There are a few testers/simulators on the market (such as our Road Tube Simulator) that can be used to test the air switches in various hose counters.
If you want to use a "classification" type layout on a tester/simulator (two hoses, with a spacing between hoses entered) then you will need to be aware of how our software works...
Previously the software (versions under 1.6.x.x) would just start with the first hose hits it saw assuming those were always legitimate hits and try to make things happen. If the first hits were bogus or were the last part of a vehicle or some other such thing it could cause the whole set of data to not get properly classified (or a large chunk of it) in some circumstances. So one of the changes to improve this (which is related to heavy traffic) is that now the software will scan the data until it sees a reasonably large gap so it knows for sure that the very next hits will without question be the start of a legit vehicle. Then it will be on track and will classify everything just fine. And due to how the software "chops" the data up into days (to save memory and time on processing) it will actually do this every day starting at midnight. In any traffic, even legitimate heavy traffic, there will be a gap large enough for the software to sync up pretty quickly. The gap has to be several seconds of no traffic. Without a gap like this (which only happens in simulated traffic) our software may take a LONG time to process and then not end up with any counts because it could not find a "legit first vehicle".
With our Road Tube Simulator we added a short burst of hits several seconds before we start sending out vehicle patterns. This is just to keep our software happy so it will process the data properly when our simulator is used with a "volume" pattern" (which is very close together traffic with no gaps), our classifications patterns and such have gaps so they should work either way. Other testers may put out a vehicles in close proximity with no gaps, so it's not at all like real world data and our software will show no counts in those cases. To avoid this you can start your tester very briefly, then pause it and wait several seconds, then start it up again and let it run... you should then get your classifications as expected. Now if you let it run over the midnight boundary you'll see the issue arise again, so we don't recommend running fast-paced no-gap tests like that in classification mode for testing over 24hrs.
If you want to run longer tests I would recommend you choose a volume type setup like a median setup with a hose on each side -- then you will have no problem at all as it won't be classifying. Doing a "volume" test really the only thing that typically needs to be done unless you are making changes to classifications schemes in the software and trying to see how those test on a tester. But the classification testing is not really giving you much of anything over just a hose hit test -- if the counter is seeing the hits properly then it's all up to the software from there, the counter is working fine. So really all you need to be looking for are the two channels being in balance on the hits they see (hopefully they're right on or with in a couple of counts of each other) with choosing a median setup at time of download it makes this really easy. You can even take a quick look at a report and see that each hose is matching up each hour or day. Trying to do vehicle classification during testing doesn't do anything more for you except maybe in a very extreme case where a air switch is not responding with normal timing... but that would be extremely rare and is not at all a typical failure.
If you want more than just comparing the two channel's hits to each other, then you could adjust the output amplitude of your tester/simulator up and down for each test while running live data view and see when your channels stop seeing hits. They should stop somewhere near the same amplitude (when the pulse is too weak to register on the air pulses). If you notice one counter has a channel that requires the amplitude to be extra high compared to most other counters then it may be worth looking at further as one of the channels may be getting less sensitive either due to stuff in the nozzle or a failing sensor.
Cleaning Your Counter