[scrps]

Not to hijack your post with a PSA but I think you'll endorse it...

If you are building a product and it has any indicator lights please dim/diffuse/lightpipe them.

It seems to be a trend these days of ultra bright LEDs for indicators, I have so many devices I've either disconnected, dimmed or taped over the LEDs because it is so bloody bright.

[godelski]

Extra helpful if you can add a photodiode to the system that can adjust brightness accordingly. It costs effectively nothing in parts and should take any competent engineer less than 5 minutes to include it. Better, use multiple because redundancy is better! I wish my car had redundancy so it's entertainment panel could go back to adjusting from being visible during the day to not being blinding at night (it has brightness adjustments but that's insufficient for a car and living anywhere outside a major city)

For people doing software, press for the love of god just make that shit adjustable. Only fucking noobs hard code variables. Practicing good habits will help everyone, including you. Unless you got a serious reason not to, expose that to the users. Even if you don't think anyone will want it, I promise you, someone does. There's a lot of people and everyone thinks differently. So only lock down what needs to be locked down.

Unless you're trying to create e-waste or piss people off. Which in that case I only have two words for you and they aren't nice

[paradox460]

You don't even need to do that. You can measure ambient light with the LED, using it as a photodiode

[godelski]

The sane one that's on? Is that light constant or blinking?

[paradox460]

You'd measure when the LED is off, but you could do this many times a second

[godelski]

Well you can't measure they brightness of the LED that is emitting. You can ambient light but not sure that's what matters in this application.

Also don't forget about response time. You may not be able to do it that fast. Depends on the led

[fecal_henge]

To reply to this and the above comment:

Cheapest light pipe on digikey: 16 cents Cheapest photodiode on digikey: 11 cents Cheapest LED (obviously that annoying blue): 625 milli cents!

Part costs matter! Its not just the BOM, its the NRE from the increased complexity. Im not saying saying its OK, just that its inevitable considering the economic conditions.

I do board level designs and drop down LEDs. If you are not specialising in indicators, its hard to visualise how bright 10mA through your diode is going to be. Add to this that sometimes you never even see the thing you designed!

[godelski]

  > 16 cents Cheapest photodiode on digikey

On DigiKey the cheapest is $0.125[0]. On Alibaba I can find much cheaper, including less than penny [1,2]. Now pretend with me that you're building in lots. You're going to get cheaper than Alibaba, though maybe not by much.

  > I do board level designs and drop down LEDs. If you are not specialising in indicators, its hard to visualise how bright 10mA through your diode is going to be.

I think if it is annoying enough that at presumably >10m away it can keep the gp awake you don't really need to be that precise in your measurement. It's basically way too bright.

But if you want to get creative you could use the diodes in the circuit or even your led if you'll blink it. But that last one doesn't tell you the led brightness, only ambient.

[0] https://www.digikey.com/en/products/detail/harvatek-corporat...

[1] https://www.alibaba.com/product-detail/1-4V-20-30-45-60_6253...

[2] https://www.alibaba.com/product-detail/JST3-3MM-receiver-pho...

[fecal_henge]

The digikey costs are more (I quoted qty=1 prices though), its the ratio between the component types which is more interesting though.

I think annoying at 10m is sufficient power to require a 'power' LED, so that shouldnt be hard to avoid in a design I would agree. Annoying in your kitchen however is much harder to eyeball a value for (excuse the pun).

[godelski]

  > I quoted qty=1 prices though

Yet on the unit I linked it is $0.14 per. 2c difference, but hey, that's a full photodiode on Alibaba.

A big part of the assumption that I was making in my original comment is that you're mass producing devices, not treating it as a hobby project. I agree, the constraints are very different in these two settings. A major difference is the cost/ratio of parts to other costs/time. In mass production the price per part is going to approach <$0.01 for standard COTS parts. Same reason Alibaba has them for so cheap (which doesn't include shipping)

  > Annoying in your kitchen however is much harder to eyeball a value for (excuse the pun).

Yes, I agree with this. Though In that setting I think you may be able to dual use the LED. Periodically blink it to detect ambient light. You can be pretty rough with this frequency and your accuracy level. But the best solution is always adding some little POT so it can be adjusted. It adds a little cost and design time, but if you're actually selling products that's very likely going to be a inconsequential amount of time and inconsequential cost.

*BUT* it is something the bureaucrats will be able to look at and say that they need to remove it because at scale it will save "a lot of money." But does it? I'm unconvinced. To me it seems to just make the consumers more upset and annoyed. Sure, you might have saved the company $100k because you are producing 10M units, but I'm pretty confident most consumers would rather pay $0.10 to have that part. The bigger problem might be a Lemon Market type of problem where consumers don't actually know/realize they want such a feature at time of purchase but randomly and intermittently realize it over the next week or couple years. It never gets priority thought because there's usually other things to complain about and well then we need to have a larger conversation about if we're doing our jobs as engineers. Are we actually providing value to people with the products we make? Can we take pride in our work? You may be making something dumb, but can you honestly tell yourself "I made the best thing I could"?

The business people need pushback because they don't care about the product, they only care about the profits. That's part of your job as an engineer. Because your job as an engineer isn't to be concerned with the profits, it is to be concerned with the product. Your engineering managers are the bridge, not you the engineer. And a good engineering manager doesn't fully align with the business people either. Because when push comes to shove you have to decide between either: 1) sacrificing the product in favor or profit or 2) sacrificing profits in favor of the product. As an engineer, your job is to have the latter.

[Razengan]

This is actually one of the best things that Apple did with their devices (except a few like the MacBook MagSafe) that PC users with their 1000 lights on everything just don't get.

[8n4vidtmkvmk]

Hey now, don't blame us PC users. It's almost impossible to buy decent hardware that isn't RGB`d up the wazoo. Do I want all that crap lighting? No. Do I have a choice? No.

[brikym]

Please forward this to the billboard advertising industry.

[Razengan]

The entire advertising industry is cancer.