Dear List, My measurement of a 1 MHz tone is about 8 dB off of my link budget calculation. Can you help? Here's my set up: HP 8656B signal generator: 901 MHz CW at -47.3 dBm The USRP2 (with the RFX900) and the signal generator are both disciplined to the same oscillator. My estimated link budget: Cable: -2 dB SAW Filter: -2.2 dB LNA: +14.5 dB Mixer: +10 dB (I'm pretty sure I've set this correctly, based on a DC measurement at VGIN) IF Filter: -1 dB (I don't think this is correct) IF Amp: 30 dB Based on this I expected to measure 2 dBm or .8 Vpk-pk at the ADC. sqrt(10^((-47.3-2-2.2+14.5+10-1+30-30)/10)*200)*sqrt(2) = 0.7962 What I got was 0.31 Vpk-pk, which is -6.2 dBm. That's a difference of 8 dB, which seems too large to attribute to gain variation. I don't think I have the IF filter loss correct, but can it be 8 dB off? It's possible the input signal is not at -47.3 dBm. I'm using the signal generator's front panel display. I don't have a power meter. Do you think I missed something? Thanks! Brian

On 08/21/2011 07:17 PM, Brian Heilig wrote: > Dear List, > > My measurement of a 1 MHz tone is about 8 dB off of my link budget > calculation. Can you help? > > Here's my set up: > > HP 8656B signal generator: 901 MHz CW at -47.3 dBm > The USRP2 (with the RFX900) and the signal generator are both > disciplined to the same oscillator. > > My estimated link budget: > > Cable: -2 dB > SAW Filter: -2.2 dB > LNA: +14.5 dB > Mixer: +10 dB (I'm pretty sure I've set this correctly, based on a DC > measurement at VGIN) > IF Filter: -1 dB (I don't think this is correct) > IF Amp: 30 dB > > Based on this I expected to measure 2 dBm or .8 Vpk-pk at the ADC. > > sqrt(10^((-47.3-2-2.2+14.5+10-1+30-30)/10)*200)*sqrt(2) = 0.7962 > > What I got was 0.31 Vpk-pk, which is -6.2 dBm. > That's a difference of 8 dB, which seems too large to attribute to > gain variation. I don't think I have the IF filter loss correct, but > can it be 8 dB off? > It's possible the input signal is not at -47.3 dBm. I'm using the > signal generator's front panel display. I don't have a power meter. > > Do you think I missed something? Baseband Filter loss is 6dB -- you have 100 ohms in series and a 100 ohm termination. Matt

Thanks Matt. I have one more question. A 'textbook' approach to the mixer equation would show that half of the amplitude is lost after the baseband filter of the I and Q components, due to the upper sideband. This would result in an additional 6 dB loss which I'm not accounting for. However the mixer provides 10 dB of (variable) gain. Is this loss already accounted for somewhere else? I think this is probably a basic RF Engineering question, so thanks a lot for your help. Brian On Sun, Aug 21, 2011 at 11:22 PM, Matt Ettus <matt@ettus.com> wrote: > On 08/21/2011 07:17 PM, Brian Heilig wrote: > > Dear List, > > > > My measurement of a 1 MHz tone is about 8 dB off of my link budget > > calculation. Can you help? > > > > Here's my set up: > > > > HP 8656B signal generator: 901 MHz CW at -47.3 dBm > > The USRP2 (with the RFX900) and the signal generator are both > > disciplined to the same oscillator. > > > > My estimated link budget: > > > > Cable: -2 dB > > SAW Filter: -2.2 dB > > LNA: +14.5 dB > > Mixer: +10 dB (I'm pretty sure I've set this correctly, based on a DC > > measurement at VGIN) > > IF Filter: -1 dB (I don't think this is correct) > > IF Amp: 30 dB > > > > Based on this I expected to measure 2 dBm or .8 Vpk-pk at the ADC. > > > > sqrt(10^((-47.3-2-2.2+14.5+10-1+30-30)/10)*200)*sqrt(2) = 0.7962 > > > > What I got was 0.31 Vpk-pk, which is -6.2 dBm. > > That's a difference of 8 dB, which seems too large to attribute to > > gain variation. I don't think I have the IF filter loss correct, but > > can it be 8 dB off? > > It's possible the input signal is not at -47.3 dBm. I'm using the > > signal generator's front panel display. I don't have a power meter. > > > > Do you think I missed something? > > > Baseband Filter loss is 6dB -- you have 100 ohms in series and a 100 ohm > termination. > > Matt > >

On 08/23/2011 11:36 AM, Brian Heilig wrote: > Thanks Matt. I have one more question. A 'textbook' approach to the > mixer equation would show that half of the amplitude is lost after the > baseband filter of the I and Q components, due to the upper sideband. > This would result in an additional 6 dB loss which I'm not accounting > for. However the mixer provides 10 dB of (variable) gain. Is this loss > already accounted for somewhere else? > > I think this is probably a basic RF Engineering question, so thanks a > lot for your help. I'm not sure I follow you, but I think you are talking about the 6dB loss you see in normal mixers. These are active mixers and their gain/loss is accounted for in the datasheet gain numbers. Matt