Function generator drops off a lot as frequency goes over 1MHz

[pierrel2]

Picoscope 5442D
It appears that the ADC does not work as expected as the frequency climbs.

Specifically, using an external calibrated scope, 1V amplitude sine wave gives:
@1MHz: 1.86V
@2Mhz: 1.56V
@3MHz: 1.36V
@4MHz: 1.38V
@5MHz: 1.12V

Since our work is focused in this range, it's a bit frustrating to have to use a lookup table to generate the correct values. Is this something that can be adjusted, or a design limitation?

[Gerry]

Hi pierrel2,

You refer to the function generator in you title but you mention the ADC in your description, so I'm not entirely sure where you think the problem is, but what I can tell you is, that is not what we find when performing measurements with either the Function Generator, ADC in the Input Channel, or both. The peak values are consistent through the range you mentioned.

Could you post a psdata file so that we can see what you are doing.

Regards,

Gerry

[pierrel2]

Gerry,
thanks for pointing this out. I retried the test with a straight BNC cable between "Gen" and "A" and the function generator behaves as expected (100mV in = 100mV out across 1MHz - 5MHz). OTOH, if I do what I did yesterday, and naively use the included probes, both set to "1x" to do the same, I get the results I shared yesterday. Never paid much attention to the data sheets included in oscilloscope probes, but I suppose it's time to start looking at that.
The probe data sheet suggests the bandwidth is much higher at "10x", so I set both probes to "10x", reconnected it and saw that with a 1V (2V p-p) output, p-p is steady across 1MHz - 5MHz and measures on PicoScope at 568mV down to 544mV p-p and 604mV down to 578mV on a calibrated scope (where I've gone and adjusted the input to label it as using a 10x probe in both cases). I like the stability across frequency but not quite sure how to understand the amplitude of the output signal. I would have guessed that I would see 200mV p-p for 2V p-p function generator output going to 10x probe.
Thanks!
Pierre

[Gerry]

Hi Pierre,

The Probe types for the channel inputs are the means to give you what the True reading is in the display, when using a Probe that will change the True reading. So when using a x1 probe type setting with a x10 probe you will see the level drop, but when using a x10 probe type setting with a x10 probe the reduced voltage is multiplied by 10 to compensate for the reduction.

Regards,

Gerry

[pierrel2]

Gerry,
I'm not disputing that - scopes have worked that way for a long time. What surprised me, and led me to abandon my initial idea, was that the "10x" , applied in the case where the probe was used to bring the output of the function generator out to the circuit, was that the behavior was something quite unexpected. Am now using a straight BNC cable with an alligator-clip adapter to get around the issue.
Thanks,
Pierre

[Gerry]

Hi Pierre,

If you're connecting the BNC connector of a probe to the Signal Generator set to a 1V output and the tip and ground to a circuit, and the Probe you're using is set to x10 then if the probe is working correctly you will have 100mV at the tip of the probe. The reason I mentioned the settings in the picoScope was that I assumed you might have had the input switched to x10 for the probe, and then connected it to the Sig Gen instead, while switching it to x1 because it's no longer connected to the input (corrected).

If you're getting the same level from a BNC cable with a clip and a probe switched to x10 then I would have to assume that there's a problem with the probe (e.g. x10 switch is broken).

Regards,

Gerry