My guess is the LED's suffer reverse bias thermal runaway when they're hot from being in a steamy enclosure and then they get a reverse 5v across them and any leakage current turns into heat accelerating the process.
This is literally evidence of stuff being designed to fail. An extra diode costs less than a cent at production scale. This was a manufacturing choice, not an error.
nah, this is just not something designer would expect to fail like that. The LED has datasheet, the datasheet have leakage current, it has no data on increased leakage over years, you plan for what you have.
What would help is not randomly planning for some of the segments to fail (they are multiplexed with other things, you'd have to put more diodes), but to just get slightly better/less cheap LED display
Only "choice" made here was sorting by price when buying components for the cheap device.
LEDs are diodes (Light emitting diode). Certainly this was a cost saving measure, but it's not a bad assumption that the LED wouldn't allow reverse current flow.
Capitalist profit motive strikes again. The invisible hand expands tech and the visible hand keeps making tech worse.
People usually respond to this by saying that it would be absurd to suggest the company did this for its own benefit, when anyone who engineers knows these are often caused by revising design to minimize costs... and increase profits.
In almost every system with failsafes there will be conditions that can bypass them. The goal is not to make it impossible for the unsafe condition to happen, but to make it so that in the expected uses the failure will not happen.
In this case it's a domestic microwave and the mainboard is housed inside the electronics enclosure, so covering the whole mainboard in salt water is not an expected occurrence in a domestic kitchen.
But there are ~1 billion microwaves in the world... I'm sure it has happened somewhere. As a designer of a billion-sold device, your job is to make sure that the expected number of people harmed by your device is substantially less than one, which gets really hard when all the risks are multiplied by 1e9.
Your job is to make sure the number of people harmed _while using the device as intended in a reasonable situation_ is as close to 0 as possible.
A domestic microwave is for use only on land, indoors, in a domestic kitchen, and in an unmodified form. In these conditions there is no conceivable way that salt water could saturate the main board, or bypass all the interlocks in another way.
Yes there are ways that all the safety systems can be bypassed, but not while a reasonable person is using the device as intended.
I noticed when tearing down an old microwave for salvage that the light bulb was part of the power circuit. If the bulb burned out, so did the microwave.
> Even with triple redundant relays, how do you know the salt water didn't just wet them all?
The design typically includes a mix of normally open and normally closed switches. If everything failed in the same direction (closed) it wouldn't satisfy the failsafe.
If you're spilling conductive liquid on the board, it's going to blow fuses anyway. It's more likely to short to ground than to short only to the precise path needed to activate.
My guess is the LED's suffer reverse bias thermal runaway when they're hot from being in a steamy enclosure and then they get a reverse 5v across them and any leakage current turns into heat accelerating the process.
Wouldn't it be more likely to be reverse-bias degradation of the LED junction causing permanently increased leakage current?
All LEDs are photodiodes too, certain degredations of parts or poor circuit design could lead to the display turning into a switch.
Prev discussion:
https://news.ycombinator.com/item?id=41480038
Articles like these are great to argue nonconformity which can get you your money back in EU. Even past the warranty period.
168 points and 116 comments at the time: https://news.ycombinator.com/item?id=41480038
This is literally evidence of stuff being designed to fail. An extra diode costs less than a cent at production scale. This was a manufacturing choice, not an error.
nah, this is just not something designer would expect to fail like that. The LED has datasheet, the datasheet have leakage current, it has no data on increased leakage over years, you plan for what you have.
What would help is not randomly planning for some of the segments to fail (they are multiplexed with other things, you'd have to put more diodes), but to just get slightly better/less cheap LED display
Only "choice" made here was sorting by price when buying components for the cheap device.
Sounds like this is far more common a problem with blue LEDs than others, and that was certainly a choice.
As if I needed another reason to detest the eye-searing blue LEDs that have infested every device.
Don't underestimate the appeal of saving one cent per unit. So long as the costs are externalised, anyway...
Eh, I don't agree.
LEDs are diodes (Light emitting diode). Certainly this was a cost saving measure, but it's not a bad assumption that the LED wouldn't allow reverse current flow.
It’s not exactly designed to fail, they just don’t care. If they could add a one-cent part that made it fail sooner, they wouldn’t do that either.
what happens when that diode fails?
Capitalist profit motive strikes again. The invisible hand expands tech and the visible hand keeps making tech worse.
People usually respond to this by saying that it would be absurd to suggest the company did this for its own benefit, when anyone who engineers knows these are often caused by revising design to minimize costs... and increase profits.
Very impressive engineering on the door switches. On the display, not so much.
sort by price strikes again
More proof blue LEDs are the devil and should have never been put into all of our electronics to be the shining beacon of "OW MY EYES" at 2 AM.
You can do an awful lot to make a device like a microwave safe with loads of failsafes...
But rarely do those failsafes protect reliably against 'the mainboard was splashed with salt water'.
Even with triple redundant relays, how do you know the salt water didn't just wet them all?
In almost every system with failsafes there will be conditions that can bypass them. The goal is not to make it impossible for the unsafe condition to happen, but to make it so that in the expected uses the failure will not happen.
In this case it's a domestic microwave and the mainboard is housed inside the electronics enclosure, so covering the whole mainboard in salt water is not an expected occurrence in a domestic kitchen.
But there are ~1 billion microwaves in the world... I'm sure it has happened somewhere. As a designer of a billion-sold device, your job is to make sure that the expected number of people harmed by your device is substantially less than one, which gets really hard when all the risks are multiplied by 1e9.
Your job is to make sure the number of people harmed _while using the device as intended in a reasonable situation_ is as close to 0 as possible.
A domestic microwave is for use only on land, indoors, in a domestic kitchen, and in an unmodified form. In these conditions there is no conceivable way that salt water could saturate the main board, or bypass all the interlocks in another way.
Yes there are ways that all the safety systems can be bypassed, but not while a reasonable person is using the device as intended.
I noticed when tearing down an old microwave for salvage that the light bulb was part of the power circuit. If the bulb burned out, so did the microwave.
Reminds me of a comment from a previous time this story was posted here:
https://news.ycombinator.com/item?id=41509748
In that situation one of the switches should short the mains voltage and blow the fuse when the door is opened.
> Even with triple redundant relays, how do you know the salt water didn't just wet them all?
The design typically includes a mix of normally open and normally closed switches. If everything failed in the same direction (closed) it wouldn't satisfy the failsafe.
If you're spilling conductive liquid on the board, it's going to blow fuses anyway. It's more likely to short to ground than to short only to the precise path needed to activate.