EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: redgear on September 09, 2020, 05:23:58 pm
-
I am looking for a relay that can reliably used in production to switch a load with 0.6 power factor and 1.85 crestfactor. The load draws 2A nominal current with inrush current as high as 30A lasting 200 microseconds.
The ambient operating conditions of the relay will be 35-40 degs. The average time between switches will be 2 minutes expecting in cases of safety cut-offs during which it can be as low as few hundred milli seconds(rare). The relay should be able to complete ~20k switching cycles.
Right now I have been looking at RZ03-1A4-D005 (https://www.te.com/commerce/DocumentDelivery/DDEController?Action=showdoc&DocId=Data+Sheet%7FRZ%7F0516%7Fpdf%7FEnglish%7FENG_DS_RZ_0516.pdf%7F1-1415899-0). It is designed to meet IEC-60335 and is UL508 approved. The relay is rated for 80A inrush and about 150k cycles at 5A. I was optimizing my BOM cost and I felt like this one is a overkill. It costs about 1.44 in Mouser and isnt available on LCSC.
I was looking for alternatives and found these relays on LCSC:
1) Omron G4A-1A-PE DC12V (https://omronfs.omron.com/en_US/ecb/products/pdf/en-g4a.pdf) @ 0.84 on LCSC -- Again I feel this is overspecced but it is way cheaper than the TE part and comes from a reputed brand.
2) Hsin Da Precision 133-1A-5DS (https://datasheet.lcsc.com/szlcsc/Hsin-Da-Precision-133-1A-5DS_C259260.pdf) @ 0.38 on LCSC -- From the datasheet they are very similar to the TE and Omron relays and also have a UL Safety approval but I am not sure if they are a good brand to rely on. Can these handled the specified load?
3) HF33F/005-HS3 (https://www.generationrobots.com/media/JQC-3FF-v1.pdf) @ 0.25 on LCSC -- same doubts as above.
Hope I get some suggestions. Thanks.
-
You mention a power factor which I'm going to assume is an inductive load at the moment. If so you need a robust relay to handle that properly, ideally their will be mitigation at the source of the inductive load. The high voltage arcing from the collapsing coil voltage can have a pretty significant impact on relay life. So you need to consider carefully how that load will behave when turning off as well as the inrush when turning on.
To be perfectly honest with you I'm never impressed by engineering that cheaps out on something like this. Working in automation I run across this frequently, some times to the point relays have to be changed on a yearly PM schedule. The worse thing you can do for the reputation of a product is to put a relay in place that doesn't last a year.
-
What exactly is the load? Can you tell us or is that a secret of your employer? I mean you're worried about $0.40 here, just pick out a Songle. But if it has a safety function, I would not cheap out.
Relays have different specs for switching motors. Many are not rated for it, like the RZ03. It's not purely contact current rating but also the open gap you may need to consider. A cheap or cramped product design with no arc suppression across the relay contacts, will ensure short life, so use a more expensive relay with larger contact spacing i.e is motor-use rated or splurge on a mov/snubber.
I find it's the arc protection that determines the lifetime and I put that in regardless of what management wants because they'll come to you months later complaining the relay isn't lasting very long. The hidden payoff is reducing EMI from the contact arc which causes the MCU to reboot...
-
You mention a power factor which I'm going to assume is an inductive load at the moment. If so you need a robust relay to handle that properly, ideally their will be mitigation at the source of the inductive load. The high voltage arcing from the collapsing coil voltage can have a pretty significant impact on relay life. So you need to consider carefully how that load will behave when turning off as well as the inrush when turning on.
To be perfectly honest with you I'm never impressed by engineering that cheaps out on something like this. Working in automation I run across this frequently, some times to the point relays have to be changed on a yearly PM schedule. The worse thing you can do for the reputation of a product is to put a relay in place that doesn't last a year.
What exactly is the load? Can you tell us or is that a secret of your employer? I mean you're worried about $0.40 here, just pick out a Songle. But if it has a safety function, I would not cheap out.
