EPROMs...EPROMs...yes I remember them.

Do you have an eraser? We used to use a small UV light.

______
Rob

It's almost certainly because the rise time of your input is too slow. Frankly the easiest way is in software, but you could add a schmitt trigger between the LDR and input (depending on the voltage levels you are receiving) or an analogue comparator with some hysterisis.

______
Rob

Farnell have them which means Newark does too

http://www.newark.com/tracopower/tsr-1-2450/converter-dc-dc-24v-5v-1a-sip/dp/08R5932?Ntt=trs+1-2450

But they seem to get them from Farnell UK anyway.

Traco don't have a distributor list on their site, so maybe you'll have to email them.

http://www.tracopower.com/Contact.201.0.html

______
Rob

Why Arduino uses 16Mhz not 20? For stability?

I can think of no reason, maybe they were cheaper at the time   Sometimes you choose a value because it will divide down nicely for buad rates etc and 16 does divide down to 1MHz well but I can think of no compelling reason.

I'll let someone else comment on those tutes but it all looks good to my relatively non-Arduino eye.

Re the regulator issue you started with, the cheapest and most efficient regulator is none at all, so if you use batteries as you suggested above that has to be the best option, at least from that point of view

______
Rob

Looks like you are correct, the graph actually indicates up to 11MHz @ 3v and 20 @ 4v5 with external xtals but if you use the internal osc you get 8MHz.

BTW Remember to program the CLKDIV8 fuse or you will get 1MHz.


______
Rob

Is it possible to use simple voltage divider made from two resistors?

Possible but way not advisable and would burn power anyway.

Have a look at the TRACO TSR 1-2450 switching regs, up to 94% efficient and same pinout as the 7805.

______
Rob

i now send the number [glow]666[/glow]

:o Uh oh now you're in trouble, you'll have to get both Arduinos exorcised.

______
Rob

As you're reading through optos I gather you are just interested in the existance of any voltage over the LED threshold, not the actual value of the voltage.

______
Rob

Sorry danman2991 I don't understand either.

With analogue MUXs you can easily select 1 of N in/outputs.

Is this a one-off just to ID cables so you can tag them or a recurring test of some kind?

If one-off there's other simple ways to test.

______
Rob

Nice, are they your boards?

______
Rob

Hi Boffin,

Not sure if you're asking anything or making a statement but let's up the ante to 20c.

// turn off the output so the pins don't light up
 // while you're shifting bits:
 digitalWrite(latchPin, LOW);

IMHO the output pins will not change until the latchpin goes high...
You can fill all the registers and then flip the latchpin low then high and then the LEDs will change.

True, the above comments don't make sense unless they are doing something wierd somewhere.

to delay the latch if it is driven high at the same time as the shiftreg clock

Yep the cap should not be there, it makes absolutely no sense to tie the latch and clock pins together unless you are desperately short of pins and don't mind having the data ripple through the SR outputs.

why not loop the last shiftregisters' output back to the micro, and compare the last byte sent previously with what comes back this time, if no bits are lost it should be the same...

If you have reason to believe there could be a fault in the chain that's a good idea. Not much point on a single PCB but if your SRs are on several modules connected by cables that's a good diagnostic tool.


______
Rob

[are]

First datasheet is quite useless, but second one says:

OK I found the second one and it's little better than the first. I haven't used AT commands for 20 years so unless someone else here is up to speed on them you'd better start Googling, better still ask the guys selling this to point you in the right direction.

On one sheet the chip is referred to as a GSM1, whereas the other sheet is about an SIM900, are they the same chip? The SIM900 talks about SPI, I2C etc but these atren't on the schematic. The SIM900 does however have a stack of UART_xxx pins which presumably are the handshaking lines on the first sheet.

Looking at a few of the pins they do appear to be the same.

But isn't enough to just connect RX, TX to arduino to control a serial-controlled device?

CTS/DCD/RTS/DTR etc are handshaking signals, presumably they are used or they wouldn't be there. If they are used then some of them will need to be tied low (or maybe high).

But apart from that you're right, you should just need Tx and Rx. BUT, what are you going to say and what do you expect from the device.

You need to know

a) What are the AT commands it needs, and
b) What (if any) handshaking lines need to be dealt with.

With a chip of this complexity there should be a programmer's manual of some kind. Get onto http://sim.com (I tried but the site was down) and see if they have a real data sheet.

______
Rob

There appear to quite a few serial handshaking signals that you may not need so I reckon the following...

RESET -> any spare digital pin, use to reset the board
ON/OFF -> any spare digital pin, use to power up/down the board
VCC/GND -> same on Arduino
RX -> any spare digital pin, use as your serial output (probably, depends on their naming convention)
TX -> any spare digital pin, use as your serial input (probably, depends on their naming convention)
CTS/DCD/RTS/DTR ? some may need to be tied to enable an internal UART
RI -> may be usable as an interrupt

It looks fairly straight forward but you really need a proper data sheet, even if you get the pins right what are you going to say to the thing?

______
Rob

[interrupts]

But if there is no way

There is a way, it's easy and the guys have mentioned it twice, interrupts.

In fact there's probably a 100 ways to do it, but here's one take.

You set an interrupt to trigger say on the rising edge of your LED, the interrupt service routine (ISR) records the fact that the event has happened by setting a global variable to the millis() value. Your loop() code tests the variable and compares it with the saved value from last time there was a pulse. If they are different a new pulse has occured so you subtract the new value from the old value to get the duration of the pulse and overwrite the old value with the new value.

At 50 pulses per second this would work but have pretty course resolution. How accurate do you need to be?

______
Rob

Do you know of an IC that would contain multiple 2n2222 or similar transistors?

Have a look at ULN2801/2/3/4/5, ULN2001/2/3/4.

Even better these chips don't need base resistors so you will lose 6 components.

_____
Rob