Yes, unfortunately debugging requires that.
Is your existing z-wave stick recognized by something else (like Home Assistant + OpenZWave running on a Raspberry Pi)? If it is the z-wave stick, as you suspect, its failure to be recognized by OZW would strengthen your conclusion.
Another possibility to consider - I assume you have a C-5 hub. If so, make sure the USB splitter cable is ok.
because majority of You are not using external z-wave stick? 
I think that this is really an issue but how You wrote not for majority. I'm using HE from ~3 months and I had to do power reset about 5 times because z-wave devices disappeared. I found some posts and very often this problem was described by users with EU version = external z-wave stick. For me this is not a problem with z-wave devices (maybe with drivers - I'm waiting for an official support) because before I was using RPi3+Razberry shield and everything was working properly.
Both of my hubs use the external stick
maybe there is something wrong with EU version? Don't know. I'm waiting for official drivers for my trvs...
What am I looking for?
Well locked up again! This time instead of shutting down and pulling power I disabled z-wave and pulled the stick out, non z-wave working fine, plugged it back in and re enabled z-wave. Everything working fine again.
God I dont want to go through having to repair everything to a new stick!
Well, if it is working fine now, then it eliminates this possibility. But, I would have made sure the cable is ok. No fraying, unusual voltage drops etc.
Externally looks fine not sure how to check for voltage drop
You'll need two things:
a) a microUSB female to barrel converter
b) a voltmeter
This will let you easily measure voltage at the barrel connecter if you connect it to the microusb from the splitter cable. The splitter cable will also be connected to the power supply.
You should see 5V DC.
I think I've got a usb to barrel, if I take out the z-wave stick I could check it on that side?
Edit.. with the other side still connected to hub?
Yes. Be VERY careful.
No, I would do it with the other side ONLY connected to the power supply. THere's no need to endanger the hub by putting in the middle. The goal is to see if there's a voltage drop across the splitter.
Tomorrow job, I'm sure it will lock tonight so probably need to power it down again
Please check your email, I just replied to your ticket. You have a few devices at the outer edge of your mesh, which can cause the repair to stall and hub to "freeze" while the repair is running.
I've respond those devices should be within the middle of the mesh, and the errors seem to be just after mode change
Ok strange thing, after I unplugged the stick and put it back in I've just run what would be a 2 hour repair in under 10 minutes, nothing else changed...
The physical location and proximity to the hub do not define where a device is in your mesh. I have a switch that is in the same room as the hub, within touching distance. It is the closest device I have, by measurement, to the hub That switch is actually three hops from the hub. I don't have anything further away by hop count, which would put it at the edge of my mesh but physically in the center.
How did you find out your hop count?
I think my problem was a device which was in the old ST mesh and hadn't been moved over
I have an Aeotec Zstick paired to the HE hub and managed by Open Zwave Control Panel running on a Linux computer. It can do many things but I just use it for gathering the neighbor table and hop counts. There are other solutions but I already had the stick and was familiar with OZWCP.
Other ways?
Download the Hubitat app
