HMIs question about head noise

[Patryk_Rebisz]

when i plug EB (as opposed to MB) into my LTM PAR head and LTM fresnel head and the heads start making high frequency noise (fresnel having DE bulb makes way more noise). I understand that it's because of the high frequency at which the ballast drives the bulb. Heads make noise both at 24fps setting and flicker free (at flicker free they make a bit more noise). Is it normal? Or is it an issue with my EB (i don't have another electronic ballast to test out)?

 

[Cool Lights]

Head noise is a very normal thing where any HMI electronic ballasts are involved. Its a tradeoff for not having flicker. The flicker free setting is a higher frequency drive of the bulb setting so it follows logically that it produces more noise. Magnetic ballasts will be quieter but of course are more prone to flicker.

Its really the bulb that's making the noise, not the head itself. I think the bulb makers could do something about this if there was more push on them to do so. When they first made HMI bulbs, the only ballasts were magnetic so it wasn't an issue back then. Some brands make more noise than others. Osram and Philips are going to be about as good as you can get with the other cheaper brands being more noisy.

 

[Patryk_Rebisz]

Thanks Richard, this is exactly what i observed:

2 heads connected to mag ballasts = slight hum from ballast's transformer and flicker when fridge turned on but no head noise.
2 heads connected to electronic ballast = no flicker but head's noise.

Do you think that lauded head noise of fresnel has to do with it housing DE bulb or because with large empty space it acts like a large drum probably exaggeration the noise?

 

[Cool Lights]

I am sure that some fixtures will exaggerate the noise more than others because of acoustics or whatever. The bulb is resonating at such a high frequency though, its hard to muffle it out. One problem you have is a fully air tight and sealed fixture is one that will reduce bulb life somewhat because of greater heat buildup in the head. The fixture really ends up being a heat sink for the bulb, to keep it alive longer. The smaller the fixture or more air tight, the greater the chance the bulb won't go as long as with a larger and more ventilated fixture.

 

[j1clark@ucsd.edu]

Head noise is a very normal thing where any HMI electronic ballasts are involved. Its a tradeoff for not having flicker. The flicker free setting is a higher frequency drive of the bulb setting so it follows logically that it produces more noise. Magnetic ballasts will be quieter but of course are more prone to flicker.

Its really the bulb that's making the noise, not the head itself. I think the bulb makers could do something about this if there was more push on them to do so. When they first made HMI bulbs, the only ballasts were magnetic so it wasn't an issue back then. Some brands make more noise than others. Osram and Philips are going to be about as good as you can get with the other cheaper brands being more noisy.

This 'noise' is present in any high frequency (within the human audio range...) circuit. Where I hear it most is when I have a number of serial ports (you know ancient communications method...) connected and operating at say 9600 baud... despite my 'age'... I can still hear them...

Back in the olden days when I made 15-20 KHz 'inverters' either driven or 'self-starting'... both types would sing, and depending on the load, have different levels pitch changes... these circuits were in the '20 mA' category... so I could only imagine what a loud noise a switcher controlling 10 amps or so would do...

Different housings would probably create resonant cavities and 'amplify' the sound...

 

[Cool Lights]

If you're talking about within a circuit, I have found the noise can be taken care of by choosing the right components. Namely inductors, transformers and especially filter capacitors. For instance, I have an 7" LCD screen for one of my cameras that makes a bit of high frequency noise when dimming it. I would bet that if I opened it up and replaced the filter capacitors with better ones (and sometimes larger ones too), the noise would most likely go away.

Engineering a bulb so the glass envelope doesn't resonate so much is a much greater challenge. We worked with a Taiwanese company back in 2008 to investigate this problem. Their bulbs were much louder than Osram or Philips so we both viewed it as an opportunity to learn something. They never were able to change it too much so I would guess it lies in some choice of materials for example that together don't resonate as much.

 

[Movie Mike]

The flicker free setting is a higher frequency drive of the bulb setting so it follows logically that it produces more noise. Magnetic ballasts will be quieter but of course are more prone to flicker.

I’m afraid Richard is wrong. Electronic ballasts are “flicker free” not because they operate at a higher frequency drive, but rather because they square off the power sine wave to the head, which causes an increase in the duration of the HMI arc so that the light is on more than it is off. Called “Square Wave” ballasts for this reason, a light powered by an electronic HMI ballast will put out 10% more light than the same light powered by the sine wave of a magnetic ballast. The down side to Square Wave ballasts is that the sharp angle of the power waveform as it starts to descend from its’ peak level causes the bulb quartz envelope to resonate. Arri ballasts offer a “Silent Mode” setting where the ballast rounds off the corners of the square wave to reduce the resonance. While still “Flicker Free” at most shutter speeds, you will see flicker at extremely high shutter speeds in “Silent Mode.”

For a detailed explanation of how HMIs operate see the companion website Harry Box has established for the Fourth Edition of “The Set Lighting Technician’s Handbook.” You can access the site, called “Box Book Extras,” with the password “set lighting.”

