Any 15A force-guided relays/contactors with non-PCB contacts?

[sparkydog]

I'm fiddling with a design that has similarities to this; in particular, I'm also trying to switch a lot of power (in my case, 120 VAC at, ideally, 20 A, minimum 15 A), and trying to do so on a PCB is challenging.

Another user on that thread brought up the reminder that contactors with off-board contacts are a thing.

In my application, however, I'd really like to use force-guided ("safety") relays for, well, safety. 🙂 In particular, with an FGR, I can tell that it wasn't welded before I flip it on, and I can detect if I've turned it off but it hasn't opened.

Right now, I'm looking at using Elesta SIP 512s. However, the contacts for these are PCB mount (through-hole). I am planning to use bus bars instead of traces to carry the load power, but that's still more complicated than connecting them directly to the mains 12ga wire that they're switching.

Does anyone know of an alternative part I could use for this application? It must be force-guided with both NC and NO sense contacts and have NO load contact(s) able to handle at least 15 A. Good inrush rating is strongly preferred. I'm flexible on coil voltage (including whether it's AC or DC), though silent operation is preferred. Ideally, the coil and sense contacts would be PCB-mount while the load contacts would not be PCB mount. Something that doesn't cost several hundred dollars is also strongly preferred.

[PCB.Wiz]

In my application, however, I'd really like to use force-guided ("safety") relays for, well, safety. 🙂 In particular, with an FGR, I can tell that it wasn't welded before I flip it on, and I can detect if I've turned it off but it hasn't opened.
Right now, I'm looking at using Elesta SIP 512s. However, the contacts for these are PCB mount (through-hole). I am planning to use bus bars instead of traces to carry the load power, but that's still more complicated than connecting them directly to the mains 12ga wire that they're switching.

That's a tough combination.

Another approach could be use what you have found, and to focus on very short, wide traces, and make your own short connections which become 'directly to the mains'.

We've used the below connectors for high current PCB connects.
We prefer right angle over vertical, as the force is managed better, and this particular model, adds a lip so the PCB edge can take the force, and the cranked forming means it stays vertical.
Surprising how many of these QC PCB tabs have given zero thought to the force paths.


   

[thm_w]

Almost every kill-a-watt style device has contacts going to a PCB with traces rated to carry 15A.
Lots of 20A IOT plugs also exist that use PCB mount relays.
https://www.elektroda.com/rtvforum/topic3987713.html

[uer166]

20A on a PCB is not challenging, but rather routine even on cheap fabs. Ignore the online calculators and realize that it's not the trace width, but rather the aspect ratio is what matters. Also ignore Jon claiming there is "no need to pass current through PCB, you can find relays with separate terminals". In the real world most devices (interconnects, relays, fuses) are designed for PCBs since it is most optimal. 4 layer with 1Oz each should be more than enough for a long-ish path.

I've designed a PCB with 160A/44kW going through the substrate no problem WITHOUT bus bars, for the love of god do not overcomplicate/overthink it. 20A is 🥜.

[wraper]

I'm fiddling with a design that has similarities to this; in particular, I'm also trying to switch a lot of power (in my case, 120 VAC at, ideally, 20 A, minimum 15 A), and trying to do so on a PCB is challenging.

And what challenging about it? 20A is not some extraordinary current. Just use wide traces from a relay to appropriately rated connector. Also you can put traces on both sides of dual layer PCB.

[PCB.Wiz]

20A on a PCB is not challenging, but rather routine even on cheap fabs.
Ignore the online calculators and realize that it's not the trace width, but rather the aspect ratio is what matters.

Certainly short/wide traces help, and even copper that does not seem to be directly in the current flow, can help cool the copper that is carrying current.
That's why copper pour / maximal area gives the lowest temperature rise.

I would not entirely ignore the online calculators, but certainly use them as a guide only. 

At some point, you will need to build and measure. 

Temperature rise comes from all parts of the system : the formed pins, the solder, the traces, and even the cables/crimped connections.

[uer166]

I use IDC2D to do initial FEA simulation of plane resistance, very easy to use. 9 times out of 10 the relay contact resistance losses are more than the rest of the whole system, including wires, PCBs, and interconnects, which is why I'm personally baffled by the hyperfocus on PCB substrate when it is a tiny fraction of I2R losses.

I agree that better build it and try, but with even some napkin math it should be obvious that PCB will not be a problem if a little care is taken.

