Solderless protoboard electrical limits

For analog noise I don't know. You could put a ground plane under the whole thing, but it would be quite a bit further away than the other side of a PCB. Probably not good in the presence of strong RF from nearby sources either - you could perhaps get rectification in oxidized spring-clip to lead connections, and can probably couple in more noise in the interconnects than you would get with point to point coax or PCB stripline.

Traditional prototype construction for small signal RF circuits is "dead bug" - put ICs, transistors, etc upside down over unetched copper, wire most connections in the air, with anything that goes to ground soldered to the board. You can also add extra high value resistors as standoffs. Not so much fun today with TSSOP IC's, but you can build little stonehenge replicas with chip resistors and caps...

Quite a bit closer, actually -- the bottoms of those spring clips are just an adhesive label away from the bottom of the board. I have one that I use for prototyping oscillators up to 10MHz or so that's on standoffs. Note that at 10MHz the only piece of information that I consider valid is that it'll actually oscillate once I get it onto a dead-bug style board.

When I use those I am careful to keep the wires tight against the board surface -- no big loop! It makes a significant difference.

Yup. I just prototype with through-hole components. If I got serious with RF stuff I'd prototype with quick-turn boards. Usually even a bad guess at a circuit is a pretty good starting point for a cuts & jumps fix.

With careful layout you can do narrowband (i.e. radio) stuff up to

I think that you could push these up pretty high with careful layout, and consideration for what you want to learn. If, for instance, you don't know if a circuit is going to work at all but you're confident that you can deal with EMI considerations if it does you can breadboard it, test it for the stuff that a breadboard won't screw up, then either scrap it or put it on a PC.

Well, fundamentally the limited frequency response is a property of the capacitance between protoboard conductors. 10-20-30pF is the ballpark depending on geometry.

And "noise", that is not a product that they put into the breadboard. Any noise you see on a protoboard can be attributed to lack of a ground plane, long Vcc and ground lines, big loops in the wiring that pick up external noise, etc. Those are all properties of the construction details (although of course you've chosen to do it on a protoboard so you can't get good short lines.)

Long wires etc. also limit frequency response by introducing inductance.

"Dead bug" construction is surprisingly quick and effective.

Tim.

In the digital domain protoboards are useful up to the 10-20MHz region without a lot of effort. Although it takes some thought. At the same time it's trivial to come up with examples that will not work at 1MHz if you put awful long wires on grounds and Vcc's etc.

If you scope out the ground pin on a TTL IC at just 20MHz on a protoboard you can see horrendous bounce! CMOS or HC isn't so bad (but it's still plainly visible on a scope). "Flyover" jumpers turn a nice square 8MHz clock into a ugly wiggly blob after just a few inches, and attempting to clock more than a few chips at once will probably be a disaster. But overall you'd have similar problems on a similarly laid-out PC board (strangely enough you will have the same problem on some of those pre-etched protoboards!)

Tim.

Layout helps a lot, though. First time I assembled the circuit, the driver transistor(s) oscillated around 100MHz. You can't even see that on a slow scope!

Tim

-- Deep Fryer: a very philosophical monk. Website:

The contact in the breadboard, yes. But the components would be a lot further away. I guess maybe that's more of a comparison to a ground (+power) plane as a low inductance supply issue than a ground plane as an RF shield.

Personally i have found that digital gets flaky around 10MHz and analog can get flaky at 200kHz and is always flaky by 2MHz.