Mixer Level vs. Reference Level

[cochenob]

I understand that the "Reference Level" is the expected integrated power over the band of interest.

What exactly is NI defining as the "Mixer Level"?  Is it the "actual" input power?  Is it the LO power?

I am trying to understand the automatic attenuation procedure in the 566x (or similar downconverters), as outlined in the NI 566x help document. 

[Andy_Hinde]

Hi cochenob,

Mixer Level (ML) is the signal level at the input to the first mixer in the signal chain. Reference Level (RL) is the signal level to the RF input of the instrument.

What lies between the RF input and the first mixer input is programmable attenuation. The ML is always < RL.

RL, ML, and input attenuation are all linked together. Any two parameters define the third with the relationship: Attenuation (dB) = RL - ML.

Regards,

Andy Hinde

RF Systems Engineer

National Instruments

[cochenob]

Ok.

So the Reference Level is set by the user in the program.

From the documents...it suggests that the programmable attenuation is a "fixed" attenuation. (i.e. 3 RF attenuators of 20, 20 and 10 dB...and 2 IF attenuators of 20 and 10dB). 

This suggests that there is an "ideal" power range that should be delivered to the input of the mixer.  In other words...based on whatever the ideal power delivered to the mixer is...combined with the Reference Level set by the user...LV figures out how many attenuators to set. (In other words...there's no way to get, say, 7.38dB of attenuation).

Is this right?  If so...this further suggests that the input to the first mixer has some sort of dynamic range, but I can't find it anywhere.  I guess my point is this...if for some reason you set the Reference Level way far away from your actual input level, the wrong amount of attenuation might be configured.  The impact of this would be dependent on the dynamic range of the mixer input.  It's possible though that you'd either under-drive the mixer (too much attenuation) or send it into compression (too little attenuation).  True?

Also...on the 5660 block diagram...only one RF attenuator is shown at the input.  Is this really the 3 RF attenuators?  Likewise...later on in the chain, there's another programmable attenuator...assuming this is really 2 IF attenuators.

---

Brandon

[Andy_Hinde]

Hi Brandon,

"This suggests that there is an "ideal" power range that should be delivered to the input of the mixer."

Yes, this is definitely the case, and the case with all spectrum analyzers and vector signal analyzers. Setting the RL too high will result in too much attenuation applied and you will get lower SNR than you could get with an optimal RL. Setting the RL too low will result in overdriving the mixer, causing non-linearity spurs to pop up in the spectrum, and also risking damage to the mixer in extreme cases.

I gather you are using the PXI-5660 - the 5660 has 50 dB of programmable RF attenuation, applicable in steps of 10 dB. You are correct that the block diagram is simplified, showing one RF attenuator block to represent the three programmable attenuators (20, 20, 10 dB). The same goes for the IF attenuators as well.

Regards,

Andy Hinde

National Instrumemnts

[cochenob]

Thanks Andy-

Two questions...

1...Somewhere in the code, LV must account for the fact that the input signal has been attenuated (i.e. - if 20dB of attenuation is automatically set...then the I and Q samples have to be scaled up somehow such that the magnitude calculated is the actual magnitude into the 5660, not the magnitude into the mixer).  However...I can't seem to find anywhere in the code where this actually happens.  It might be burried in a code interface node.  What insight can you provide here?

2...What is the dynamic range of that first RF mixer? (JUST the mixer, not the entire front end).  In other words...if you happen to set the reference level far away from the actual input, such that the attenuation is set incorrectly, how much of a "wiggle room" do you have with regards to the input power to the mixer before some of the issues you mention above start to appear?  I know there are some other variable that might impact this (like RBW), but just trying to get a feel.  I have some data I'd like to send you regarding this that we're trying to make sense of that I will include in another post.  Wanted to get your thoughts first.

Thanks again.

---

Brandon

[cochenob]

Perhaps another, more simply stated way to ask my second question is, "What IS the ideal mixer level?".  My guess is that it's somewhere in the middle of the mixer's dynamic range (which happens to be ??????).

[Andy_Hinde]

HI Brandon,

1. This occurs in the driver software somewhere. This means in a driver DLL function call, or a Call Library Function node inside a subVI.

2. The PXI-5660 documentation provides some guidance on this. The documentation states that a ML of -20 dBm be used for best thermal noise performance (with the slight added risk of non-linearities) and a ML of -40 dBm be used for best linearity performance (with a slight degradation in thermal noise). A -30 dBm ML is the best compromise, and indeed the best setting in general. This is the default ML the driver uses for all RLs > -30 dBm. By using the driver defaults for ML, which means just setting RL and never setting ML or Attenuation manually, the attenuation will be off at -30 dBm RL and stay off for all RLs < -30.

RBW has no effect on this, as it is a digital filter at the end of the signal chain, whereas the attenuation and mixer are at the front of the signal chain. The onlu things pertinent to the mixer performance are your input signal level and the amount of RF attenuation applied.

My best advice is to let the driver handle everything, and not to set ML and/or attenuation directly.

Regards,

Andy Hinde
National Instruments

[cochenob]

Thanks Andy-

This is beginning to make more sense...but let me probe a bit deeper with an example.

Based on your last email...the "default" mixer level is desired to be -30dBm.  Let's assume that my expected RF input power is -20dbm, and this is what I set my RL to.  Therefore...the attenuation would be set to 10dB.

