Question About UV Light Source

[photomc]

Have done the required reading on Unblinkingeye and other sites. But have not found why most light source have 8-12 bulbs. Can you build a smaller unit, say with 4 bulbs, for nothing larger than 8x10? I plan to build one, but thought I would ask, since there are a few here that do alt. printing.

The bulbs I plan to use are 24 in., blb - or 20 watts and will be using mag. ballast in place of electric. Will I be wasting my time? Should I just go ahead a get at least 4 more?

Thanks,
Mike

p.s. - Yeah! I know, this is just starting something else...but using the sun for the Ziatypes just isn't working out. Darn clouds

 

[ThomHarrop]

I built a unit with 3 UV lights in a closet drawer set. The bottom drawer holds the lights and a power strip. The top drawer has no bottom and has a rack to hold an 8x10 frame. The lights are hooked to a Gralab and the whole thing is pretty light tight. I love it. I do Cyanotype and Van Dyke Brown and get times of 8 to 10 minutes. This is around the same time as our high end burner at my school.

 

[Joe Lipka]

Four bulbs would barely cover the area. You need to be able to get the image "centered" on the bank of bulbs. Need to avoid the ends of the bulbs in both directions. Go for six minimum, and then with 8 bulbs you can go at least to 11 x 14.

 

[clay]

Definitely use more than 4 bulbs. You will want the coverage to be as even as possible, and if you are going to spend the time to do this anyway, I would recommend just doing it 'right' from the beginning.

 

[Jorge]

Listen to Clay, he knows.....he has saved me a lot of money, if that is an incentive..

 

[Michael Mutmansky]

Here's the issue with using too few bulbs: uniformity problems.

If you make a printer you generally want the lamps as close together as possible while still allowing room for your fingers to get the lamps out to replace them. Most people space the lamps about 2" OC (Whoops... that means 'On Center' for all you non-engineering geeks).

2" OC will result in a 1/2" spacing between the lamps, assuming you are using the standard T-12 UV lamps that are available. I wouldn't recommend going any further apart than that.

The reason for this spacing is that you generally want to place the print frame very close to the lamps to increase the printing speed of the unit. If they are spaced too far apart, you may see banding on the print from the non-uniform output associated with the lamp and the band between the lamps.

If you place the print frame glass about 3" or so from the lamps, a 1/2" spacing between lamps will not be visible in my experience, but you need to make sure that you have a good UV reflector behind the lamps (this goes without saying).

So, with 4 T-12 lamps, you cover an area of about 7-1/2" in width, but there is an additional problem. The area at the very edge of the printer will have less UV light than the middle, and so you will still have a uniformity problem.

This can be explained through fairly simple light integration or magnetic field theories, which are way too afield for this discussion. However, you can think of it this way: If you were to place your eye at the center of the location of the print frame, looking at the lamps, you will see a bunch of lamps in all directions, and there may be one directly in front, about 3" away.

If you place your eye at the edge of the printer looking at the lamps, you will see a lamp directly ahead, and more lamps to one side, but none to the other side.

This lack of lamps on both sides at the edge of the printer will cause less exposure in this area. This is an inherent property of every fluorescent exposure unit. The easiest way to avoid this is to make the unit with enough lamps that when an 8x10 is place under it, it still 'sees' lamps on both sides at the edge of the print.

This problem will occur at the sides of the lamps, but it can be even worse at the ends because there are some things about fluorescent lamps that can cause problems, one of them being that over time the lamps may have a tendency to get 'end wall blackening' in the last inch or two, which will severely reduce output. The lamps also inherently have reduced output in the last inch or two because of the cathode and anode that are built into the lamp. So basically, avoid the lamp ends if at all possible.

My recommendation when people design a UV unit is to make the unit at least 4" larger in every direction than the largest print you think you are going to make. It's really better to be 6" from the ends if at all possible. That's the best way to ensure that you will have good uniformity everywhere. So for 8x10, I would recommend at minimum a 16x18 exposure bank, preferably a 16x22 bank minimum. That equates to a 16x24 bank using 24" lamps, which equals 8 lamps 2" OC.

If you use 8 lamps, you may or may not be happy with the uniformity of the print if you happen to move up to an 11x14 image, because you are right on the edge of where things will begin to be noticeable.

