Dodgy switches?
By David Stonier-Gibson
Quite some time ago (2 years, maybe) I bought a bunch of cheap miniature toggle switches from China. They've just been sitting there, waiting for a use. This evening I went to use one in the exhaust system for my 3D printer. I decided to "buzz it out" with my multimeter to determine which way the toggle goes, so I could get it working in the direction I want.
These are Double Pole Double Throw (DPDT) switches, so there are 4 make contacts.
Well, what a drama that became! The continuity function gave totally sporadic results, either taking a long time to beep or not beeping at all. So I switched to the 200 Ohm range, and got readings from a few ohms to >200.
A dud switch, maybe? So I grabbed another one. Same thing. And another. And another. All wrong!
Silently I start cursing some dodgy operator for selling me rejects (oh yes, that's not unusual for China/eBay/AliExpress).
Then it struck me: Maybe it's contact contamination?
Start mini-lesson on wetting current: Switch and relay contacts can build up oxides and sulphides (from sulphur dioxide in the air) on the surface. These act as an insulator and increase the contact resistance. This is a very real phenomenon. Any members who worked for Telstra back in the type 3000 relay days will know about it. One counter-measure is to ensure enough voltage across and current through the contact as it switches to blast away the gunk. It's called wetting current. For very low currents you use gold (plated) contacts.
To test my theory I set up my linear bench power supply for 2A current limit and 10V open circuit. I would have gone to 30V, but there's enough capacitance in the output of the power supply to produce a significant spark when the leads are shorted. At 10V it was not too bad (the spark energy will go up with the square of the voltage, and at some point will do more harm than good). (I also have a switchmode bench power supply. Its spark at 30V was massive, probably enough to destroy a switch).
For the test I decided to operate the switch "dry", i.e. with no power, to see if it would mechanically clean the contacts. So I flicked the switch 20 times "dry", then 10 times for each contact with power (5 closures), recording the contact resistance before and after. Here are the results:
"Virgin" switch (it's been lying around untouched for so long that "old maiden" might be a better description!):
Contact 1: > 15 Ohms, erratically up to 70
Contact 2: 6 Ohms
Contact 3: >1 Ohms erratically up to 10
Contact 4: 0.6 Ohms
After 20 dry flicks (10 closures per contact)
Contact 1: 0.2 Ohms
Contact 2: 0.7 Ohms
Contact 3: 0.5 Ohms
Contact 4: 0.2 Ohms
After 5 closures 2A/10V
Contact 1: 0.1 Ohms
Contact 2: 0.1 Ohms
Contact 3: 0.1 Ohms
Contact 4: 0.1 Ohms
My resistance measurements have 0.1 Ohm resolution, with ~0.2 Ohm lead resistance subtracted. So the final 0.1 Ohm could be 0 or 0.2. No way of knowing without proper 4-wire measurement (which I can't be bothered doing, I'm spending the time writing this instead)
You can see that just the mechanical exercise went a long way towards clearing the gunk. The power switching really sealed the deal (a pun some teleco guys might get!)
These are Double Pole Double Throw (DPDT) switches, so there are 4 make contacts.
Well, what a drama that became! The continuity function gave totally sporadic results, either taking a long time to beep or not beeping at all. So I switched to the 200 Ohm range, and got readings from a few ohms to >200.
A dud switch, maybe? So I grabbed another one. Same thing. And another. And another. All wrong!
Silently I start cursing some dodgy operator for selling me rejects (oh yes, that's not unusual for China/eBay/AliExpress).
Then it struck me: Maybe it's contact contamination?
Start mini-lesson on wetting current: Switch and relay contacts can build up oxides and sulphides (from sulphur dioxide in the air) on the surface. These act as an insulator and increase the contact resistance. This is a very real phenomenon. Any members who worked for Telstra back in the type 3000 relay days will know about it. One counter-measure is to ensure enough voltage across and current through the contact as it switches to blast away the gunk. It's called wetting current. For very low currents you use gold (plated) contacts.
To test my theory I set up my linear bench power supply for 2A current limit and 10V open circuit. I would have gone to 30V, but there's enough capacitance in the output of the power supply to produce a significant spark when the leads are shorted. At 10V it was not too bad (the spark energy will go up with the square of the voltage, and at some point will do more harm than good). (I also have a switchmode bench power supply. Its spark at 30V was massive, probably enough to destroy a switch).
For the test I decided to operate the switch "dry", i.e. with no power, to see if it would mechanically clean the contacts. So I flicked the switch 20 times "dry", then 10 times for each contact with power (5 closures), recording the contact resistance before and after. Here are the results:
"Virgin" switch (it's been lying around untouched for so long that "old maiden" might be a better description!):
Contact 1: > 15 Ohms, erratically up to 70
Contact 2: 6 Ohms
Contact 3: >1 Ohms erratically up to 10
Contact 4: 0.6 Ohms
After 20 dry flicks (10 closures per contact)
Contact 1: 0.2 Ohms
Contact 2: 0.7 Ohms
Contact 3: 0.5 Ohms
Contact 4: 0.2 Ohms
After 5 closures 2A/10V
Contact 1: 0.1 Ohms
Contact 2: 0.1 Ohms
Contact 3: 0.1 Ohms
Contact 4: 0.1 Ohms
My resistance measurements have 0.1 Ohm resolution, with ~0.2 Ohm lead resistance subtracted. So the final 0.1 Ohm could be 0 or 0.2. No way of knowing without proper 4-wire measurement (which I can't be bothered doing, I'm spending the time writing this instead)
You can see that just the mechanical exercise went a long way towards clearing the gunk. The power switching really sealed the deal (a pun some teleco guys might get!)