Can auto gain/attenuation/pre-amp settings change during Spike recording?

[andrewclegg]

I am seeing what looks like discontinuities in DANL during a Spike recording session with a BB60C. I’ve attached an example showing a seven-day capture, and it appears the DANL goes up around the 10/31/2023 08:00 mark, stays higher until around 11/5/2023 20:00 mark, and then goes back up again just before 11/6/2023 00:00. (Apologies if the waterfall is slightly fuzzy, I had to compress it down to get under the file size limit).

The discontinuities I’m referring to are where the background level changes from dark blue to lighter blue, or vice-versa. I’d estimate about a 10 dB change in baseline (DANL) level between the two.

I used the auto setting for gain/attenuation/pre-amp. My hypothesis is that Spike is changing one or more of these settings during capture for some reason, even though only one set of values is recorded in the .shr header. Is that hypothesis possibly correct?

FWIW, the data capture method is max hold over one-minute intervals (11,309 sweeps shown here = 11,309 minutes = 7.85 days). There are very few ADC overloads during the capture (a few dozen sweeps near mid-day 11/1; not near in time to when I see any discontinuities in the DANL).

I do not have any strong nearby out-of-band transmitters that I know of. These particular data were acquired in a rural area. The nearest strong transmitter is a 162 MHz NOAA weather station several miles away, which is typically about -40 dBm at this location.

The BB60C is indoors (the antenna is outdoors, about 25 ft cable run away), so there should not be any sudden dramatic swings in temperature, humidity, etc. I am using a minicircuits pre-amp (approx 10 dB gain) near the antenna feedpoint, and that is outdoors in a weatherized enclosure.

I’m currently recapturing using manual settings as an experiment but it will take a few more days to complete.

Any thoughts on what I’m seeing would be appreciated.

Thanks.

• This topic was modified 10 months ago by andrewclegg.
• This topic was modified 10 months ago by andrewclegg.

Attachments:

You must be logged in to view attached files.

The gain/atten settings getting changed is very possible and likely the cause. Based on your other responses, it’s safe to assume this unit is exposed to the elements? (at least temperature variation) The Spike software will auto reconfigure the device when it experiences a 2C shift in internal temperature. That shift might necessitate different temperature corrections as well as a different gain/atten to maintain optimal measurements. If either the gain or atten changed due to that automatic update, it would result in ~5dB shift in the noise floor. The signal power of an input signal should not change (other than correcting any small amount of error that might have been present with the old temperature correction).

At room temp, I would not expect this auto reconfigure to occur once the unit has stabilized in temperature.

I know you are on an older version of Spike, but newer versions do allow you to disable the auto recal in the preferences menu. If you do this, you subject yourself to measurement errors if the devices drifts too far in temperature. You can also force the gain/atten, which would still perform the recalibration, but might not update the gain/atten to eek out more dynamic range when it could, or may not provide enough room for a full scale input. Both of these would be minor repercussions.

Thanks Andrew. This particular installation is not using an old version of Spike like my other one, so besides trying out the manual settings for gain/attenuation/pre-amp, I will also experiment with disabling the auto recal, just for good measure. This BB60C is indoors, but it is in a cabinet with a laptop, so I can imagine there may be temperature swings more than 2 deg C under some circumstances.

My next task was to see if the measured signal strengths are changing during the discontinuities, but as you note, they shouldn’t be, just the noise level.

Thanks for confirming that I may be on the right track.

If the signal level changed by a non-trivial amount, I would consider that a bug and we could investigate further. It’s hard to tell from your picture since a lot of the signals are broad band. If you were ever able to inject a simple known signal like a CW for a test run, that would certainly make it easy to determine.

[Author]

Posts