Best way to attach an signal generator to a breadboard circuit

[petert]

Hello,

I was experimenting with filters for lower frequencies (less than 1Mhz) on a breadboard. To test them I used waves generated by a signal generator. And an oscilloscope to view the output. Nothing special so far.

The question is: what is the best way to connect the signal gen to the breadboard circuit? So far I abused oscilloscope probes for that, but I assume it may have some undesired effects.

Besides soldering a circuit in dead bug / manhatten style (and adding coaxial connectors), so that I can use coaxial cables directly, what is a good way to connect a signal generator to a breadboard filter circuit?

Any specific recommendations?

[bob91343]

Are you concerned with the mechanical connection or the electrical interface?  Electrically, a signal generator ought to be terminated in its output impedance, often 50 Ohms.  One must be aware of the dc path as well, as some generators can be damaged by certain kinds of loads.  A coupling capacitor is usually a good idea.

Mechanically, a coaxial connector is good.  Deoending on size and power level, perhaps BNC or SMA would be appropriate.

In short, there is no 'best way' but there are some appropriate ways.

[tautech]

These or ones with croc clips:

[Whales]

The ones in the last photo are cheap enough too that you can cut the ends off and connect them to jumper wires instead.

[CDaniel]

I use very cheap chinese oscilloscope probes ... that are not very good anyway for the original purpose .

[petert]

I was looking mainly for mechanical connections, and wondering if scope probes have unexpected side effects, compared to simple coax cables, since they are meant as inputs, not outputs.
I wonder if cables with crock clips or other connectors cause a (significant) change in the impedance of the cable.

Electrically, a signal generator ought to be terminated in its output impedance, often 50 Ohms.  One must be aware of the dc path as well, as some generators can be damaged by certain kinds of loads.  A coupling capacitor is usually a good idea.

I looked up again how termination is done, to see if I remembered right:
https://www.eetimes.com/author.asp?section_id=36&doc_id=1225569

Where would you place a coupling capacitor (e.g., in the first circuit)? Are there other things to pay attention to?




(Related videos, for future reference:

)

[T3sl4co1l]

I use a BNC connector with pigtail leads on it.

Tim

[vk6zgo]

Input & output impedances of filters will affect the resultant response, especially if they are low frequency LC ones.

Many years ago, I had to build a filter to prevent interference between a local transmitter & a distant  one which was being used to provide backup programme material for the local one.

i happily designed an appropriate filter for this job, for 50 Ohms input & Output impedances, using design tables.
When built & tested, it was a disaster!
What I had neglected to account for was the resistance of the quite high value inductors, which approached the design impedance.
Too late, I found I should have designed it for much higher impedances, with matching sections at Input & Output.

[bob91343]

The coupling capacitor is best at the load, in series with the 'hot' lead, or center conductor.  That way the load can do anything at dc without causing a problem with the termination or the generator.  The capacitor needs to be a low reactance at the frequency of interest, less than perhaps 5 Ohms in a 50 Ohm system, and with low ESR and ESL.  And short leads, in terms of wavelength.  And a voltage rating commensurate with the dc the load might supply.

A friend got lucky when he connected his signal generator to the plate lead of an IF stage.  With perhaps 100 Volts dc there he could have blown the generator output circuit.  He actually might have done so because his testing wasn't conclusive but he said it still worked okay.  He also may have damaged the IF transformer but again, he said it's okay.

[fourfathom]

I use very cheap chinese oscilloscope probes ... that are not very good anyway for the original purpose .

I assume you have 1x probes, but just so nobody gets the wrong idea, the typical scope probe is a 10x probe and the circuit looks more or less like this:


This is *not* appropriate for coupling a signal generator output to a circuit.

I have a box of BNC and SMA connectors and usually solder one (or use a pigtail) to my breadboards.  I also have a big "jewelry box" full of SMA/BNC/N/PL259 adaptors, and terminations and attenuators, which come in very handy for interconnecting my gear.

[CDaniel]

Of course I use 1X , and the schematic is not like that , there is nothing in series like you have "inside probe connector" . The trimmer and/or the resistor in parallel could be removed , for 10$ a pair of ( bad) probes is not a big loss .

