The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
Mar 28th, 2024, 3:59pm
Pages: 1
Send Topic Print
NF of mixer (Read 12306 times)
new_guy
Junior Member
**
Offline



Posts: 10

NF of mixer
Oct 23rd, 2007, 3:35am
 
Hi all,

Been simulating a single balanced mixer, however i'm confused about the NF, i getting somewhere aroung 300dB of NF.
I suspect that there is a mistake in my simulation setup. DC points are checked with all transistors is saturation. Conversion gain of 15dB.

First my rf input is PORT with type set to SINE impedance is 50 Ohms
LO is VSIN
The if input is PORT with type set to DC

As the output is differential, i used ideal_balun to convert the differential signal to a single ended signal before connecting to
the output if port with large off chip capacitors. NF was simulated using PSS and PNOISE. RF is 1.5GHz while LO is 1.504G.
Beat frequency is therefore 4M.

Ouestion 1) In PSS i set the output harmonics to 10, is this reasonable, or do i have to set to 400. Would this parameter affect the noise simultion?

Question 2) In PNOISE, sweep type is absolute. If i change it to relative the noise figure is better. However i not really sure if it is correct.

Question 3) From the noise figure graph, i always noticed a hugh spike near the RF/LO frequency.

Question 4) I would be able to get reasonable noise figure, when i set the RF port type to DC, beat frequency is PSS is automatically 1.504G. Is this meaningful. Got this technique from some mixer cadence lab session online.

Question 5) There would be a LNA in front of my Mixer, however my RF port is set to 50 Ohms, how could i bring the simultion setup to closer to my LNA output? I could sim the LNA output S22, which has a real and imag part, can i replace the 50 Ohm by this value?

Apologize for this being such a long question. Appreaciate anyone's insight and comments

Thanks a lot

Regards
Newbie.








 
Back to top
 
 
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: NF of mixer
Reply #1 - Oct 23rd, 2007, 6:06am
 
Hi.

Show me portion of netlist regarding analysis setting,
for example, pss, pnoise, options, port, vsin statements.

And show me output log of Spectre about input and output frequency.
Maybe you set wrong setting in pnoise.

Your conditions are:

RF : port, 50ohm, large, 1500MHz
LO : vsin, large, 1504MHz
IF  : port, 50ohm, dc, 4MHz

Quote:
Ouestion 1) In PSS i set the output harmonics to 10, is this reasonable, or do i have to set to 400. Would this parameter affect the noise simultion?

Fundamental Frequency is 4MHz, so harmonics of 10 is too small.
Rather set "Select from range" or "maxacfreq"

Quote:
Question 2) In PNOISE, sweep type is absolute. If i change it to relative the noise figure is better. However i not really sure if it is correct.

If you set IF frequency in pnoise, you have to set sweep type as absolute.

Quote:
Question 3) From the noise figure graph, i always noticed a hugh spike near the RF/LO frequency.

This is natural.  

Quote:
Question 4) I would be able to get reasonable noise figure, when i set the RF port type to DC, beat frequency is PSS is automatically 1.504G. Is this meaningful. Got this technique from some mixer cadence lab session online.

http://www.designers-guide.org/Forum/YaBB.pl?num=1184111670/1#1

Quote:
Question 5) There would be a LNA in front of my Mixer, however my RF port is set to 50 Ohms, how could i bring the simultion setup to closer to my LNA output? I could sim the LNA output S22, which has a real and imag part, can i replace the 50 Ohm by this value?

If my memory is correct, current Spectre support "port" with reactance in HB PSS.

Although it is not restricted to you, read manual or contact with technical support of EDA vendor.

Back to top
 
« Last Edit: Oct 24th, 2007, 3:15am by pancho_hideboo »  
View Profile WWW Top+Secret Top+Secret   IP Logged
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: NF of mixer
Reply #2 - Oct 23rd, 2007, 6:09pm
 
Hi,


you can set your RF port to DC. However, if you are using an integrated LNA and mixer, you should not use a 50 ohm port. Maybe you could just use an ideal voltage source, and take into account the mixer input impedance when you design your LNA.


cheers,
Aaron
Back to top
 
 

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: NF of mixer
Reply #3 - Oct 23rd, 2007, 9:29pm
 
Hi.

Quote:
you can set your RF port to DC. However, if you are using an integrated LNA and mixer, you should not use a 50 ohm port.
Maybe you could just use an ideal voltage source, and take into account the mixer input impedance when you design your LNA.

