Any thoughts or general concerns?

Added by Aaron Harper about 11 years ago

This will be a very rapidly developed project. If there are issues, please speak up! Waiting an hour may mean that a good idea is not implemented or a critical change will involve hours of re-engineering.


Replies (6)

RE: Any thoughts or general concerns? - Added by Jeremy Wright about 11 years ago

I'm set up to watch this forum and will give input where I can.

RE: Any thoughts or general concerns? - Added by Aaron Harper about 11 years ago

Need assistance in setting up the Wiki navigation side bar. Thx

RE: Any thoughts or general concerns? - Added by Jeremy Wright about 11 years ago

I went ahead and added a skeleton sidebar. I've also added the "delivery/deployment" step.

RE: Any thoughts or general concerns? - Added by Aaron Harper about 11 years ago

Due to unforeseen issues (our local school district's network is having seizures), the project will be delayed. The block diagram will be updated and schematic posted on Monday, and this will complete everything necessary to generate the documentation necessary for a design review by peers and the client.

RE: Any thoughts or general concerns? - Added by Aaron Harper about 11 years ago

We are back on track. Bill of materials and preliminary budget will be complete by tonight. The circuit board layout should be completed by Monday, along with the detailed design portion.

Engineering change - Added by Aaron Harper about 11 years ago

The current range has been increased to 0-20 amps with survival to 30. This change happened in conversation with the client yesterday afternoon.

The basic design is good to far beyond that point (150 amps), but good engineering practice dictates that we should redesign the current sensors to be external boards to keep EMI to a minimum. This means that rather than the large shunt resistors being on the main board, they occupy the majority of the real estate on the sensor boards, which allows us greater space on the controller board. I am considering placing five small cable connectors on the controller board to connect to a total of four current sensors and one temperature sensor in addition to the one on the controller board. This is the maximum of the components I2C addressing, and would represent the most utility and flexibility in design. It also means that destructive events should be isolated to the specific sensor board with the rest of the system remaining not only unharmed, but also running.

Specific changes that will need to be made and values checked include changing the resistance and power capacity of the shunt resistor, checking the series limiting resistor on the zener diode (over-voltage protection), adding 2 resistors and a 2 position dip switch for address selection of the current sensor, creating a simple temp measurement circuit, moving components which are specific to current sensing to their own board, and specifying the module connector and cable system. It sounds difficult but will be done in a few hours. Also, by moving the sensors to an outboard PC board, we lose a few capabilities. The first is the ability to draw from power parasitically, and the second is the fault condition interrupts.

The ability to draw controller power parasitically though the monitored circuits is a convenience feature which was not specifically requested by the client. This feature may be manually achieved by connecting the power directly to the monitored circuit which may be anywhere from 6-30VDC. The reason this can't be used is that the remote sensor modules will all send their power through the thin control circuit cable to the controller, but only the selected input will power it. This means that the controller will be powered by a 24-28ga wire, and while this is possible, it would be better from an isolation standpoint to draw power from only one source selected at the time of installation.

The fault condition interrupts would require an additional connection, making the number of conductors five. Five pin wires are fairly easy to find, but submini 5 pin connectors are not. This would disproportionally raise the cost of the project well beyond what benefit it would bring. The control board may be programmed to identify a fault condition, and it could even be made adjustable by the user. Doing so in software is only a couple of clock cycles slower than the interrupt method, and according to calculations, the difference would be about 187.5nS which is significantly faster than the refresh cycle of the web information query at 250mS. The bottom line is that it would make no real difference in the readings, problem identification, or resolution.

(1-6/6)