Forum Replies Created
- Posts
Mehdi,
Signals present when the input is terminated are “residual responses” or sometimes “residual spurious”.
The specification for the BB60D is -120 dBm. This is one of the strengths of the BB60D.
The specification for the SP145 is -103 dBm, similar to the BB60C.
This is not directly related to the preselector, and is more closely related to filtering and isolation differences in the architectures.Roman,
We have some customers who use the PPS from an external GPS as the trigger input. This marks the sample in an I/Q stream at the top of the second. The time stamping requires a little more work on your part, as you would need to interface the GPS yourself and generate the time stamp based on trigger position.Mehdi,
The standard approach is a high pass (often with an attenuator) for checking harmonics without a preselector. There is still a bank of low pass filters before the mixer, but not before gain control, so you end up with harmonics that may be higher than they actually are. I think you can expect -40 dBc or better, but if you need 60 dB a high pass filter would be needed. Something like this would work for 2.4 GHz harmonic testing:
https://www.minicircuits.com/WebStore/dashboard.html?model=ZHFG-K4000%2BMehdi,
That’s a good summary. A preselector really helps eliminate spurious responses from out-of-band signals, but adds noise figure.
Our products with preselectors (BB60D and SM200) are ideal for spectrum monitoring and harmonics measurements.
Our low-IF architecture products (SM200 and SP145) have the best dynamic range and flattest IF, making them ideal for modulation analysis.
For many applications, preselection can be accomplished with an external band pass filter. Mini Circuits is a good source for these.Each product line is a bit different. But for all of them we have a “spur reject” mode (under settings) where two different LO frequency results are masked together to eliminate most spurs. For measuring CW signals this is a good approach, but you lose the real-time nature of measurements in this mode.
The BB60D, being a double conversion superhet, will probably have the most consistent behavior across frequency.
Also keep in mind the BB60D spur shown is -77 dBm. Your bottom picture has a spur of -78 dBm, you just don’t notice it because of the noise.
Edit: I should add that all of our spectrum analyzers, in real-time mode, use a larger set of LO frequencies than sweep mode.Mehdi,
I think this is a spur from the real-time mode. Sweep mode lets us choose optimal LO frequencies, but real-time mode does not. This can result in additional low-level spurs, like this -77 dBc spur you found.
If you suspect a spur in real-time mode, set your span to, say 20 MHz, and step your center frequency by 5 MHz. This will configure the LO for a new frequency and result in a different set of potential spurs. If the spur doesn’t move, it is much more likely to be real than if it does.Jay,
Our Audio Demod tool in Spike would be able to give you the exact frequency of the audio pilot tone, but unfortunately it clamps to 20 kHz. In normal swept mode your accuracy will be limited to your RBW. Marker peak, delta, next peak would give you the audio frequency +/- your RBW. An RBW of 1 Hz would give you a 1 Hz accuracy. If this is not accurate enough, the next step would be to either put in a request for us to increase the maximum audio frequency in the Audio Demod mode, or use the API and do the FM demod yourself.
The peak deviation can be viewed in zero span mode FM, as you did. Measuring a local max, then a local min, with markers, should give you a peak-to-peak deviation.I believe the SA124B uses a digital correction for its OCXO frequency (it is not a voltage-controlled OCXO like the SM200). For zero frequency error, put both the SA124B and the VSG on a common external 10 MHz reference.
If you selected an external reference on the VSG, and generated a 1 GHz signal, I would expect the signal to be off by more than 10 kHz if the external reference was not found.APetrovic,
We do not have any public plans to go to higher frequencies at this point. But we would love to hear about your application / use case if you’d be willing to share details. Email justin (at) signalhound.com
Thanks!Tim,
I believe the trigger resolution is 122.88 MHz, or roughly 8.14 ns. This is 1/2 sample resolution at the highest data rate.Andrew,
The BB60D was a complete overhaul of the C. Every part was selected to optimize dynamic range for the power budget. The first thing people notice is that full sensitivity is reached with a -30 dBm reference level instead of a -50 dBm, so signals have to be 20 dB stronger to overload. The second is with preselection, signals at 1/2x or 2x your signal of interest are filtered before any amp or mixer.
Being that close, the jammer won’t be filtered out by the preselector, but the extra 20 dB before overload will make a big difference.
It looks like the preselector band breaks didn’t make it into the manual. They are at 130, 210, 310, 450, 650, 950, 1350, 1950, 3010 and 4210 MHz.I suppose if the problem is locating an interfering transmitter, a directional antenna and walking around your job site might do the trick. You can get a log periodic directional antenna for $70 or so from Digi Key. https://www.digikey.com/en/products/detail/te-connectivity-linx/ANT-DB1-LP-RM-01-N/2402472
Ethan,
This is a tricky problem. The first solution that springs to mind is that we are developing a phase coherent multi-channel system. This would be how this problem is generally solved (e.g. MUSIC algorithm, etc).
Short of that, if you had two systems, and arranged the antennas such that the signals arrived at different amplitude ratios or with different phase offsets, a little clever math could separate the signals.The other thing that can invalidate noise measurements is having spur reject mode on. And, of course any averaging done in log units (e.g. averaging points as dBm instead of mW).
Can you confirm that when you “averaged” the sweeps it was in power units, not log units?Quadsat,
Can you please confirm that your trace type is Clear / Write? The log averaging when trace averaging is active will invalidate noise measurements.About 10 years ago, we offered raw IF output. The challenge is that without corrections, the data is not very useful for most tasks. And those corrections are built from real-time temperature data. We don’t have any IP for generating these after the fact. It ends up being a pretty big filter to compensate for the IF SAW filter ripple.
Long story short, we recommend using a CPU with a bit more horsepower, rather than trying to get around the real-time processing.
You could also look at the SP145, which has a much lower CPU load when streaming.
I can’t provide schematics, and opening the case will void the warranty. But there is a 10 MHz TCXO on the board. We have a software correction for its measured frequency though, which might make it challenging to do what you want.
Matt,
Unfortunately we do not have a STEP file for our SA-series. The closest would be the BB60C, but the SA44 is 1 inch shorter, uses a USB 2.0 type B, and has a slightly lower SMA output.
https://signalhound.com/wp-content/uploads/2024/02/5146-9941_BB60C_defeature_artwork.STEP_.zip