Presence of a spike below my TX frequency on USB

[JJ4SDR]

While on 20 meters today, a FLEX user reported seeing a consistent spike below my TX frequency. I believe it has nothing to do with Thetis, but rather the convoluted way of processing audio . For grins (I thought that it would make no difference and it did not!), I also engaged Pure Signal with the same results. Not that anyone can necessarily HEAR that spike as it is fairly low in amplitude, but I really would want to understand why that spike appears! I get many great unsolicited reports on the quality of my audio, but I guess, I am wanting to make the TX bandwidth perfect !

Note, the difference (the "spike" is MUCH lower when I TX 4.45k wide compared to TX at 3.3k wide) in the amplitude of the spike. Note that I only use CFC inside Thetis which works great. Main audio processing takes place outside Thetis (see below).

The main processing takes place in a 1073 Rupert Neve Mic Amplifier/EQ clone (Heritage Audio Elite Series, HA73 EQ) that comes in 1RU (19") and a Digital Audio Workstation (DAW). The only difference between the 2 TX bandwidths is that I use a different chain of VST plug-ins inside the DAW. When DAW is not in the audio path (chain #3 below), I use the processing available in the UI-24R unit, thus bypassing the HA-73EQ completely in which case the audio chain is a little shorter. I tried that as well, but no changes and I can still see the spike (only when transmitting 3.3k wide) below my TX frequency.

Below, see the audio path(s) for all of my TX chains (all 3 of these paths share the spike below my TX frequency, when transmitting 3.3k wide only):

1. Condenser Mic -> HA-73EQ -> Soundcraft UI-24R (Digital Mixer) -> DAW -> Thetis (CFC)
2. Condenser Mic -> Soundcraft UI-24R -> DAW -> Thetis (CFC)
3. Condenser Mic -> Soundcraft UI-24R -> Thetis (CFC)

Awaiting words of wisdom!

Cheers!

Juha
NI2M

TX at4.45k wide.jpg (372.9 KiB) Viewed 326 times

TX at 3.3k wide.jpg (358.73 KiB) Viewed 326 times

[w-u-2-o]

There's not anywhere near enough information here to figure out what's going on.

This energy falls outside the selected transmit passband. Therefore it seems likely that it is occurring either in Thetis, or after the DAC in the RF chain.

A few more measurements would help narrow down what's going on and where.

1. Can you identify a test tone or tone pair that reliably causes the issue? If so, then that will greatly help to make the testing easier and consistent.

2. What does the audio spectrum look like prior to being sent to Thetis? You could put a spectrum analyzer plug-in into your DAW configuration to look at this (and then stick with this config. moving forward). If the spur is there then it is happening before Thetis (but I doubt that this is the case).

3. What does the non-DUP spectral display in Thetis look like compared to DUP? The fnon-DUP display is an FFT of the DAC data stream. If the spur is there then it's happening before the DAC.

4. What DSP filter type are you using in Thetis? Linear? Min-phase?

[JJ4SDR]

There's not anywhere near enough information here to figure out what's going on.

This energy falls outside the selected transmit passband. Therefore it seems likely that it is occurring either in Thetis, or after the DAC in the RF chain.

A few more measurements would help narrow down what's going on and where.

1. Can you identify a test tone or tone pair that reliably causes the issue? If so, then that will greatly help to make the testing easier and consistent.
Need to do this later. I think I will be able to swing it.

2. What does the audio spectrum look like prior to being sent to Thetis? You could put a spectrum analyzer plug-in into your DAW configuration to look at this (and then stick with this config. moving forward). If the spur is there then it is happening before Thetis (but I doubt that this is the case).
See images of spectra (both TX bandwidths). Peak-hold engaged as well. I don't see a spur in either one like the one I see in the Thetis Panadapter.

Audio Spectrum at 4.45k wide Pre-Thetis.jpg (209.81 KiB) Viewed 288 times

Audio Spectrum at 3.3k wide Pre-Thetis.jpg (234.62 KiB) Viewed 288 times

3. What does the non-DUP spectral display in Thetis look like compared to DUP? The fnon-DUP display is an FFT of the DAC data stream. If the spur is there then it's happening before the DAC.
See spectral display (after my signature) in Panadapter for both non-DUP and DUP for 3.3k TX bandwidth.
NOTE: I changed the filter sizes to the same I use for the 4.45k TX bandwidth and the spur got reduced dramatically or at least it got reduced significantly.
I guess some of my settings are not correct as the "amplitude" of the TX between the DUP and non-DUP spectra are very different. The DSP Option form below reflects numbers when the spur is there!





4. What DSP filter type are you using in Thetis? Linear? Min-phase?

Somewhat surprisingly, looking at the DSP Options form, when I change the Filter Size to the same that I use when TX bandwidth is 4.45k the spur appears to get very much reduced!! NOTE: the RX and TX filter sizes are dissimilar when using 4.45k bandwidth! I thought that, ideally, they need to be the same. When I change the filter size for 3.3k bandwidth to reflect the numbers for the 4.45k bandwidth, the spur appears to get reduced and almost completely eliminated.




Cheers!
Juha
NI2M

Non DUP TX 3.3k wide.jpg (399.54 KiB) Viewed 288 times

DUP TX 3.3k wide.jpg (354.99 KiB) Viewed 288 times

DSP Option at 3.3k wide.jpg (100.26 KiB) Viewed 288 times

[w-u-2-o]

Not seeing any energy on the opposite sideband in any of your tests.

This implies it is happening after the DAC. Hence further testing will want to look at DUP display. We can also rule out it happening in your receiver, since both you and another station were able to notice it.

Are you running an amp? What happens if you run barefoot?

What happens into a dummy load?

[JJ4SDR]

Hi Scott,

Sure value your consistent feedback and support during this experimentation !

Mark with whom you communicated a lot concerning the different options on the DSP form in the past also provided some very good information. It appears that increasing the filter size under Options further contributed to the spur getting less pronounced and less energy spilling over into the opposite side-band. Bouncing between Low Latency and Liner Phase didn't bring much relief though, nor changing from BH-7 to BH-4, but that may be due to the fact that I am using an old Thetis version. So, filter size appears to have the most positive impact as of now.

Whether I run an amp or not does not make any difference whatsoever. Furthermore, running into the dummy load makes no difference either.

Cheers!
Juha
NI2M

P.S. When I have more time, I may upgrade to a later Thetis version, assuming that it will not call for any hardware changes in terms of how I currently process my audio.