Sciencemadness Discussion Board

Ace Hanovia Photochemical Equipment

Sauron - 11-4-2008 at 10:37

I am buying a second hand Ace power supply model 7830-60 for $400. This power supply is for the Hanovia high pressure quartz-Hg lamp model 7825-34 or -35, both 450 W output, the -35 has a longer arc length and overall length.

My intention is to buy the lamp new from Ace and use it in a 5 liter prep scale Kriel type photochemical reactor with a quartz immersion tube for the lamp.

Anyone have any hands-on with this sort of kit?

I am trying to get a price for a new 7830-60 from Ace but so far they have been uncommunicative.

I have had good relations with them in the past but it has been several years.

Yes I know all about the hazards of UV and will build an enclosure so I won't have to have my chemistry books translated into Braille.

[Edited on 12-4-2008 by Sauron]

57393.jpg - 32kB

jokull - 11-4-2008 at 12:35

Good acquisition I must say.

I have an Oriel (nowadays Newport) system, with 200W Hg and 500W Hg-Xe lamps, it is not intended for immersion, buit they also sell such kind of lamp.

I was trying to construct an immersion system and got some information on lamps from:

www.heraeus-noblelight.com

Maybe they may provide you the right lamp.

Sauron - 11-4-2008 at 17:52

Thanks, I know that company because they also make D2 lamps for UV detectors (HPLC) - deuterium lamps.

Unfortunateltely they do not appear to have anything comparable to a Hanovia for this application.

My remaining question is more of an EE issue than a chemical one.

This supply is 115V 60 Hz.

I live in a 220V 50 Hz locale. So I would have to operate it on a stepdown (auto)transformer.

Since the power supply itself is simply a transformer with a capacitor-stabilizer circuit, do I need to fret about inductive coupling? I assume I will lose 16% of the nominal output wattage due to the lower Hz. Any other issues?


[Edited on 12-4-2008 by Sauron]

bio2 - 12-4-2008 at 14:26

No, If you can find a 2:1 transformer rated at least 1/2 the load then wire that as an
auto-transformer by connecting the primary and secondary windings.

The "common" line needs to be the neutral conductor.

Sauron - 12-4-2008 at 18:27

The large autotransformer I have is rated 3 KW. It is not a variac, and not a little hardware store brick. It is massive.

bio2 - 13-4-2008 at 08:26

The basics of autotransformers.

http://ecmweb.com/mag/electric_basics_autotransformers/

Sauron - 13-4-2008 at 12:45

Thanks.

I know what an autotransformer is. The company that manufactures the one that I have, makes both fixed and variable autotransformers. The one I have is a large fixed 2:1 autotransformer, cased, fused, 220V in, tapped for both 100 V out and 110 V out, and rated at 3000 W. It weighs about 50 lbs. Would you like a photo?

I have used it for a long time and operated a lot of lab equipment off it, including vacuum pumps, water baths, rotavaps, variacs, heating mantles and temperature controllers from Glas Col and Ace. Never have had any problem.

But, I have not operated a UV Hg lamp PS off it before. Hence my question.

In the end if I am unsure I will buy a 220V 50 Hz power supply from Ace and be done with it.

The Ace power supply is a reactive transformer of the core and coil type, equipped with a capacitor-stabilizer circuit. It is a model 7830-60 for which the primary voltage, mamp voltage etc are published in Ace catalog online. It is a 2.5:1 step-up and I am quite certain, does not come close to the wattage rating of the one I want to connect it through.

The PS has a primary at 100-125 V. an O.C. of 285 V and is powering an Hg arc lamp 7825-34 or -35 with an amperage of 3.6 and a lamp voltage of 125 -145.V. The map is nominally 450 W.

Based on the above I estimate the PS at 1 KVA or less. This is 1/3 of the rating of my stepdown 3 KVA so that is OK.

