To Switch + Or - Supply?

If it has other connections to the outside world, they are probably referenced to the (-) supply and so switching the (-) supply may cause other weirdness. Probably best to switch the (+) side.

Cheers, yeah, switching the + was what I was expecting but couldn't really see a good reason why.

Switching the gnd allows me to use an existing mosfet and save inventory, not that it matters much, but sort of has an elegant simplicity. In fact that's what I tried first as I had the part (soldering surface mount fets by hand... grrrr ftzzz hisss). Got some P types coming in today to try.

'Other weirdness', sounds sort of fun
Del

If you switched the (-) side and were it disconnected, allowing all the connections on the RFID side to float to (+), how would those connections behave? Like forward-biased diodes to whatever loads they're connected to externally? What sort of loads are they, and what would be the consequences of their all suddenly being asserted to (+)? Basically, what kind of 'circuit' could unwitting result from disconnecting the RFID's (-) supply pin? It all depends on those connections and how they're implemented internally.

Switching the (+) allows the RFID 'float' down to (-), and so, for input connections what sort of 'circuit' could unwittingly result? You don't want to create a condition that looks like forward-biased diode from a low impedance source, ie, high-current conditions that could fry things.

Open-drain output stages would be off and so their pull-up resistors (if external) would float their drains to (+) - no danger there of overcurrent conditions, but what of other output stage configurations? Totem-pole outputs would be floating, I suspect, and so no overcurrent conditions there, although both output transistors being off would allow the connections to float, making them susceptible to noise which might be read by whatever inputs they're connected to and, lord knows what that might be telling the rest of the system: "Red Alert! Red Alert! Del's playing with electronics again! Judas Priest! Why doesn't he just stick to making world-class bows!"

Cheers, yeah the 2 day week doing electronics is getting a bit much.

Del

BTW. I've tried the RFID module in my cup of coffee and it didn't float at all

What are you using it for, anyway? What kind of RFID module?

Two days doing electronics? I'd do a lot more if I had my lab gear here. It's all in storage down in Texas. Everything. Got a sh!tload of projects and no tools to do them with.

It's an SL031 module from 'Stronglink' it's detecting the presence of a chemical container and counting how many days it's been in use. Nice little module, mind none of 'em are cheap.

My P type mosfets turned up this morning and I've got it working fine from a spare output on the micro . I've made the change on the PCB too... can I go home now?
I've got two bows to wax before they are collected tomorrow!

Del

Switching the + lead sounds like the correct idea to me, too. Some 3.3 V micro are fussy about a well regulated 3.3 V supply. So I'd put the MOSFET switch between the power supply and the regulator instead of the regulator and the micro.

Yup that's how it is... The 3v3 supply is going out from the regulated supply on the main board (which feed the micro) down and inch and a half of thick solid core ribbon cable to the module which can fend for itself as it has plenty of onboard 'C'

Del

Another CR4 software defect discovered. We can't do a GA and a OT at the same time without the net result being nothing.

Now that I think about it maybe nothing is the correct response.

I think what we need is a second scoring metric to allow for a better distinction.

We already have the Good Answer - Off Topic axis, which indicates the relative scale of 'pay attention to this post' to 'ignore this post'

What we need is a second voting axis of 'This amused/pleased me' to 'This Offended/upset me or made me re-evaluate my worldview.'

That way we get a better feel for the type of post that is being evaluated. So we could have a post rated 'Amusing but off topic,' (OT & Amuse) or 'this is important and correct and now I feel bad for reading it' (GA & Offend) or 'I love this idea, and I'm excited to be a part of it. (GA & Amuse) or 'this post is Troll Bait' (OT & Offend).

(Self-marking as OT because this has nothing to do with which leg of a power supply needs to be disconnected. Personally, I always disconnect the side that is NOT used as 'ground'/'common'/'return.')

(Removing the self-OT, since I just gave my answer on the OP's topic.)

It sounds as though the design is sadly lacking both hard and firmware/software development.

Such a "reset" should be standard to my mind, but with good firmware/hardware design, should never be needed.

You should be looking either to shield it better from static or similar, or even to look around for a better design. Maybe some leads, external to the unit, are picking up noise/static and feeding it into the unit....Finding such problems are not easy...some designs of equipment allow "logging" as to when and how the problem happens

Have you contacted the manufacturer yet?

As to whether you switch the + or the - to reset the device, I cannot see a difference.......electrons are stopped from flowing......

Best of luck.

Something is happening with your power supply. Voltage ought to be very stable with no possibility for spikes to develop. Add a big electrolytic on power rail near the board, and if voltage is within specs and it locks again, it is defective. I'd forget mosfets etc. If you think you have to do this power reset, use a NC momentary switch. S.M.

Maybe, but a manual reset is no good when the unit will be installed somewher inaccessible with no indication of if it's working or not. (It operates once a day when the premises has finished work).
The main board accesses the RFID module periodically, and if it is switched on and off every time it will be getting a hard reset and also saving power.

I suspect the cuprit is static discharge/esd during handling and shipping. A manual reset is fine for the final assembly stage but no use in the field (should a lock up occur). I can't reproduce the lock up in the lab so it's hard to investigate, thus a self correcting fix seems like a good idea.

Del

Not being able to reproduce lock-up, ads to the possibility the supply being the problem. If you don't solve the real issue, one of these days the thing will NOT wake-up. Add the power rail capacitor (2200-4700uf) NEAR or better ON the board, and see how it goes. If something not supply related is remotely induced to the circuit and is strong enough to make it lock, the whole thing will also NOT survive for long. S.M.

Then "self resetting" or Watchdog code can be written/designed quite simply. This will reset the chip(s) or whatever, if it hangs for longer than the timer is set for.....

You can easily design a simple hardware reset that needs say an occasional pulsed signal to NOT reset.

It can be a signal sent every few milliseconds, seconds, or even minutes or hours, whatever your design needs....depending upon power usage and "up" time required, from a properly running program.

It can be as simple as maintaining a charge on capacitor for example.....

If the software/firmware hangs, it is unlikely that it will keep sending the required pulse and a hardware reset will then be automatically performed....

I do believe that nowadays you can buy a chip to do it....

Its nothing new, we had it in our equipment more than 30 years ago.....I am pretty sure that the idea is even older.....though no facts!!

I found on Wiki a good article:-

A watchdog timer (WDT; sometimes called a computer operating properly or COP timer, or simply a watchdog) is an electronic timer that is used to detect and recover from computer malfunctions. During normal operation, the computer regularly restarts the watchdog timer to prevent it from elapsing, or "timing out". If, due to a hardware fault or program error, the computer fails to restart the watchdog, the timer will elapse and generate a timeout signal. The timeout signal is used to initiate corrective action or actions. The corrective actions typically include placing the computer system in a safe state and restoring normal system operation.

Also using a 7555 chip as here:-

Watchdog Schematic

Watchdog Schematic 2

Being a Dog lover, I found this website fascinatig too:-

Watchdog Timers

I hope this gives you a few ideas.....

You are missing the point.
The main board, which I designed is fine, there are no issues with it at all.
the module is a bought in board over which I have little control. By using the mico on the main board to power it up and down periodically it is ensures a clean wake up when needed.

I hate RF with a passion and thought RFID was a bad idea in the first place. It is by definition going to be susceptible to RF interfence.

I don't think there is actually a problem in thefield, but I'm just trying to be belt and braces with the design.

Thanks for all the input everyone.

Del

when malfunction fluid is exhausted a battery bypass occurs discharging the capacitor requiring a restart