RF Measurement Difference between 2945 and 2944

[Namq2948]

Hi all, 

I have a question about RF measurement using USRP-RIOs.

Of course, I know the USRP products aren't calibrated devices, but this difference is kind of huge, so I really want to know the reason.

I tested USRP-2944 and USRP-2945.  

First of all, my hardware connection is that RF generator's RF out port is connected to USRP's RF0 Rx1 terminal. The cable is typical SMA-SMA cable, and I'm using the 30 dB RF Attenuator for protecting the devices.

The RF signal is 1MHz CW signal in baseband, and the carrier frequency is 2GHz. The result is below.USRP-2944 Measurement (Reference Level : 0)

USRP-2945 Measurement (Gain : 0)

The measurement value from 2944 is quite exact compared to what I expect. But, the value from 2945 isn't. What is the reason for this difference? In the configuration, the only difference is Gain and Reference Level. To acquire the same measurement value from two different devices, I have to adjust the Gain parameter in 2945's example project. I think that there is some reason the default value "Gain" parameter of 2945 is 40. Could I check this reason? And does gain and reference level have different meanings what I don't know well?

In summary, I have three questions.

1. Why do the 2945 and 2944 RF measurements differ significantly?

2. Does gain and reference level have different meanings what I don't know well?

3. Why is the gain parameter of USRP-2945 set to 40 by default?

Thanks for your help.

[Chase.E]

I would imagine the different performance is due to the different daughter boards inside the device. While they two USRPs are very similar, they have different specs if you compare the spec sheets:

http://www.ni.com/pdf/manuals/375724b.pdf  | http://zone.ni.com/reference/en-XX/help/373380J-01/usrphelp/2944_block_diagram/

http://www.ni.com/pdf/manuals/376706b.pdf  |  http://zone.ni.com/reference/en-XX/help/373380J-01/usrphelp/2945_block_diagram/

[quote]Does gain and reference level have different meanings what I don't know well?[\quote]

Im not quite sure what you mean here, what is your understanding?

[quote]Why is the gain parameter of USRP-2945 set to 40 by default?[\quote]

Does that change depending on the example you open and USRP you set it to? Or is it just defaulted in the example?

[Namq2948]

Hi, Thanks for your reply.

I know these USRPs have the different type of daughterboard. In USRP-2944, daughterboard is UBX and In USRP-2945, daughterboard is TwinRx. Right? Now I compared the specifications as you said. In answer to your last question, the 2945's project default gain value is 40. (I set to 0 before running the project.)

The specification sheet states that the gain range of USRP-2944 is from 0 dB to 37.5 dB, and I understand that this is the reference level for the USRP Simple Streaming project. This is because there is no part that determines the receiver gain of USRP-2944. Is this understanding right? Of course, I already know that the meaing of "reference level" is originally the maximum RF power expected from a device that measures RF signals, but I think a little bit differently only about the USRP because the reference level of USRP Simple Streaming Project is set to 0 by default although most USRP's maximum input power is -15 dBm. (Please see the specification document.) So, I carefully assumed that the reference level parameter on the USRP project receiver section is the receive gain. I want to know if this understanding is right.

Here is one thing more: 2945 has a maximum input power of 10dBm, and USRP-2944 has a maximum input power of -15dBm. And I was using RF CW generator to generate -10dBm and I put a 30dB attenuator, so the expected RF signal measurement was -40dBm. USRP-2944's measurement result is quite similar to my expectation but USRP-2945's measurement isn't. Let's say a -40dBm RF signal comes into the USRP. If the receiver gain is set to 0 dB, USRP should show the measurement value similar to -40 dBm because it doesn't amplify this signal. Is it also right? (I know USRP is not a calibrated device, but this result is too different from my estimate.)

I can observe the RF measurements I expected by adjusting the receive gain to about 35dB. But I don't think this configuration makes sense. If -40dBm RF signal is entered into the device and it is amplified by 35dB, should it be -5dBm? Of course, since USRP is not an instrument not like VST or RFSA in NI Product, it should be a fairly large value even if USRP doesn't measure the exact value.

In summary, my question regarding the spec sheet is: (You can see actual measurement value from the captured picture listed on my previous post.)

[USRP-2944]

Input RF Signal Power : -40 dBm

Reference Level : 0 dBm

Actually Measurement value : -42 ~ -43 dBm (Similar to expectation)

[USRP-2945]

Input RF Signal Power : -40 dBm

Receiver Gain : 0 dB

Actually Measurement value : -71 ~ -72 dBm (Totally not similar to expectation)

I originally thought it should be calculated like this: -40 dBm (Input RF Signal Power) + 0 dB (Receiver Gain) = -40 dBm (Expected RF Measurement value by USRP-2945)

One more question about the spec sheet: Suppose that the -15 dBm RF signal, the maximum input power of the USRP-2944, is coming into this device. The specification document said the gain range is from 0 to 37.5 dB, so the expected observable value seen on the host side is -15 dBm to 22.5 dBm. Is this calculation correct? (Of course, it is mentioned that the gain range may vary depending on the frequency band or device, but let's assume that the gain can be set from 0 to 37.5 dB exactly.)

Sorry for the many questions, but I want to know.

Thanks.