We are using VGA IC ADL5330 at 2.2-2.6GHz range & observed that at frequency 2.44GHz , VGA control voltage =0.95V, I/P of VGA = -17.2dBm & at O/P of VGA ( after balun) = -44.3dBm .At pg. no. 4 of IC datasheet it is mentioned that in 2.2GHz - 2.7GHz range higher side gain (i.e. at Vgain=1.4V) 16dB to 10dB resply.At pg. no. 11, fig. no. 28, in 2.2- 2.7GHz range, I/P return loss is -15dB & fig. no. 30, O/P return loss is -5dB approx.On page no. 15, it is mentioned that “The useful lower limit of the gain control voltage increases at high frequencies to about 0.5 V and 0.6 V for 2.2 GHz and 2.7 GHz, respectively.” Can’t we use VGA gain control from 0 to 1.4V? or should it be from 0.5V to 1.4V?
As per data sheet of ADL5330 & observations it seems that VGA is not going to offer any gain in dB above 2.2GHz. Will you pl. suggest another VGA or any other alternative to achieve gain in dB? Your comments pl.
There is no danger in applying less than 0.5 on the GAIN pin. At the higher frequencies, the gain does not change much between 0V and 0.5V for the GAIN in. Please see figures 6 and 7 in the datasheet.
There is gain above 2.2GHz, but the frequency response is rolling off. Please see Figure 25.
All of our VGAs and DGAs are listed here:
You can sort by frequency, min gain, max gain, etc. to choose the VGA that suits your needs.
Thanks for your reply. My further query is at frequency =2.44GHz, when VGAIN is varied from 1.1 to 1.4V range output in dBm increases but side bands appears in the spectrum . Upto VGAIN = 1.0V no side bands appear in the spectrum. How to remove these sidebands ? Pl. comment.
PFA for spectrum analyser waveform at 1.0V & 1.4V .
You are probably compressing the part. Check out Figure 12 in the datasheet. You'll notice Input P1dB stays relatively flat with VGAIN, but Output P1dB changes with the gain level.
What is your input signal level with the associated gain values? You if could proved the following info in tablular format, that would help diagnose this problem:
VGAIN (V) Input Power (dBm) Output Power (dBm) Visible Sidebands(yes/no)
0.8 ? ? ?
0.9 ? ? ?
1.0 ? ? ?
1.1 ? ? ?
1.2 ? ? ?
1.3 ? ? ?
1.4 ? ? ?
I have note down readings for Input power & output power in dBm for VGA ADL 5330. Pl. refer following table for the same.
Input signal Level (dBm)
(I/P of balun)
Output signal Level (dBm)
(O/P of balun)
Visible side bands (Yes/no)
Now instead of evading main query of ours, you are requested to answer following:
1) Assuming you are NOT compressing the input (i.e. keeping the input level below +3 dBm approx.), I believe you are compressing the output: the act of compressing the part causes the side bands. Note that side bands may start to appear at levels below the output P1dB. You must lower the signal level to remove the side bands.
2) I assume you mean the input P1dB vs. Gain plot? This is because the gain control element (the ladder attenuator) is after the input gm stage. See Figure 31 in the datasheet. Thus the input gm stage has its own compression number, that varies between 1 dBm and 3 dBm, based primarily on Frequency and somewhat on Gain.
I notice that your gain at VGAIN = 1.4V is 6.7 dB. This seems a bit low for 2.44 GHz. Compare this to 2.7 GHz, the gain is 10 dB at VGAIN = 1.4V. Do you have a lossy network on either the input or the output?
I also notice that you do not observe the side bands with the conditions in the table above, including from Vgain voltages between 1.1V and 1.4V. Did you change something in your testing between the data presented in the spectrum analyzer pictures, and the data presented in the table above? Before, there were side bands at VGAIN = 1.4; now there are no side bands observed at 1.4V.
Let me respond to your comments point wise-
1)Correct, input level stays well below+3dbm, we are trying to compress output. It seems that we can increase output level at following PA output pin to +3dBm when Vgain=1.1V at ADL5330. Therefore this seems to be upper limit. Isn’t there anyway to increase this by another 10dBm?
2) Unnumbered point (4?) of your reply –you queried whether you have lossy n/w on either I/P or O/P? ADL5330 being differential device, we have employed balun (ETC-1-1-13) to apply single ended signal & connect to PA following it which is single ended. Those balun cause loss of 3dB each & that explains 3+3=6dB loss.
3) Unnumbered point (4?) – These are based on SA readings at balun I/P & O/P being single ended & the amplitude is very small -47dBm & no sidebands The sidebands are observed at O/P of PA only. This table was presented to you since you wanted to know conditions at ADL5330 I/P & O/P. I thought actual data will mislead/ confuse you, since you cannot check at O/P or I/P of ADL5330. If you try to connect SA there then SA’s 50Ω I/P impedance & existing devices like PA’s input impedance will come in parallel & SA will show very low values like -47dBm but no sidebands. Hence we gave values at O/P of ADL5330 by subtracting 16dBm (gain of PA) from amplitude shown on SA at O/P of ckt. BNC.
“No sidebands” comment according to me is obvious because by connecting SA you will get very low value of -47dBm then obviously sidebands will be still lower than this & hence not be seen on relative vertical scale of SA.
4) Return Query: Does that mean that we can never reach an O/P level of 13dBm?
Conclusion:-Still the main question remains unanswered-
At 2.44GHz frequency, when VGAIN is varied from 1.1 to 1.4V range output in dBm increases but side bands appears in the spectrum . Upto VGAIN = 1.0V no side bands appear in the spectrum. How to remove these sidebands?
Do you have access to an oscilloscope with a high impedance probe? If so, at these higher signal levels, I would probe the output of the PA, input of the PA, output of the ADL5330, and possibly the input of the ADL5330 (the input signal level could be hard to see on the scope). You could determine all the gains and losses of the VGA, PA and the baluns.