Flash watt.second and capacitor size?

[Chan Tran]

I have this Novatron 240 which is supposed to be 240WS. I found 4 capacitor of 480μF and rated at 400V. So assuming that the voltage will be 350V (I think they don't charge the capacitors to their rated voltage) and using the formula E=1/2CV² so it's only 117J or 117Ws wich is about half of what is rated? Did I make wrong calculations somewhere?

 

[Focomatter]

Is that 120J per capacitor?

 

[Chan Tran]

Is that 120J per capacitor?

my calculation is 120J for all 4. Each capacitor is 480µF so 4 of them in parallel is 1920µF. So expression is:
Energy in Joules = (Capacitance in Farad x Voltage ²) / 2

Energy in Joules= (0.001920 x 350²)/2= (0.001290 x 122500)/2 = 235.2 /2 = 117.6

 

[koraks]

Your calculations seem to work out. 1920uF would be 240Joule only at 500V - way above the rating of those caps. So yeah...sounds like a 120Ws flash, really.

 

[maltfalc]

sounds like they pulled an alien bees, making their model numbers double the capacity.

 

[mmerig]

Can you measure the voltage across the caps when fully charged? That way you don't have to assume 350 volts. Of course be careful of a shock, and use a meter that can handle that voltage and doesn't load the circuit.

 

[maltfalc]

Can you measure the voltage across the caps when fully charged? That way you don't have to assume 350 volts. Of course be careful of a shock, and use a meter that can handle that voltage and doesn't load the circuit.

why bother? whatever the exact voltage is, it's obviously nowhere near the over 700v it would need to be.

 

[Chan Tran]

why bother? whatever the exact voltage is, it's obviously nowhere near the over 700v it would need to be.

I will try to measure it later today. I have to be careful because removing the caps out of the case to measure it giving a lot of chance for something to touch something accidentally. However, using the formula to have 240W.s it only has to be charged to 500V but the capacitors are only rated for 400V.

 

[DWThomas]

Funny, I hardly ever use flash, so I've not thought too much about all this in recent times. The calculated value is obviously the input to the tube, but is it possible the "rating" looks at light output which could be possibly enhanced by the reflector -- or tinkering with the gas in the flashtube to get more light on the target surface? (Somebody out there must know more than I do about the current technology.)

About 65 years ago myself and a nerdy friend photographed .22 rimfire slugs in flight using a spark gap in air held in some glass lab tubing. Not a lot of light, but we got some results (which may be buried in the archives "somewhere" in a dusty pile. That was fun. Charging caps up to thousands of volts, etc. That was in the late 1950s when there were "surplus stores" in Philadelphia with all manner of radar related trinkets for sale cheap. It was enough to win a science fair prize; likely because the judges couldn't fully understand what we accomplished!

 

[Chan Tran]

Well the W.S must be rated at the power to the tube even if there is no light coming out of the tube. That's it.
I know some manufacturers said their low power flash is so efficient that they can produce the same amount of light with less power output so they rate their flash with higher W.s rating but that's cheating.

 

[mmerig]

why bother? whatever the exact voltage is, it's obviously nowhere near the over 700v it would need to be.

The voltage could be much less than the assumed 350, and that could point to a circuit problem. Also, 500 volts would give the 240 W.S, not >700. The caps could be under-sized regarding supply voltage (like close to 400 V), and that would be good to know.

By the way, there could be more convenient places to measure the supply voltage than at the caps. Unless access is really awkward, it's like a 30 second job.

 

[koraks]

The calculated value is obviously the input to the tube, but is it possible the "rating" looks at light output which could be possibly enhanced by the reflector

That reminds me of how computer speaker sets used to be marketed in the 1990s. You could buy a 250Watt set of speakers only to find a tiny 8W voice coil and a 5W amplifier inside. They'd call it "PMPO" output power ("Peak Momentary/Music Power Output") and then get very creative with the maths to arrive at a ridiculously high number.

The voltage could be much less than the assumed 350, and that could point to a circuit problem. Also, 500 volts would give the 240 W.S, not >700. The caps could be under-sized regarding supply voltage (like close to 400 V), and that would be good to know.

Theoretically, maybe. But realistically, nobody, not even a cheapskate engineer or penny-pinching production manager with the ethics of a war criminal, is going to put 400V caps in a 500V circuit. And if the voltage turns out to be much less than 350V, well, it's even much less than a 120Ws flash...the question then would be: is it by design, or is it because of a defect? 99.9% it would be a design choice, because a flash unit doesn't magically settle on a significantly lower voltage and otherwise continues functioning perfectly. That's just not how these circuits work.

 

[Chan Tran]

I got over 900 volts at the connector to the flash head. But the cap is only rated for 400V. I know the caps are wired as 2 banks. You can use 1 or 2 banks. So I guess hey put 2 of the caps in series for each bank. So the total capacitance is back to 480mFD total. With 960V it could get close to the 240ws rating. But over charging the 400V caps to 480V each.
At 960V it's reaching the limit of my meter without a special high voltage probe.

 

[Chan Tran]

Ah. That's a bit of very critical information you left out.



Did you actually measure the voltage over the caps, or just the trigger voltage? Did you trace the schematic (no, because you missed the fact the caps are parallel-series connected...)?



Two caps of 480uF in series is 480uF, but with the voltage rating of both added up provided the voltage balances out, which is generally done with a couple of resistors. Two of those arrays in parallel makes 960uF. 960uF at 700V is about 240J.