[radiolistener]

Of course I use 1X , and the schematic is not like that

if you're using 1x, it still has 331 Ohm resistor connected in series with the probe, also there are some capacitors. So, technically the schematic is a little different, but still has a lot of components 

[CDaniel]

No , this cheap probes at least don't have any more components ... like in David's schematic . Anyway anything is easily removable from the probe connector ( the square box near the BNC ) . Measured from probe to BNC is about 1 ohm , the central cable resistance .

[SilverSolder]

I like the idea of using an el cheapo scope probe for signal injection.  You could even put a coupling capacitor in the square box near the BNC if you are taking out the compensation adjustments anyway, so making the probe safe for the signal generator if it doesn't like external DC on its output.

[macboy]

No , this cheap probes at least don't have any more components ... like in David's schematic . Anyway anything is easily removable from the probe connector ( the square box near the BNC ) . Measured from probe to BNC is about 1 ohm , the central cable resistance .

"Real" probes use lossy cable, with over 100 Ohms (maybe several 100's Ohms) resistance end to end, due to the cable alone, not a discrete component. The conductor in this cable is so fine it is barely visible to the naked eye, much thinner than a hair. The lossy cable reduces reflections from the scope input (remember the signal is not terminated to a matching impedance) and the fine conductor also has much less capacitance to ground (shield) than a thicker one, reducing higher frequency losses.

All that is to say that if you measure only about 1 ohm end to end, it is a garbage probe, so don't feel bad about abusing it.

Back on topic, I made up a couple of BNC jacks with short 24 AWG leads to plug into the breadboard. Even at just 1 MHz, consider terminating at the proper 50 Ohms. If you don't the signal will be 2x the expected amplitude, as most signal generators expect this termination. At these low frequencies, just an ordinary 51 Ohm resistor, or two 100 Ohm resistors in parallel works fine.

[Plasmateur]

I've bought some "BNC to bread board" adapters off ebay in the past for ease of use.

Just googling it now, I've come up with these guys

https://www.whiteboxes.ch/shop/pre-assembled-female-bnc/

https://www.marutsu.com/Other-kits-c-50/BNC-JACK-Board-p-8609.html

https://elabbay.myshopify.com/products/elabguy-bnc-f-bo-v1a-bnc-female-connector-breakout-board

https://www.atlas-scientific.com/product_pages/components/bnc.html

Which then might be able to be used with the stripboard below

https://www.amazon.com/Busboard-ST2-Traditional-Stripboard-Prototyping/dp/B00LLQFRAU

To increase your frequency range by using three strips.  Ground the two outer strips, and you can modify the middle inner strip as you see fit - you may need to use a razor blade to remove some of the copper before you solder your capacitors and inductors that need to be in series.

[radiolistener]

No , this cheap probes at least don't have any more components ... like in David's schematic . Anyway anything is easily removable from the probe connector

even if you remove all from the probe, the probe uses special coax cable with high loss (it is required to reduce ringing in the cable), it has resistance about 331 Ohm. So, you're also needs to replace coax cable and connector (which has impedance match circuit with trimmer capacitor inside).

[CDaniel]

What "high loss"  ? I just wrote that cheap probes have 1ohm resistance anyway . Any coax cable you could use has some capacitance and inductance and I didn't said it is good for GHz range ... By the way , even for audio you should allways measure the real signal , not rely on what a generator tells you it is . This should be standard practice if you care about accuracy ( when is needed ) 

[rf-loop]

No , this cheap probes at least don't have any more components ... like in David's schematic . Anyway anything is easily removable from the probe connector

even if you remove all from the probe, the probe uses special coax cable with high loss (it is required to reduce ringing in the cable), it has resistance about 331 Ohm. So, you're also needs to replace coax cable and connector (which has impedance match circuit with trimmer capacitor inside).

All probes do not use "lossy" coaxial aka "R cable". Specially many cheap "crap" probes use coaxial where center wire is as normal wire.  These can easy mod for OP's purposes and example  CDaniel talk about probe what use low resistance coax.

But then. Many good probes use special coax with low capacitance and/or where center wire is highly resistive, example around ~300 ohm/m  (center wire example nickel-chromium) and shield have normal low resistance. Using these for OP's purposes, yes but then user need really know what he do and how it affect.
These things are well explained in old (not so old but 1969) and still true Tektronix probe "Bible" pdf. (page ~15 )

[radiolistener]

Specially many cheap "crap" probes use coaxial where center wire is as normal wire.

I have some Chinese cheap probes and they use lossy coax cable. But I can believe that these crap probes really exists