It seems to be a meme - a saying passed down from engineer to engineer, with no logic as to why.
I had it again today. The aircon was being maintained and the engineer (who I think I have upset now) said not to use in auto and I was rude enough to ask why?
He said it is not good for it to keep changing from cooling to heating. OK, that is probably a true statement, wasting time and energy and maybe even increasing wear, who knows, it is plausible. But we were looking at the panel which someone had set, rather oddly, to auto heat to 16°C and cool to 22°C. I asked how it would be switching between cooling and heating exactly? Surely, if it is warm outside (as today) it will cool to 22°C and stop, and then heat up from outside so cool again back to 22°C and stop. When would it switch to heating? Similarly when really cold outside (and 16°C is daft and needs changing) it will heat to 16°C and stop, cool down from outside, so heat to 16°C and stop. It would not keep changing between heating to cooling.
Well, he had no answer to that, but said that the issue now is that it "stops". He say it would cool to 22°C and "stop", where as in cooling mode rather than auto mode, it would cool to 22°C and then keep fans going circulating air rather than "stop".
Well, (a) why would that make a difference as it still does not keep changing between heating and cooling and would get to required temperature, (b) why would it behave differently in auto and cooling mode anyway, that makes no sense, and (c) look, the fan is running when in between temperatures not heating or cooling now, so that is clearly untrue!
He had no answer and resorts to "well, if it that is what you want then fine!" and refused to actually explain his advice. He reiterated that they tell people not to run in auto mode, as if the fact that they tell people is, in itself, reason to tell people. He simply had no reason and seemed quite upset that I wanted him to back up his advice.
This sort of thing really pisses me off. If there really is a valid reason, then tell me, or better still tell the manufacturer - why on earth would they provide the feature if it was not advised to use it? Or better still, if there is a reason not to use it, why don't they fix that?
What I suspect happened is that once upon a time some air con units were somehow a bit dumb and would overshoot cooling or heating enough to trip the aircon to go the other way and get in a cycle. That is clearly broken behaviour, but instead of complaining to the manufacturer, installers just advised not to use auto. These days there really is no excuse - it is all microcontroller based and a computer can be sensible enough to ignore any overshoot anyway, or better still, avoid an overshoot in the first place.
In fact, my aircon at home is excellent. It is set to auto. When it is hot outside, it gently blows some cold air briefly and stops, maintaining the temperature. It does not switch to heating mode. When cold outside it gently blows some hot air until temperature, and stops. I can set the hysteresis to a wider or narrower setting. It is not as flexible as the office one where you set two temperatures. But it works nicely and just works, maintaining a temperature.
Oddly, even one of my colleagues tried to defend the "don't use auto" argument (because he is an argumentative sod himself) saying some people do not want a fixed temperature. That argument is, of course, completely irrelevant. It is true, sometimes I may want it hotter or colder - so I set the temperature a degree or two higher or lower. I don't care how it gets to that temperature.
Oddly, our industry has the same problem. The meme I am thinking of is not setting auto-negotiation on 10/100 Ethernet ports. There really is no reason for this advice. auto-neg just works, and can be set to auto-neg specific settings only. It seems that maybe, once apon a time, some kit, in the early days of the feature, would get it wrong, hence the advice. Now it is just advice that is a "known thing" so gets repeated, and ironically causes no end of issues.
I've seen some network equipment fail to auto-neg correctly.
ReplyDeleteSo instead of "don't use equipment X as it is crap" is it "don't use auto-config"? Which do you go for?
DeletePossibly, In terms of networking kit it's usually when you have 2 different vendors kit that sometimes auto-neg doesn't work correctly. Fortunately it's fairly rare these days.
DeleteSounds like the people who installed my condensing boiler: "You don't have any young children so just set it at the max 84 degrees and forget about it" !!
ReplyDeleteI've certainly seen some network kit fail to negotiate (frequently leased line CPE supplied by ISPs, so no choice but to use it!) However, I certainly wouldn't advise "always turn off autoneg" - just turn it off if it happens to be a problem because the vast majority of the time it is fine.
ReplyDeleteI work somewhere where setting aircon to auto is genuinely a bad idea.
ReplyDeleteWe have a large open plan area, which is too large to heat/cool with a single unit. As a result, we have two units.
If they're both manually set to heat or cool at the same temperature, we end up with one unit doing slightly more work holding equilibrium, but otherwise things are OK. If they're both manually set to heat or cool, but have different temperatures, we get a breeze in the office (or one area hot, one area cold), and know to investigate.
If one unit is set to heat, one to cool, they rapidly (under 30 minutes elapsed) go to flat out and fighting each other. This is obvious, and thus easy to fix.
If one or both units is set to auto, they appear to work fine until you hit the edge condition that has one cool, one heat. At that point, they flip into both running flat out until someone fixes them. Because this takes a while to appear, they may flip into this state when no-one's around to spot it and fix it, costing us a lot of money.
This has proven tricky to explain to people in the past - not just aircon engineers, but also new hires in R&D. Thus, it's easier to just say "don't use auto mode".
Of course, the proper fix would be a ganged control system for the units that end up fighting - but that's also pricy, and therefore not going to happen.
Interesting case - and yes - solution is single control for the area. In both of my cases it is simple controller. But none the less a case I had not considered and perhaps one of the sources of this mis-advice.
DeleteReminds me of the old Unix advice of "run sync three times before rebooting"...
ReplyDeleteI've seen many service outages due to ports "negotiating" half duplex 10 mbps wronly. Glad to
ReplyDeletehear it if the defective kit isn't still out there, but honestly I'm skeptical.
Sometimes one is not in a position to dictate new kit there and then.