND current is flowing through either of the current sensors, THEN shunt the breaker. Fault Relay but No current flow = Overload, no need to Shunt.

"Will work for (the memory of) salami"

 

[waross]

Thanks for the explanation, Jeff.

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

 

[bentov]

jraef, that's a great idea!

Remember these are the 2 phase starters so there's no survivable 1 SCR short scenario - I'm thinking maybe just one current relay, put it on the middle (pass through) leg, maybe gain a better ground fault response time in the bargain (compared to the breaker trip curve only). And maybe a current relay with appropriate available trip delay adjustment could eliminate the trip timer and the slave relay (so fault relay powers current relay coil, delayed current relay output contacts supply Shunt Trip) - improve protection and simplify the schematic, a positive outcome for sure. I'll rig it up and experiment, let you know how it turns out.

Thanks again, beats the heck out of the full size extra OL (or contactor) solution.

Thank you also LionelHutz for the capacitor advice, saved me some time and trouble there. I started down that road a little (just need a rectifier, figure out the sizing on the RC charging circuit, then a discharge/inverter scheme to go around the rectifier to re-power the AC side, some sensing stuff to know when that's called for and when to switch back to charging mode, and then, and then, . . .). That avenue is certainly risky since no parts are available for these . . . spark a board, buy a starter.

 

[jraef]

I forgot that the 3RW40 was 2 phase, so you're right, a single sensor on the center pole would work. That delayed contact is also an elegant condensation of my logic, one that I would not have come up with until 2 months after installing the superfluous hardware, and it would have come to me in a dream that would wake me up at 3:00 AM...

http://www.nktechnologies.com/current-sensing-switch/asl-current-switch.html

This one has a built-in 2 second delay, should be enough. Current flow from sensor because starter is running, but no fault trip, no shunt. Fault trip on overload, unit shuts off, no current flow after 2 seconds, no need to shunt trip. Fault trip, unit shuts off but current flowing anyway 2 seconds later, need to shunt. I like it.

"Will work for (the memory of) salami"

 

[bentov]

You know I have that very one in inventory, really like it cause it's self-powered . . . which means it won't work here (will be on all the time), have to track down one of the SSACs I used to use that require external power instead. It's always something!