Bench Programming & Testing

There are going to be those who have this idea that time spent testing a system under 'lab' conditions is not going to prove it working and will regard such time spent a waste. We fully disagree!

Now that the record is set straight let us get down to actually saving the most unbelievable amount of time by setting up the system while it is all under one roof before spreading it far and wide. If you still have a problem with this section please put the book down now and start climbing the hill to that reservoir where you forgot to program in the one obscure little command - Ah! I see you are now convinced.

The time saved comes from two angles, the first is ensuring the inputs and outputs actually match the system layout plan (as discussed under the section "Achieving the Requirements") and the second being ensuring the radio channel is going to carry all the required data.

Setting up such a test should be done in as many phases as there are components. Many have tried to do it in one phase i.e. all together and found more time wasted chasing their own tails than actually getting the system operational.

Identifying the components is relatively simple e.g. the radio telemetry and the SCADA are two components, and are the most common. Many tools (mainly software) are available to help testing at an accepted "division of responsibility" e.g. the MODBUS link between the telemetry system and the SCADA. It is far easier to manipulate a software package than it is a hardware input.

Two superb packages for testing MODBUS interfaced devices come from WinTech (www.wintech.com), one acting as a RTU master, the other as a slave. Armed with these two packages all testing of each component can be done separately (and simultaneously) without each component having to be ready for the other. Also, any dispute as to which component may be at fault is very quickly determined with the aid of such tools.

Other tools not to be discounted are the debug options within the various devices and packages themselves. However, thought must be given that these may slant towards proving the component being tested as fully working (designers are just like this!). Having independent testing tools will invariably prove or disprove the worth of the debug options offered within various pieces of equipment or software packages.

Testing of radio telemetry stations should not be done without the radios operating into a proper 'load'. Operating a radio transmitter into an open circuit puts a strain on the transmitter components which leads to premature failure. A suitable load for LPR shown here. The short non-resonant antenna portion allows some of the radio energy to radiate allowing the testing to be done over a number of metres. An alternative for LPR is 50 ohm LAN terminations but the distance over which the units can be tested is significantly reduced.

Bench testing is begun by programming all the units with the required configurations. Once done it is best to connect the radio telemetry host to a package as mentioned above and determine that each and every input is being relayed to the required output or image point. Testing analogue inputs is simply done with a resistor from the power supply. Accuracy is not being tested here as this is assumed to be correct, unless of course this has been an area of failing before. If so, change your radio telemetry supplier.

Once the radio telemetry system is fully operational and all I/O configured and proved a very important test needs to be conducted - assessing the radio channel usage. With all units powered up monitor the radio Tx and Rx lights to assess how busy the radio channel is and whether you are likely to run into trouble when you finally flick the big switch.

Once you are satisfied the telemetry system will hold its own it is time to connect it too the SCADA package, if there is one.

The exercise here is to ensure the SCADA has all the tags and registers allocated correctly. There is nothing more frustrating than trying to repair an incorrectly defined tag when the input generating the value is many kilometres away.

Only when the full bench test is complete is everything moved to site.