using max7219 for seperate switching

[isabella]

hi brian,

I bought myself the WiZ-C software and have been using it for a while now, it works great! Can you send me that sample code for the shift register application we've been discussing?

thanks

 

[betwixt]

Hi Isabella.

It's been two months since you wrote so I thought you had solved it by yourself!

Before I go into software routines, are you still using the design in "back panel.pdf", the one with 74LS164 and 165 ICs?

Brian.

 

[isabella]

I got caught up in other projects.
yes, im using the same design as before.

thanks for the help.

isabella

 

[betwixt]

Sorry for not replying sooner, I'm trying to install a 70 square metre solar panel in terrible weather conditions so the work is going very slowly. We have had very heavy rain here for several weeks and much of the area is flooded.

Before lookng at code, can I ask some questions:
1. Is it safe to assume that all possible fault conditions can be detected by sequentially 'walking' a voltage across each pin at the source end?
2. Assuming a fault is detected at the destination end, how much information has to be presented to the person performing the test. For example, do they need to know every fault detected in a full test or should it stop at the first failure?
3. Is it possible for a pin to be wrongly wired so it goes back to the same end it started from instead of the other end?

Brian.

 

[isabella]

oh no! that sounds terrible! I hope your luck with the weather changes soon.
Allright, let me answer your questions:

1. All fault conditions can be detected, simply by stepping the voltage trough all of the pins, and then determining if the correct pin is on and if there are any other pins that are on when its not suppose to be.
2. My idea was to test each "group" of pins individualy, ex. if you press the button for sequence 1, it starts to test the group. and so on (i hope that it make sense) the after each sequence all the fault in that sequence will be displayed.
3. the pins on the plug is wired to the pins on the board, sometimes 1 plugpin is connected to more than 1 boardpin, but there will never be a plugpin connected to a plugpin or a boardpin connected to a boardpin. in case of a short we can determine that by checking which pin is on when it is not suppose to be.

i hope that answers your questions.

thanks again!

isabella

 

[betwixt]

Sorry for the late reply again - I'm still struggling with works here!

The idea I have is to connect the 'sending' shift registers so they cascade, in other words, the bit pattern on one is shifted along into the next one in line. To test the wiring, the shift registers first have to be cleared, either by electrical reset or by clocking enough '0' in to them that the whole chain has to be full of them. Then a pin counter is set to a value of 1 and a single high is clocked into the first shift register. The input to the shift register is then set low again. Next, the counter is incremented and another clock pulse is sent to the shift register, this continues until all the pins have had the '1' walked across them.
So at any time the counter holds the same number as the pin being driven high.

At the other end, after each time the sending end has placed a '1' on one of the wires, the entire shift register is read by clocking it once for each pin. The resulting incoming bit stream is compared with the expected pattern to check for missing or extra signals. Knowing the count at the sending end and the number of clock pulses at the receiving end when the error was detected, lets the PIC caculate where and what the wiring error is.

The expected patterns can be stored in a table with one bit per pin so for your 150 pin system it would only need 150/8 = 19 bytes per pin of storage or 2850 bytes in total..

How the result is reported would be the next task. I'm not sure if this is connected to a PC, in which case a serial link would be the best method or if it is to have a built in display. Both are fairly easy to implement. I'm visualizing something saying "Error: Driving pin 12, signal found on pin 12 and 13" if there was a short between the two pins. Is that acceptable?

Brian.

 

[isabella]

Brian, that exactly what I had in mind...I think the fault must be displayed on a lcd, because the "tester" is going be portable. Let's take the following as an example to work with
:
Plugpin 1 to boardpin 6
Plugpin 2 to boardpin 5
Plugpin 3 to boardpin 1 and 4
Plugpin 4 to boardpin 2 and 3


Thanks brian

 

[isabella]

Hi Brian,
How is your project going? I hope well.
I know you are very busy, but have you by any chance had time to look at my project idea?

Thanks

Isabella

 

[betwixt]

It's progressing but painfully slowly!

As this will not be of interest to others on Edaboard because of it's very specific software, can we move to direct email to continue please. It will also let me communicate while I'm on the move rather than when I'm at my desk. If you are OK with this, click on my name at the left of this message and look at my profile. It lists my own web site (one of many!), then visit that site and on the home page above my photograph you will see a link to contact me. It doesn't show my email address but it will reach me, give me your email address and I'll get in touch.

Apologies to other readers - I normally like to share information but in this instance, several files have to be transferred and they won't be of use to anyone else.

Brian.