PDA

View Full Version : Help building a Photostability Chamber



Prashun Patel
05-22-2013, 2:34 PM
A PS chamber is used to expose products to UV light and heat in a controlled manner. The temp range is between room temp and 140F.
My company is in need of one to replace the one that we fashioned out of plywood and lightbulbs 20 years ago, but it is not easily regulated and doesn't get as hot as we'd like - as uniformly as we'd like. It's just a big wood, easy-bake oven.

I can spend $10-20k for one of these, but if we can make one it'd be great. I don't need fancy electronics. Just a heating element, fan, and a reliable method of regulating the heat.

Does anyone have plans for building a crude oven like this?

Jeff Bartley
05-22-2013, 3:26 PM
Prashun, have you tried a rheostat to regulate the intensity of the bulbs? What about a small fan to help uniformity of heat? I don't know what temperatures plastics are annealed at but you might check that avenue of information.

Dan Hintz
05-22-2013, 5:00 PM
If you want to regulate, look for a PID controller... you should be able to pick up an industrial controller with the typical bells/whistles for <$1k. Everything else stays the same (you'll need a thermistor in there somewhere so the PIC knows what to do).

Stephen Cherry
05-22-2013, 5:21 PM
On about the third day of my becoming a real engineer, an experienced guy called me into his office and told me, "Steve, don't think too much; if you can, just copy what someone else did".

Great advice then, and now. Can you look around at similar devices and see how it was done.

Jim Matthews
05-22-2013, 7:52 PM
Not to be flip, but could you do this with a refrigerator?

I'm thinking that it's highly reflective inside, insulated outside and could be perforated to install a fan to exhaust excess heat.
TekSupply (https://www.teksupply.com/contractor/supplies/ExternalPageView?pageKey=EXTERNAL_PAGE_3034)has off the shelf rheostats used to trip ventilation fans. Perhaps these could be adapted to this task.

I'm unclear of the application, but it sounds a lot like incubating chickens from eggs...

Pat Barry
05-22-2013, 8:28 PM
Lots of used manufacturing equipment like ovens on ebay. I would do a search there for an oven. How big does it need to be (interior dimensions)

John Sanford
05-23-2013, 5:35 PM
yeah, size is a big factor. Rather than making the new one with plywood, make it with insulated panels, aka Plywood/insulation/plywood sandwiches. Make it a CUBE, i.e. don't forget the floor. 360 degree insulation will make regulating the temperature much, much easier. The other thing you may want to consider is humidity. Controlling the humidity will make it easier to control the temperature.

Depending on size, you may be able to go cheap with a used refrigerator (or oven), commercial freezer/walk-in, or even a used insulated shipping container if you need something big.

Myk Rian
05-23-2013, 5:47 PM
Omega makes some good, and inexpensive ($200 or so) temp controllers. http://www.omega.com/toc_asp/subsectionSC.asp?subsection=P03&book=Temperature
I used this company extensively for instrumentation.
Look for something with a "K" (chromel/alumel) thermocouple input. Pretty universal thermocouple.
EDIT: Red is negative with thermocouples.

John Coloccia
05-23-2013, 6:49 PM
I'm very familiar with Omega products, and that's the first company that came to mind when I saw the thread title. Be aware that the thermocouples are NOT easy to solder to. There is a very specific way to do it, and woe to you if you take any short cuts.

That said, there are curing chambers designed for UV curable finishes. Depending on the size you need, I would look into this first. As a wise person once told me, and as I used to tell the guys that worked for me when I had a real job, "Don't build what you can buy, and don't buy what you can steal". If I could find COTS product to do this, there is no way I would ever build it myself because it will cost me FAR more. I think a UV curing enclave is your best bet.

Dan Hintz
05-24-2013, 7:44 AM
If I could find COTS product to do this, there is no way I would ever build it myself because it will cost me FAR more. I think a UV curing enclave is your best bet.

I'm generally of the same mindset, but all too often the item is WAY overpriced because of the market type. If I could make a big box for $500 and a weekend, I'd likely choose that method over purchasing a $40k COTS box. Some markets are priced that way. When items start getting down into the sub-$1k price range, I typically just buy it outright.

