Ghetto Monolight, Part III

Due to lack of correct non-polar capacitors, I've decided to revamp my inverter circuit for the Ghetto Monolight. Rather than muck around with scary analog stuff, I decided to drive the power transistor using the trusty ancient LM555 timer chip, my good friend since elementary school. This is the circuit I decided to use:





The driving frequency of roughly 120 Hz is not optimal, since the transformer is a 60 Hz variety, but I didn't have the right combination of resistors and capacitors to produce the right frequency.





In any case, the damn thing works. Not fantastically, but it works. Based on my informal tests, the bare bulb has a Guide Number of about 10 (meters, at ISO 100). Pitifully low, but that's bare bulb (spherical light pattern); it will certainly go up with reflectors.



A more crippling drawback, I measured the cycle time, it takes about 2 minutes to charge the 1100 uF flash capacitor to 270V which is the minimum necessary to fire the flash tube. That's unusably slow.

I'll try to use a bridge rectifier on the high-voltage output, this should halve the charging time because both half-cycles are used. I'll also get closer to 60 Hz for the inverter, and use a power MOSFET like an IRF530 instead of a BJT for switching the transformer (less voltage drop).

But even with all that, I don't expect charging time to go much below a minute. Still unusable.

I thought of this cheap but potentially lethal solution: drive the flash head (tube, capacitor, and dangly bits which trigger it) directly off the 220V line. That's how traditional studio equipment works.

By putting a light bulb (a.k.a. power resistor) in series with the rectifier diode, the current draw is limited to a reasonable amount. And still guarantee quick charging times. But it's scary: circuit failure equals maiming or death. Whereas by powering it off a battery, circuit failure just equals really nasty shock.

Ghetto Monolight, Part II

Last night I built an inverter for my ghetto monolight. Tonight, I've put together a trigger circuit which produces the 4kV pulse to fire the U-shaped flash tube. I got the flash tube and trigger transformer from Electronics Goldmine. Here's a trigger circuit which I scrounged from somewhere off the intertubes:



Here's the completed construction, along with a 1100 uF, 350V capacitor from the broken Vivitar 283:



I also recycled the low-voltage trigger circuit that I constructed some months ago, and the hot shoe foot from the Vivitar. This allows me to fire the Ghetto Monolight from my Pentax K10D DSLR safely.



And here's the flash firing (photo taken with the Pentax K10D):



There's just one, huge problem. The one-transistor inverter circuit is unaccountably inefficient (it's supposed to be powerful enough to drive a 15W fluorescent tube, but this is not what I experienced). The bad news: it takes about a minute for the Ghetto Monolight to charge!

I need to build a more efficient, class-B inverter. But building a working flash is immensely confidence-building.

Ghetto Monolight, Part I

My little experiment with the basket of teeny tiny flashes hasn't worked out. It's impossible to trigger all the flashes using a single trigger circuit (only one or two of the flashes fire). I don't feel motivated to assemble ten trigger circuits worth $3 each (and my labor) to fire tiny GN 7 flashes that cost $0.80 each.

What I've decided to do is use the trigger transformers and U-shaped flash tubes that I bought with the basket of tiny flashes, and make my own (full-power) monolights, using the huge capacitors from the non-working Vivitar 283 hot shoe flashes.

First step is to build a high-voltage power supply.

I used the "Super Simple Inverter" design from Sam's Strobe FAQ Components by Sam Goldwasser.

I made a few minor changes to the design: I substituted an MJE3055T for the 2N3055 in the design (TO-3 transistors are just so bulky).

And since Alexan Commercial was incredibly, out of 1 uF capacitors when I visited last week, I changed the timing circuit from a 1 uF / 4.7K combination, to a 2.2 uF / 2.2K combination. RC time constant is pretty much the same. I used the smallest 220V to 6-0-6 V transformer that Alexan had.



Prototyped it on a breadboard, got about 250V on the output of the transformer when powering it off a 6V gel cell, not too shabby:



I then built one copy on a piece of veroboard. I'll build the other one tomorrow night when I have more copious free time.



Next step, building the trigger circuit to produce the 4000V pulses to fire the flash. Coming soon!

The Quest for Better Bokeh

Beginning lens "devotees" often mistake shallow depth of field for good bokeh. When I first used the Jupiter-9, I was bowled over by its "wonderful bokeh," as compared to, say, the Canon 50mm f/1.8 II or my cheap Canon zoom.

However, it turns out that "good" bokeh, particularly in the OOF highlights, is really hard to get: you want undercorrection of the background, so that the OOF highlights do not have hard edges. Erwin Puts of Leica fame claims that more highly-corrected (read: modern) lenses have harsher bokeh than "old" ones. Certainly, Veijo Vilva has gotten amazing results with an ancient meniscus Kodak VPK Anastigmat from the 1920's. But at the expense of flare galore and low contrast.

I decided to do a simple test on some lenses I have to while away the time.

First a couple of 50mm lenses: the Carl Zeiss Jena Pancolar, and the SMC-A Pentax.

	Pancolar	SMC-A Pentax
wide-open
at f/5.6

Both of them have slight bright-line bokeh wide-open, it's hard to choose between the two, and both have clean but six-sided OOF highlights when stopped down, due to their six-blade irises.

13 December 2008:I've added the 58mm f/2 Helios-44-2 (a Zeiss Biotar copy) which has an eight-bladed iris:

wide-open	at f/5.6

The Helios-44-2 also shows characteristic oval ("cat's eye") OOF highlights off-center when wide-open, and bright edges on the highlights. It cleans up and presents neutral bokeh when stopped down, but the edges on its eight-blade iris are still visible; it doesn't have as many blades as the Jupiter-9 or Jupiter-37AM. In other words, a perfectly serviceable but not noteworthy lens.

Now for two 135mm lenses, the Jupiter-37AM and the lens it was copied from, the Zeiss Sonnar:

	Jupiter-37AM	Sonnar 135
wide-open
at f/5.6

Both have very similar bokeh characteristics, as to be expected, as they are essentially the same lens. I think the Jupiter-37AM has better bokeh wide-open. Stopped down, both are clean, but the Jupiter comes out ahead (again) due to its circular iris.

Finally an unmatched pair, the "famous" Jupiter-9, and the SMC-Takumar 200mm f/4:

As confirmed by other reviewers, the Jupiter-9 has rather nasty bright-edge bokeh wide-open, cleaning up as it is stopped down. So much for being a "Sonnar copy," it certainly doesn't behave in the same way as the Jupiter-37AM.

	Jupiter-9	SMC-Takumar 200mm
wide-open
at f/5.6

The 200mm SMC-Takumar has fairly good bokeh both wide-open and stopped down, but as it has a six-bladed iris, the highlights are hexagonal when stopped down.

For now, I think the Jupiter-37AM is the best of the bunch.