Relays have different specs for switching motors. Many are not rated for it, like the RZ03. It's not purely contact current rating but also the open gap you may need to consider. A cheap or cramped product design with no arc suppression across the relay contacts, will ensure short life, so use a more expensive relay with larger contact spacing i.e is motor-use rated or splurge on a mov/snubber.
I find it's the arc protection that determines the lifetime and I put that in regardless of what management wants because they'll come to you months later complaining the relay isn't lasting very long. The hidden payoff is reducing EMI from the contact arc which causes the MCU to reboot...
Thank you both.
It is no secret. It is just two electronic ballast that power two T5 lamps. I agree with your view of not stuffing cheaper components to cut costs. I am just afraid if this will be a overkill for our application. The relay will be operated at max 10 times a day. We would be very happy if the product lasts for 2 years before failing. The load behaves normally when turned off. When it is turned on the inrush currents lasts for a maximum of 200 microseconds and after which the current draw quickly drops to 2A. I have zener diodes on the relay coils to handle to surges when the relay is turned off. Should I still worry about arcing and suppression? The MCU will be on a separate PCB so EMI. Would the EMI still create problems?
Edit: Will it be better to replace the Zener with a MOV?
-
I think electronic ballasts are a largely capacitive load for many msec so the 0.2msec spec seems a bit odd for inrush time. It takes longer than that for the tube to ionize. There should be very little inductive kickback from the electronic ballast, compared to a magnetic ballast. I would take any open relay or switch or even loose wires, and on the bench observe the spark when you connect the ballast on/off. It pays to confirm the small 2A relay has a chance at a long life.
A zener across the relay coil or just the usual back EMF diode deals with the coil's kickback fine.
-
I think electronic ballasts are a largely capacitive load for many msec so the 0.2msec spec seems a bit odd for inrush time. It takes longer than that for the tube to ionize. There should be very little inductive kickback from the electronic ballast, compared to a magnetic ballast. I would take any open relay or switch or even loose wires, and on the bench observe the spark when you connect the ballast on/off. It pays to confirm the small 2A relay has a chance at a long life.
A zener across the relay coil or just the usual back EMF diode deals with the coil's kickback fine.
Thank You. We have confirmed the inrush current and time from the manufacturer.
And if I were to go with the chinese relays, which one of the two you think is good? HSin Da Precision or the HongFa ones?
-
Price and contact rating are not the only considerations for relays.
Especially when you switch mains voltage, and need proper isolation barriers.
I've bought a few of the Ali Ebay relay boards and found them horrible.
it has (among others) the text string: "10A 250VAC" printed on it and it might even be capable of switching this, but I opened one to look at the internals and the distance between the coil wires and the switch is less then 2mm (measured 1.92mm).
I do have much more confidence in the Omron G2 (or clones?) These have a much wider separation and even have a labyrinth between the coil and the contacts as extra safety during faults from gross overloads, melting contacts, arc an ionized gasses and such things. The relay will also fail under such circumstances, but it is much more likely to fail in a safe way compared to the blue "Songle" cube type relay.
If you're worried about USD 1.44 for a relay, that suggests you are using lots of them, probably have a factory and that sort of thing. This also implies you should now much better then the average bloke on an anonymous forum like this.
So I assume this is for production and you need 1000's of these relay's. In this case my suggestion is to go to LCSC and compare datasheets and build sizes, then buy a few samples of at least 10 interesting relays and test them yourself. Open one of each to inspect what's inside, and make a test rig to test longevity in your application.
If you want 10 times / day for 2 years, that's about 7000 switches, so switching once each second is only 2 hours of testing. (Do note that contacts will heat up more when switched so often). But I'm with the others here. If USD1.44 is too much for your relay and you want to cheap out, then you start to gamble with the reputation of your company within a few short years.
But then:
Why is this posted in the "beginners" section? As a beginner you should be worried more about quality / reliabily then about a few dimes. If you buy a relay outside of china and it has the proper testing stamps, then you're probably already on the safe side.
