Lighting-Gallery.net
General => General Discussion => Topic started by: sol on September 25, 2017, 09:17:11 PM
-
What would make an electronic MH ballast whine or whistle ? My famous table lamp, fitted with either a Universal or Vossloh Schwabe (I don't remember, but it is essentially the same manufacturer) full size electronic MH ballast (50W) recently started whistling and it is getting very annoying. I'm even considering fitting a magnetic ballast as a replacement.
When time permits later this week, I'll have to try a different lamp to see if it changes.
I also recall one guy (I don't remember who, he is/was a member here) fitted an electronic ballast in an old MH floodlight, and planned on using it as an uplighter in an office with a CDM lamp. The plan was scrapped because his sister was bothered by the whistling ballast.
-
There must be some component vibrating at HF. Power the ballast open and try to listen to each component (most likely coils or capacitors) separately using a PVC pipe
-
Normal operating frequency of these things is far supersonic, co in no way that could be heard.
But when some fault occurs (mainly worn out or failed components; could well be even just the lamp), the main high frequency gets some parasitic modulation and that then turns onto some components well within the audible range and this then causes the whistling noise.
So Whistling ballast definitely means there is something wrong with the system.
Given how most of the PFC circuits in these things are designed, the cause could be the PFC stage being overloaded. And overload of the PFC stage is in fluorescent ballasts very frequently caused by too high voltage drop across the lamp (ballast itself is very close to a constant current source, so higher load voltage due to higher load resistance means higher power). And too high voltage across a fluorescent lamp is either too high cathode drop (so cathode EOL) or inner atmosphere contamination (by an air leak around some seal).
Other reason for the modulation may be the instability of some regulation loop (tend to be caused by some worn out, failing electrolytic capacitor). But that is more plausible for a DC power supply and way less likely for a fluorescent ballast (except dimmable ones) because the only regulation loop in a fixed power ballast is in the PFC stage and that is designed as slow, so drying tank capacitor can not cause it to oscillate (the pole on that node is anyway way beyond the crossover frequency, so its shift up due to capacitance loss can not influence the loop stability anymore)
And sometimes the 100Hz from the mains rectification tends to be the noisiest modulation source. This is either due to a bad design (I have one such fixture) or again a capacitor failure. In any case it does a similar buzz as mains frequency operated magnetics (maybe only a bit "sharper").
-
Thanks to both of you. I will first try with a different lamp and see what I get.
-
MH ballasts put out low frequency wave, on the order of 150..300 Hz
I guess the ballast is made like a switching power supply --- DC bus --- H bridge --- lamp. And if so, the power regulation in the switching part might be based on "feedback ---> impulse / skip impulse" (similar to the usual TL431/optocoupler controlled isolated supplies), so the noise is not from the clean HF but from the modulated work of the supply
-
Could be, I'm not really sure...
Anyways I tried a different lamp and it still whistles. Since I'm not really into ballast surgery, I think I might switch to magnetic and fasten it really tight to the junction box to minimise hum. Although I haven't tried it long term, I think a soft low pitch from a magnetic ballast would be more bearable than the higher pitched whistle from the current one.
-
I guess the ballast is made like a switching power supply --- DC bus --- H bridge --- lamp. And if so, the power regulation in the switching part might be based on "feedback ---> impulse / skip impulse" (similar to the usual TL431/optocoupler controlled isolated supplies), so the noise is not from the clean HF but from the modulated work of the supply
I don't think any ballast is using the pulse skipping method. The main reason is, the ballasts operate with rather high and narrow range power levels and there the pulse skipping has no advantage anymore, it even bring more losses than a regularly operating PWM, mainly when that could well be optimized for quite narrow range of power levels.
With normal power supplies the pulse skipping is used to maintain reasonable efficiency at very light load levels with magnetics sized for way higher power (e.g. a 100W rated TV power supply operating at below 1W when the TV is in stand-by, or an electronic electricity meter internal supply feeding just 100mW, when the smallest practical hf magnetics yields 1..2W power capability). With higher loads the most efficient and in fact cheapest BOM operating mode is some breed of a soft switching semi resonant mode (in fact the simple cheap single transistor selfoscillating charger supplies inherently operate exactly in such mode), a fixed frequency PWM is then the easiest one for a supply design using an intelligent controller IC. Either of these has just the HF components in the spectrum, so normally can not generate an audible noise. Of course, unless their operation is somehow modulated by something with strong audible components, but that is again usually a result of some loop oscillation.