All that listed above is for a 3" spacing from the lamps to the print frame glass. You can use fewer lamps if you increase the spacing from the lamps to the print frame glass. You'll be fine with a spacing from about 2" to 4" from the lamps with the spacing listed. There's really no point in increasing the distance, as it reduces printing speed, and there's no point in trying to reduce it because you have to be careful with uniformity issues. That's why 3" is about ideal.

If you go to 6" away, then you could probably increase the lamp spacing to
4" OC and still have solid uniformity. However, be aware that your print times will be substantially longer (greater than 2x longer) due to two factors. The first is you are using a lot fewer lamps, and the second is there are 'system losses' in the unit, and as you move further away, they typically increase. This is another issue of light integration, and a bit too complex to explain here.

There are many ways to solve the UV printer problem, but I think that the best way is to do lamps spaced 2" OC and make a bank large enough to cover the largest print you want.

I also think that you should consider using 48" lamps if you have the room. It will probably not cost any more (it may actually cost less because the lamps and ballasts are more common), and you will have the room to get two 8x10 prints going at any one time. If you only are printing one image, you can use the very large sweet spot in the middle with impudence.

I also recommend electronic ballasts if you can get them, as they are better for the lamps, have higher efficiency, and produce less heat, which does have a negative impact on the output of the lamps. With 48" lamps, you can get 4-lamp ballasts, which will cut the cost difference down a bit. Make sure that all your lamps and ballasts are identical, because they will often vary in output, which could cause striping problems.



---Michael

 

[Michael Mutmansky]

Oh BTW, can you tell I'm an illumination engineer by training?

If you have any other questions, feel free to post.


---Michael

 

[wmlaven]

Michael's reply is very comprehensive. I would add only a couple things. First, I have found that the distance from the bulbs to the print frame can equal the distance from one bulb to another OC (in contrast to M's making the bulb-to-print distance slightly longer than bulb-to-bulb OC). One wants to watch for uniformity, but getting the print frame as close to the light source without banding (ie with uniformity) means faster print times.

Second, consider, too, the type of UV light you use. Many people obey the majority of Alt Process literature and buy black light bulbs when they may not be the best fit for the material they are using. You always want to match your light source to your print material. My experience, and that of others (plenty of posts on this topis on the web) is that Super Actinic (SA) bulbs result in faster print times and slightly more contrast with both Pt/Pd and POP. (SA's output most their light at 420nm rather than the 360nm of BLB's) I don't know how Van Dyke or Cyanotype respond to SA's, though someone out there does.

 

[photomc]

Michael,

WOW!! Well I knew there was a reason to ask, just had never seen any "Good" explanation. The summary you just listed looks like a great article for the article section. Have done quite a bit of searching (readgogglee) and have not seen anything like what you just posted. Would say your training is an asset for us, because no where have I found this type of information, and so well explained - Even I understand .

Guess I will go and pick up the additional lighting and parts to use at least 8 bulbs, seemed like overkill for 4x5 negatives, but you have explained it very well. Wish I did have room for the 48 in. but right now, no room. Will see how this works out for me and if I move up to a larger format, or start doingsufficientt prints (read have more than one processing at a time) then I will want to go that route.

Thank You for such a concise explanation.

Mike

 

[wmlaven]

As a p.s. to my own post, the KISS philosophy (Keep It Simple, Stupid) works for UV lights as it does for much else. For years, I took my Pt/Pd dog-and-pony show on the road and taught workshops in California, Oregon and Washington, hauling printers, trays, etc with me. For cheap simple printers I bought these cheap little flourescent fixtures designed to be mounted under cabinets in kitchens for local light. I got them at Home Depot type stores for about $10 apiece. I'd yank out the tube that came with them and replace them with 15" Super Actinics (bought from a good light supply store or an aquarium store). I'd sit about 5 or 6 of these next to each other and the 8x10 print frame would lie on top of them (face down of course). The units have their ballasts built in and I'd just plug them all in a power strip and way you go. So, it cost about $125 for an 8x10 printer and while the bulbs were only 15 watts, it was 15 watts over 15 inches of bulb so the power per inch was the same as larger 20 watt bulbs. These little printers print as fast as the sun (we tested them) and as fast as most commercial units. And if you want to get really fancy, you set the lights inside a wooden box designed around the dimensions of your print frame and so when you put the print frame on top, you have no light leaks. It's so easy it's dangerous. No wiring, no special ballasts, nothing, And though these under counter units are made really cheaply, mine have kept burning like the Energizer Bunny for decades of teaching.

