But I'm getting good at it

​

Sad thing is, I have to patch a few dozen of those boards .

 

But I'm getting good at it

View attachment 332038es​

Sad thing is, I have to patch a few dozen of those boards .

1. Get yourself some insulated wire-wrap wire.

2. Don't leave those bare leads touching the board over copper (regardless of soldermask).

 

1. Get yourself some insulated wire-wrap wire.

2. Don't leave those bare leads touching the board over copper (regardless of soldermask).

Thanks for the advice. Wire-wrap wire is a bit too thick for my taste ... maybe I should've used magnet wire instead?

 

Wire-wrap wire is a bit too thick for my taste

Solid 30 AWG is too fat?

Compared to what? The work in the photo?

 

For the work in the photo, I used 28 AWG split in half. So yes, insulated 30 AWG would be fat in comparison. But you have a point in saying (implying) that it's perfectly workable.

Gonna get myself some 30, the thinnest I have with me right now is 28.

 

For the work in the photo, I used 28 AWG split in half. So yes, insulated 30 AWG would be fat in comparison. But you have a point in saying (implying) that it's perfectly workable.

Gonna get myself some 30, the thinnest I have with me right now is 28.

Regardless, my most serious concern is the uninsulated wire lying on the traces. Soldermask is not intended as an insulator.

At a minimum, apply scotch tape to the board under the wires to give some protection. Kapton is even better.

 

I usually use Kapton to insulate the pads.

 

This patch isn't mine, it's from someone else's website. But I found it hilarious.

​

 

This patch isn't mine, it's from someone else's website. But I found it hilarious.

View attachment 332945​

I can tell you what probably happened: the design engineer chose to use the MCU's internal oscillator without reviewing the accuracy specification.

 

I can tell you what probably happened: the design engineer chose to use the MCU's internal oscillator without reviewing the accuracy specification.

That's a possibility ... I wonder what the LED is for, anyway

 

That's a possibility ... I wonder what the LED is for, anyway

Is that an LED? If so, it's not attached to anything.

 

Is that an LED? If so, it's not attached to anything.

If it is a connected LED, it is probably serving as a clamp diode. Sometimes those are required.

 

If it is a connected LED, it is probably serving as a clamp diode. Sometimes those are required.

For an MCU crystal oscillator?

Never heard of such a thing (over a hundred years of doing this). Enlighten me.

 

After taking a closer look, I don't think that's a LED. It's way too small. Most likely, it's a reflection of some sort.

 

This patch isn't mine, it's from someone else's website. But I found it hilarious.

View attachment 332945​

It would be interesting to know if that actually worked. Depending on the frequency of the crystal, those are some pretty sweet antennas, and a relatively very long way from the pin on the chip.

 

It would be interesting to know if that actually worked. Depending on the frequency of the crystal, those are some pretty sweet antennas, and a relatively very long way from the pin on the chip.

Yes, it'd work. Probably a 32.768 kHz crystal for time keeping. You can't use the internal oscillators for that.

 

1. Get yourself some insulated wire-wrap wire.

2. Don't leave those bare leads touching the board over copper (regardless of soldermask).

Magnet wire also works well, 30awg or smaller. It's covered in lacquer for super thin insulation. Dip the ends in molten solder to clean the lacquer right off, sometimes you can get away without cleaning the ends first.

 

Magnet wire also works well, 30awg or smaller. It's covered in lacquer for super thin insulation. Dip the ends in molten solder to clean the lacquer right off, sometimes you can get away without cleaning the ends first.

I do that all the time with magnet wire. I super heat my soldering iron to 400°C (750°F) and then tin the magnet wire's tips with it. The high temp chars the insulation away. After that, I lower the temp and proceed to normal soldering.

 

Magnet wire also works well...

For prototypes, maybe. But I wouldn't trust it for production.

 

For an MCU crystal oscillator?

Never heard of such a thing (over a hundred years of doing this). Enlighten me.

I did not mean that a clamp diode would be reasonable. And it might also be a QC inspector mark. If somebody is able to magnify the image a whole lot they might see something better.