Stephen Cherry
05-24-2013, 8:38 AM
Personally, if I were building this for my hobby shop, I would first decide if you can tolereate some range of temperatures during operation- say between 138 and 142. If you can, you can use some simple on-off control to operate the oven, like a normal oven. It seems to me that pid would require an elecrionic device to either modulate power to the heaters, or to modulate the duty cycle with a base frequency faster than the step response of the oven chamber. Nice if you have a budget, but the thermal mass of the oven would tend to smooth things out anyway. Speaking of thermal mass, it would be nice to reduce the mass of the oven, as well as the THERMOCOUPLE, which uses thermocouple cable matching the type of thermocouple (type K would be fine, or an rtd).

Here is what I would personally consider:
http://factorymation.info/catalog/fmcat_autonics_temp.pdf
or
http://www.plccenter.com/Shop/RED%20LION%20CONTROLS/TCU11001/NSFP?source=GoogleShopping&gclid=CLuvo_nfrrcCFYGe4AodGE0Auw

Dan Hintz
05-24-2013, 10:35 AM
The duty cycle on anything larger than a bread box would be measured in numerous seconds... the PID is the device, just a smart thermometer/controller. Like a VFD for spinny things.

Myk Rian
05-24-2013, 10:51 AM
Be aware that the thermocouples are NOT easy to solder to. There is a very specific way to do it, and woe to you if you take any short cuts.
It's called arc welding.
We had a small wooden box with many turns of a resistance wire in it. An electrician made it back in the 1930s. 2 nails sticking out the front were the terminals to lay the TC wires across. It was powered by 240 VDC taken from the crane rails. Now-a-days an Argon gas arc welder is used.

Stephen Cherry
05-24-2013, 10:53 AM
The duty cycle on anything larger than a bread box would be measured in numerous seconds... the PID is the device, just a smart thermometer/controller. Like a VFD for spinny things.

The problem with the pid is that it will not typically offer the transient speed of a simple on off controller. If you think about it, an on off controller offers the fastest possible warmup time because it operates the heaters at 100 percent until it is up to temp. A pid presents an output that tapers off as the proper temperature is approached. PId is the ticket for precise process control (not really, but that is another story), but it requires some sort of an ampifier from the pid controller to the heater because a pid controller typically delivers a low power 4-20 ma or 0-10 vdc output. Something like this:

http://www.titus-hvac.com/utility/getfile2.aspx?fileid=5754

Another advantage of an on-off, with hysteresis control is that the driver could be a contactor or suitable relay. (a cheap part) So basically, it would be a k type thermocouple to k type thermocouple cable to an el cheapo factorymation on-off controller to a contactor (relay) to the lightbulbs. These are just my thoughts of how I might approach this if I wanted to make some sort of oven for my personal use.

Prashun Patel
05-24-2013, 10:54 AM
Jeez, what a fantastic resource this forum is. This is exactly the kind of discussion i was hoping for.

To answer some questions: we don't need precision. My company makes fragrance oils for consumer products. We can certainly tolerate +/- 5 degree variance and don't need to pay 10x the price to get ultimate precision.


Keep the thoughts coming; I'm all ears.

Stephen Cherry
05-24-2013, 11:02 AM
If I were hypothetically doing this in a commercial setting, I would throw in what would be basically an on off motor contoller box- start swithc, stop switch and contactor. The output of the contactor would supply the contactor for the light bubs.

In series with the stop switch I would throw in an FM approved high temperature limit switch set well below the combustion temperature of the oven. This would trip the power on the heaters if the temp got dangerously high, and it would require manual operator intervention for restart.

All just hypothetical discussion, take it for what it's worth.

Myk Rian
05-24-2013, 11:15 AM
PIDs (Proportional, Integral, Derivative) is handy when the heating sources lose their temperature slowly when shut off, and take their time coming back up to temp. when on.
Many controllers have a 10 amp relay that can be used for light loads, or to activate an external relay.
In a pinch, a simple thermocouple can be made by just twisting the TC lead wires together tightly. In a liquid bath, a protection tube will be required. Usually a 1/4" stainless steel tube with one end welded closed. Instrumentation can be done very simply.