[jonpaul]

use relay or contactor with screw or 0.250" AMP tab terminals, use #16..#14 AWG 600V rated stranded wire, not a PCB

j

[sparkydog]

Y'all... the problem is, at least with my current design, my relay-to-terminal distance isn't short. I could manage 5mm traces at 2oz (i.e. using both front and back) easily enough, but those would be dropping... 5 W. That's a very non-trivial amount of heat, and that's on top of whatever I'm losing in the relays.

The relays themselves drop 1.3 W (each) just from the coils, and there are other heat-generating components kicking around. Now, I am planning on active cooling, but I'm also a bit paranoid how things are going to be arranged in the box that I'm actually getting useful airflow.

use relay or contactor with screw or 0.250" AMP tab terminals, use #16..#14 AWG 600V rated stranded wire, not a PCB

Thank you, Captain Obvious. If you'd read the original message, you'd note I already cited exactly this advice, and that the whole point of this thread is that I don't know of / can't find any such relays/contactors.

Also, I'm pretty sure 20A is going to fry 14ga stranded. NEC specifies 12ga solid for that, and that (or even 10ga) is what would be carrying the load to/from the box.

[ywara]

From an industrial point of view, you would most likely use a contactor with aux contacts rather than a "force-guided relay" in this application.

Devices sold as force-guided relays generally do not handle much current. I don't have much cause to use them often, but it seems they are mostly intended for adding additional contacts to safety relays or injecting a signal into a safety circuit from an external source. They are not used to switch large loads. Certainly you are not going to switch motors or heating elements with these.

It sounds like you are in North America, so you may want to use a definite purpose contactor with an aux contact. The aux contact will be, effectively, force-guided by the contactor and can be used for monitoring (i.e. to detect welded contacts). These are often available in 1-pole configurations for single phase loads, so no need to waste the 2 extra poles on a conventional 3-phase contactor or the second pole the rest of the world insists on having because their electricians are too dumb to put neutral in the same place every time.

[uer166]

How long are the traces? The usual solution to your problem is to add a pair of layers to your PCB, and use those entirely for the power path. Instead of 5mm they become 50mm (or whatever), unless your PCB itself is somehow super narrow and long.

Fabs are cheap nowadays!

[thm_w]

Y'all... the problem is, at least with my current design, my relay-to-terminal distance isn't short. I could manage 5mm traces at 2oz (i.e. using both front and back) easily enough, but those would be dropping... 5 W. That's a very non-trivial amount of heat, and that's on top of whatever I'm losing in the relays.

5 watts doesn't sound like a lot given that these traces would work out to be ~20cm long, no? Unless its a really skinny PCB.

and that is 5W without any busbars or parallel wires soldered on top or anything.

[sparkydog]

From an industrial point of view, you would most likely use a contactor with aux contacts rather than a "force-guided relay" in this application.

Any specific suggestions? That was the point of the thread...

How long are the traces?

About 25-30 cm total, though I'm probably going to redo the board to try to shorten that. (Unfortunately, that means shoving the relays right up against each other and having to make the entire enclosure bigger. Other options for relays/contactors, especially ones that don't need 17 cm² of PCB space, are of interest.)

The usual solution to your problem is to add a pair of layers to your PCB, and use those entirely for the power path. Instead of 5mm they become 50mm (or whatever), unless your PCB itself is somehow super narrow and long.

That's going to add $20, which isn't awful... but I'm not sure I can practically do that. There's all manner of other pins breaking up any plane I'd try to pour and areas I need to stay away from that, even though the board is plenty big, it would be hard to make a pour much wider.

I'm attaching what the current board might look like trying to go that route. Mind, this is also trusting the FR4 to be sufficient insulator, which AFAIK, it isn't. Realistically, I don't think there's an alternative to ending up with an 'L' shaped pour from the outsides to the bottom, and the pour between relays being more-or-less as shown.

Right now, my guess is I'll wind up smashing the relays together and running 50×8 mm pours, top and bottom, to T5. And I may just go with 2-layer 2oz boards. Again, it's an extra $20 or so, but that's $20 that I don't need to spend on bus bars.

5 watts doesn't sound like a lot given that these traces would work out to be ~20cm long, no? And that is 5W without any busbars or parallel wires soldered on top or anything.

The original plan was to use bus bars instead of traces, yes. Throwing that 5 W number out was reacting to the people saying not to bother... and that's 5 W for one board of a system that's triple-redundant. That's about 12-14 W difference between "don't worry about it" and having a decent volume of copper.

