Bring back wirewrap!

[maluse]

Modern surface mount chips are all very well, but prototyping with them is a sod. I need to somehow solder 8 bits of wire (two of the contacts are on the back) onto this little thing (a dual MOSFET).

Comments

[aeia]

Eeek. That's tiny! I don't think my hands would be steady enough for that! You probably need a vascular surgeon to do it :)

[maluse]

I managed it! But unlike a vascular surgeon I was able to accidentally kill the first one I tried to work on...

[evil_c]

and yet you have such poor eye sight you can't even tell when I've had my hair coloured... hrm... :)

[hazeii]

Bit of veroboard, dash of solder (Pb included) each end and a delicate touch in the middle with some WW wire. OK up to 8 pins or so, but above that SMT->DIL adapters are your friend.

One good thing about modern disposable electronics is there's a lot of broken junk about to practice on!

[maluse]

That's what I ended up doing, I superglued the chip upside down on a bit of veroboard and then looped bits of wire from the vero tracks onto the pins and pads.

I've got a bunch of SMD adaptors but chip manufacturers seems to be in a rush to invent new packaging forms and make them incompatible with the adaptors. I won't even mention ball-grid array packages...

[hazeii]

BGA's a necessary evil in the quest for greater speed (which most go for, 'cos speed is a better marketing message than greater efficiency).

Even clocking slowly, the rise/fall times are now so fast that everything's analogue again; at 5GBps a PCB via is an issue (e.g. when connecting buried layers, the unterminated stubs above and below can ring, and need to be ablated).

Never thought I'd view 3mm of interconnect from Vdd to die as heavy inductance; how things change!

(On the other hand, this 20Mhz DIL PIC I'm currently messing with is working fine, despite being overclocked to 24Mhz).