integrated circuit shells are no longer doomed to explode (#95523)

## About The Pull Request

Integrated circuit shells now have an early return for setting their
`power_used_cooldown` if the cooldown already exists, allowing the power
usage cap to reset itself properly and not doom all circuit shells to an
explosive fate, as reported by @Acer202 et al.

## Why It's Good For The Game

Perhaps we shouldn't be making our circuit shells out of explosive
composite alloys like BattleTech missiles.

## Changelog

🆑
fix: Integrated circuit shells now only explode if they draw too much
power in a minute, rather than being doomed to explode if it uses too
much power over its entire lifetime, following complaints from at least
one guy. Nanotrasen representatives claim that this is entirely
unrelated to missing shipments of explosive alloys.
/🆑

---------

Co-authored-by: Hatterhat <Hatterhat@users.noreply.github.com>

code/datums/components/shell.dm

@@ -284,6 +284,7 @@
 /datum/component/shell/proc/override_power_usage(datum/source, power_to_use)
 	SIGNAL_HANDLER
 	if(COOLDOWN_FINISHED(src, power_used_cooldown))
+		COOLDOWN_START(src, power_used_cooldown, 1 MINUTES)
 		power_used_in_minute = 0
 
 	var/area/location = get_area(parent)
@@ -297,7 +298,7 @@
 		return
 	location.apc?.terminal?.use_energy(power_to_use, channel = AREA_USAGE_EQUIP)
 	power_used_in_minute += power_to_use
-	COOLDOWN_START(src, power_used_cooldown, 1 MINUTES)
+
 	return COMPONENT_OVERRIDE_POWER_USAGE
 
 /**