I am using Powercab and Bowser F7A with ESU Loksound Select decoder factory installed.
In DecoderPro when I ask for a decoder selection by reading the decoder the response is to allow selection of several ESU decoder types eg Loksound V4 (all subsets) or Loksound Roco decoders, but not Loksound Select versions.
As the Boweser material advises it is a Loksound Select decoder I have overidden the selection of a decoder and used Loksiound Select.
In reading the decoder CVs DecoderPro reads many of the indexed CVs and then writes to unindexed SI? CVs.
Errors in reading include reading CV1 as 3 when I have changed it to four digit address 922 (don't know the bit value).
Also the functional mapping is all over the place eg F0 is headlight operation but the function mapping suggests it is a fan operation.
Generally I think the ESU decoder reading and writing is problematic.
I have several other brands of decoder and have little trouble reading them correctly.
ESU provides no CVs we can read to determine the decoder type or the sound project loaded into it. All we can determine is the manufacturer. There is nothing we can do about that.
That is the correct and only way to do it.
That is correct behaviour. In order to read an indexed CV JMRI (or your Power Cab) has to write values into the unindexed PI & SI CVs, then read the value in the indexed CV. That applies to indexed CVs from other manufacturers, not just ESU.
It is not possible to store a value greater than 255 in any single CV. When you program a long address, the address is stored in CVs 17 & 18 and the value of CV 1 is unchanged. Bit 5 of CV29 is set to tell the decoder to ignore the value in CV 1 and use the values in CVs 17 & 18 instead.
The original function mapping pane for ESU V4/Select was quite incorrect and was a significant job to fix. It is now fixed in JMRI Version 3.7.1, just released.
If you are using JMRI with ESU V4/Select decoders, you MUST use JMRI Version 3.7.1 or later.
Dave Heap