[Edited on 14-4-2008 by Sauron]

Twospoons - 13-4-2008 at 14:44

Dropping from 60Hz to 50Hz means the inductive ballast will be down to 83% of the required impedance, meaning more current through the lamp. You could counter this reasonably well be using the 100V tap on your transformer. Given the usual safety margins in these sorts of things, I think you'll get away with it - it would be about 10% over nominal current, but mains line voltage can vary by more than that anyway.

Sauron - 13-4-2008 at 16:16

Right. I am awaiting a quote from Ace for their 220V 50 Hz version of this power supply. For some reason they don't give prices on their website and thus far I have not been able to get their online ordering registration system to respond. They do a good job of hiding their email addresses, while touting their fax numbers and WATS line. Someone should shoot their webmaster and drag them kicking and screaming into the 21st century.

Other major players in scientific glassware like Kimble/Kontes and Chemglass publish their prices online. So what's so different about Ace?

Kontes even has an agent here, so I order through them. But Ace has got a monopoly effectively on certain kinds of specialized glassware and associated kit, mainly pressure reactors and photochemical reactors. Some of the glassware I can have duplicated locally. (I have done for years and have never had a failure. The Thais know how to stress relieve borosilicate.) To some extent that reduces the importing to some specialized glassblowing components like special ground joints and Ace-threads.

But with the Hanovia lamps and power supplies there's not much room to manuever. This is also true of the quartz immersion wells. The logic of having the lamp inside the reactor is pretty compelling, because as soon as you go to external UV irradiation you either have to accept the losses associated with borosilicate flask walls or, procure special and costly quartz flasks. Ugh.

Internal lamp also simplifies enclosing the reactor assy in order to avoid eye damage.

Sauron - 14-4-2008 at 11:43

Ace quoted. The lamps are about $600

The 240 V 50 Hz version of the power supply goes for $1900.

Ouch. I think I will stick with the used one in 115 V 60 Hz for $400.

evil_lurker - 14-4-2008 at 12:19

That's ridiculous... Ace is by far the most expensive glassware company out there.

I'm headed down to Florida this week, and thinking of visiting a place that does glassblowing.

Best I can tell, a de-bulbed 20,000K metal halide lamp should fit down into a 45/50 tube nicely.

Going to see if I can have them remove the jacket from some bulbs and rig up some tubing with a cone joint on top.

That way I can use my standard reaction flask heads to convert them over to photochemical apparatus.

Sauron - 14-4-2008 at 18:52

It isn't quite that simple. A 45/59 joint is not large enough, you will need a water jacket round the immersion well. The immersion well and jacket really ought to be quartz. Ace uses a 60/40 joint for this purpose and they do not sell 60/40 joints as glassblowing components. Fortunately, other companies (at least one, that I found, of the majors) do sell 60/50 joints and obviously those will work. Well, I have not checked the ID at bottom of taper, but I am pretty sure they will pass a standard Ace jacketed immersion well.

Alternatively, skip the ground joint and take a page from Ace's book -literally. Use a #60 or #80 Ace-Thred and bushing w/o-ring to hold an immersion well without a joint. This allows vertical adjustment of position.

I do not mind the price of the lamp so much. Ever price a D2 lamp for an HPLC detector?

Once I get my hands on the used power supplies, I reckon I can reverse engineer the 220-240V versions. And if they work all right in 115V on a stepdown, I won't have to.

The cheap alternative is as you say, a quartz halogen bulb. Those bulbs are about $1.50 here, and would fit down an immersion well. Or, in the usual sheet metal holders, they could be mounted to opposite sides of a containment with the glass plates removed, and simply used to illuminate a borosilicate flask. The glass plates in these indoor/outdoor floodlights are for filtering out the UV portion of the output. I use these for inexpensive video lighting, mounted on tripods. The ones I have are 300 W and 500 W and I have spares. The whole damn thng only costs about $15. I am given to understand that 1000W shop lights and work lights can be had. But two 500 W halogen lamps ought to put out enough UV to photolyze a reaction. The literature is full of such setups prior to the advent and dominance of Hanovia.

These things put out a lot of heat. That would be a bit of a hassle for a rxn that needs to be held to RT or below.