Mike Morgan

 

[Michael Carter]

Hmmm... while we Have Mike and Richard - our two resident HMI fiends - on hand...

I think of HMI as ballast, cable, connectors... ignitor, wiring, bulb-base, bulb... and fixture. A bunch of integrated parts.

How adaptable are the components of the fixture itself? Can, say, a fresnel head be disassembled and made into something like a space-light for instance? If the ignitor (and safety switch and base) are discreet components, wired together... can they be disassembled and made into "something else"? Could a fresnel be built into a large par or open face, with a decent/sturdy fixture that could be modded? (I really do like the idea of an HMI bag or space light...)

Not like I have a master-plan in mind... something I think of as I drift off to sleep.

 

[Patryk_Rebisz]

The follow up question:
it seems that the bulb affects how loud the buzz gets, would the ballast have any effect too - namely would one EB ballast produce more buzz than ballast of another maker?

 

[j1clark@ucsd.edu]

The follow up question:
it seems that the bulb affects how loud the buzz gets, would the ballast have any effect too - namely would one EB ballast produce more buzz than ballast of another maker?

I think that was the 'take away' message from Richard's statement. While I've had a lot more experience with electronic components, and in many cases a choice of a 'few cents' on a part can make a difference... and some manufacturers select things on 'a few cents worth of difference'... apparently the materials that make up the 'bulb' may also affect how sensitive the resulting bulb is to resonance.

For Michael Carter...

As for 'frankencircuit' building... sure if the components are components, you can 'build yourself something'... one piece at a time... However, I tend to not advocate this for 'high powered' circuits that are used to power lights... Most reputable light manufacturers have a slew of safety certifications, which... well... take the completed device as a whole... taking it apart and reassembling would most likely result in something that was 'unsafe' unless one was very much aware of the safety requirements.

 

[slondon]

The beauty of the square wave supply to an HMI is that it eliminates flicker (really strobing) by keeping the current flowing to the globe almost constant instead of the big peaks from a choke type (magnetic because the have a big inductor in them with an iron core) ballast. The problem with a square wave is that it contains all the odd harmonics of the fundamental frequency so it mechanically excites resonances (meaning parts vibrate like a tuning fork) in the audible range all over the place in the head with each rise and fall. Silent mode does try to tame these with a slight round-off of the shoulders of the wave but the tradeoff is re-introducing flicker which is what is being engineered out in the first place.

Make sure everything in the head is tight and if you run in silent mode look at some test footage to be sure you don't have uneven illumination of some frames.

 

[slondon]

I've got the books around here somewhere but I can't seem to get into Box's extras site anymore. Anyone know what's up with that?

 

[Michael Carter]

As for 'frankencircuit' building... sure if the components are components, you can 'build yourself something'... one piece at a time... However, I tend to not advocate this for 'high powered' circuits that are used to power lights... Most reputable light manufacturers have a slew of safety certifications, which... well... take the completed device as a whole... taking it apart and reassembling would most likely result in something that was 'unsafe' unless one was very much aware of the safety requirements.

Indeed, the safety-mindset. That's how I ended up with three kids instead of two. (But hey, number three's in film school and loves watching sci fi flicks with his dad...)

 

[Cool Lights]

On most ballasts, ours included, there are 3 settings: 50hz, 60hz and what is commonly called "Flicker Free". The 50 and 60 hz are sine wave outputs that mimick a magnetic ballast running in a 50 or 60 hz environment. These are quieter mostly because they are lower frequency. The Flicker Free setting, on our ballast and many others I know of runs at 75hz and is louder not only because its running a square wave output but also because the frequency is higher. Yes the square wave insures the bulb is on longer and yes the square wave guarantees a louder noise as well in the bulb. But...

If someone were smart enough to know what to change and they changed the output frequency of the flicker free setting, they would find that its even higher pitched and noticeable bulb singing. We've done this on other ballasts running as high as 300 hz and there is definitely a noticeable difference, at the same wattage (I'll cover how wattage output affects things in a moment), in noise from the bulb. And not only does square wave guarantee longer on time and less flicker for the bulb but higher frequencies guarantee even more on time so frequency does indeed not only affect whether your camera sees flicker or not and also guarantees a more noticeable bulb noise.

In terms of wattage output, you can think of that as amplitude in terms of noise. The higher the wattage, the more the resonance on the bulb and therefore louder noise at any given frequency.

Patrick: Its possible that one ballast would produce more noise than another. Certainly the ballast itself may have some noise of varying kinds: fan, hum or buzz from internal components like transformers, etc. For the most part though, I would say that if you have two different ballasts running the same type and wattage of bulb at the same frequency, you'll probably not notice much difference as long as they both have similar wave outputs, sine or square, as well.

