:: Re: [DNG] Information request (mayb…
Página superior
Eliminar este mensaje
Responder a este mensaje
Autor: Simon
Fecha:  
A: Devuan ML
Asunto: Re: [DNG] Information request (maybe OT - - - dunno)
o1bigtenor via Dng <dng@???> wrote:

> I am in the beginning throes of an embedded system (SoC, and/or
> microcontroller) design.
>
> Have been reading somewhat on the use of RTOS vs and OS (thinking
> perhaps FreeRTOS and Devuan). My initial problem is a weighing 'cell'
> where I'm weighing liquids being deposited in a 'jar'.
> (A second would be items deposited (and then removed in stages) from a
> 4 wheel cart.)
>
> How does one decide when a RTOS is more appropriate than an OS?


As already says, it comes down to “if I want X to happen at time Y, how reliable do I want Y to be ?”
These days, something as cheap as a Pi runs so fast that many tasks that would have needed an RTOS back when I used to be involved with this stuff can be done “reliably enough” without going “hard” RT. A secondary question is “does someone die or lose limbs if it goes wrong ?” - which also influences the decision since you’d never pass a safety case for something where the answer is “yes” with a raspberry pi and a generic OS.

And as part of the mix, you have trade-offs. So while you could just chuck in (say) an ARM SoC and use brute force to create near-enough RT operation, you might decide that the reliability downside of running millions of lines of code to emulate a few TTL chips is too much - and so you’d be better using something smaller & simpler, and then having to do more low-level stuff to make things work.

In other words, there isn’t “an” answer. What you’d have to do is make a list of options, filter them to remove any that wouldn’t meet some non-optional requirement, and then list the pros & cons of each so you can decide which is the optimum choice.



>> PoE over CAT-5 can deliver 5V or 12V to multiple remote locations. Other options include solar and battery
>
> Hmmmmm - - - just did a little playing with ampacity charts
> seems at 5V with 2.5 A (not that big a draw imo) well I need to run 10
> ga if I'm running 5 m.


I wouldn’t run 5V over any distance - you’ve little tolerance for volt drop, and it’s hard to provide any form of protection against being connected to the wrong voltage.
Using passive PoE (basically just stick power on two of the wires) you can use 12V or 24V - with 24V you’d be down to just under 1/5A which is a lot more manageable. Not only that, but (assuming you a switching down-converter) you get to compensate for any volt drop in the conversion.
Active PoE (e.g. 803.11AF) uses 48V. Also, the power sourcing equipment does not apply power unless it sees the right signal (basically a resistor), and provides protection against connecting stuff that wasn’t expecting power coming down the wires.

>>
>> If you standardise on CAT-5 then you can run most protocols and power in 1 cable.
>
> Given the above - - - - there is a lot more thinking through that will
> need to happen!


Yes, lots to think about.


Simon