There's an electromagnet sitting next to me that's allegedly 2 W. Let me tell you, that thing gets toasty, and that's in open air. I have a healthy respect for what a few watts of heat will do in an enclosed space.

[PCB.Wiz]

Right now, my guess is I'll wind up smashing the relays together and running 50×8 mm pours, top and bottom, to T5. And I may just go with 2-layer 2oz boards. Again, it's an extra $20 or so, but that's $20 that I don't need to spend on bus bars.

It's correct to focus on lowest added-watts.
I would build and measure something, and certainly increase the pour areas. Even tho some copper does not carry current, it still gives heat spread cooling. 

You can also make 'simpler planar bus bars' from PCB material, for very low added cost, with holes and copper exactly where you want them.

Addit: the connector seems to be only 25A rated, with single PCB pins.
I would bump to a higher current rating, and either dual pins, or wider tags that go into slots. This is low hanging fruit stuff... 
A quick scan at lcsc finds some at 3mm x 0.7m tags and some 4.0 x 0.9mm tags.

Addit2: Also, do not rely on the plated solder joints alone carrying current thru the board, add some redundant vias too.

[ywara]

From an industrial point of view, you would most likely use a contactor with aux contacts rather than a "force-guided relay" in this application.

Any specific suggestions? That was the point of the thread...

No. Absolutely not. You may be looking for specific suggestions, but you gave no specific requirements. I can only provide general advice. If an off-board solution for ~$10-20 in BOM cost is acceptable then use your favorite search engine and find a definite purpose contactor meeting your actual needs and go ham. If not, then hire a consultant and give them some real information.

[uer166]

FR4 to be sufficient insulator, which AFAIK, it isn't

Since when?! FR4 can be considered a reinforced isolated cemented joint in the UL world given a few minor conditions, which is way overkill for phase-to-phase insulation per your use case! You only need either basic or functional depending on your fusing scheme. If you add 2 internal layers, the spacings can also be massively reduced, so the vias/TH don't make huge blow-holes to your internal planes. Looking at the layout, 20A should be a piece of cake, there is plenty to optimize as-is even with 2 layers. I agree with the previous poster, you have provided very little relevant information, so any specific advice is impossible.

[sparkydog]

Addit: the connector seems to be only 25A rated, with single PCB pins.
I would bump to a higher current rating, and either dual pins, or wider tags that go into slots. This is low hanging fruit stuff... 
A quick scan at lcsc finds some at 3mm x 0.7m tags and some 4.0 x 0.9mm tags.

WDYT of Keystone 8196 or something similar? I don't love how exposed it is, but it's small, and the PCB connection is certainly beefy.

I have no idea what a "tag" is, at least since I assume you aren't talking about any of these meanings... Also, nothing here seems relevant.

Any specific suggestions? That was the point of the thread...

You gave no specific requirements.

I didn't?

It must be force-guided with both NC and NO sense contacts and have NO load contact(s) able to handle at least 15 A. Good inrush rating is strongly preferred. I'm flexible on coil voltage (including whether it's AC or DC), though silent operation is preferred. Ideally, the coil and sense contacts would be PCB-mount while the load contacts would not be PCB mount. Something that doesn't cost several hundred dollars is also strongly preferred.

If that isn't specific enough, or isn't the right specifics, please clarify. (Note: the "force guided" requirement is specifically that it must guarantee that no NC contact is closed if an NO contact is closed, and vice versa.)

If an off-board solution for ~$10-20 in BOM cost is acceptable then use your favorite search engine and find a definite purpose contactor meeting your actual needs and go ham.

My "favorite search engine" (general purpose, anyway) does a terrible job identifying contactors that meet my specific needs, and half of the ones it turns up are >$100. I haven't found any $20 contactors that meet my originally specified requirements.

Octopart doesn't really catalog contactors. What search engine do you use that is able to filter by contact current rating and number/type of contacts? Is there even such a thing as an affordable (sub-$60) force-guided contactor?

I agree with the previous poster, you have provided very little relevant information, so any specific advice is impossible.

Well, then, ask questions because I have no idea what else I can say. The load-switching part of the circuit a) should be obvious from the previous picture (seeing as it consists of a whole three parts and three connections), and b) is as simple as it possibly can be. All it does is sit on a 20A mains wire and interrupt current. It's literally this thing but with a) with a magnetic coil to actuate the switch and b) aux contacts so I can read some information, electronically, about what position the switch is in. That's it. (Okay, pedantically it's several switches wired in series.)