EDIT:
+1 on the over-temp safety. An Aquastat can be used for that purpose. In fact, Aquastats can be used for the entire control system. They are a simple bulb operated controller. No power is needed to use them.
http://www.omega.com/ppt/pptsc.asp?ref=AR&nav=

Stephen Cherry
05-24-2013, 11:30 AM
THese days, you just use the screw terminals on the parts, with the correct thermocouple cable. It's not complex because all the brains are built into the electronics.

Tom Stenzel
05-24-2013, 10:17 PM
In the steam powered box making machine video that was posted earlier, the digital temperature device show was an Omron DIN 1/16 temp controller, a functional equivalent to the Omega that Myk Rian linked to. Where I worked we would have replaced either with a Dwyer/Love 1500 controller. It's what we carried in our stockroom. There's no shortage of temp controllers out there.

For the low temperatures that the box will be run at a J type thermocouple would be preferred over a K type. I would use a K type if the temps ran over 1000 degrees F.

For that low a temp use to make a thermocouple from extension wire I would have silver brazed the thermocouple junction. We used a material with cadmium in it to insure the iron lead (J type) would wet. I think it was Allstate 101 (?). I don't know if it's on the market anymore, we just had a lot of it.

Where we used K type thermocouples they had to welded as the temps they saw would melt silver braze.

If you decide to twist the wires together to make a junction, tap it with a hammer to help "fuse" the metals together. But over time a twisted connection will become unreliable.

Whatever controller you decide to use, check the rating on the output. You may need to add an interposing relay between the controller and the contactor used to apply power to the heating elements.

What do you have in mind for the heating elements. We used these, they were pretty reliable unless they got wet:

http://www.chromalox.com/productcatalog/Component+Technologies/Cartridge+Heaters/Cartridge+Heaters/SCB/model-details.aspx?m=3256


-Tom Stenzel

Stephen Cherry
05-25-2013, 9:50 AM
In the steam powered box making machine video that was posted earlier, the digital temperature device show was an Omron DIN 1/16 temp controller, a functional equivalent to the Omega that Myk Rian linked to. Where I worked we would have replaced either with a Dwyer/Love 1500 controller. It's what we carried in our stockroom. There's no shortage of temp controllers out there.

For the low temperatures that the box will be run at a J type thermocouple would be preferred over a K type. I would use a K type if the temps ran over 1000 degrees F.

For that low a temp use to make a thermocouple from extension wire I would have silver brazed the thermocouple junction. We used a material with cadmium in it to insure the iron lead (J type) would wet. I think it was Allstate 101 (?). I don't know if it's on the market anymore, we just had a lot of it.

Where we used K type thermocouples they had to welded as the temps they saw would melt silver braze.

If you decide to twist the wires together to make a junction, tap it with a hammer to help "fuse" the metals together. But over time a twisted connection will become unreliable.

Whatever controller you decide to use, check the rating on the output. You may need to add an interposing relay between the controller and the contactor used to apply power to the heating elements.

What do you have in mind for the heating elements. We used these, they were pretty reliable unless they got wet:

http://www.chromalox.com/productcatalog/Component+Technologies/Cartridge+Heaters/Cartridge+Heaters/SCB/model-details.aspx?m=3256


-Tom Stenzel

All good stuff, in my opinion.

I just do not understand all the brazing, welding and twisting, when you can buy the parts already made. I would just buy a thermocouple from pyromation, omega, etc, which has thermocouple terminals (wiring terminals of the thermocouple material) in it, and then use thermocouple wire to the controller. In this particular app, it may be possible to wire the thermocouple right to the controller if leads are ordered long enough.

And as Tom said, heating elements, then whatever drives them would be the question. In my opinion, that would be my first question, heaters, then contactor. Also, for j vs k, I have to admit, I always ever used k because I always ever used k. Maybe it's not the best.

Also, you may be able to get an FM approved high temp limit switch built into the controller. Why FM, because you never know who down the road will want to see approvals, and it's cheaper to buy them now than to buy them later. All just personal opinions.

