Andy,
250 kHz is the magic number. Sweeps that start below this cannot use the preamplifier, which raises the noise floor.

The TG sweeps now use two sweeps to build the final result. One covers the low loss region (approx. 0 to -40 dB for a passive device, or -20 to 20 for an active device). The other covers the high loss region (approx. -40 to -80 dB or -20 to -60 dB for an active device).

When the start frequency is below 250 kHz, the software selects the active device range since it cannot use the preamplifier.

Andy,
These might be from steps in input / output impedance at those frequencies. You can remedy this by adding a 3 dB pad to the output of the TG and input of SA before “store thru” and keep them on for all measurements. This improves return loss by 6 dB and cleans up impedance mismatch-related measurement error.
Another thing you can do is “store thru” with a similar length cable to the circuit you will be testing.
If these steps occur with 30-50 dB insertion loss, you may need to “store 20 dB pad” to get rid of them.

KBONE, thank you for the bug report. I’ll see if we can duplicate it. Do you mind sharing details about your CPU and GPU so we can try to duplicate it?

Roger,
What application do you use the trace math for? We’ve been discussing how to implement it in a robust, maintainable way. Would something like (Trace 1 – Trace 2 – offset), shown on the same graticule work for you?

Ockhams, Thanks! That’s what we love to hear.
Andy, we duplicated your problem. We’ll work on it and see if we can find a fix.

FYI: This discussion continues post-release here: https://signalhound.com/support/forums/topic/spike/

Andy,
The zip file you download contains the drivers, if you don’t already have them from a previous install. You still have to install them separately as a second step. We will continue trying to simplify the process, but for now it is a two step install if you don’t have the drivers already installed.

Cliffy,
This poorly worded error code means the temperature correction data file, typically downloaded from our web server, was inaccessible.
This can be because:
1) The computer does not have an internet connection
2) The application cannot use the connection (e.g. company firewall), or
3) The file does not exist on our server.
If you email support@signalhound.com with your serial number, we can email the file to you, or you can manually copy it from the working installation folder. Copy two files from the working installation (both have your serial number in the file name), to the non-working installation.

With our new software release tomorrow, this error description will be a thing of the past.

Nice! Yeah, type N connectors definitely have some advantages…

The precision 10 MHz option is handy if you don’t have access to a GPS-disciplined 10 MHz. As long as your external 10 MHz is clean, an external reference will generally perform about as well as an internal reference, and will perform much better for close-in phase noise (think 10 Hz offset). If your 10 MHz input is not clean, it may actually degrade your system performance.

I should mention as well that for the SA124 broadband channel power measurements, you would need the signal to have a 100% duty cycle.

[Andrew]

Our new software release (Feb 2015) will enable the SA124B to make occupied bandwidth, channel power, adj. CP, harmonic distortion, as long as all signals analyzed fit within 40 MHz, and are roughly “white” when viewed as narrow-band. Our current software supports everything but occupied bandwidth.

You can easily measure devices with any bandwidth using the TG and SA. However, these would generally be swept scalar measurements (no vector information for impedance matching). We have single-frequency vector capabilities with our Smith Chart utility, but it does not have all of the VNA length correction features, etc, so you might need a calculator and a few complex multiplies to get a useful vector impedance measurement.

• This reply was modified 10 years, 3 months ago by Andrew. Reason: Changed 2014 to 2015

The current TG software is really touchy. It sounds like you are trying to make sensitive measurements. Before you enter TG mode, try setting your reference level to -30 dBm, then attenuator to 0 dB. Then enter TG mode, set reference level to 0 dB, store thru. This should keep the 0 dB attenuator setting and give you a flat response.

The new software will be so much better for this.

We are planning on adding tracking generator compatibility with the BB60C, but we have not established a timeline yet. The down side is that the speed would be limited to 300-700 points per second, about the same as the SA-series. I’ll post more information as it becomes available.

You’d need an ESD-safe work area with hot air. Be sure to get an SMA end-launch connector with the longer barrel. The internal shield will have to be removed–it is soldered down with low temperature tin-bismuth solder. And, of course, it will void any warranty you may have. If you’d like, you can send it to us for a quote instead.

I’ll look into what it would take to add sweeping for these devices. Would a 50-100 point sweep be sufficient? Would a minimum bandwidth of 150 Hz be OK? Would a minimum span of 1 kHz and maximum span of 100 kHz (or 200 kHz) work? Would a dynamic range of 80-90 dB be enough? With 80-90 dB of dynamic range (vs 100+ dB) I think I could display the updated points in real-time so you wouldn’t have to wait for the complete trace…

Andy,
We are moving to the “fast, high dynamic range” sweeps. This should work well for any filter with at least 10 kHz of bandwidth, group delay of less than 100 microseconds, and a starting frequency of > 10 kHz. It will give you about 90 dB of useable dynamic range. Do you have filters that fall outside these specs? If so, we are planning on including an enhanced TG CW utility that allows you to, for example, tune to a single frequency and adjust for maximum insertion loss. You should be able to measure well below 100 dB insertion loss this way, and manually step across your frequency of interest. Similarly, for filters with less than 10 kHz bandwidth, you can manually step across them in increments as small as 10 Hz.

[Author]

Posts