Trueform and some guick checks...

This is the first blog post of my Agilent 33622A arbitrary waveform generator Roadtest review. Initial impressions review is at: http://www.element14.com/community/roadTestReviews/1759

One of the prime features of Agilent's 33600 and 33500 series of arbitrary generators is the proprietary Trueform technology. According to Agilent claims, it bridges the world of DDS (direct digital synthesis) and PPC (point per clock) arbitrary generators, bringing the almost PPC like signal generation, with low noise, no point skipping and improved signal integrity, to a DDS price range.

Agilent claims it achieved this through „virtual variable clock“ and „advanced filtering“. I had a feeling that there is some heavy duty signal processing happening in the digital domain, and yes, after reading the excellent white paper „Complex Modulation Generation with Low-Cost Arbitrary Waveform Generators“ by Joan Mercade for Agilent (5991-1100EN) , my initial opinion was confirmed.

In the block diagram (taken from the above mentioned white paper) it is clearly visible that there is an additional element involved in signal generation chain, compared to DDS architecture, the „Interpolarator“, whose task it to take samples from the waveform memory using oversampling, probably at full 1 GS/s, digitally low pass the signal and then decimate the signal to match the required sample rate before feeding it to the DAC. This ensures minimum uncertainty period between samples, and provides for an excellent true RMS Jitter specs.

Low level performance

One of the advertised capabilities of 33622A and other generators from the series is excellent low voltage level performance, again thanks to the Trueform technology and 14 bits of resolution.

I wanted to verify this with a few quick measurements.

Sine, 1 mV RMS at 100 Hz:

Sine, 10 mV RMS at 100 Hz:

Sine, 100 mV RMS at 100 Hz:

Sine, 1 mV RMS at 10 kHz:

Sine, 10 mV RMS at 10 kHz:

Sine, 100 mV RMS at 10 kHz:

During the measurements, output load impedance on channel 1  was set to High Z. It it worth mentioning that the output load impedance of 33600 series generators can be easily set to most common values, or to a completely arbitrary value. However, and this is a trap for young players (Dave Jones TM ), this will NOT change the actual output impedance of the instrument, but it will adjust displayed values of amplitude and offset levels on the instrument to match those measured at the actual load.

AC bandwidth of the Fluke 287 is specified  at 100 kHz, and while being an order of magnitude above the max test frequency of 10 kHz, it is safe to assume that it had no influence on the measurements.

In the next blog post I will show the easy way to create IQ baseband signals on this unit, without the IQPlayer option.