1
\$\begingroup\$

I'm aware that breadboards have a limit of 1-2MHz. I think this is because of capacitance.

Are there similar limits for a stack of stripboard connected by through-board sockets/pins? Would breaking all the tracks, to make them as short as possible, help?

\$\endgroup\$
  • 1
    \$\begingroup\$ I think the inductance is affecting more than the capacitance on breadboards. Personally I've made a 220 MHz oscillator on a breadboard. \$\endgroup\$ – Harry Svensson Sep 9 '18 at 8:46
3
\$\begingroup\$

A 200 MHz 10:1 probe is limited to 20MHz with a long ground clip due to excessive ringing. But with a short clip its ok or better with no clip and tip/ring.

I have run analogue logic with harmonics into the 200MHz region too but localized current loops must be kept small, so decoupling is needed and attention to details on layout since there is usually no ground plane.

For high-speed signals, I preferred twisted pair 24AWG wirewrap wire. With 8~16 turn per inch gives a low impedance for reducing crosstalk and ringing that can work well on protoboards from 150 to 75 Ohms. But for high impedance logic like CD4xxx that also adds capacitance and can reduce rise time and add prop delay and reduce speed. But it can look neater for single jumpers with precision tight corners with flush jumpers. I would have a jar of them to choose from.

But if you look at modern DVD players, some use single sided boards.

But above being said does not mean you will have success above 1 MHz on a breadboard. Even a professional design on a 6 layer PCB can fail for crosstalk and ringing long tracks and split grounds.

Final comment

So ultimately it is not the board but rather the user skill at understanding physical wire inductance and EMI , crosstalk, E-fields and H-fields,that limtis performance.

Rule of thumb is to keep jumpers < < 2cm, if speed matters and use twisted AWG24 pair jumpers using thin insulation wirewrap wire.

other info

If you want to run 200 MHz it would have to be in a small zone with good RF decoupling and it will have some unintentional radiation, but OK for proto's.

From my experience, professional designers will design, build and test on an in-house etched board same day rather than use a protoboard and get it working usually 1st time with minor value tweaks. This would be to evaluate a new chip. Then a complete board design turned around in 2 to 5 days with Getek, FR4 or 2 weeks for ceramic hybrids. Either using iron-on boards for Op Amps or lithographic film prints same day for RF, etch and make it work the same day.

\$\endgroup\$
  • \$\begingroup\$ Remember if your logic has a 1ns rise time but you operate < 1MHz , don't think you are ok. You are dealing with 300MHz of bandwidth. \$\endgroup\$ – Sunnyskyguy EE75 Sep 9 '18 at 11:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.