ok, so I’ve been running a Pi based Telraam for over 3 years, have been helping with my local council (Parish level in England) on a medium term traffic survey using ~10 Pi based devices, & have recently put in a Telraam S2.
First impressions: Wow. It’s a big step forward. Well thought out - from packaging (shows you how to attach the mount) to cable (3m to reach from socket to window) to camera (wide field of view to suit more locations). Power consumption is around 1W - about 1/3 or less of the S1 (measured using an inline USB device as my socket testers were not sensitive enough to detect this!!) Set up is straightforward (it just works*), configuration is simple. It seems to be very reliable - much more so than the Pi based solution, albeit mostly due to form factor (no camera to knock / wi-fi to glitch). Counting seems to be improved over the S1 / Pi based device, due to a number of reasons:
- improved field of view - S1 in the same location could not ‘see’ the pavement below AND all of the road - the S2 improves this (location is right above a ~1.5m wide pavement)
- It appears to differentiate between different vehicles more reliably - fewer cases of cars with shadows being counted as bicycles (though this location is not the most challenging for that issue).
There are some UK specific points:
-
The device shipped with the standard EU PSU & cable, but also with a UK PSU - ideally the EU one would have been removed & re-used (!); the UK PSU has a fixed, shorted cable, I purchased a USB-C to USB-C cable to lengthen it (a cheap upgrade, but may be necessary). No complaints here, just observations.
-
The location I wanted to install this ‘didn’t work’. Working through this with Kris & the team it turns out the S2 uses a specific mobile telephony service - LTE-M - intended for IoT devices. This is NOT the same as LTE… Anyway, the only LTE-M provider in the UK is O2. It looks like the O2 deployment is not heavily used, and certainly is not well documented. It does not work at the mast that I needed to use (possibly not some of the nearby masts either). O2 have confirmed this, though no eta for a fix.
Speaking to Telraam, a fix was pushed out (I tried the device in other locations and got good signal) that should allow the device to also use NB-IoT (another mobile telephony service, in the UK apparently provided exclusively by Vodafone), anyway as far as I can see this also does not work… So… I tried in alternate locations (using a USB battery pack to power) & replaced a problematic S1 device. I suspect that I’m VERY unlucky, and that this is probably mostly a UK thing (NB-IoT & LTE-M seem to be at early stages in the UK - it looks like the next generation of smart meters will drive take up & reliability of these services, but not yet!). The only way I’ve been able to check for reception with a Telraam device is by using the S2… Good O2 LTE reception did not correlate with LTE-M reception. The UK LTE-M coverage map was little help (it’s now working fine in an area that their maps suggests it won’t). I’ve raised a ticket with O2 which seems to have gone no-where - they confirmed the problem, called me to suggest it was a mast capacity issue (I was really not convinced) & then closed the ticket after a month because it had been open for a month.
I really see this reception / LTE-M issue as being specific to ‘my’ location - I did not chase up NB-IoT reception but suspect with a little chasing I’d have sorted something (note that the mobile data contract is via Telraam - they did get some information, but was easier for me to chase O2 directly, as I have my own albeit separate contract with them).
-
I struggled a little with installation - how to place the device without seeing an image, but it ended up well placed based on guessing 45 degree angle down & distance from side of window. The first image took around 1.5 hours to upload (e.g. it uploaded after ~1.5 hours), which was a bit of a surprise but not really an issue. Seeing the ROI was easy. If you need to move the device you will probably have to replace the double sided tape, it does not come off without quite a fight. A small roll of 5M VHB helps here (required for S1 devices anyway). Night counting was giving weird speed / V85 statistics, but has been removed for now & so has settled down.
-
At present I’m on the non-subscription version, I do intend to upgrade to the subscriber edition to get higher granularity - which highlights another downside - the ongoing subscription. It’s good value for what it is, but adds a lot of cost for community projects. Telraam are absolutely doing their best to find a decent compromise.
-
For a professional, easy to install & reliable solution, the S2 devices are amazing - very good at what they do. One thing that this also highlights is now good the S1 devices are - their accuracy is still incredibly good for such a small, cheap device (most ‘failures’ I see with these seem to be related to small curious children, a particularly curious cat, or house visitors unplugging the device!). The suggestion from Telraam about improved accuracy of S2 over S1 feels about right, albeit I’m not able to check yet without subscribing to get that 15 minute granularity (I can’t face counting for an hour…)