The Noodling Process is a deliberative approach to solve a problem involving a fault tree analysis. The old Italian Mechanic is merely describing a symptom when he says: "She don'ta work". That gives no clue to a solution or where to start. Thus job #1 is to describe the problem. But accurately formulating a problem statement is a result of gathering data and analyzing that data t0 arrive at something more substantive than the observation of our old Italian Mechanic.
Our somewhat real-world crass analogy: It took some time to figure out that Mary Jo, our 300-pound close friend, had to have two Bob's Big Boy hamburgers first, before she could be enticed into the back seat of a 57VW Beetle.
.jpg)
Try to image 450 pounds of raging teenage hormones in the back seat of this 57VW. Scary?
To formulate a real problem statement requires assembling known facts.
* The mainboard with the ADE-1's is a known design and works with QUISK. Other rigs using this same board design + QUISK do not have the issue.
* The MHST works on Receive using the VWS package and the receive function works perfectly with the Pi Pico and the two sound cards.
* The QUISK software has functionality to abate image signals in both transmit and receive. I was able to null out any images on the receive side, but that same internal QUISK process for transmit has no effect on the mystery signal. (Could this be a clue that it is not an image but an extraneous spur?)
* I can see the signal on the Half Moon Bay SDR receiver site which is 300 miles away from me. Thus, this is not something to be ignored as we know the 85th Harmonic of N2CQR/KK4DAS DCR project was picked up by the FCC monitoring stations in Bali and the Northwest Cape in Australia.
* The problem is only observed in transmit and since the two ADE-1's are bidirectional, wiring errors do not seem to fit into the equation. Working perfectly on Receive also tends to eliminate a bad set of ADE-1's.
* The offending signal appears stronger than what is intended as the primary signal. Using a station receiver with a 1-inch antenna located about a foot away reflects a 10dB difference with the stronger being the undesired signal. Incidentally, the frequency spacing is about 30kHz away and observed higher in frequency.
(Mental Note set up the Siglent scope for FFT and observe the pattern on transmit.)
* In looking at signal mixing we have the quadrature LO signals at 7.2 MHz and the I and Q streams which are in the audio range and limited to maybe a max of 6 kHz. So, a 5th harmonic of the I and Q streams mixed with the LO signal would give a signal substantially removed from the intended signal. But past experience would say the 2nd or 3rd harmonic of the I and Q would be more of a problem but not observed.
* This is when apples to apples comparisons are important. A fellow knowledgeable ham sent me an email this morning that I should turn off any unused clocks on the Si5351. Shazam a lightning bolt from the Radio Gods. In most of my use of the Si5351 based on some of the earliest applications, only CLK0 and CLK2 were used for the LO and BFO. It was suggested to avoid using CLK1. Initially I did note that the VWS package used CLK0 and CLK1. So aside from the mystery issue I am going to go back to the Arduino sketch and put the quadrature signals on CLK0 and CLK2 and turn CLK1 to off.
* The working HB SDR uses a Si570 with a single output a 4X which is run through a 74AC74 to produce a quadrature signal at 1X. So, this is not apples to apples. This opens the possibility of running the Pi Pico at 4X and running a single output through a 74AC74 which is already on the MHST board. This would provide another check on just where the issue is originating from.
The problem statement is now leading us to the following.
The MHST transceiver has an issue with transmit is that a spurious signal is being generated approximately 30kHz from the desired signal. The nature of the signal does not respond to abatement using the embedded tools in the QUISK software and is identified as being potentially spurious and not related to the I and Q Stream. (However, the I and Q stream may still be an issue.) The problem is the signal, and the solution is to initiate progressive steps identified from this analysis to observe the effects if only to eliminate potential issues and narrow the scope of possibilities.
The Isley Brothers said it best
Them that know can make it go.
73's
Pete N6QW