Michael: If you know what you're doing, you can transform a fresnel to whatever, but its a lot of trouble and you really do have to know what your doing as we are talking about thousands of volts at startup and that takes special wire, bulb sockets, connectors, etc. Not to mention UV shielding issues for whatever bulb you are using. All that to guarantee no problems and safe running around your sets.

 

[Patryk_Rebisz]

Thanks guys - lots of wisdom.

Last question, i promises, is there something specific that would cause mag ballast to cause the light to flicker? (besides not shooting with mag ballast's flicker-free frequencies)

 

[Cool Lights]

Probably not likely but anything is possible if there was some malfunction in part of the circuitry. Normally though they'll run pretty solidly at 50 or 60hz and therefore you proceed accordingly with whatever shutter speeds work in those frequencies.

 

[Movie Mike]

The Flicker Free setting on our ballast … runs at 75hz and is louder not only because its running a square wave output but also because the frequency is higher.

I guess that is another reason not to buy Cool Lights HMIs – they are noisier in flicker-free mode than other HMI electronic ballasts.

On most ballasts, ours included, there are 3 settings: 50hz, 60hz and what is commonly called "Flicker Free". The 50 and 60 hz are sine wave outputs that mimick a magnetic ballast running in a 50 or 60 hz environment. These are quieter not only because they are sine wave but because they are lower frequency.

I can’t imagine what ballasts Richard is referring to since, outside of special Power Gems and Arri high speed ballasts that operate at 300Hz and a 1000Hz (respectively), Arri, Power Gems, and Power-to-Light (who between them manufacture 90% of the ballasts used) all generate a Square Wave at the supply frequency (60 or 50 Hz) in Flicker Free mode.

Silent mode does try to tame these (bulb resonance) with a slight round-off of the shoulders of the wave ….

In silent mode, the ballast does not “mimick (spelling error is Richards) a magnetic ballast running in a 50 or 60 hz environment (sine wave outputs)” as Richard erroneously contends, but rather because the ballast slightly rounds-off the shoulders of the square wave as Slondon correctly states above.

If anyone questions the veracity of Slondon & my claims, I would suggest they see the companion website Harry Box has established for the Fourth Edition of “The Set Lighting Technician’s Handbook.” On it, he has several good articles on how HMIs operate and how to operate them reliably on small generators. You can access the site, called “Box Book Extras,” with the password “set lighting.” Arri also makes available on line a good pdf written by John Gresch, Vice-President of the Lighting Division of Arri USA and frequent instructor on HMI lighting technology for IATSE Local 728. It is available at
http://www.arri.com/fileadmin/media...lasts/Examining_A_Powerful_Light_Source_4.pdf

… we have a special on DVXUser: Everyone is Right. .

If we are all right (Richard, Slondon, Harry Box, John Gresch, & Myself) then Cool Lights HMI ballasts are in fact louder in flicker-free mode than most ballasts because they operate at a higher frequency of 75Hz.

Movie Mike

 

[Patryk_Rebisz]

Guys, are we dealing here with difference in terminology: rounded off square wave = sine wave or altogether different technology on how to create silent mode on certain EBs?

 

[j1clark@ucsd.edu]

Guys, are we dealing here with difference in terminology: rounded off square wave = sine wave or altogether different technology on how to create silent mode on certain EBs?

There's 'rounding' and then there's 'rounding'...

But a square wave comprises a 'large number of sine waves', by 'filtering' one can obtain the sine wave that matches the square wave frequency (let's not quibble on the details...)...

One can also create a 'triangle wave', where the slope is constant on the intervals... or a 'trapezoid' wave, where the slope rises to a maximum, the maximum is held for a period, then slopes down... This is achieved by limiting what is known as the 'slew rate' of the driving circuit.

In any case, a 'sharp edged square' wave will contain more energy in the various 'sine waves' that make up the 'square' wave, and a device that resonates at those frequencies... will resonate... by filtering the square wave less energy is available for resonance, except at frequencies that are related harmonically to the fundamental, or harmonics that are not filtered out.

Here's a little animated gif showing the result of adding frequencies, starting from a 'sine wave' and ending up with a square wave...

 

[Movie Mike]

Guys, are we dealing here with difference in terminology: rounded off square wave = sine wave or altogether different technology on how to create silent mode on certain EBs?

A rounded off square wave is not the same as a sine wave. While I am very impressed with j1clark@ucsd.edu’s animated gif, I am afraid it only confuses the issue. The square wave and slightly rounded square wave of flicker free ballasts that drive the lamp head are very different than the “flat topping” of the supply voltage power waveform that the harmonic currents these ballasts kick back into the power stream can create (for a detailed explanation of the difference see Guy Holt’s excellent on-line article on the use of portable generators in motion picture lighting.) Essentially a choke coil, magnetic ballasts feed the lamp a sine wave of the frequency of the power supply. Electronic ballasts are a type of Switch Mode Power Supply (SWPS) that takes the supply power sine wave, converts it to DC, and then uses IGBTs to convert it back to a square wave or slightly rounded square wave in the case of silent mode.

Movie Mike