 

[Michael Mutmansky]

First, I have found that the distance from the bulbs to the print frame can equal the distance from one bulb to another OC (in contrast to M's making the bulb-to-print distance slightly longer than bulb-to-bulb OC). One wants to watch for uniformity, but getting the print frame as close to the light source without banding (ie with uniformity) means faster print times.

Bill,

This is why I stated that there is a range of acceptable distances. However, consider that the distance from the lamps to the print frame glass is not a significant factor in the speed of the printer. Most people think that the illumination (and therefore printing speed) will fall off to the square of the distance (similar to how flash illumination falls off). That is not correct in this case. In fact, with a large enough bank of lamps, the distance from the lamps to the glass is irrelevant (within limits).

This is related to the integrated illumination that a point receives, and is somewhat complicated to calculate, but you can be assured that the difference in exposure time for a 2" OC lamp bank at 3" vs. 2" from the glass will essentially be negligable. However, as you get closer, you run the risk of uniformity problems. That's why I recommended 3" and not 2". It may be safe at 2", but 3" is definitely safe, and there really is no penalty to the greater distance.

I know that this sounds like a bit of a contradiction from what I stated earlier about system losses, but it has to do with near-field calculations of light sources (or magnetic fields, or heat sources, etc.) that have a large dimension relative to the distance between the emitter and the receiver. It's called the 'parallel plate' or 'parallel plane' theory. It's applied in many ways, like in the design of magnetic fields in capacitors, as well as illumination engineering.

Essentially, as long as the distance between the lights and the glass remains small relative to the overall dimensions of the light bank, the average illumination on the glass will remain constant. Where this approximation fails is on the edges of the bank, but I think we've determined that we want to stay away from the edges as much as possible anyway, so we avoid that problem.


---Michael

 

[sanking]

Just a couple of more opinions.

First, about printing speed in relation to tube spacing and distance from printing frame. My own UV bank has the bulbs spaced 1.5" on center. With this spacing there is less than 1/4 stop difference in printing speed between bulb to printing frame distance of two inches and six inches. I settled on four inches assuming that I would get more uniform illumination at this distance but lose virtually nothing in terms of speed.

About type of tube. As William notes, there is a lot of discussion out there about the printing speed of various types of lights, say comparing BL, BLB, SA, Aqua, etc. and many Pt./Pd. printers report that the SA tubes print faster. However, an equally large number have found that the BL tubes print as fast or faster. I have personally compared printing speed of BL and SA tubes on at least four occasions and in every tests the BL tubes were actually slightly faster. Not enough to get really excited about, but nevertheless faster in every test. This was very controlled testing, i.e. using the tubes in the same fixture and carefully controlling all over the variables that go into printing. I had a conversation on this subject a couple of years ago with the folks at Edwards Engineering and they told me that their own tests had given the same results. However, the difference in speed between BL and SA tubes of the same wattage in fixtures of the same ballast is minimal, certainly not worth the expense of changing out one type for another.

The above observations applies only to Pt/Pd printing and the other iron-based processes, i.e. VDB, kallitype, cyanotype, etc. If you print with any of the dichromated colloid processes the BL tubes print quite a bit faster.

So my opinion is this. If you print with only Pt./Pd. or the iron-based processes the BL and SA tubes give virtually identical results, with some slight differences in curve shape. For printing with most of the other alternative processes the BL tube is clearly superior in terms of printing speed. In fact, for these processes even the BLB tube is much better than the SA.

Sandy

 

[wmlaven]

"So my opinion is this. If you print with only Pt./Pd. or the iron-based processes the BL and SA tubes give virtually identical results, with some slight differences in curve shape." (from Sandy)

Sandy and I have discussed our tests offlist on several occasions and we get different results. At several workshops, I brought printing units with both BLB and SA bulbs and we always found the SA bulbs faster by about a stop. We tested with new bulbs so bulb age wasn't a factor. Similarly, I had a friend test two untis with POP and she found the SA's faster and more contrasty. We were very careful in our testing to limit all other variables. God knows why the difference. In my 20 years of working with Pt/Pd I've found that it sometmes behaves predictably and sometimes throws one a curve ball. Perhaps a new bulb buyer could buy bulbs from an understanding store and explain they want to do only a few exposures for tests so they can return the set of bulbs they don't like and see what works for them waking up on their side of the bed.

