Ringing on waveform, frequency multiplier / squarer

[Chris Wilson]

I have built a circuit from a schematic to double a 10dBm 136kHz sine wave to 272kHz square wave. It works but I am unsure of why the input sine wave shows ringing when the circuit is in operation, and what its implications are. I have taken photos of the input signal with the circuit powered off, and with it powered on, as well as the generator settings, the counter display and the waveforms with the circuit powered on.

Powered on there are two scope photos, one with the output unterminated to the scope, probes set X1, and one with the scope probe terminated at 50 Ohms. Is this a probe issue,or what please?
The output will be driving a 4013 IC.

Photos are at :


http://www.gatesgarth.com/scope2/unpowered_input.jpg

http://www.gatesgarth.com/scope2/ringing.jpg

http://www.gatesgarth.com/scope2/unterminated_output.jpg

http://www.gatesgarth.com/scope2/50_ohm_terminated_output.jpg

http://www.gatesgarth.com/scope2/counter.jpg

http://www.gatesgarth.com/scope2/generator.jpg

http://www.gatesgarth.com/scope2/dBm.jpg

http://www.gatesgarth.com/scope2/circuit_board.jpg

Note some shots have different volts / div settings.

Thanks

Schematic should be attached.

[jlmoon]

how about return path ground currents from that Class D output stage going into cutoff or saturation creating some really nasty harmonics? 
where is the ground attached on your signal generator, what point in relation to the ground of your 3904/3906 driver pair, power supply?
more decoupling on driver pair supply / grounds?
maybe a filter network on that input (T1 primary)  to snub the back emf created by inducing a changing voltage on the input of T1...

so many possibilities

 

[T3sl4co1l]

No ground plane on that poor perf board is a likely candidate...

How about... ground currents.  What's it do when you put some big ol fatty ferrite beads on any of the cables (including probe and power supply leads)?  Don't forget to try all combinations.

Tim

[Chris Wilson]

My probes are ebay el cheapos, to be honest, and I also modded one of them with a 12 inch ground lead when working on a TIG welder. I found some better probes with normal length ground leads, and ran them X2 and got these traces below. If they are too big they are also at:

http://www.gatesgarth.com/scope2/X2.jpg

I am sure with some decent probes (Santa let me down...), things would be better still. I may recreate the circuit on a ground plane, I believe the creator built it like this, the quality of his point to point can be seen here:

http://www.gatesgarth.com/Phasingexciter2.jpg

I was hoping at a max of 272kHz I could get away with perforated "Veroboard" type material. I have never tried point to point on a ground plane, perhaps I should. Learning a lot, that's 60% of the fun for me. Thanks for the reply!

[PA0PBZ]

+1 for the ground. Hard to create a ground plane but try a star ground connection and/or show us the reverse side of the board.

[Chris Wilson]

+1 for the ground. Hard to create a ground plane but try a star ground connection and/or show us the reverse side of the board.

Hmm, no sniggering at the back please!

[PA0PBZ]

Hmm, no sniggering at the back please!

I looked with my eyes closed, so no worries. But yes, that layout would create the problem you spotted. Get some wires and bring all your ground connections to one single point like your negative power connection.

[w2aew]

My probes are ebay el cheapos, to be honest, and I also modded one of them with a 12 inch ground lead when working on a TIG welder. I found some better probes with normal length ground leads, and ran them X2 and got these traces below. If they are too big they are also at:

You should check out my video on probe ground lead length, and it's affect on signal integrity...

I was hoping at a max of 272kHz I could get away with perforated "Veroboard" type material. I have never tried point to point on a ground plane, perhaps I should. Learning a lot, that's 60% of the fun for me. Thanks for the reply!

Your fundamental might be at 272kHz, but with a square wave like you have, there is frequency content at odd harmonics that go MUCH higher than that!  The faster the edges, the higher the frequency content.

[Chris Wilson]

I have rebuilt this X2 multiplier using ground plane methods and things look a lot better. I am not using probes. I am taking the output as a sine sine wave at 10dBm from my sig gen direct to the multiplier using BNC terminated 50 Ohm coax, and the multiplier now has proper BNC sockets for input and output. Similarly I am taking the X2 square wave output from the X2 multiplier direct to my old Tektronix 7854 scope with BNC terminated coax. I am using a T piece at the scope end with a 50 Ohm termination BNC resistor. Whilst I am told this will now be fine to run the rest of my project I am curious as to why part of the square wave is rounded off.

I also took the 136kHz output from the sig gen direct to the scope, again terminated in 50 Ohm, but as a square wave. I see rounding on two sides of the square wave like that. Temporarily I have to have the scope and sig gen, which is rack mounted, some distance apart so am using maybe 12 feet of coax in total, in and out of the multiplier. I think this may be the cause. Comments welcome, thanks.


Alan, your video was very useful! Thank you for such helpful YouTube posting.