I suppose a fan cooling arrangement can be rigged up to remove some of that heat and a half jacketed reactor vessel would do the rest and still have the minimum amount of borosilicate (in upper half) to pass enough of the UV.

Sauron - 14-4-2008 at 23:02

I may have another option.

My transformer manufacturer locally makes variacs in 220V, with amperages up to 30A. The outputs are 0-260 V.

Why can't I use one of these to power the Ace Hanovia 1200 W Hg arc lamp directly?

That lamp has a voltage of 270-300 V and 4.7 A. (Nominally 1200 W).

With the variac say of 5A rating and set at 260V I iught to be really close to 4.7A and 270 V.

http://www.thaitran.com/

Or maybe I can get these guys to build a PS for me with the capacitor-stabilizer circuit Ace uses, and the exact output (285 V +/- 15V, 4.7 A) and I bet it will not cost me the $3700 Ace wants. Maybe slice off a zero.

The Auto-30 is the stepdown I have, largest one they make in 1Ph.

The 5A variac is what I am wondering about.

Photo below

[Edited on 15-4-2008 by Sauron]

variable%20trans.jpg - 11kB

Twospoons - 15-4-2008 at 16:10

Powering arc lamps is not so trivial as simply applying volts. You must be aware that arc lamps exhibit negative resistance once lit ( resistance falls as current rises) so low impedance, constant voltage sources like variacs are out: the moment the arc strikes the resistance drops, allowing more current to flow, dropping R further ... until something breaks. The other point is that arc lamps require a short high-voltage pulse to get them started, on the order of a couple kV or so.
So the simplest supply for an arc lamp is a ballast inductor ( which behaves more or less like a current source) and something like a fluoro lamp starter. The capacitors are usually included for power factor correction - so the grid sees "resistance" rather than "reactance" - which power utility companies dislike for a number of reasons.
The other OTC solution that comes to mind is to use an HID lamp ballast - the sort used for metal halide lamps for stadium lighting. An HID lamp is very similar in its electrical characteristic to your Hg arc lamp.

Sauron - 15-4-2008 at 18:52

Thanks. I received same advice last night from a close friend with experience with short-arc lamps and their power supplies.

So the variac is out - well, it was a happy thought while it lasted.

I am still buying one or two of the second hand Ace supplies, and I am confident that an examination of their design will allow design and fabrication of a 220V primary 285 V O.C. reactive step-up transformer for 50 Hz with the correct capacitor stabilizer circuit, by a local specialist manufacturer, for this application.

Failing that I can just fall back on using these original Ace PS on the 100 V tap of my heavy duty 3 KVA stepdown. As we discussed previously.

Thanks again. I will let you know how this works out.

I just obtained some technical info from Ace.

They describe the power supplies as "constant current reactor ballast" transformers which regulate the amps out to the lamp.

When power is turned on, and before lamp arc strikes, a higher voltage spike of c.325 V builds up across the lamp. This breaks the lamp (Hg vapor) down to light. The voltage across the lamp drops to 25 V @ 4 A and the lamp starts to heat up. It takes 5-10 minutes for the quartz wall to reach c.600 C or more, at which point the voltage across the lamp has reached 135 V @ 3.8 A. This is the normal full operating voltage of the 450 W lamp.

Clearly, this is just what you were talking about, to counter the negative resistance (reactivity) of the arc lamp.

The performance of the power supply can be tested with a digital multimeter (no lamp involved) and compared with the nominal open circuit (O.C.) VAC and short circuit Amps AC (S.C.) out per a supplied chart. The power supplies are matched to particular Hanovia lamp models and are not interchangeable.

The ones I am buying are 450 W and 550 W.

Ace tech brochure attached.

[Edited on 16-4-2008 by Sauron]

Attachment: PhotochemPacket.pdf (752kB)
This file has been downloaded 951 times

Sauron - 16-4-2008 at 02:42

Constant current reactor ballast basics:

http://www.hilltech.com/ballst.html

Standard stuff for Hg medium pressure and HID lamps as it turns out.