HID ballasts have actually two loops, where the audible oscillation may happen: The PFC stage (usually a semiresonant "critical conduction" PWM mode) and then the final current regulating buck stage (usually fixed frequency or fixed t-off hard switching PWM). The temperature response (making it a constant power or something more advanced) is usually either in a microcontroller firmware (as it does not need to be fast, but needs to process a multiplication to get the real power - very difficult in analog, but easy in digital signal processing) and that cannot fail in a way to just oscillate (the DSP either works or does not at all, but it can not change the frequency response).
Being it HID (this I overlooked in the last post), the noise could well be just the harmonics of the low frequency commutation (the thing, which makes the rectangular output), mainly if it is operating at some higher frequemcy (400Hz or so). Then the culprit could be either broken, or somehow loose core in some of the output coils (either the filter, or the igniter). If the inductance remains within reasonable limit, the circuir remains operating without any external parameter change, just the noise.
-
Could be, I'm not really sure...
Anyways I tried a different lamp and it still whistles. Since I'm not really into ballast surgery, I think I might switch to magnetic and fasten it really tight to the junction box to minimise hum. Although I haven't tried it long term, I think a soft low pitch from a magnetic ballast would be more bearable than the higher pitched whistle from the current one.
I hear ya there... honestly i just let all my lamps Hum, i don't fasten alot of the ballasts, my desklamp,70w flood,1000w, etc etc all have the ballasts just kindof in there, however it is a snug fit manytimes so they dont rattle around, but they do buzz... thats for certain, and i love it, the only ballast I've ever fastened down in anything I've built is currently in the room lamp. in a place i go occasionally, there is a bunch of electronically ballasted fluro fixtures on the ceiling, they whistle annoyingly... i hate whistling fixtures, they hurt my ears, but huming/buzzing is fine by me, it is calming.
-
Yes, a magnetic ballast hum is more tolerable than an electronic ballast whistling. I prefer really quiet fixtures, though. Zero noise is the target, although I like magnetic ballasts and tolerate the hum if it is very low.
In this case, remote ballast is not an option. I believe this is the frequency commutation that Medved refers to.
-
You could always fill the ballast with thermally conductive potting compound that would make it quite.. like MG chemicals 832TC but give up all hope of ever fixing it if any thing goes wrong it's almost impossible to remove later in life, but you can mask off the fuse with a soft silicone which is easy to remove..
-
I have two of these Philips MH electronic ballasts (http://www.lighting-gallery.net/gallery/displayimage.php?pos=-116437) and they both whine. I've always assumed it was normal, though I believe they were both used.
-
What i had observed, the whining sound often comes from lamp itself. Try to change lamp orientation or turn it upside down if it have G12 pin base or try another lamp. The LFSW square wave current have many harmonic frequencies that cause lamp metal parts resonating by magnetic forces. I had also many times being irritated by whine sound and had solved the issue by working with lamp as i wrote above.
-
What i had observed, the whining sound often comes from lamp itself. Try to change lamp orientation or turn it upside down if it have G12 pin base or try another lamp. The LFSW square wave current have many harmonic frequencies that cause lamp metal parts resonating by magnetic forces. I had also many times being irritated by whine sound and had solved the issue by working with lamp as i wrote above.
This can also happen with magnetic gear, when the lamp vibrating at 50hz, causing a mains frequency humming from the lamp. Seen this in a lot of L_V and other lighting collectors Youtube videos.
My tracklight don't hums at all with my unaided ears. Its humming can be heard only through my camera, and during hot restrikes: http://www.lighting-gallery.net/gallery/displayimage.php?pos=-129566
When we lighted a Tungsram TC-100 100W HPS lamp on an Eltam EHID electronic ballast, at Eltam factory, I didn't heard any humming with my unaided ears, but the lamp humming can be easily heard through the camera: http://www.lighting-gallery.net/gallery/displayimage.php?pos=-75282
-
My par38 25watt MH self-ballasted screw in lamp, also whines. Very, very high pitched whine, but it can be heard
across my bedroom easily, and even across my living room, back when I kept it there.
-
I thought that the US Philips CDM-I PAR38 have a HF ballsat, to make it compact so it can't whine since the frequency is higher than the human hearing range (It operates at 20KHz similar to a CFL ballast.
-
Under normal conditions, yes. However, this thread seems to lean towards very small defects in said electronic ballasts which make this whistling sound without otherwise affecting performance. Of course, they're not supposed to emit audible sounds but that would be in a perfect world.
Interestingly, I have started to take notice of this in a more active way and I found something. I have several MH electronic ballasts in use, two of which are full case and three that are small case (intended for small track lights). The bigger ones whistle (although not always) and the smaller ones don't. I wonder if the larger case amplifies the sound, or if it is just a coincidence.
-
I don't know why my PAR38 does this high pitched whine, but it does... although it silences itself a little
bit once the lamp has warmed up to it's original color temperature. Makes it all the more weird to me.
My 3 tracklight MH's all seem to be silent.