Breadboard Woes

[Bryan]

Well, I guess you get what you pay for and there is a reason why 3M breadboards cost a arm and a leg. Spent two days trying to debug a project on a far east EBay breadboard. Pulling my hair out trying to figure out why the power supply was being dragged down from 5v to 4.6v, which component could it be??. Turned out it was nothing more than the breadboard, guess the connections are so crappy and either corroded or adding resistance. Geez the project only drawed 200ma!! I suppose part of the issue could also be the breadboard wires. They certainly don't have the quality that they had many years ago. Used to be a time when DuPont breadboard wires meant you were buying wires made by DuPont, not any more.

[jdraughn]

Since I have had similar issues, I always double up all power connections, using second sets of wires AND holes and have never had troubles since. I can pick up and bump and move the breadboard around and the circuit never seems bothered. I have only ever bothered doubling up the power/ground connections, not for data connections. One of the problems I always had was my ICD resetting while debugging my microcontrollers.

[setq]

I spent many years being shafted by this, even on a nice 3M breadboard.

Our forefathers were right: If it's not soldered you can't trust it.

Ergo, I've got a policy of building very small circuit elements on a solderless breadboard and testing them, then transferring them to a deadbug "master prototype". Deadbug being copper clad board with all the parts soldered upside down. When I do breadboard, I use a very small (and therefore relatively cheap), high quality Wisher WB-501 board [1] and hand cut 1/0.6 jump wires to order. Very few problems.

You're pretty much screwed with a solderless board past a few MHz anyway as the strip capacitance, lead inductance and absolutely terrible ground plane gives you trouble. Also if you're putting more than a couple of hundred mA through one they tend to melt. Plus when something goes phut they melt.

Fugly deadbug (simple lm317 voltage regulator). Uses cheap 4mm sockets. Will lose a good 6-7 watts:





[1] http://www.wishmaker.com.tw/cubecat/front/bin/cglist.phtml?Category=8

[KL27x]

I just stick long, 20mm pin headers into the breadboard. Never just a single pin, but at least sections of 3 or more, so they don't wiggle. Then wire wrap between the pins rather than use jumpers.

The flexible reusable jumpers are more trouble than they're worth. It takes longer to wrap the wires, on the surface. But debugging and reworking is much more stress/error free over trying to follow a piece of spaghetti without dislodging something. You don't have to second guess if the other 2 or 3 wires poked into the breadboard all actually being connected. With the wire wrap, they're all wrapped to the same pin, scope probe is attached to that pin, nothing wiggles, and it's money in the bank.