What is the resolution of the ADF4350/ADF4351 devices? People asked and I verified...

Hi,

Some people using the VMA Simple Spectrum Analyser Software asked me about some strange signal shapes and in the past I thought they were simply looking at Double Sided Amplitude Modulation (https://en.wikipedia.org/wiki/Amplitude_modulation).

But as the inquiries increased, I sat down and did some simple tests.

First off, here are the specs as published by the Chinese sellers on eBay:

• Output frequency range: 35MHz - 4.4GHz
• Stepping: 1kHz
• Output power (for reference):
• 150MHz -3.2dBm
• 250MHz -3.4dBm
• 500MHz -3.4dBm
• 750MHz -2.8dBm
• 1000MHz -3.4dBm
• 1500MHz -3.6dBm
• 2000MHz -4.0dBm
• 2500MHz -2.8dBm
• 3000MHz -1.8dBm
• 3500MHz -3.6dBm
• Spectrum of the input frequency range: 35MHz - 4.4GHz
• Frequency Accuracy (after calibration): +/- 1KHz @ 1GHz
• Low-noise input signal amplitude:-70dBm @ 0.5GHz

The alone means, that the device cannot "look" at anything under 1KHz of resolution.

Here are some pictures, to show what happens:

Picture 1: This shows an analogue TV transponder. The right peak shows the video carrier, the second peak the audio carrier. The modulation is of course AM - Amplitude Modulation (https://en.wikipedia.org/wiki/Amplitude_modulation), so there shouldn't be the inverted spikes in the middle of the peaks, as in Sideband Modulation (https://en.wikipedia.org/wiki/Sideband)!

Picture 2: Is it a problem of the VMA Simple Spectrum Analyser software? I guess not, as WinNWT, the original software for these devices, shows exactly the same spectrum!

Picture 3: Fortunatly I own a Siglent SSA3021X spectrum analyser, so I fed the same signal, using the same parameters. This is how the transponder looks like. Quite different!

Picture 4: Zooming into the video carrier, we can see the video carrier exactly how it should look like in a spectrum analyser.

Picture 5: Zooming into the the same video carrier with the VMA Simple Spectrum Analyser, we see a totally different picture. Clearly we are beyond the resolution capabilities of the device! When it is sweeping, it will use a too wide RBW - in fact we can only change the sweeping steps, but not the RBW.

Picture 6: If we select a combination of span and samples, so that the (equivalent) RBW is 0.1MHz, we get a correct representation of the spectrum.

Picture 7: Same spectrum settings as in previous picture, but using the Siglent SSA3021X. This confirms that at 0.1MHz (equivalent) RBW the ADF4350/ADF4351 devices are showing a good representation of the spectrum.

Conclusion: Do not expect to resolve higher than specification of the ADF4350/ADF4351 with the VMA Simple Spectrum Analyser software, even though you can input the relevant settings!

This problem results from the requests I had to implement three decimal points to all frequency input fields - before the software only used 1MHz resolution. This opened the door to usage beyond the hardware capabilities. Nothing wrong with that, just always question what you are measuring!

Hopefully future SMA/NWT devices will come along with higher resolution.

Regards,

Vitor