Twospoons - 16-4-2008 at 13:41

I have total confidence that you can find a much cheaper, and equally good, lamp supply compared to the Ace one. It is clear they are banking on chemists wanting a turn-key solution, hence the ridiculous price. Your average sparky would just laugh and go elsewhere, with prices like that.

Lumenlab sell a bunch of stuff for people building their own projectors, using standard HID lamp. They have a number of ballasts for sale, at around the $250 mark.

[Edited on 17-4-2008 by Twospoons]

jimwig - 16-4-2008 at 13:42

i was needing too many voltage/current combinations to live in the real hahahaha world.

so took four diodes that i happened to have -each rated 400amp @1600 volts and strapped them to two aluminum heat sink fins and connected them up so as to form a full wave bridge.

that gives me 1600 amps at 1600 volts capacity.

the idea was to use a 20amp variac to dial up an input voltage to achieve the correct (unfiltered) DC voltage out.

i also have been dumpster diving large electrolytic caps for a few aeons and have several high voltage and high capacity models. so that a smooth DC is possible.

i built it but have not used it.

i also have an AC stick welder that will put out around 250amps @ 36VAC to input into the diode array for some real current beyond the variacs 20amp limit. i get excited just thinking about it.

one day i will use it. on what i have no idea. a cosmic battery charger for the light ship maybe.

craZy jiM wGGns

Sauron - 16-4-2008 at 17:22

I looked at a pile of HID ballasts but they all seeme to be ANSI standardized and none was quite what I was after.

Will check out lumenlabs.

I found the original Hanovia part numbers for the Ace lamps which means I can query Hanovia directly about prices for the lamps and whether or not they sell power supplies. Also Hill Technical Supplies in Illinois sells lamps and Hg lamp ballasts, including made to order, so I can RFQ them. Some of their lamps have specs identical to Hanovia;s in terms of lamp voltage and lamp wattage, but generally seem to be higher watts per arc-inch.

A few interesting tips re operation:

It takes 1-5 minutes for lamp to get up to full power.

Once you turn lamp off, you have to wait 15 minutes for lamp to cool back to RT before you can get it to strike arc again.

Lamp life is partially a function of number of on/off cycles. This one I am familiar with from HPLC detector D2 lamps, ugh!

You must cool the lamp but if you overcool it you can mess it up. Even cause catastrophic failure of the main capacitor in the ballast. (Read: BANG!)

As you probably know, the main appln for these Hg arc lamps is curing polymers (plastics industry) and they use arc lengths from 1" to 120" with correspondingly HUGE ballasts (and generally they use a lot of lamps.) These plain vanilla Hg lamps are rather old hat, as they now have Hg "additive" lamps, Super UV lamps and extended life lamps. Also used for curing are metal halide and HID lamps.

The additive lamps are modified UV spectra.

I went ahead and purchased the two used Ace power supplies (450 W and 550 W) because the prices were right. $900 the pair. The originals, $2000 each. But now I have my teeth into this and we will see what we will see. I will report back and let you know.

A typical Hg lamp ballast (constant current reactor power supply) for curing lamps, standardized from 400 W to 40,000 W lamp power.

[Edited on 17-4-2008 by Sauron]

balst.jpg - 15kB

Sauron - 17-4-2008 at 08:42

Turns out there is a Thai manufacturer of Hg lamps, lamp ballasts, UV curing equipment etc in the northern suburbs of Bangkok about 30 minutes away from me. I talked to the boss, he is willing to rebuild my 110V 60 Hz power supplies (ballasts) to 220V 50 Hz. Also to manufacture direct replacement lamps for the Hanovia models Ace uses.

In looking at the Hanovia specs I can tell that these are obsolescent lamps, based on the lamp watts per unit of arc length. The state of the art for standard Hg vapor arc lamps now is 200 W/inch (80 W/cm) while these are typically one fourth of that. In fact I can't find these models in Hanovia's website, even though I got the Hanovia model numbers from Ace. Ace might be selling old stock or else Hanovia is making them up special for Ace periodically based on demand.

