One thing that has impressed me about the collapse OS radio concept is an aggressive sense of adaptability in the implementation. I gather its hard to think in terms of parts lists for a radio designed to accept any one of several parts for various key components in the platform.
This highly flexible approach to component sourcing has got me thinking; for instance it occurs to me that under this approach one could
extend this loose definition of required parts to entire functioning radios.
The usdx + springs to mind. It is built on the same design pattern that is inherited by the the Collapse OS radio.
Would it be possible to pare down the feature set of the existing usdx + sufficiently to add code back in that implements scamp?
It has a lot of code dedicated to CW; that code, for instance could be removed were the radio to be intended to be an exclusively scamp device.
There are a lot of such radios around, including the (tru)sdx.
I have both, and have successfully reimaged the firmware on the usdx+ several times with an arduino uno and a home made cable. I need to do similar with the (tru)sdx, though I am reluctant as I fear getting support for CAT and virtual audio over a single USB C cable might cost me the SWR functionality.
I digress :)
What do y’all think? could the usdx + run scamp?
@KI5SMN @k2za@mastodon.hams.social asked on @OpenResearchIns Slack and will bring comments back here.
@KI5SMN @k2za@mastodon.hams.social @OpenResearchIns
Dr. Marks writes:
I've looked plenty at the uSDX source code and it is quite possible to add something like #SCAMP to the #uSDX. The SCAMP demodulator would have filters for both in phase and quadrature phase and then select the sideband by adding or subtracting the two filters. It might be easier to port the #RFBitBanger source code to the uSDX hardware than adapt the uSDX source code to SCAMP...
@KI5SMN @k2za@mastodon.hams.social @OpenResearchIns
The major difference between the #uSDX and the ##FST3253 or similar commutating quadrature demodulator rather than a single phase demodulator. I decided not to use such a commutating demodulator because it requires a high speed analog switcher which is not necessarily easy to find in readily available parts...
@KI5SMN @k2za@mastodon.hams.social @OpenResearchIns
But if the #RFBitBanger had a commutating mixer then the method that the #uSDX uses to demodulate #SSB, which is the phasing method using the Hilbert transform, could be used. The main problem with the #uSDX I see is that the noise level from this method without any sort of manual or automatic #gain control seems to be quite noisy (S5/S6 noise floor). ...
@KI5SMN @k2za@mastodon.hams.social @OpenResearchIns
Getting matched variable #gain on in-phase and quadrature might be difficult. I think having a lower noise floor with a controllable gain #DSB #receiver is better than a #SSB receiver with only crude gain control.
Microcontrollers have only 10 or 12 bit ADC and so if one wants a good quality digital filter one has to ensure the dynamic gain is fully utilized so that a given signal does not only modulate only the bottom few bits of the signal. /fin
@abraxas3d @k2za @OpenResearchIns
Interesting, and at least a few points above my current reading level ;)
That said, I'm by no means against a complete adaptation of rfbitbanger to the usdx; if we're in a critical situation, and a scamp radio can improve our chances, then it is safe to assume (in my estimation), that the functioning usdx+ approximates 'good junk' until repurposed accordingly.