In response to Michael's claims that the difference between 3 and 6 inches isn't that great, I guess I'm wrong. I've always copmpared greater distances, say 2" versus 14" (which is how I found one printer's light set up when I suggested a change) and with those margins the results were different.

 

[Joe Lipka]

I've read everything. I like Clay's response the best.

 

[sanking]

I've read everything. I like Clay's response the best.

Joe,

Damn, if I had know the responses were going to be rated I would have tried harder!!

But for what it is worth I agree with Clay. Built the largest unit that fits in your workroom since you may want to print larger at some time in the future. The economics of building UV banks is that it may actually be less expensive to build a twelve tube unit of 48" tubes than a twelve tube unit of 24" tubes. This is due to the fact that the 48" size is much more commonly used in construction than the smaller size. But if printing speed is important you should definitely use electronic ballast. My tests indicate that units with electronic ballast print up to 25% faster than units with magnetic ballast, which amounts to more difference in speed than you would normally expect to see with different type tubes, SA, BL, BLB, etc. And, not all electronic ballast is the same either for that matter.


Sandy

 

[mark]

Can we get a dumbed down version please. I have just read the thread for the third time and I am no where near being clear. Do I have this right?

use BLB or SA bulbs
Use more than four and go larger cuz you might want to go bigger later
Put them 3-6 inches above the printing frame

Mike mentioned some kind of ballast-what is wrong with just using the ballasts from a shop light or one of those under the cabinet ballasts that wmlaven was talking about. My head hurts.

 

[clay]

Can we get a dumbed down version please. I have just read the thread for the third time and I am no where near being clear. Do I have this right?

use BLB or SA bulbs
Use more than four and go larger cuz you might want to go bigger later
Put them 3-6 inches above the printing frame

Mike mentioned some kind of ballast-what is wrong with just using the ballasts from a shop light or one of those under the cabinet ballasts that wmlaven was talking about. My head hurts.

Summary: There is no single Right Way to do this. Just a few wrong ones.

 

[photomc]

LOL - Mark, I feel your pain..and even worse, I requested this information from the individuals .. Michael, Clay and Sandy - you guys are great. William, thanks for your input as well. Now how do I explain to the wife that the darkroom needs to be larger so I can use a 12 light 4 foot light source? Maybe I can tell her it's a tanning bed.... Now correct me if the vision is off, but that is quite a bit of space. Guess I'll just have to stick with the 24 inch bulbs. Thanks guys, this was fun.

 

[Michael Mutmansky]

Mike mentioned some kind of ballast-what is wrong with just using the ballasts from a shop light or one of those under the cabinet ballasts that wmlaven was talking about. My head hurts.

There are two reasons I don't recommend the cheap magnetic ballasts. The first is something called ballast factor, which Sandy alludes to in one of his posts. Electronic ballasts will typically output more light with the same lamp, which makes printing faster. This in not always the case, but in general it is correct.

The second is energy efficiency. While I know that we're only talking about a few ballasts here, the electronic ballasts are much more efficient than the magnetic. It is part of my professional and personal beliefs to attempt to conserve energy wherever possible. I don't specify magnetic ballasts, and I don't use them personally in my residence.

Mike asked 'why', not 'how'. 'How' is easy, do 8 lamps, 2" max. OC, and put the print frame glass about 3" below (assuming max of 8x10, and may be suitable for 11x14 depending on your tolerances. Personally, I would do 10 for an 11x14 to be safe.). Put a fan in the unit. Shield the lamps from normal viewing angles. Put a bright white (not specular) metal reflector behind the lamps and tie it into the ground of the fixture. Wear safety goggles when exposing. Adjustments to these recommendations are acceptable within limits.

'Why' is much more complicated.


---Michael

 

[sanking]

" At several workshops, I brought printing units with both BLB and SA bulbs and we always found the SA bulbs faster by about a stop. We tested with new bulbs so bulb age wasn't a factor. Similarly, I had a friend test two unties with POP and she found the SA's faster and more contra sty. We were very careful in our testing to limit all other variables. God knows why the difference.

Ahhhhh, I just noted that your comparison was SA to BLB, not SA and BL? If that is so then I would agree with you that the SA tube prints faster and gives more contrast than the BLB tube with all of the iron processes, though my tests indicate that the difference is only about 1/4 stop in both speed and contrast. The BLB tube is basically the same tube as the BL but it has a filter that cuts out all radiation above about 420nm, and as we know the iron processes have considerable sensitivity to radiation in the near visible.

