Controller Design for the Complicated HVAC setup

[wdc1160]

I wanted to open a discussion about controllers to help organize in my mind what the pro’s and knowledgeable DIY’ers think about modern controller cards. And, if they have short-comings what they are.

My experience with controllers for boilers is limited to what people discuss this board, my own, 1 university, and two large grade schools. I do have daily experience with controllers in general, but their use in HVAC is what I would like to talk about in this thread.

My own controller was intended to be a based on Nofo’s.

So I went to the drawing board on what a complicated HVAC system requires in a controller for the AVERAGE JOE.

REQUIREMENTS
1. Must be able to log data from forms of outputs from all HVAC sources.
• Vents
• Water Pipes
• Water Tanks
• Chimneys
• Fireboxes
• O2 sensors
• Pumps
• Solar equipment

2. Must be able to scale to the number of inputs and outputs it can receive.
3. Must run on < 350 watts per hour
4. Must run for an average of > 120 days before requiring a reset.
5. Must recover from power loss in a completely automated way
6. Must take < 2 hrs to setup software and OS (by a novice)
7. Must be able to use thermistors in conjunction with inputs
8. Must cost < $300 for software and hardware.
9. Must have the ability to log data for a period > 90 days without intervention
10. Must be able to see the log data without bringing down the controller.
11. Must be able to control all forms of HVAC equipment automatically
• Pumps
• Dampers
• Blowers
• Valves

12 During hardware/power failure the controller must still allow the system to run in a way that is not dangerous or overly inefficient.


Does such a controller exist? Does their need to be additions to the controller’s requirements?

 

[heaterman]

The only thing I have ever worked with that would cover all those bases is a PC based HVAC management system. We've put a couple in smaller commercial buildings (<30,000 sq ft) and they work well. You can use something like that along with sensors, valves, actuators to program any number of if/then scenarios you want. Honeywell has some products available that are scaled for smaller applications. Not cheap but it would do everything you listed in a nice integrated package. We did a system that controlled a boiler, 9 AHU's with hot water and AC coils, 6 zones of baseboard, 9 fresh air dampers with drives and IIRC the total material cost was about $12,000 + the old PC that was used to run the whole thing with. The only thing I don't think it would handle is the O2 sensor. I'm assuming that you want that for monitoring flue gas which in turn would provide feedback to your boiler control. That is going to be a pricey piece all by itself.

Do a search on Honeywell building management systems or HVAC management controls. There's tons of stuff out there. Tekmar is getting into that area a bit also.

 

[Nofossil]

Comments embedded in italics....

I wanted to open a discussion about controllers to help organize in my mind what the pro’s and knowledgeable DIY’ers think about modern controller cards. And, if they have short-comings what they are.

My experience with controllers for boilers is limited to what people discuss this board, my own, 1 university, and two large grade schools. I do have daily experience with controllers in general, but their use in HVAC is what I would like to talk about in this thread.

My own controller was intended to be a based on Nofo’s. Not for the faint of heart. I'd be glad to help, but be aware of the complexities.

So I went to the drawing board on what a complicated HVAC system requires in a controller for the AVERAGE JOE.

REQUIREMENTS
1. Must be able to log data from forms of outputs from all HVAC sources. Yes, as long as the sensor can be interfaced. I haven't done O2 yet. My logging is to an SQL database on another server remote from the controller, but it could easily log to on-board SD card or USB flash drive.
• Vents
• Water Pipes
• Water Tanks
• Chimneys
• Fireboxes
• O2 sensors
• Pumps
• Solar equipment

2. Must be able to scale to the number of inputs and outputs it can receive. Mine is based on PC-104, which allows additional I/O cards to be stacked easily.
3. Must run on < 350 watts per hour CPU board is < 5 watts, I/O uses a bit more, but not much. Relay power comes from external source.
4. Must run for an average of > 120 days before requiring a reset. Has never required a reset. Not Windows ;-)
5. Must recover from power loss in a completely automated way Yup
6. Must take < 2 hrs to setup software and OS (by a novice) Not even remotely close. Major geekage required at this point. I've dabbled with the idea of a web=based configuration, but at present it's all hard-coded in C
7. Must be able to use thermistors in conjunction with inputs Yup. Thermocouples too.
8. Must cost < $300 for software and hardware. Yup, though a decent enclosure and power supply might get you close to the limit.
9. Must have the ability to log data for a period > 90 days without intervention Yup
10. Must be able to see the log data without bringing down the controller. Yup
11. Must be able to control all forms of HVAC equipment automatically Maybe. I use almost exclusively relay outputs at present.
• Pumps
• Dampers
• Blowers
• Valves

12 During hardware/power failure the controller must still allow the system to run in a way that is not dangerous or overly inefficient. Yup, though that takes a bit of planning.