-jeffB - 17-4-2008 at 10:58

Quote:
Originally posted by jimwig
i was needing too many voltage/current combinations to live in the real hahahaha world.

so took four diodes that i happened to have -each rated 400amp @1600 volts and strapped them to two aluminum heat sink fins and connected them up so as to form a full wave bridge.

that gives me 1600 amps at 1600 volts capacity.


Um, no. Only two diodes are conducting at any time, and they're in series, not parallel. Four 400A diodes in a full-wave bridge gives you 400A capacity. And you'll need more than a couple of fins to dissipate the hundreds of watts you'd lose at those current levels (assuming their forward voltage drop is around 0.7V).

Sauron - 18-4-2008 at 20:41

I reviewed the Org.Syn. preps that use the Hanovia photochemical system, there are quite a few.

Almost all used the 450 W lamps. I saw one using 550 W lamp and one using the little 100 watter while checkers used 200 W and reation time was not shortened. But in the last instance they both used the borosilicate version of the immersion well, rather than quartz. Go fidure. A 100 W Hanovia only puts out about 11-12 W of US, so borosilicate double walled well would have effectively reduced that to a few watts.

In a number of these old preps little attention was paid to UV safety. Typically, aluminum foil was used to wrap the reaction vessel. I do not consider this anytrhing like good enough. I am already blind in one eye and want to keep the other in some semblence of working order as long as I can. So a totally light-tight containment is in order.

I see in New Ace catalog Hanovia's name has been dropped Maybe Hanovia quit supplying these lamps (they are pretty ancient models) and Ace had to go elsewhere for replacements.

Just supposition on my part.

The quartz immersion wells cost about the same as the lamps (or more.) The one to get is the no-joint but modified, 450 mm OAL version. This has 31mm ID for lamp, and 48mm OD so it will fit through a #50 Ace-Thred. This assembly allows adjustment of well height. The well is of course jacketed for cooling. Cooling water is recirculated at 20-50 C with a thermostatted chiller. Overcooling is very bad idea. For reactions involving lower temperatures special immersion wells are offered.

While Ace also makes immersion wells with 60/40 joints, they are costlier, lack height adjustibility and offer no compensating advantage. Also a quartz/borosilicate joint interface is an invitation to freeze.

Ace offers larger RB reactors, up to 12 L, but, for my money the Kriel 3 L and 5 L reactors are best. These are "european" pattern tapered flat bottomed flasks with #50 Ace Thred center neck for immersion well, two 24/40 side necks and two #7 Ace Thred necks for thermocouple, gas inlet etc. The tapered straight sides keep the substrate proximate to lamp and allow full length of arc to be below liquid level. Reaction times are typically 20% below spherical or cylindrical reactors.

The three lamps that correspond to the power supplies I have are the 7825-34, -35 and -36. The first two are 450 W, last 550 W. The difference is deceptive as the actual difference in UV radiative output is only 25 W or so. When you study the spectral data you will quickly see that total radiated energy is about half of the nominal wattage and of that, about 49% in UV, an equal amount is visible light and the balance IR. So for a rough number, divide the nominal wattage by 4 to get the real UV output. Of course for more accurate data see the spectral chart for particular lamp.

These lamps run HOT. That is why the lamp jacket is quartz, the immersion well is quartz, the electrodes are tungsten etc. The lamp jacket reaches 600 + C in operation.

The quartz must never be touched with the fingers (even when cool) or residual oils can cause jacket failure. Expensive failure. They are cleaned with isopropyl alcohol and dried.

NEVER NEVER operate lamp where it can be seen, unless your ambition is to learn Braille. This is not head shop black light! Not a toy and not benign.

Photocatalyzed chlorinations often monitor end point by appearance of the famous Cl2 yellow green. Old Org.Syn. preps talk about Al foil shrouds on the UV reactor and making a small window in the foil so the appearance of the color can be seen. I wonder how many good chemists got cataracts or retina damage like that?

