Actuator wear from not having PID deadband? HW Spyder.
Regarding valve and damper actuators PIDs, I have been leaving the PID deadband at 0. I wonder what sort of affect that has on the actuator over time, since even a well-tuned loop will have regular small corrections (tenths of a percent).
When I first started learning, it seemed that having a deadband resulted in more hunting when the PID went out of idle mode when the deadband was exceeded, so I concentrated on tuning the throttling range and interval so that the corrections would be within a deadband range without actually having a deadband. Now I wonder if I'm saving valves at the long-term expense of actuators.
I haven't seen an unusual number of actuator failures. The ones we've installed are lasting for years, but they do seem to fail more than other things, I guess any moving part will fail more than non-moving parts. But if an actuator should last 10 years I want to do whatever I can to help that happen.
Belimo has some information on this.
See control loop tuning on page 28.
Personally I try to stick with 15 sec minimum between repositions, tune loops good, and have a appropriate deadband on just about everything. On VAV dampers I like to use 10% of the min flow or 15-20CFM, largest of the two, as the deadband. There are cases that the actuators need to move more often, but that's rare in HVAC control. If you treat the actuators right they will easily last over 10-15+ years.
Its pretty easy to know if you have issues with too much movement on the actuators (at least after the fact). First a small handful will die, then it turns into a tidal wave of failures. You can easily wear out actuators this way in well under 5 years. IMHO, many of the times the actuator is blamed for failing early, when in reality the controller was driving it like a RC drive motor...
Thanks for the response and the document. Now I know that the term for what I'm concerned about is "dithering". I'll be using deadband from now on.
Yep, I pretty much agree with Orion.
I've known and worked with controls techs/engineers who insist upon tight control/fast response. Which means frequent, small adjustments.
And I've replaced a LOT of their stuff after it failed.
OTOH, typically the stuff I and others I work with who think similar to myself, usually easily lasts 15+ years. Granting you're using decent, quality actuators. That's not a guess, based upon experience. We've got a LOT of installed stuff out there that's been in that long and longer.
Some manufacturer's, Belimo is one of them, deliberately build in a small "deadband" within the electronics of their actuator control circuitry for the whole reason of helping to minimize frequent, minor adjustments (jittering). In order to extend the life of the actuator.
Besides, the truth is such tight, jittering control and response is not really beneficial to comfort HVAC control.
For instance, take a conference room that's first empty and then quickly fills. It'll take perhaps 5 minutes for actual air temp to increase 1'F. Another couple minutes, perhaps, for the average room space temp sensor to register the change enough to cause control response to kick in. So, okay, why have TIGHT control and frequent repositioning of reheat valve and/or air damper in a VAV serving the room? Especially given that very small repositioning of valve or damper has little effect upon increasing/decreasing DAT from VAV?
Besides, humans typically don't detect ambient air temperature shifts around them if its less than 2'F. UNLESS such changes are rapidly fluctuating. They will detect shifting air temps much quicker if its changing rapidly.
Just my opinion. Not worth more than that.
Chuckle, we have one guy at work who INSISTS that no deadband is the way to go. I've talked to him time and time again. Have never been able to change his mind. So I gave up trying. Nowadays I just wait until I get a complaint call from the customer he serviced and then go in and change everything to my way. Problem solved, everybody is happy.
Originally Posted by fxb80
You mention Spyders.......you using custom programming or a Venom app?
If a Venom.....what PID are you referring to?
I have found with the Venom app there is alot of hunting. You need to dial in the AIA block feeding the actuator.
Also, I have seen alot of high failure rates on the Spyders with actuators. The actuators fail due to lack of stops set accordingly. Also, the AIA block doesn't change with stroke timing. This needs to changed manually.
Been away for a few days doing the "it's not a career, it's a lifestyle" thing converting a church from Trane to Honeywell.
I have done custom programming to this point but am trying to become familiar with the Venom apps.
The PID I refer to is the block called PID, I haven't used an AIA. The only Venom PIDs I've used as is are the space temp blocks, 3F throttling range with 3450 seconds interval, and they don't seem to hunt.
I sounds like you've done a lot more Venom apps than I have, probably a lot of VAV?
What stops are you referring to not being set? Physical stops or the min/max output limits? You mention stroke timing; I use modulating actuators whenever possible, not floating. Whether modulating or floating I do try to match actuator travel to valve or damper travel by setting the physical stops on the actuator if needed and checking full travel more than once.
Great information on Belimos valves.
I saw Spyder, Actuator, and Failure......and assumed we where talking Spyder VAV controllers with actuators.
Originally Posted by fxb80
But you should look into AIA blocks......I prefer them over straight PID for actuator control.
As for VAV Spyders, are most of you guys just using the Venom apps or writing your own. The Venom app is ok, but I have been toying with the idea of writing my own to make things a bit more simple for what we typically use them on. Anyone want to share what they are using?
I do both venoms and custom. Th eimportant thing is to build your library and migrate it from version to version so you dont have to redo every project. Most of the venoms get modified to suit, simply save the modified versions to your library for future reuse.
Originally Posted by chadtech
As far as actuator failure do to hunting, i agree with Orion/ Belimo. I have worn out actuators from being to picky with the tuning, we all like to see tight control however if you really think it out its not necessary. If you have customers that want to see 72.0* and not 72.5*, remember you have control over what they actually see and there is no way the can "feel" the difference. In fact most of the time I give zero precision to points, no need to be that precise.
"It's always controls"
Yep, showing that extra precision can be problematical. And makes no real difference to human comfort. OTOH, with some occupants, if they can see that value to the right of the decimal place, it can make a MENTAL difference. Some will start complaining about that half degree (or whatever) difference. Just human nature.
Originally Posted by skwsproul
One customer we have had been converting their buildings to digital controls for years. Previous space temp sensors we'd put in for them didn't have local displays at all. Then they had us do yet another building for them and this time their specifying engineer specified space temp sensors with local displays AND specified we show tenths of a degree. <Shrug> So we put those in. A year later their building maintenance guys were so disgusted with the increased number of complaints (hot and cold calls) that we were asked to remove the LED display sensors and replace them with blank faced ones. We did nothing else, but the complaints decreased drastically.
Another large customer did the same. Had the same issues. But decided to keep the local display type sensors, but had us reconfigure them to show only the whole numbers. Again, complaints decreased a lot.
Now certain customer activities might require close and tight control. Lab areas, certain manufacturing or processing areas, etc. If there is a valid reason to do it, the customer is usually okay with the increased costs entailed, including increased actuator failure rates.
But just for human comfort? Humans are pretty durable, sturdy, and adaptable. And most humans on this planet live daily lives, healthy ones, with far less continuous, tight climate control than we're used to providing in this country. As I've mentioned before, keep the temperature (and room air velocity) swinging slow, and you can easily get by with an up to 2 degree swing in the room temperature without discomfort to the occupants. Typically, they won't even notice. 99% of them.
As concerns the rest, probably not a darn thing you can do for them. They're too picky to please. It's like one guy at one customer site. Unfortunately he was the boss ... the owner. And was always complaining. Finally, after we and a design engineer tried everything we could think of and none of it was good enough for him, we stripped out his existing HVAC system (VAV system). And provided him with his own, separate from the rest of the building, unit and a local control next to his desk. And let him play with and adjust things to his heart's content. THEN he was happy.