Fault and acknowledge

The wall heater in the lab holds 18 to 20 °C. One morning the sensor shows 40 °C – the furnace next door has heated up the room, maybe something is even burning. The heater switches off. An hour later it is cool again and the heater starts as if nothing had happened. Nobody noticed anything, until it happens the next time.

Real heating systems have a safety temperature limiter for this: it switches off and stays off until someone resets it by hand. That is exactly what you build today.

Latching the fault#

A register cause records what happened. 0 means “no fault”, every other number is a reason:

Every tick the chip works out whether there is a reason right now. A new cause is only taken over if none is stored yet:

move now 0
sgt tmp temp MAXTEMP
select now tmp HOT now               # 1 = too hot
snan tmp temp
select now tmp NOSENSOR now          # 2 = sensor missing
select cause cause cause now         # first cause stays

The last line is the latch: if cause already holds something, it stays, otherwise now goes in. When the temperature drops again, now becomes 0 – but cause does not.

Acknowledging with a button#

Only a button may clear the fault, and only when there is no reason any more:

seqz tmp now            # cause gone?
and tmp tmp pressed     # ... and button pressed?
select cause tmp 0 cause

If it is still too hot, pressing does nothing. That way nobody can push away a fault whose reason still exists.

Showing the cause#

The heater only runs without a fault, a flashing light makes it visible, and an LED display names the number:

seqz clear cause        # 1 = no fault
and tmp heat clear
sbn HEATER HASH("Heizung") On tmp
snez tmp cause
sbn LIGHT HASH("Stoerung") On tmp
sbn DISPLAY HASH("Ursache") Setting cause

The hysteresis in heat keeps calculating undisturbed. After acknowledging, the heater carries on exactly where the temperature is right now.