The proper way is to run a small recirculating loop out of the reactor using a peristaltic pump with a UV-Vis spectrophotometer in line and a flow cell. The chlorine color is easily monitored this way, it's a technique documented in a patent for prep of SO2Cl2. (That reaction does not call for UV.) In the case of the UV chlorination, the loop is long enough that the UV-Vis spec is outside of the containment cabinet.



[Edited on 19-4-2008 by Sauron]

Sauron - 21-4-2008 at 19:33

The Thai company which manufactures UV curing equipment including power supplies, medium pressure Hg vapor arc lamps, capacitors for lamp ballasts etc., has quoted to me for a power supply and a 400 mm lamp so this is directly comparable to the Ace 7830-60 PS and 7825-34 450 W lamp.

$700 US.

Ace wants almost that much just for the lamp, and $1900 more for the power supply.

So I will buy a set from them.

I will also have them rebuild my Ace Hanovia power supplies to 220 V 50 Hz and supply 450 W and 550 W lamps to match these.

This amply illustrates how overpriced and out of touch with reality Ace Glass is.

The Hanovia lamps they sell are not only overpriced, they are obsolete. Medium pressure Hg UV lamps are rated by arc length and watts per inch (or cm) or arc. The bottom end of the range these days is 200 W/inch arc. The old type Hanovia lamps have identical services life (1000 H) but much longer arcs and much lower watts per arc length. This makes for longer lamps of a given power.

Anyone interested in getting into photochemistry would be well advised to buy Ace Glass but look elsewhere for the UV lamps and power supplies.

Also I surveyed the 23 photochemical preps in Org.Syn. that utilized Hanovia lamps. Many used borosilicate immersion wells, rather than the much costlier quartz wells, and many of those that did use quartz wells, used them with borosilicate or vycor filter sleeves (to control which spectral portions predominate). So my conclusion is that it may not be absolytely necessary to start with a quartz immersion well, it depends on the substrate and the reaction.

----------

A few hours later

I took a taxi ride up to see TAS UV Ciring Co.,Ltd. in Pathum Thani and examined their power supply and 400 W portable UV lamp w/medium pressure Hg vapor lamp. They manufacture their own lamps using General Electric quartz tubing from USA. The owner is knowledgeable and experienced, company is 10 years old. They export worldwide. I am quite satisfied and placed an order for one power supply and lamp. I won't be using their handheld fixture (which is for woodworking and flooring adhesive applications) but they will dismount the lamp and replace the leads so I can use the lamp vertically in an immersion well with cooling jacket.

Not bad for $700, new.

[Edited on 22-4-2008 by Sauron]

Attachment: http___www.tasuvcuring.com_equipment.pdf (302kB)
This file has been downloaded 949 times

Sauron - 22-4-2008 at 08:35

It has occured to me that I really do not need three photochemical mercury lamp power supplies...and there is some uncertainty as to the cost effectiveness of retrofitting these 110 V 60 Hz supplies to 220 V 50 Hz.

SO while I will go ahead and buy them, bring them here and get quote on the retrofit, if the cost is not competitive with having new local power supplies built for lamps somewhat more powerful than the 400 W one I just bought, then it makes more sense to just leave these Ace power supplies as they are and have lamps made for them in 450 W and 550 W (direct replacements for Ace/Hanovia 7825-34 and -36) and then sell them off in USA.

If and when that happens I will let the members have first crack at them; failing that I'll sell them on LabX.

Sauron - 30-4-2008 at 21:05

As it transpires, the 550 W Ace power supply turned out to be dead so the seller trashed it.

I had already paid for both, so he refunded the $500 for the 550 W model. The 450 W supply is now on its way.

Meanwhile I bought the 400 W new supply and its lamp from the local manufacturer after visiting them in Pathum Thani last week.

So when the Ace one arrives I will get them to quote on rebuilding it to 220V 50 Hz, we we can take it from there. Maybe I will send it back to USA and sell it with a new lamp; maybe I will convert it and use it here. All depends on the conversion cost.