-
Price and contact rating are not the only considerations for relays.
Especially when you switch mains voltage, and need proper isolation barriers.
I've bought a few of the Ali Ebay relay boards and found them horrible.
it has (among others) the text string: "10A 250VAC" printed on it and it might even be capable of switching this, but I opened one to look at the internals and the distance between the coil wires and the switch is less then 2mm (measured 1.92mm).
I do have much more confidence in the Omron G2 (or clones?) These have a much wider separation and even have a labyrinth between the coil and the contacts as extra safety during faults from gross overloads, melting contacts, arc an ionized gasses and such things. The relay will also fail under such circumstances, but it is much more likely to fail in a safe way compared to the blue "Songle" cube type relay.
If you're worried about USD 1.44 for a relay, that suggests you are using lots of them, probably have a factory and that sort of thing. This also implies you should now much better then the average bloke on an anonymous forum like this.
So I assume this is for production and you need 1000's of these relay's. In this case my suggestion is to go to LCSC and compare datasheets and build sizes, then buy a few samples of at least 10 interesting relays and test them yourself. Open one of each to inspect what's inside, and make a test rig to test longevity in your application.
If you want 10 times / day for 2 years, that's about 7000 switches, so switching once each second is only 2 hours of testing. (Do note that contacts will heat up more when switched so often). But I'm with the others here. If USD1.44 is too much for your relay and you want to cheap out, then you start to gamble with the reputation of your company within a few short years.
But then:
Why is this posted in the "beginners" section? As a beginner you should be worried more about quality / reliabily then about a few dimes. If you buy a relay outside of china and it has the proper testing stamps, then you're probably already on the safe side.
Thank You.
The datasheets says the relays have a 8mm creepage distance. I do understand that chinese components are not very reliably but after reading from this forum I came to know that they have few good parts too and LCSC is a legit distributor. I just want to make sure I don't spend 1.44 when the 0.35 relay would fit my requirements. No I don't have a factory that needs 1000's of relay for production. I am just out of college and luckily work as an Intern in a small factory that uses about 500-600's of these relays monthly. I am still a beginner and have a lot to learn. I am not from a STEM background either. I compared datasheets from LCSC and these were really similar to the Omron and TE relays. They also have UL and TUV certifications.
-
The Omron from LCSC would be a good choice--it saves quite a bit of money but still will be of good quality, even the Chinese version. Nobody will criticize the choice of a properly specced Omron relay if things do go bad.
Electronic ballasts are pretty easy loads overall. I would ordinarily not want to use a part that specifies its first USP (feature) as "Extremely Low Cost" as does your 25-cent choice, but if you are only shooting for a two year lifespan ( :wtf:) I suppose either of the cheapos will do.
-
Building something as cheap as possible, you will eventually learn is a bad idea when you have any morals. Factor in the warranty costs and shipping etc. and the pennies saved is gone after just a few returns.
You could design the PCB with a dual-footprint to accomodate more than one make, assuming you can keep the creepage/clearance spacings meeting regulatory requirements. You have a fuse somwewhere, right? Usually one in the ballast.
The cube relay has been around for decades, many companies have offerings. But... HongFa JCQ-3FF has no inrush rating. Omron G5LE (https://omronfs.omron.com/en_US/ecb/products/pdf/en-g5le.pdf) the same, just a 10A max.
The Omron looks stronger. Sometimes the price is very low because the parts are sold in high volumes in appliance use.
-
Building something as cheap as possible, you will eventually learn is a bad idea when you have any morals. Factor in the warranty costs and shipping etc. and the pennies saved is gone after just a few returns.
Bad reviews and damage to the company's reputation will be far larger than the savings. How many units are going to be manufactured? What is the price target or, more to the point, how much profit per device? Does an extra 1$ per unit even matter?
Grab some representative relays and life cycle test them. Turn them on and off with double the load as fast as possible and let the test run as long as possible. Reality trumps specs every time! I would run at least 10x expected life.
Engineering is a 'no excuses' profession.