Dan Hintz
05-27-2013, 11:23 AM
The problem with the pid is that it will not typically offer the transient speed of a simple on off controller. If you think about it, an on off controller offers the fastest possible warmup time because it operates the heaters at 100 percent until it is up to temp. A pid presents an output that tapers off as the proper temperature is approached. PId is the ticket for precise process control (not really, but that is another story), but it requires some sort of an ampifier from the pid controller to the heater because a pid controller typically delivers a low power 4-20 ma or 0-10 vdc output. Something like this:

http://www.titus-hvac.com/utility/getfile2.aspx?fileid=5754

Another advantage of an on-off, with hysteresis control is that the driver could be a contactor or suitable relay. (a cheap part) So basically, it would be a k type thermocouple to k type thermocouple cable to an el cheapo factorymation on-off controller to a contactor (relay) to the lightbulbs. These are just my thoughts of how I might approach this if I wanted to make some sort of oven for my personal use.

What gave you the impression a PID is less capable than a bang-bang controller? The last PID I designed (ran hot enough to melt high-temp plastic but kept the temp within about 5 millidegrees of desired) ran the heaters at 100% until it was within a couple of %, then it started cycling the power down to keep the overshoot below 1 degree. No bang-bang could ever hope to have that sort of control.

A PID is simply a controller, not a power supply. You don't run the heating elements off of the PID itself, but you do connect it to the relay that connects the heating element to power, just as with the bang-bang controller. There are more PIDs out there than just industrial-level 4-20mA loops.

Stephen Cherry
05-27-2013, 1:27 PM
What gave you the impression a PID is less capable than a bang-bang controller? The last PID I designed (ran hot enough to melt high-temp plastic but kept the temp within about 5 millidegrees of desired) ran the heaters at 100% until it was within a couple of %, then it started cycling the power down to keep the overshoot below 1 degree. No bang-bang could ever hope to have that sort of control.

A PID is simply a controller, not a power supply. You don't run the heating elements off of the PID itself, but you do connect it to the relay that connects the heating element to power, just as with the bang-bang controller. There are more PIDs out there than just industrial-level 4-20mA loops.

Dan- I never said that a pid is less capable than an on-off controller; each has it's own pluses and minuses; and in my opinion in this application an on off controller is a better option. First it does not require very much knowledge to tune. Second, a pid is much better as a regulator than a controller. Remember, the original poster may not have much experience with loop tuning.

Also, from my pharma experience, I'm not a big fan of slapping setpoints into the pid and watching the output bounce around. Remember, a pid does have mathematical realities that need to be addressed. It's much better to control to a desired trajectory of system states than letting the chips fall where they may. This is analogous to a flight plan for an airplane which places the aircraft at a particular location, speed and altitude at any particular time, rather than just saying I want to be in Baltimore, and then I want to be in Cleavland. Of course we were dealing with rather high value products, so there was less of a sense of humor from the clients.

All just my opinions though.

Tom Stenzel
05-27-2013, 3:48 PM
With the control that's desired a simple on-off t-stat will work. But the problem is the temp sensor placement and the heater sizing become more critical. Mis-sizing, and in particular oversizing, the heater can cause the temp to roller-coaster and never settle down.

A temp controller will pulse width modulates the heater and give you a lot more forgiveness. Many of the temp controllers also have an auto-tune feature, I've found that they give good results. That can get you out of trying to figure out cycle times and other control factors that are unknown to the box-builder.

-Tom Stenzel

Tom Stenzel
05-27-2013, 4:11 PM
Of course I forgot to finish my post.

Where I worked the furnaces were required to meet the NFPA-86 code.That requires a safety system that is independent of the control system. That is it can't share the temp sensor or any other part of the normal control circuit.

If I had to build this box I would add a temperature fuse like the ones found in Mr. Coffees and electric fan motors. The temp rating would be something that is higher than the highest temp the box could see in normal use and open at a temp below that would damage the box.

http://www.cantherm.com/products/thermal_fuses/sdf.html

I've never dealt with the company I linked to, it's just an example of what I'm talking about. Radio Shack sells them online for less than $2 each but interestingly doesn't mention what the temp rating is for the different ones they carry(!). Clothes dryers also use them so an appliance repair store might conjure up something useful.

With the current rating of most of these fuses they can be wired in series with the power and shut down the whole system if it opens.

-Tom Stenzel