What's the load? I don't know. That's the point; it's intended to be able to handle anything that you could safely put on the basic switch on a 20A circuit. I'm probably not going to fuse the individual PCBs. I could fuse the whole system, but its unclear how that would be other than redundant with the upstream breaker. (Most of the other stuff in the system is individually fused, just not the load-switch leg.)

[PCB.Wiz]

Addit: the connector seems to be only 25A rated, with single PCB pins.
I would bump to a higher current rating, and either dual pins, or wider tags that go into slots. This is low hanging fruit stuff... 
A quick scan at lcsc finds some at 3mm x 0.7m tags and some 4.0 x 0.9mm tags.

WDYT of Keystone 8196 or something similar? I don't love how exposed it is, but it's small, and the PCB connection is certainly beefy.

Yes, that style has good PCB connection, but as you say is exposed.

More are here
https://www.lcsc.com/products/PCB-Welding-Terminal_11512.html
some have captive nuts, for a bit more strength.
https://datasheet.lcsc.com/lcsc/2106070034_XFCN-TT44080080-0820_C2689180.pdf

I have no idea what a "tag" is, at least since I assume you aren't talking about any of these meanings... Also, nothing here seems relevant.

Sorry, I was a bit unclear
Parts like this, that would use a PCB slot, and have much larger cross section on terminals intended for solder tags.

https://www.lcsc.com/product-detail/Barrier-Terminal-Blocks_Cixi-Kefa-Elec-KF45H-9-5-2P_C707433.html 
https://www.lcsc.com/product-detail/Barrier-Terminal-Blocks_Cixi-Kefa-Elec-KF48H-9-5-2P_C707675.html
3 x 0.7mm

and this one is larger, with 4 x 0.9mm cross section.
https://www.lcsc.com/product-detail/Barrier-Terminal-Blocks_Cixi-Kefa-Elec-KF65H-11-0-2P_C707592.html



 

[ywara]

Any specific suggestions? That was the point of the thread...

You gave no specific requirements.

I didn't?

It must be force-guided with both NC and NO sense contacts and have NO load contact(s) able to handle at least 15 A. Good inrush rating is strongly preferred. I'm flexible on coil voltage (including whether it's AC or DC), though silent operation is preferred. Ideally, the coil and sense contacts would be PCB-mount while the load contacts would not be PCB mount. Something that doesn't cost several hundred dollars is also strongly preferred.

If that isn't specific enough, or isn't the right specifics, please clarify. (Note: the "force guided" requirement is specifically that it must guarantee that no NC contact is closed if an NO contact is closed, and vice versa.)

If an off-board solution for ~$10-20 in BOM cost is acceptable then use your favorite search engine and find a definite purpose contactor meeting your actual needs and go ham.

My "favorite search engine" (general purpose, anyway) does a terrible job identifying contactors that meet my specific needs, and half of the ones it turns up are >$100. I haven't found any $20 contactors that meet my originally specified requirements.

Octopart doesn't really catalog contactors. What search engine do you use that is able to filter by contact current rating and number/type of contacts? Is there even such a thing as an affordable (sub-$60) force-guided contactor?

I agree with the previous poster, you have provided very little relevant information, so any specific advice is impossible.

Well, then, ask questions because I have no idea what else I can say. The load-switching part of the circuit a) should be obvious from the previous picture (seeing as it consists of a whole three parts and three connections), and b) is as simple as it possibly can be. All it does is sit on a 20A mains wire and interrupt current. It's literally this thing but with a) with a magnetic coil to actuate the switch and b) aux contacts so I can read some information, electronically, about what position the switch is in. That's it. (Okay, pedantically it's several switches wired in series.)

What's the load? I don't know. That's the point; it's intended to be able to handle anything that you could safely put on the basic switch on a 20A circuit. I'm probably not going to fuse the individual PCBs. I could fuse the whole system, but its unclear how that would be other than redundant with the upstream breaker. (Most of the other stuff in the system is individually fused, just not the load-switch leg.)

As the designer, you need to decide what loads you are willing to support, what duty cycle this will operate at and a number of other things. Look, I'm not a power EE or anything close but I can tell you're missing information. Starting a 2HP motor under load, turning on an 1800W space heater, and switching some office electronics are all very different scenarios.

