Demo Arbitrary Waveform Generator ARB unit brought in by Tristen Tan <tristen.tan@triotest.com.au> for us to test.
The unit was taken to David Reilly's lab, USYD for noise testing and comparison with Tektronix 5014C AWG (1.2 GS/s). Same SMA cable from minicircuit was used.
The recent testing result on Tektronix 5014C is shown in this link
Tektronix5014C
Rise time testing
Both AWGs were connected to Agilent
DOS90254A (2.5GHz&20GS/s)
Square Wave f=10MHz Vpp=0.5mV
Tab has 1ns rise time but more ringings.
Tek has 2ns rise time but less ringings.
Harmonics testing
Sine 10MHz Vpp = 0.5V
Connected to Agilent
E4407B spectrum analyzer with bandwidth set to 300kHz
|
Tab WX2182B |
Tek 5014C |
1st harmonic |
55 dB lower |
55 dB lower |
2nd harmonic |
35 dB lower |
50 dB lower |
Clock signal |
60 dB lower |
75 dB lower |
Comment |
clock amp change with Vpp |
clock amp does not change with Vpp |
Changing waveform to square does not affect clock signal amp
It seems Tab uses attenuator to change Vpp so clock signal is also changing with Vpp.
If zoom in at the clock signal, carrier modified peaks can be found with 10MHz higher/lower than the clock. For Tab, the two peaks are almost symmetric while for Tek the higher frequency peak is almost invisible.
Then set Vpp = 2V, similar result can be found on both AWG
clock = -54dBm, carrier = -2dBm, huge harmonics
Dynamic range testing
Spec comparison
|
Tab WX2182B |
Tek 5014C |
Vertical Resolution |
14 bit |
14 bit |
Vpp range |
50mV to 2V |
40mV to 9V |
Set Vpp Tab's minimum value = 50mV and the output looks similar on both AWG.
Jitter testing
Jitter is the undesired deviation from true periodicity of an assumed periodic signal.
For Jitter test, the OSC trigger was set at the raising edge of the wave. The after half period (falling edge) or full period (raising edge), a small measurement window around 0V was opened (+/- 100uV) and the measurement point falling in that window was recorded. Then the histogram was plotted and the standard deviation (std) calculated. For good periodicity, the std should be small and the histogram should be close to standard distribution.
For this testing, 10MHz 0.5V Vpp square wave was used
|
|
Std |
σ (68%) |
2σ (95%) |
3σ (99.7%) |
Raising edge |
Tek 5014C |
3 ps |
77% |
99% |
100% |
Tab WX2182B |
0.9 ps |
87% |
100% |
100% |
Falling edge |
Tek 5014C |
2.8 ps |
72% |
99% |
100% |
Tab WX2182B |
2.6 ps |
63% |
100% |
100% |
Low frequency noise testing
Standford SR760 FFT Spectrum Analyzer was used.
The FFT is recorded between 0 to 100kHz.
Tek 5014C only shows 1/f noise.
Tab
WX2182B has two channel. Each channel also have DC and AC output. When CH1-DC is connected and turned OFF, a broad peak around 80kHz can be seen. When then channel is turned ON, peaks with ~6kHz spacing can be found. When CH2 is turn on, cross-talking can be recorded.
When CH1-AC was connected, no cross-talking can be recorded. But few peaks with higher amplitude can be found
Tektronix and Tabor AWGs on SR FFT spectrum analyser
Factory test data
The Tabor technician tested
WX2184A (similar specs but 4-channel) and
WX2182B with the latest firmware.
Low frequency noise
They measured the low frequency noise (1 to 100kHz) and found several peaks at similar positions. But the amplitude is significantly smaller.They measured the spectrum with Vpp = 2V / 500mV / 50mV (left to right)


Please note that the unit is dBm. Using 50 Ohm impedance and converting that to voltage, Vpp ~ 20 microV
Then they tested the
WX2182B and WX2184 again (left
WX2182B, right WX2184)
4 channel unit has more spikes but the noise floor is lower. Vpp ~ 18 microV for both of them

Clock noise
They got similar result. (left: Vpp = 2V, right: Vpp = 500mV)

Harmonics of carrier signal
They got similar result. (Vpp = 2V / 500mV / 50mV left to right)


Access Control
Previously the default template was set up so that all pages were visible by the whole world, and we relied upon authors to use access control to limit access to sensitive pages. That hasn't worked, so now pages are by default only visible to QED group members. Delete the following line if this page is meant to be visible by the whole world.