Function Generator question - frequency vs. voltage

[Electro Fan]

I only have one function generator so it's hard to compare but any reason why changing the frequency (from say 20MHz to 20kHz should cause the peak to peak amplitude to change (from say 8 volts to 6 volts)?  Changing from 2MHz to 200kHz to 20kHz to 2kHz or from 2MHz directly to 2kHz doesn't seem to change the amplitude.  I'm using a BK Precision 4040A; maybe it's a limitation of this particular model? 

The spec for sine waves is:
Flatness
+/-5% (0.45 dB) 10Hz to 10MHz
+/-20% (2.0 dB) 10MHz to 20MHz

So maybe I'm just seeing the flatness curve?

Thanks

[Electro Fan]

Might have solved my own problem.  The function generator has a 50 Ohm output; the scope has 1M Ohm input.  When I insert a 50 Ohm terminator between the scope input and the cable going to the function generator sine wave amplitudes remain constant across all frequencies.  (And square waves clean up nicer too.)  Maybe that was it?  Thanks for any info.  EF

[w2aew]

Might have solved my own problem.  The function generator has a 50 Ohm output; the scope has 1M Ohm input.  When I insert a 50 Ohm terminator between the scope input and the cable going to the function generator sine wave amplitudes remain constant across all frequencies.  (And square waves clean up nicer too.)  Maybe that was it?  Thanks for any info.  EF

Yes, that is why. You were seeing the roll off due to the high impedance scope load. Lesson on high speed signal integrity and the need to properly terminate 50 ohm transmission lines - complete!

[Electro Fan]

Kind of a follow-up related question.

I've been experimenting with an Arduino Uno.  I made a simple sketch to pulse out square waves with patterns designed to represent ASCII characters.  I haven't figured out how to adjust the speed overall.  I think it's fixed based on the output of the Uno.  According to my scope the signal is only 100Hz and it seems to require adjusting the baud rate on my scope decoder to about 200 baud (anything from about 190 baud to about 210 baud correctly decodes the characters - slower or faster decodes the signal into incorrect characters).

In the process of getting things sorted out I tried connecting some breadboard wires to alligator clip test leads with a BNC connector on the other end connected to one of the scope inputs.  This rendered only modestly clean (somewhat noisy) square waves, and the decoder didn't consistly decode the correct characters; every few characters were decoded incorrectly. 

On a hunch on put a 50 ohm terminator between the BNC connector on the test leads and the scope input and sure enough, the signal cleaned up very nicely.  It was only then that things stabilized to the point that I could change the baud rate value on the decoder and see the range within the decoder that would accurately decode (which seemed to confirm that the baud rate needed be 2x the frequency; I think this is correct?).

The question is:  how was I supposed to know or determine that this test setup needed the 50 ohm terminator?  I'm starting to wonder when I wouldn't want to use the 50 ohm terminator.

Fyi, I have a Rigol 2072 (sorry about that w2aew, you're still in my scope and test equipment Hall of Fame).  I'm wondering if the 2072A might have a substantial advantage if it automatically adapts between 1M ohms and 50 ohms.

[Zero999]

If you used a compensated x10 'scope probe, you could have easily measured the output of the function generator without any attenuation or distortion.

In the case or your experiment involving serial transmission. I suspect it's mains pick-up which is the problem, rather than transmission line reflections as 100Hz is too low for that to be a problem. The wires you used weren't shielded, so picked up noise from the mains which was removed by connecting 50 Ohm to 0V as the source impedance of the mains hum would be quite high (many Mega Ohms), yet the source impedance of the serial signal would be very low (under 50 Ohms). Try using shielded cables and this should improve, without the need to add a resistor.

[Rigby]

Can someone explain to me why a 10,000,000 ohm load is 10M impedance, but a 10,000,050 ohm load is 50 ohm impedance?

[BravoV]

Can someone explain to me why a 10,000,000 ohm load is 10M impedance, but a 10,000,050 ohm load is 50 ohm impedance?

That 10 Mega Ohm is in "parallel" with the 50 Ohm, NOT in series.

[robrenz]

Can someone explain to me why a 10,000,000 ohm load is 10M impedance, but a 10,000,050 ohm load is 50 ohm impedance?

The 50 ohm feedthru terminator is not 50 ohms in series with the 10M input impedance of the instrument. It is 50 ohms to ground which puts the 50 ohms in parallel with the 10M input impedance which gives an effective 49.99975 ohm input impedance.

BravoV you beat me to it

[Electro Fan]

Can someone explain to me why a 10,000,000 ohm load is 10M impedance, but a 10,000,050 ohm load is 50 ohm impedance?

The 50 ohm feedthru terminator is not 50 ohms in series with the 10M input impedance of the instrument. It is 50 ohms to ground which puts the 50 ohms in parallel with the 10M input impedance which gives an effective 49.99975 ohm input impedance.

BravoV you beat me to it

Any practical (significant) difference between 50 ohms and 49.99975 ohms if I'm just trying to decode some ASCII characters?

[BravoV]

The 50 ohm feedthru terminator is not 50 ohms in series with the 10M input impedance of the instrument. It is 50 ohms to ground which puts the 50 ohms in parallel with the 10M input impedance which gives an effective 49.99975 ohm input impedance.

Any practical (significant) difference between 50 ohms and 49.99975 ohms if I'm just trying to decode some ASCII characters?

And that 49.99975000124999 Ohm total is just a 0.00049999750002% deviation or 4.9999750002 ppm off from the original 50 Ohm impedance, laymen term ... It doesn't matter !   

BravoV you beat me to it

See that ? I rather say that we're complementing each other, rather then beating each other.