The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Measurements >> RF Measurements >> LNA gain and input matching
https://designers-guide.org/forum/YaBB.pl?num=1281853392

Message started by yixiusky on Aug 14^th, 2010, 11:23pm

Title: LNA gain and input matching
Post by yixiusky on Aug 14^th, 2010, 11:23pm

Hello everyone

I would like to ask one question

I designed one LNA works frome 1.8 to 2.2 GHz. After meausrement, the input and output matching are ok, but the gain shiift to 1.5GHz. I could not think the reason. Could you help me?

Thank you very much.

Title: Re: LNA gain and input matching
Post by pancho_hideboo on Aug 15^th, 2010, 3:40am

yixiusky wrote on Aug 14^th, 2010, 11:23pm:

After meausrement, the input and output matching are ok,
but the gain shiift to 1.5GHz.

What do you mean by "the gain" ?
Do you want to mean "a peak of gain or S21" ?

Show me frequency characteristics of S11, S22 and S21 as LNA.

yixiusky wrote on Aug 14^th, 2010, 11:23pm:

I could not think the reason.

I think it is due to characteristics of device intrinsic S21.

Show me frequency characteristics of S11, S22 and S21 of intrinsic device.

Title: Re: LNA gain and input matching
Post by vp1953 on Aug 15^th, 2010, 10:44am

That seems like a large shift (assuming you are referring to peak gain). How was the measurement done - was the LNA output buffered? If not, what was the load on the LNA output for the gain measurement?

Title: Re: LNA gain and input matching
Post by yixiusky on Aug 30^th, 2010, 12:12am

Thank you very much for reply, I did more measurement, and i attach the result here.

(1) input matching S11 (both magnitude and smith chart)

(2) Output matching S22 (both magnitude and smith chart)

(3) Gain and NF

My LNA design is 1.8GHz to 2.2GHz, the peak gain should appear at 2 GHz.

However, the inputmaching peak is around 2GHz, but the output peak is at around 1.5GHz.

Then the peak Gain is also 1.5GHz. and the gain decrese around 5dB compare with post layer simulation result.

I do not know the reason why the Peak Gain is shifted to 1.5GHz, and also decreased around 5dB. Could you please help me? thank you so much

Title: Re: LNA gain and input matching
Post by yixiusky on Aug 30^th, 2010, 12:18am

I also attach the gain and NF response here. We can see the peak gain is at 1.6GHz, not 2GHz. My target gain is 15dB, now is 11.9dB.

Title: Re: LNA gain and input matching
Post by pancho_hideboo on Aug 30^th, 2010, 6:52am

yixiusky wrote on Aug 30^th, 2010, 12:12am:

However, the inputmaching peak is around 2GHz, but the output peak is at around 1.5GHz.
Then the peak Gain is also 1.5GHz.
and the gain decrese around 5dB compare with post layer simulation result.

Show me circuit topology of your LNA especially regarding output load configurations.

Title: Re: LNA gain and input matching
Post by yixiusky on Aug 30^th, 2010, 7:58pm

I use resistive feedback type LNA , I attach the schematic here

Title: Re: LNA gain and input matching
Post by pancho_hideboo on Aug 31^st, 2010, 6:08am

yixiusky wrote on Aug 30^th, 2010, 7:58pm:

I use resistive feedback type LNA , I attach the schematic here

Simply output matching frequency moves to lower frequency due to parasitic capacitors for Rf, drain of M2 and Lload.

Title: Re: LNA gain and input matching
Post by RFICDUDE on Sep 1^st, 2010, 2:49am

Your circuit has two resonant circuits that are coupled. The resistive feedback couples the tuned output load and the resonant input matching network, so it is very difficult to simultaneously match the input and have the output tuned for maximum output at the same frequency. It is even more difficult if you are trying to maintain a controlled (matched) impedance at the output too.

Some systems can tolerate some amount of gain variation over frequency, so having the gain at a maximum may not be as important as maintaining a good match and low noise figure.

Alternatively, you could decouple the output and input by removing the shunt feedback, but then the input matching network would have to be redesigned for the impedance change and the gain would be higher (presumably) from input to output. But the increased input/output isolation would decoupled the two tuned networks such that they will not interact (unless there are significant parasitic coupling paths).

Title: Re: LNA gain and input matching
Post by vp1953 on Sep 2^nd, 2010, 4:08pm

Hi RFICDUDE,

Why in your opinion, did not the measurement data correlate with simulation results? Assuming that there are no bugs with the PDK and the simulations were done correctly, the difference between simulations and actual results seem quite large.

Title: Re: LNA gain and input matching
Post by yixiusky on Oct 2^nd, 2010, 1:24am

when i measure this LNA, i found one problem.

In this LNA, the inductor is connected to VDD. So when i meausre the chip, the VDD Pad is easily die. And then the DC condition is totally wrong.

What should i do to protect the vdd pad?

Title: Re: LNA gain and input matching
Post by yixiusky on Oct 2^nd, 2010, 8:39pm

when i measure this LNA, i found one problem.

In this LNA, there are two VDD pad. One is connect to bias circuit, while another one is connected to LNA core (inductor), as shown in schematic.

When i meausre the chip, among all pads, only the VDD Pad (connected to inductor) is easily dead. And then there is no DC flow from this VDD pad.

What should i do to protect the vdd pad?

Title: Re: LNA gain and input matching
Post by yixiusky on Oct 2^nd, 2010, 10:19pm

pancho_hideboo wrote on Oct 2^nd, 2010, 9:50pm:

[quote author=yixiusky link=1281853392/0#11 date=1286077142]
I think gate of M2 are broken actually.
Insert RC filter between Vdd and gate of M2.

Thank you very much.
All pad are ESD PAD
Actually, there is RC filter between Vdd and gate of M2. i did not draw in the shematic here.
So i think the gate of M2 has no problem.

Title: Re: LNA gain and input matching
Post by yixiusky on Oct 2^nd, 2010, 10:36pm

Thank you very much

if the ESD diode is broken, what should I do?

Yes, i set the current limit for Vdd regulator.

Title: Re: LNA gain and input matching
Post by yixiusky on Oct 2^nd, 2010, 10:56pm

thank you very much

i will try it right now

Title: Re: LNA gain and input matching
Post by yixiusky on Oct 2^nd, 2010, 11:59pm

pancho_hideboo wrote on Oct 2^nd, 2010, 10:46pm:

As another possibility, unless impedance of Vdd is enough low, extraordinal oscillation could occur between Vdd and drain of M2.
Then it breaks drain of M2.

Try to place 10pF, 100pF, 1000pF and 100kohm between Vdd and Gnd.

I have question

You mean place all 10pF, 100pF, 1000pF and 100kohm in parralel? or place just one by one, and test?

My frequency is wideband, from 1.8~2.7Ghz.

From VDD to ground, when I place the capacitor, which value should I choose?

Title: Re: LNA gain and input matching
Post by pancho_hideboo on Oct 3^rd, 2010, 2:13am

yixiusky wrote on Oct 2^nd, 2010, 11:59pm:

You mean place all 10pF, 100pF, 1000pF and 100kohm in parralel?

Yes.

10pF must be placed in most near of Vdd.
100pF is next. And 1000pF is next to 100pF.
100kohm is near of 1000pF.

BTW, what value do you apply as stimulus power in Network Analyzer ?

Title: Re: LNA gain and input matching
Post by snaildr on May 7^th, 2011, 7:52pm

Why do you use resistive feedback and also inductive degeneration? I though the point of using resistive feedback and thus making the NF a bit worse is to go without any inductor and save a ton of area, right?