#95 is up and running, had some led issues

[Bill E]

3 leds were bad. I tested them after I pulled them to confirm.

Seems the vendor doesn't pay much attention to yield, clearly doesn't test. Annoying, but enough spares to solve the problem.

I'm also a bit dubious about the spring clips in the rack version. I might 3d print some solid locking blocks.

Bill

[Oscar Vermeulen]

Bill,

On Thursday, October 2, 2025 at 6:25:44 PM UTC+2 wjegr...@gmail.com wrote:

3 leds were bad. I tested them after I pulled them to confirm.

That drives me nuts. We add spares to deal with that eventuality, but I really had hoped never to see bad LEDs again. I did not, for 8 years, and then we got a bad batch just when the PiDP-10 came out. That caused a lot of hair-pulling.

Are you sure (I know you are, I'm asking because of neuroticism) that the LEDs were bad and it was not a soldering problem/cold-solder joint?

We're testing all LEDs for all kits for a couple of days (which is no fun at all!) to see if we see this occur.

Seems the vendor doesn't pay much attention to yield, clearly doesn't test. Annoying, but enough spares to solve the problem.

Glad the spares were there! There's only one vendor for these 'diffused flat top LEDs' that we could find so far. Jose is trying to find the actual manufacturer. With 124 LEDs in each kit, we need loads of them and I'd rather we buy from the manufacturer directly. 

Logistics -- another example: JLCPCB has had QC problems for the last three months. That never happened before, ever, I always called them the Perfect Manufacturer. But we rejected 300 Enigma Touches, and part of the second PCB order for the PiDP-1. And the order time should be one week, we've now been waiting 6 weeks. 

I really get the feeling the world of manufacturing and shipping is degenerating, not improving as it did for XX decades in a row.  Earlier this year, we sent back 60,000 switches. 

Testing parts takes so much time these days... There's always something. Whilst 5 years ago - almost never issues.

Kind regards,

Oscar.

[Oscar Vermeulen]

Bill,

I'm also a bit dubious about the spring clips in the rack version. I might 3d print some solid locking blocks.

Don't worry! These springs turn out to be perfect. I was doubtful too at first, even went to a specialist company nearby before ordering the first prototypes. Who said they were good and kind of the industry standard for this type of job. 

They do the job perfectly - you'll see :-)

[Bill E]

Yep, solder joints checked first thing. Then some hardware debugging to verify the leds were actually connected to the select lines they were supposed to be connected to, inserted in the correct orientation, etc.

Finally pulled them, tested each one. Dead. Tested the replacements before I put them in. :)

Bill

[Oscar Vermeulen]

Harumph. We've started to test all leds for all kits now. No fun but apparently necessary...

Thank you for the feedback, much appreciated. It is the only way to ensure things are as they should be!

[Whit Turner]

Quick testing hack I used on my PiDP-10 LEDs. Put them in series with a 1K current limiting resistor and a bench supply (mine only gets to 30 Volts, so I could only do 16 at a time).

Whit

[Phillip Porch, MD]

I tested all the diodes and leds and none were bad. I did have some bad ones when I build the Pidp-10 but had more than enough spares to complete the kit without problems.

[Oscar Vermeulen]

This is precisely the setup that currently use for all LEDs in a PiDP-1 before packing up a kit. Which is no fun but apparently is necessary. QC is such a pain.