However...now say that my estimate of what the RF input power was completely wrong....for example, instead of the -20dBm I set my RL to...it was orders of magnitude higher or lower (say, 0dbm or -50dbm).  How far "off" could I be until I see problems?  As you mentioned in a previous message, there may be non-linearities in the mixer.  I also notice an amplifier at the end of the 5660, which I'm guessing, in addition to the attenuator settings, helps to optimize the IF output level such that it's in the sweet spot of the digitizer dynamic range.  However, if the initial estimate of the input power (i.e. the RL) is off...then the amplifier dynamic range could come into play as well as that of the mixers. 

I'm prying here because the signal we're looking to measure will have a huge dynamic range that we're looking to measure over time.   Given this info, we'd like to have some idea as to where to set the RL such that the dynamic range of the 5660 (and hence, our system) can be maximized.   

Is a measurement or spec of this nature available? 

Thanks again!

---

Brandon

[Andy_Hinde]

Hi Brandon,

First here is a link to the PXI-5660 specs. Pages 8-9 address dynamic range.

http://www.ni.com/pdf/manuals/371320e.pdf

Given your example in the previous post, if you have 0 dBm signal level into 10 dB of RF attenuation, you'll hit the mixer at -10 dBm. The 1 dB compression points in the specs I linked above are around 0 dBm, and you don't want to hit these. You'd have 10 dB headroom relative to the compression point, but this doesn't mean you're not adding any non-linearities.

This is where the chart on page 9 is useful. At -10 dBm ML, you'll get approximately -60dBc 2nd harmonic distorion and more importantly, -45 dBc third-oder distortion. Third order distortion is more undesirable as it is in-band.

Regards,

Andy Hinde

National Instruments

__________________________________
This is a system generated mail. Please do not reply to this email ID. If
you have a query or need any clarification you may:
(1) Call our 24-hour Customer Care or
(2) Email Us care@irctc.co.in
_____________________________________________________________________________________

Dear SHINDE SHINDE,

Welcome. We thank you for your registration at IRCTC' s Online Rail Ticket Reservation website.

Your user id is 'SHINDE143'.

You will use this user id and the password given by you for all your transaction on www.irctc.co.in. The user id cannot be changed and hence we recommend that you store this email for your future reference.

If you ever forget your password in future, you can use the 'Forgot Password' link available on the Home page of the site.

We understand that you have read and agreed to the Terms and Conditions as applicable for transactions on our site. You can now make up to 10 rail reservations (up to a maximum of 6 passengers per ticket) in one calendar month.

You can refer to any of these convenient links given below to assist you in your transactions on our site.

1. How to book tickets online? 2. Information about courier delivery. 3. Cancellation Procedure. 4. Frequently asked questions. 5. Great Complimentary Gifts Every Week .

We hope to offer you a uniquely pleasant experience in planning and booking your journeys with the Indian Railways. We look forward to having you use our services regularly.

In case you require any further assistance, please mail us at care@irctc.co.in or call us at 24*7 Hrs. Customer Support at 011-39340000 , MON - SAT(10 AM - 6 PM) 011-23344787 ,Chennai Customer Care 044 - 25300000.

Presenting Shubh Yatra Scheme for Frequent Travellers (SOFT)

A first of its kind initiative from Indian Railways, Shubh Yatra Scheme for Frequent Travellers (SOFT) is brought to you by Indian Railways (through IRCTC), to reward you for your continued patronage of Indian Railways services.

Through this revolutionary scheme you can now get back up to 10%* of your AC (except 3 AC) ticket fare as Railway Points on tickets booked at www.irctc.co.in. Use these Railway Points to purchase free railway tickets at www.irctc.co.in for yourself and your family.

To know more about Shubh Yatra log on to www.irctc.co.in

We hope to offer you a pleasant experience in planning and booking your journeys with Indian Railways, through your Shubh Yatra Card.

Also Introducing the SBI Railway Credit Card

We are also pleased to bring to you the first ever SBI Railway Credit Card. This unique credit card is brought to you by IRCTC in association with SBI Card. The SBI Railway Credit Card is the only credit card that gives you FREE* Railway Tickets to make your journey even more rewarding.

Multiply your benefits with your SBI Railway Credit Card with unmatched privileges:

- FREE* Membership to Shubh Yatra

- 0% Transaction Charge on tickets booked at www.irctc.co.in

- FREE* tickets for every rupee spent

- 0% Transaction Fee* on fuel purchases

- 0% Balance Transfer* for 60 days

- FREE Rs 10 Lac Personal Accident Insurance

Apply TODAY and get FREE 350 Bonus Railway Points!

In case you require any further assistance, please mail us at shubhyatra@irctc.co.in or call the 24 hr Shubh Yatra & SBI Railway Card Customer Helpline at 39 02 12 12 (1800 180 1295 for BSNL/MTNL users). Our customer care associates will be happy to assist you.

***************************Information******************************

For any enquiries or information regarding your transaction with IRCTC, do not

provide your credit/debit card details by any means to IRCTC. All your queries

can be replied on the basis of 10 digit IRCTC Transaction id/ PNR no./User id.

IRCTC does not store the credit/debit card information in any form during the transaction.

******************************************************************

Warm regards, Customer Care Internet Ticketing IRCTC

* Terms & Conditions Apply

[cochenob]

Thanks Andy.  This is quite useful.  I had some documentation from when it was shipped, buta it wasn't this detailed.  Not sure what I was looking at.  Didn't think to search around on the NI site for something more. 

We'll digest this.  Thanks.

---

Brandon