In RF mixer, using ideal voltage source for RF signal is not recommended.

(1) Although you can evaluate output noise, you can not evaluate NF
(2) RF input nonlinearity is dependent on RF source impedance.
(3) LO leakage to RF could cause amplitude at RF input, but it can't be included if you use ideal voltage source.

See the following.
http://www.designers-guide.org/Forum/YaBB.pl?num=1052660826/1#1



Back to top
 
 
View Profile WWW Top+Secret Top+Secret   IP Logged
new_guy
Junior Member
**
Offline



Posts: 10

Re: NF of mixer
Reply #4 - Oct 23rd, 2007, 10:24pm
 
Hi,

Thanks for your comments and feedback. Have rerun the simulations with the suggestions, still having a pretty high NF at 4MHz.
Just wondering if it's correct.

Thanks

Regards
Newbie
Back to top
 

nf_of_mixer.jpg
View Profile   IP Logged
new_guy
Junior Member
**
Offline



Posts: 10

Re: NF of mixer
Reply #5 - Oct 24th, 2007, 12:39am
 
Hi,

Thanks for your replies, i think i figured it out. THe noise figure on the right is due to blocking signals, while the noise figure on the left was determined by RF being small signal. Thanks again

Regards
Newbie
Back to top
 
 
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: NF of mixer
Reply #6 - Oct 24th, 2007, 1:26am
 
Hi, new_guy.

No, you don't understand. This is not NF degradation due to RF blocking. Sad

Quote:
pss  pss  fund=4M  harms=500  errpreset=moderate  annotate=status
pnoise  pnoise  sweeptype=absolute  start=1M  stop=1.6G
+       maxsideband=10  oprobe=PORT1  iprobe=PORT0  refsideband=-1
+       annotate=status

When RF is dc and only LO is large, "refsideband=-1" in pnoise is correct.
But when both RF and LO are large, "refsideband=-1" in pnoise is not correct

You have to set  "refsideband=-376".
And "maxsideband=10" is too small.
Back to top
 
« Last Edit: Oct 24th, 2007, 2:56am by pancho_hideboo »  
View Profile WWW Top+Secret Top+Secret   IP Logged
aaron_do
Senior Fellow
******
Offline



Posts: 1398

Re: NF of mixer
Reply #7 - Oct 24th, 2007, 2:02am
 
Hi panchi_hideboo,


Are you sure about not using a ideal voltage source? I was thinking you should not find NF anyway, you should look for input referred noise. Also, although you can look for LO_RF leakage with a 50 ohm port, it is not the correct amount of LO-RF leakage. The proper amount of leakage can only be viewed when the LNA is hooked up. In fact since the mixer is not driven by a 50 ohm load, there's no guarantee that other performance measure will be correct either...


cheers,
aaron
Back to top
 
 

there is no energy in matter other than that received from the environment - Nikola Tesla
View Profile   IP Logged
pancho_hideboo
Senior Fellow
******
Offline



Posts: 1424
Real Homeless
Re: NF of mixer
Reply #8 - Oct 24th, 2007, 2:22am
 
Hi, Aaron.

Please see the following surely.
http://www.designers-guide.org/Forum/YaBB.pl?num=1052660826/1#1  


NF is defined as 1+Neq/Nin. Here Neq is input referenced noise. Nin is noise from source.
Idealy this Neq is not dependent on source impedance, but actually source impedance could affect periodic operation point.
So Neq is dependent on source impedance.

Evaluating Neq using ideal voltage or current is possible for relative low frequency.
This is classical spice simulator's method(.NOISE statement). I use this method for IF amplifier, OPAmp, low-freq-Mixer, etc.

Of course such Neq(determined by ideal source or 50ohm drive) is useful for rough estimation of noise characteristics.
Especially when studying level diagram, I use input referenced noise for calculating interstage S/N in cascaded system.
I don't use NF for this purpose since actual stages are composed of various impedance and their signal could be current,
voltage, single, differential, etc. So it is difficult to study system using NF.

Also you should not think of NF as ratio of S/N especially for interstage of cascaded system.
http://www.designers-guide.org/Forum/YaBB.pl?num=1192525133/15

Back to top
 
« Last Edit: Oct 24th, 2007, 5:07am by pancho_hideboo »  
View Profile WWW Top+Secret Top+Secret   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2024 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.