Arduino 555 watchdog timer

Hi John.

Agreed. The diode mangles the waveform slightly on the rising edge and it's very unlikely that's causing Steve's symptoms. There's no substantial difference between a 1N4148 and a 1N4001. Here's the 4001 pictures:

Before Diode:

After diode:

POR. The Arduino family of boards all wrap the processor up with the components needed to manage Power on Reset and the other stuff needed to make the CPU user friendly. That's why Steve's circuit is so simple.

Arduino boards include a loader in firmware. As a result an Arduino reset isn't quite a simple CPU hardware reset. The process also checks the serial line to see if the IDE has a new image to download and installs it if necessary. Only after doing that does the CPU re-start the users program.

I know exactly what you mean about the 555: lots of us have a love-hate relationship with it!

Steve will have to look at his code to confirm that the reset pulse coming out of his Arduino lasts long enough to reset the 555. As Neil points out the pulse has to be at least 12mS long. On the output side the 555 also generates a 12mS pulse and that would explain why the relay doesn't work - not enough electrons can pass in 12mS to work the coil.

I can now confirm that the circuit reliably resets a Arduino Mega. As this afternoon's domestic task turned out to be much easier than expected so I might bunk off and write the code needed to generate a software heartbeat with it.

I paid about £270 for my Siglent and prices have dropped since then. The advantage of the CML models is that they have more memory to capture extra waveform detail, but they all pretty good. If money is no object Siglent have just released a new "go-faster" range that''s even more impressive. Bit pricey though.

Cheers,

Dave

I think it must be a long time since you looked at AVRs , John.

The AVRs have programmable startup delays that you can choose to suit the oscillator type (from a bewildering array of types and speeds) and power supply rise times (up to 65ms).

They also start up with all outputs hi-z which avoids the worries of things randomly switching themselves on at switch on.

You can tell I like them

Neil

Dave and everyone else who contributed I really have hit the jackpot with the responses

I have only had chance to have a quick squint through all the posts tonight and am absolutely delighted with what you have all written. I have a Siglent SDS 1102CML to play with. Last time played with a 'scope was nearly 50 years ago. It is fair to say they have changed a little. Before I only had a multimeter and couldn't see what the circuit was doing at all until I bought the thing. That was how the 1M resistor got changed; I intended to buy an old analogue device, but decided that decent ones were too expensive to risk spending the money on if it was going to get killed in the post or pop next week

I checked the 22uF capacitor with a Peak Atlas ESR70 - comes up 23.83uF and ESR of 1.60 Ohms. According to the table provided, that is a low ESR for a 25V capacitor. That leads me to think it is ok.

Tomorrow I may get the chance to print the responses and sort them out. I'm not suspending respiration though. Today I have been helping a younger member of the Sheffield Club sort his Sweet Pea valve gear out. Last week I made him some new surface ground face valves. Today we have sharp valve events and it notches up in reverse. It is great to help people, I enjoy it a great deal personally - I really hate to be the duck egg that I am with this lark.

Thanks again to all

Steve

Hi Neill,

It is indeed a long time since we had anything to do with the Atmel range, after the scare with the glitches we detected. Changed to the Motorola GP32 and never looked back. The GP32 also goes high-z with the ports on boot, which can be used for some useful tasks! On the fuel injection system I used this to momentarily turn on the fuel pump for priming before the processor had run through its start up routine. Just made the POR duration about 1-second by selecting a larger capacitor and pull-up resistor combination and it worked out fine.

Hi Steve,

Now I'm jealous! You have the scope I wanted. And a ESR70. I wonder what other goodies? It took me a long time to realise that having a decent set of tools saves a lot of frustration and false leads. Making do with a multimeter and a slightly faulty set ot second hand test equipment held me back I'm sure. It's why I'm the sort of person who diagnoses electronic faults best when the broken bit is actually smoking.

Last night I was able to write a simple sketch to prove that an Arduino really can reset the 555. The code was harder than I expected, but it does work.

I had to use a 30mS LOW pulse to reliably discharge the capacitor. One thing I hadn't thought of is that the Arduino and 555 aren't synchronised. If the 555 has only just started to recharge the capacitor, a very short pulse from the Arduino will discharge it. On the other hand the same LOW pulse arriving when the capacitor is nearly fully charged won't fully discharge it, and the Arduino gets an unwanted reset shortly after.

This screen-shot shows a successful heartbeat followed by a reset. I pulled the heartbeat lead on the Arduino immediately after the successful heartbeat tick and then watched the 555 reset the Arduino when it didn't get another heartbeat. The blue line is the reset line and the yellow line is the capacitor.

I'm pretty sure that you were right early on when you suggested that something simple is wrong with the circuit, or perhaps with the code. It's odd that 47k works when 1M doesn't. On the code side I had to do a bit of a hack to stop the heartbeat back charging the capacitor. You can't let the heartbeat pin go HIGH after you've forced it LOW, which is what normally happens with digitalWrite(). Here's how I did it, there may be a better way!

void heartbeat()
{
// Sends a LOW pulse to discharge the 555's timing capacitor.
// Note the pin is reset to be an input at the end, ie neither high nor low.
// Done because making the pin HIGH rather than LOW charges the capacitor
// and messes up the 555's timing!!!

pinMode(HeartbeatPin,OUTPUT);
digitalWrite( HeartbeatPin, LOW ); // Start the pulse
Serial.print("Tick\n"; // only needed to show heartbeat is working
delay(30); // 30mS needed to reliably discharge 22uF via 560 ohms
pinMode(HeartbeatPin,INPUT); // End the pulse
}

I like helping people too. At the moment I've had much more help from the forum than I've returned. It's not only that my direct questions get good answers, it's the wealth of knowledge I've harvested from all the other threads. I'm pleased you found my comments useful and you sound exactly to be the kind of guy who helps solve my multitudinous mechanical issues.

By the way I found this example of fun with an SDS1102CML on the web. I wish I knew how he did it!

Cheers,

Dave

Edited By SillyOldDuffer on 23/04/2016 11:39:34