variable colors / shades from LEDs?

[Charles Ess]

Hi all,

I'm happily soldering away on my second PiDP-8 kit - thanks again, Oscar!

The first stages have gone well - but I noticed after getting everything up to the "3. Quick check"  stage that while everything checked out perfectly (whew!), there is considerable variation in the LEDs that weren't present in my first build. (Just checked again: they're perfectly uniform in terms of brightness and color.)
The photo doesn't fully capture it - but while a couple in this build are bright yellow, several are a darker, heading into almost orange shade, and three have a distinct greenish tinge. 

Obviously, this is purely an aesthetic issue - everything is functioning just fine. Nor do I have the slightest interest in desoldering them all and replacing with a batch of more consistent ones. So more as curiosity than complaint: was I just lucky the first time around, and this sort of variability the second time around is more or less common and/or the luck of the draw?
And/or: is this something of an issue / problem in fact that should be corrected in subsequent kits?

Again, couldn't be more pleased with having a second one to build and tinker with, but thought this worth calling at least passing attention to.

Many thanks - 

- charles

[Obsolescence]

Charles,

Yes, I had the same when I built an extra kit for myself a month or two ago. If it is bothering you, replace the wrong-tinted LEDs with one of the spares. There does not seem to be a reliabiity issue with the LEDs, but there's been some variation in brightness.

On mine, I did not think much about it other than, 'just like the incandascent lamps, they're all a bit different too'. But that is a bit of a lame story.

Anyway - it does not seem to impact reliability. And as the LEDs flicker in real use rather than are permanently as in the test program, you don't notice it in normal use. 

Kind regards,

Oscar.

[Charles Ess]

Hi Oscar,

and thanks for getting back to me on this (and apologies for my first response getting away from me after just the first sentence).

Right: it had occurred to me to replace at least the three that are more greenish - but, given my soldering skills at least, this is a procedure I'd rather avoid if possible. 

While not optimal in some ways - and not what I was expecting after the first build's LEDs were very uniform in brightness and color - yes, as you suggest, it's not as if the original lamps were perfectly uniform in the first place and I can certainly live with the minor color differences.

I know I sound sound as if I'm complaining: that's not my intention - just reporting in hopes that some of the information and feedback may be helpful in some way or another.

In terms of reliability, however: in fact, one of the LEDs that passed the first tests began to flicker and then died a day later, so I had to replace it in any case. 

(For others like me who are still working up their chops in these directions: I've learned to gradually "walk" the two leads of the LED out of their holes by softening up one (clipped) lead and then the other with the soldering pencil-tip I use. After it's out, removing the remaining solder can be done first with a vacuum pump and then desoldering braid. But these holes are so small that I have to dig out the last bits with a very sharp point. If anyone has any other tips or tricks to suggest for future reference, please do!).

And then a second one has developed an odd crack in its outer case. I first thought there must have been a loose bit of dirt on the cover, as it shows up as a black spot - the one just over the Q in "Quotient":

Again, not really anything to complain about aesthetically - but I've never seen an LED develop a crack like this before, so it makes me wonder a bit about reliability in the long run.

Again, just reporting in case it might be useful in some way. Perhaps these are just the casualties of the first few days of operation and the rest will be fine. Hope so. In the meantime, having a ball with delving further into the computer and its capabilities - and if I can ever find some way of justifying it, will be delighted to build another one or two down the line.

Again, many thanks and all best, 

- c.

[Obsolescence]

Charles,

On Friday, May 3, 2024 at 8:15:22 AM UTC+2 charl...@gmail.com wrote:

I know I sound sound as if I'm complaining: that's not my intention - just reporting in hopes that some of the information and feedback may be helpful in some way or another.

It is. QC on parts is quite a thing actually. The problems I've had over the years, almost 50% of the various components have had a bad batch at some point. So we do check.

In terms of reliability, however: in fact, one of the LEDs that passed the first tests began to flicker and then died a day later, so I had to replace it in any case. 