$20 is pretty optimistic in retrospect for a DP contactor with aux contacts. Maybe in quantity, but at quantity 1 I'd guess $30-75. This will probably meet your switching needs, but will definitely not be quiet. No electromechanical relay will be silent.

Eaton:
https://www.supplyhouse.com/Eaton-C320KG3-1NO-1NC-Auxiliary-Contact-Side-Mount
https://www.amazon.com/Eaton-C25ANB125A-Contactor-Terminals-Inductive/dp/B009R3ECVE

No-name:
https://www.automationdirect.com/adc/shopping/catalog/motor_controls/definite_purpose_contactors/definite_purpose_contactors/wdp25-1l-120
https://www.automationdirect.com/adc/shopping/catalog/motor_controls/definite_purpose_contactors/definite_purpose_contactor_accessories/wdpaxfc11q
Disregard this, aux contacts require a 2- or 3-pole contactor

But, the more I think about this the less I understand why you want force-guided contacts. If you detect welded contacts, where the load is powered when the coil is not on, what are you going to do about it? Open another contactor/relay upstream? Fire a shunt trip on a circuit breaker? Or, just alert the user?

If you're only monitoring and the upstream side of your switching device is permanently connected to the line - which is what I suspect - then you really only need to check the downstream voltage from your switching device and feed that back to your controls. The extra contacts provide no real benefit.

[sparkydog]

As the designer, you need to decide what loads you are willing to support, what duty cycle this will operate at and a number of other things. Look, I'm not a power EE or anything close but I can tell you're missing information. Starting a 2HP motor under load, turning on an 1800W space heater, and switching some office electronics are all very different scenarios.

The scenario is "it should be okay for anything that I would plug directly into a 20A outlet wired directly to a breaker". Continuous duty (it's going to be energized for many hours on end).

$20 is pretty optimistic in retrospect for a DP contactor with aux contacts. Maybe in quantity, but at quantity 1 I'd guess $30-75.

Hey, you dropped the $20 figure. I knew that's unrealistic. $30 would be a steal. Ideally I'd want something around $40-$50, i.e. similar in price to the SIP 512 I'm otherwise considering. (It can be a bit more, especially if the coil is just plain 120VAC, due to needing fewer components and smaller PCBs vs. the SIP 512.) My searches tend to turn up parts that are well over $100, though.

This will probably meet your switching needs, but will definitely not be quiet. No electromechanical relay will be silent.

Ah! Sorry, I should clarify; I meant quiet while energized. It can be (almost) as loud as it wants when switching, as long as I don't need hearing protection if I'm nearby 🙂. (I was under the impression that an AC coil would "hum", but at least some sites claim that even AC coils shouldn't make noise while energized?) Anyway, that's a preference, not a hard requirement.

But, the more I think about this the less I understand why you want force-guided contacts. If you detect welded contacts, where the load is powered when the coil is not on, what are you going to do about it? Open another contactor/relay upstream? Fire a shunt trip on a circuit breaker? Or, just alert the user?

If you're only monitoring and the upstream side of your switching device is permanently connected to the line - which is what I suspect - then you really only need to check the downstream voltage from your switching device and feed that back to your controls. The extra contacts provide no real benefit.

Shunt-trip breaker. Also, monitoring (which is supposed to work with no load energized), though I agree knowing the contacts aren't welded before engaging the load only really helps after the fact.

A downstream relay (or whatever) only tells me if the system as a whole is energized. One advantage of aux contacts (besides that I'm not convinced that's more reliable than depending on a separate system to detect if the system as a whole is energized) is that I know if any relay has welded (and which one). That means the fault detection will trip even if one of the redundant relays did its job.

[sparkydog]

Well... the only thing I've found that looks viable is https://www.se.com/us/en/product/LC1D25G7. Or is that https://www.se.com/us/en/product/LC1D25F7? Do I need to take a multimeter to an electrical outlet to find out if I have closer to 110VAC or 120VAC?

There's also 933210 a.k.a. LC1D32 from https://www.electrodepot.com (whoever they are) for somewhat cheaper, though ED's idea of a data sheet is not exactly confidence-inspiring.

In either case, those make a SIP 512 look small...

[ywara]

I've not been paying attention so I am sure this is quite late, but if you are talking about shunt trip breakers and IEC contactors, cost is already out the window. Take a look at WEG series CWBS, they are safety rated and slightly cheaper than Schneider. Still, they are 3-phase and you don't need it.

I'm out of my depth at this point, it sounds like either a novel problem or an (standards-wise) unnecessary self-created problem.