rush job compromises...

Projects General Guidance
1

For my sins, I have to supply 6 telemetry units with 15 digit displays tomorrow.
The whole rush job has been compromised all the way by suppliers suddenly not being able to supply, wrong size holes for 15000 LED legs that had to be redrilled, etc etc

For my simple telemetry in the scoreboards that I make, I have always used the simplex Radiometrix 433 MHz wireless modules, which are brilliant, I get over 250m solid range with them ( OK, that said , I have perfect line of sight , and the receiver has an antenna several metres off the ground, in the middle of a field with no interference - which is one of the reasons I don't multiplex the LEDs - but that's another story )
I normally use VirtualWire which works 100%, but :-

The agent for Radiometrix here is useless, and let me down again, and again, I even tried switching to the TI 2.4GHz modules from them, which they did supply a sample of, but then couldn't supply in the time they promised. ( after I had done all the experimenting and pcb layout for )

I wasted a few weeks of timeline due to them, so last week, I switched to the promising looking SIM20 transceivers from another supplier, only about US$ 16, which look brilliant, but the technical literature is hopeless !

So as a compromise, I tried using a pair in default mode, as a simple TX / RX with NewSoftSerial, and sending the first bit as an identifying address ( which I call a PIN so the user can understand ) and if that matches, the next 15 bytes are decoded as data for a display ( of 15 LED 7 segs )

I am feeding the Rx ( from the SIM20 - running at 3v3 Vcc - straight into the RX pin of the NSS ( another compromise )

The prototype with 4 digits worked perfectly, I walked up the road with the remote, and ran out of road before I ran out of range !

The display was solid, so I thought I would hook these transceivers in with short hookup wire in this batch, and after a weekend of assembling, all works fine except that with all 15 digits , the last digit ( the first one sent ) flickers between whatever the remote is set to , and zero, randomly a couple of times a second.

I only have a one byte " PIN " number, so I thought it could be that number randomly turning up in the receivers " no signal" hash ? But the other displayed numbers stay solidly locked... ( and I have set the baudrate to 9600 which is much slower anyway )

I could McGyver a couple more address bytes, or send a phoney first byte and ignore it on the Rx end - the second digit sent is solid .

I would like to get it more kosher , even if I only have 18 hours :slight_smile:

For another ( rush ) project , that thankfully went out this morning, I was recommended by one of the forum members to use "size of buffer" , which is also used in the VirtualWire buffer decoding, but I am not sure how to implement it.

I need to send 16 bytes ( one PIN and 15 numbers ) via NSSerial, and dig out those bytes without resorting to my bodges.

For the next batch I want to get to know the serious modes that the SIM20 can do, if I get a chance..

2

I have to supply 6 telemetry units with 15 digit displays tomorrow.

At least you're not sitting next to the phone waiting for a customer to ring :slight_smile:

Rob

3

At least you're not sitting next to the phone waiting for a customer to ring

AAghhh , I wish the customers ( thank you, thank you ) could turn up one at a time when I am ready ?

I do have an order for a scoreboard upgrade kit ( where I supply a set of LED "tiles" for the school/club to screw onto an old manual scoreboard ) the tiles are connected by ribbon cable, so they can either drill a small hole behind each tile on an existing board , or even screw the tiles to a wall, and paint over the ribom cable.

Apart from the financial anxiety, it would be nice to have a couple of days break, to catch up on the SIM20 wireless modules modes ( the application docs are pathetic, but I like the 1500m range )

4

Customers are like buses, none for hours then 3 at once :slight_smile:

Rob

5

Too true Rob,

Heres the bodge I had to do to make the SIM20 modules fit to the pcb I had made for the TI modules,

and after all the rush to get everything out by Wednesday, the customer said Friday is fine ! heres the displays ready to be packed :-

Every led hole had to be redrilled ( 31000 holes ).

I solved the problem of the 15th digit flickering between 0 and the right number - heaven knows what caused it, but as it is actually the first data out to the shift reg, I simply stuck an dummy digit in front of the data, which simply spills out on the floor at the end of the 15 registers :slight_smile:

6

Heres the bodge I had to do to make the SIM20 modules fit

Nice, if you hadn't pointed it out you'd never know :slight_smile:

Rob

7

Boffin's Electronic Scoreboards - Now with 10% more copper!

8

Now with 10% more copper!

seems such a waste, but then again this batch was full of hassles - now I am ready for the next ! ( I hope )