2 480uF caps in series is only 240uF. 2 banks of that in parallel is back to 480uF. So it needs 1000V to be at 240ws. No I didn't measure at the cap. To risky, to many thing exposed. I measure between to 2 largest wires to the flash tube. There are 5 pin at the connector. 2 are for the 120VAC for the modeling light. I measured between the 2. When I measure the 5th pin to 1 of the pint connected to the tube it fires the flash so I think that the trigger wire.

 

[koraks]

Yeah, sorry, brain fart.

If they really overcharge those caps it's a piece of c**p design alright.

You'd have to chart the schematic and do some more measurements to figure out what's going on exactly. But I wouldn't bother. Does the unit work? If so, leave it, or sell it off if they really overcharge those caps!

 

[Ian C]

I read an explanation long ago in the Paul Buff literature that indicated that the model numbering system that was used on the White Lightning monolight flash units reflected the idea that compared to a power pack, cable, and head units, the actual light output of the Buff units was about the same.

For example, the Ultra 1200 has an energy storage capacity of about 600 joules but produced about the same amount of light as a 1200 joule power pack, cable, and head unit. The reason for this is that the distance from the capacitors to the terminals of the flash tubes is very short on a monolight. It’s much greater on a power pack and cable system. In fact, on those older systems, power reductions was often accomplished by adding another length of pack-to-head cable.

But the Novatron is a power pack, cable, and head unit, so labeling it as 240 joule doesn’t seem reasonable, based on the information given.

 

[mmerig]

That reminds me of how computer speaker sets used to be marketed in the 1990s. You could buy a 250Watt set of speakers only to find a tiny 8W voice coil and a 5W amplifier inside. They'd call it "PMPO" output power ("Peak Momentary/Music Power Output") and then get very creative with the maths to arrive at a ridiculously high number.



Theoretically, maybe. But realistically, nobody, not even a cheapskate engineer or penny-pinching production manager with the ethics of a war criminal, is going to put 400V caps in a 500V circuit. And if the voltage turns out to be much less than 350V, well, it's even much less than a 120Ws flash...the question then would be: is it by design, or is it because of a defect? 99.9% it would be a design choice, because a flash unit doesn't magically settle on a significantly lower voltage and otherwise continues functioning perfectly. That's just not how these circuits work.

I should have been clearer, but did not think the supply voltage was 500, given the cap rating of 400, but was referring to a statement that the voltage would have to be 700 for a 240 W.S output (post #7). I agree that having 400 volt caps in a 500-volt circuit would be a very poor design, and probably would fail right away. There are cases where even reputable manufacturers have used caps that were over-stressed by voltages near, but still below, the cap rating (most commonly with tantalums).

The OP did not mention whether it was functioning perfectly or not. And he measured much higher voltages than assumed.

Generally, there could be a voltage multiplier circuit that isn't working quite right, or a leaky transformer (inductance, not oil), that could "magically" lower the supply voltage. In old equipment, electrical characteristics can change. Some posts in the large format forum mention the caps in Novatrons need to be matched for ESR and C.

Without a schematic, it takes some measurements to figure out what's going on, and the OP is doing that, and carefully.

Some cautionary info in this post: https://www.largeformatphotography.info/forum/showthread.php?95753-Fixing-a-Novatron

Info on ESR in Novatron caps: https://www.largeformatphotography.info/forum/showthread.php?147903-Novatron-440-Plus-Schematic


Not so simple, and quite dangerous.

 

[Chan Tran]

My bad! I got it right now. The caps are rated at 500V not 400V as I said. The step up transformer secondary is about 350VAC RMS with 120VAC input from the outlet. This voltage is rectified and charge 1 or 2 banks of 2 caps. It would charge 1 cap for half of the cycle and charging the other cap the other half of the cycle. It's a voltage doubler circuit. The fully charge caps has the voltage of 487VDC (this because the peak of the AC cycle is 1.414 times the RMS value). These 2 caps are connected in series and thus gives the total voltage of 975VDC. There is a choice of 120ws or 240ws. If the choice is 120ws then only 1 bank of caps is charged. So the 2 480uF caps in series makes them 240uF. Parallel that to the other bank makes it 480uF again but the voltage rating is now 1000V instead of 500V.

So at 975VDC and 480uF the energy is 228J which is close to the rating of 240ws.

 

[mmerig]

My bad! I got it right now. The caps are rated at 500V not 400V as I said. The step up transformer secondary is about 350VAC RMS with 120VAC input from the outlet. This voltage is rectified and charge 1 or 2 banks of 2 caps. It would charge 1 cap for half of the cycle and charging the other cap the other half of the cycle. It's a voltage doubler circuit. The fully charge caps has the voltage of 487VDC (this because the peak of the AC cycle is 1.414 times the RMS value). These 2 caps are connected in series and thus gives the total voltage of 975VDC. There is a choice of 120ws or 240ws. If the choice is 120ws then only 1 bank of caps is charged. So the 2 480uF caps in series makes them 240uF. Parallel that to the other bank makes it 480uF again but the voltage rating is now 1000V instead of 500V.

So at 975VDC and 480uF the energy is 228J which is close to the rating of 240ws.

Thanks for figuring all of that out. Makes sense. Measurements are usually worth the time.

 

[Kilgallb]

If you want to measure near 1000VDC, find an old CRT HV checker probe. Fluke sold them for years and the work with any 10Mohm DVM.

 

[Chan Tran]

If you want to measure near 1000VDC, find an old CRT HV checker probe. Fluke sold them for years and the work with any 10Mohm DVM.

The meter is rated for 1000VDC so it's OK but it does get too close to the limit. I often use the meter to measure the DC bus voltage of an inverter drive or servo drive that run on 480VAC. The DC bus is normally over 600VDC but at times they can get close to 800V due to the regenerative braking.