As to reported differences in finding between the BL and SA tube I can not comment because there is no way of knowing the exact circumstances of the tests and many factors other than the light source itself can have an impact on printing speed. I only know what I personally have found in my tests, which could have some flaw though I have spent a lot of time trying to standardize them. But for anyone interested in doing this kind of tests here are some rough guidelines. Everyone knows the importance of keeping such things as exposure time and distance from the printing frame to the tubes constant and of using the same paper for the test, RH, etc. but beyond the obvious there are a number of other factors that must be considered.

1. Age of bulbs -- Must be the same because there is a significant reduction of light output as tubes age.

2. Same ballast -- Any tests of tubes should be done using the same ballast and the same fixture.

3. Tubes should be cooled with fan -- The output of tubes drops significantly as they heat up to around 95-100º F. This may have more impact on radiation with some types of tubes than on others.

4. The exact sensitizing mix of must be used. Palladiuim is much slower than platinum and even a couple of drops difference in the amount of platinum in the 2ml that is typically used for a 8X10 size print can have a big impact on printing speed. I did some tests of this recently for the appendix on UV lights that I wrote for the new edition of Dick Arentz' book on palladium and platinum printing and I was quite surprised to see what a big difference even minute amounts of platinum can have on speed.

I want to also point out that many of the original claims made for the faster printing speed of SA tubes was based on the use of high output (HO) and Very High Output (VHO) units, which are rated at from 2X to 4X more in wattage than normal output tubes. A regular output 24" tube is rated at 20 watts, while a 24" high output tubes is rated at 75 watts. And of course you need special ballast for the HO and VHO tubes.

In any event my own testing has found the Aqua, SA and BL tubes to be about equally efficient in printing with the iron processes, including cyaonotype, Pt/Pd, kalliype, VDB. For printing with dichromated colloid processes (gum, carbon, etc.) both the BL and BLB tubes print faster than the SA or Aqua.



Sandy

 

[cjarvis]

I would agree with Michael's estimation for number of tubes. I personally built something similar to the rig in Sullivan and Wesse' book: 12 20w BLB tubes, well-ventilated. Perfectly sized for a 16x20 print frame.

My own $.02 worth: build a large box. The convenience of being able to print multiple prints or test strips is worth using a 16x20 frame, and you can always put switches on ballasts to control two bulbs at a time. I print on two 10x12 sheets of paper simultaneously...a huge time and energy (thanks for pointing out the eco-stuff, Michael) saver.

 

[photomc]

Safety Goggles

Wear safety goggles when exposing. Adjustments to these recommendations are acceptable within limits.

---Michael

You know it may be common sense, but this is the first time I have seen any comment about Safety Goggles while using one of these units. Thanks for one more positive Michael. Also, helps explain why an enclosed unit might be safer than an open frame.

 

[Curt]

To bring this long lost thread to date, can I ask what the exposure times are for this method of exposure? Any times for Carbon Transfer? I'm still looking for a deluxe unit but may have to build one of these. I was going to get a single bulb but the wattage was discontinued.
I have seen the Edwards Engineering unit which looks well built, still holding out for a NuArc though.​

 

[sanking]

To bring this long lost thread to date, can I ask what the exposure times are for this method of exposure? Any times for Carbon Transfer? I'm still looking for a deluxe unit but may have to build one of these. I was going to get a single bulb but the wattage was discontinued.
I have seen the Edwards Engineering unit which looks well built, still holding out for a NuArc though.​

There is really no way to answer your question because exposure depends both on the output of the light, the thickness and color of the carbon tissue, the density range of your negatives, and the strength sensitizer that is used.

If you use a NuArc and the B&S regular renaissance black tissue your exposure times should be five minutes or less. Homemade tissue is usually thicker and gives more relief but exposure times are generally around 8-15 minutes with a light source like the NuArc.

I use thick homemade tissue, print with very high density range negatives of log 3.0 or higher, and sensitize with a strong dichromate solution (6-8% dichromate diluted 1:1 with acetone) and print with a ULF-28 platemaker. My exposure times average about 300 units, where one unit = 1.5 seconds.

Sandy King

 

[JLP]

Curt, send me a PM if interested in a unit with 7 24" tubes. I build one a couple of years ago and it has seen no use since i bought a Nuarc.
Comes with an extra tube.