I have a brand new ESU Loksound 5.0 decoder. It has an 8-pin connector. I took it out of the box, wired it up to the speaker, and plugged it in to a DCC ready Rivarossi H-8. I put it on the test track and it worked perfectly on address 03, just as you'd expect. The volume was low, and the starting voltage a slightly high, but these are things that can easily be fixed with a little programming.

OK, so I reconfigured the DCC system (a Pi-Sprog) to operate as a dedicated programming track with JMRI. I launched JMRI (it's a pretty current version - I think the production version before 4.18). This config has been running reliably since installation in November and has programmed a couple of other ESU 4.0s and one 5.0, although that one was in a friend's factory-installed loco. It's got my 12 other DCC locos done properly.

I created a new roster entry and had it probe the decoder type, which indeed turned up Loksound 5.0. I proceeded to make a couple of very basic changes - long address 1629, 128 speed steps and I changed the starting voltage a bit.

The trouble came when I wrote the changed pages. After a few seconds, it timed out or stopped responding. I forget precisely what it said, but it was something to that effect.

Now the decoder does not respond at all. It's not on address 03, and it's definitely not on 1629. I obviously cannot do the standard reset of setting cv8=8 since there essentially is no address.

I sent an email to ESU USA but they haven't responded in two weeks, and they don't answer the phone. I presume they're affected by the virus.

How can I reset this decoder?