Does such a controller exist? Does their need to be additions to the controller’s requirements?

I'd suggest that remote (Internet based) monitoring and management would be very desirable.The hardware should be capable of running for years without significant maintenance, in a potentially dirty / dusty environment. That means no cooling fans, in my never-so-humble opinion. It also implies a decent semi-sealed enclosure.

There are commercial controllers that are very good, but I doubt they're very close to your price point. PC based can be pretty cheap and there's decent software, but I/O tends to be expensive and reliability is pretty dismal, both hardware and software. An internet connected PC is vulnerable to infection and attack as well. Power consumption and space requirements are very high for a PC compared to other options.

I went the way I did because I'm a hard-core C programmer and a systems engineer. I've built copies that are in use by non-geek types, but I'm not prepared to commercialize what I have at present. Now, if an electrical engineer wanted to team up with me, anything's possible.

 

[wdc1160]

Heaterman, Please do show me where to find that on the honeywell page. The honeywell page is like reading greek to me. Is there a good model number. How much does itcost?

Nofo, consistent with nearly all hardware now, I figured it would have a web-based config tool / monitoring tool. I didn't want to make it a requiment. Only because it could be done otherways in acceptable fasions.
I do think I have good sealed enclosure taht are very cheap. They can't take direct hits from large equipment, but you couldn't smashem with your hands either.

Nofo, if someone doesn't point out a closer fit, than the ts 7200, I am going to try to piece together something with MYSQL and C. Proabably in conjuction ethernet harddrive.
Maybe an attached internal HDD.

I will also likely use my ts7800 for work. It is clear that it isn't as stable and uses a slightly different different os as the 7200. It is also still under heavy deveopment and is always changing, so thier wouldn't be any hope for having a homogenius array of cards. Getting the newest and best controller was a huge backfire.

But the TS7200 is the only option I have seen for meeting this requirments. If something else comes close to meeting these requirments I may comprimise in the requirments. But, I am all eyeballs for suggestions.

I have even looked at controller cards for dummies, Cards made for students, and a radioshack version of a learning card. None of these would have the ability to scale, but would work for most else.

 

[Nofossil]

Nofo, if someone doesn't point out a closer fit, than the ts 7200, I am going to try to piece together something with MYSQL and C. Proabably in conjuction ethernet harddrive.
Maybe an attached internal HDD.

I will also likely use my ts7800 for work. It is clear that it isn't as stable and uses a slightly different different os as the 7200. It is also still under heavy deveopment and is always changing, so thier wouldn't be any hope for having a homogenius array of cards. Getting the newest and best controller was a huge backfire.

But the TS7200 is the only option I have seen for meeting this requirments. If something else comes close to meeting these requirments I may comprimise in the requirments. But, I am all eyeballs for suggestions.

I've used the TS-7250 and TS-7260. I like the 7260 better in part because it's very forgiving about supply voltage - anything between 5 and 20 volts works, so I can run it off a car battery or the same DC supply that I use for my 12V relays.

I have a pair of TS9700 A/D boards and a DIO64 plugged into it. If you're brave enough to try this, I'll gladly share all the code that I've developed. I use MYSQL, but only a tiny part of the client library is on the 7260. If self-contained operation is important, I'd consider using a USB flash drive rather than a hard disk for data storage. The OS comes with it, and includes a very functional Apache web server. You will most assuredly want a desktop Linux box as a development system, but once you've put the code on the 7260, it's a fine standalone piece of hardware.

I've got plenty of working code samples for it. My weak spot is the electronic design for interfacing to sensors. I've got stuff working, but I'm not convinced it's as elegant as it could be.