Actually, that is almost always not the LED but a bad solder point! From recent experience: this week I had to resolder 2 of the 126 LEDS on my PiDP-10. They had worked fine for months. Reflowing brought them back.

I don't know if you have the time for it, but testing the desoldered dead LED would be interesting to me. Is it really dead?

The test rig is just a 5V-12V voltage, a resistor of 200-1000 ohms between it and the LED.

 

(For others like me who are still working up their chops in these directions: I've learned to gradually "walk" the two leads of the LED out of their holes by softening up one (clipped) lead and then the other with the soldering pencil-tip I use.

Yes! On the replacement LED, clip the legs shorter to reduce the walk.

Solder wick or a solder sucker are very handy if you have them, they're only a few dollars. But without them, walking the LED is the trick.

And then a second one has developed an odd crack in its outer case. I first thought there must have been a loose bit of dirt on the cover, as it shows up as a black spot - the one just over the Q in "Quotient":

You're sure that's not a speck of dirt on the acrylic panel?

Please let me know. We test 10 LEDs out of every bag of 1,000 LEDs but if there is an issue, we can investigate.

Kind regards,

Oscar.

[Charles Ess]

Hi again, Oscar - 

Again, many thanks. I hear you on the occasional necessity to reflow the solder - I in fact tried that with the LED that I'm pretty sure ended up burning out. 

A bit of background: call me neurotic, hyper-perfectionist, whatever (I've been called worse ...) but I just decided with my first build (PiDP11) that it was better to test the LEDs before inserting them into the spacers and then onto the PCB for soldering. (And thanks for the picture of your test rig. I have a breadboard set up as a small experimentation / test board in connection with my Ben Eater 8-bit that I also use for this purpose, FWIW.)

So when one went out, I was pretty sure it wasn't the LED but a problem with soldering (or worse). So I first tested it while it was in the board: yes, lit up. Reflowed. Yes, works in the case. But only for a day or so. Checked it again while in the PCB - nope, dead. 

Managed to find it in the trash this morning. When current is applied, it lights up briefly - but in more of a dark orange that then fades away entirely very quickly, starting from the center to the edges. And it gets quite warm in the process. 

Perhaps the reflowing stressed it? Dunno. But hope this is indeed interesting and helpful in some small way.

And thanks for the reminder re. soldering braid and soldering sucker - what I called a vacuum pump - and yep, got two rolls of soldering braid when taking on desoldering the 40-pin header that I placed on the wrong side of my PiDP11. (I think you'll remember this - thanks again, both to and several others on the PiDP11 list for your help on that one as well). As I said, in this case,  the pump and braid were sufficient to get most of the solder out/off. I have a bottle of flux as well, but flux didn't help with getting the last bits out, FWIW.

Yes, that was exactly my first thought when I noticed it, i.e., dirt on the acrylic panel - and now (thanks for the nudge to pursue further) I can see why the crack (or shatter?) shows up as a dark spot.

(And sorry for not sending these along earlier and thereby avoiding some confusion.) 

Here, you can see the crack / shatter in the middle LED from just below its center and then downwards, slightly on the right.

Here's what it looks like lit and with the brace panel in place (right-most, lower row):

(without the brace panel, the effect of a dark spot doesn't show up - some sort of refraction going on here, I guess)

So there you have it.  It's probably not surprising to get a few duds / issues like these in a lot of a thousand, even if you've tested a random 10. 

Thanks again on all fronts - and again, hope some of this is useful. Certainly was for me!

best,

- c.

[Obsolescence]

Charles,

That picture is very useful, thank you. So if you or anyone else needs replacement LEDs, now or later, we can always mail them.

Hopefully though, it's an isolated incident. But if not, let me know.

Kind regards,

Oscar.

[Charles Ess]

Hi Oscar,

thanks in turn - glad to know the picture was useful. 

And yes, in my experience (and that of everyone else I've been in conversation with in these groups), you're a gem about helping out in these ways. And so will holler if need be - but in the meantime, maybe I can say I'm in possession of a pretty unique LED?

Again, many thanks and all best,

- c.