Post by thread:[OT maybe] Zero Crossover Light Dimmer Impossible?

I wanted to build a PIC controlled zero crossover lighting dimmer, for our church and briefly discussed this a few months ago. I was using a 555 timer triggered on the zero crossover to get my delay. The problem was I wanted 0-10 volt remote control so I could locate my controls in the balcony and the power unit under the stage. (The other problem is, we have no money). The standard way of dimming lights is to turn on a triac some time after the zero crossover of the ac waveform. The problem with this is, the triac turning on causes rf noise (which you then have to filter with an inductor and capacitor). By the way, the TEA1007 chip does a real nice job for this. See the Vellem K8003 DC controlled dimmer kit, about $25 (change triac to 25 amp & add larger current inductor). I then thought, using a PIC I could hit the light with (only one) 1/2 sine wave, say once per second, and increase the number of 1/2 sine waves every second until the light was full on. I could then send (serial 422 from a pc) a light number and a count of 0-256 to tell the PIC how dim or bright. I tried this with a zero crossover triac and a signal generator that I could vary the pulse period and the frequency and found that even when I had 1/2 cycle every say 10 seconds, you could still see the flash of the filament on a 100 watt bulb. I know that a larger bulb, say 1000 watt would not respond as quickly, but I am trying to dim 1500 watts of xmas tree bulbs. (Other than a variac) is there any (cheap) way of zero crossover dimming for this project - am I missing something very basic? I don't mind using the TEA1007 phase shift solution, but I thought industry lighting control for theater, used zero crossover, and I was trying to understand how it worked. tks Mike Montaigne Atomic Energy of Canada Ltd. Station 18, Chalk River, Ontario K0J 1J0, Canada Phone (613) 584-3311 Ex. 4005 Fax: (613) 584-4040 email: spam_OUTmontaignemTakeThisOuTaecl.ca

part 0 1607 bytes content-type:text/x-vcard; charset=us-ascii; name="vcard.vcf" (decoded 7bit) Per your second question (from the other Email) I eagerly design products for people all over the western PA area. My claim to fame is re-developing products that have become aged and parts are getting hard to find. By this point in a products life, the original engineers are long gone. But in general, I design electronics with specialties in processor design (Real big on the PIC obviously), along with skills in analog electronics, switch mode circuit design, and general product design. Meanwhile, I am designing a product or two in the background for eventual sale to the market. Aren't all consultants? Three kids and a wife round out my life to a fast paced clip. Of course, it is only polite if you tell me what you do. Montaigne, Mike wrote: > I wanted to build a PIC controlled zero crossover lighting > dimmer, for our church and briefly discussed this a few months ago. I > was using a 555 timer triggered on the zero crossover to get my delay. > The problem was I wanted 0-10 volt remote control so I could locate my > <etcetera, etcetera> May I ask, tell me something about what sort of stuff you do? > Mike Montaigne > Atomic Energy of Canada Ltd. > Station 18, Chalk River, Ontario > K0J 1J0, Canada Content-Type: text/x-vcard; charset=us-ascii; name="vcard.vcf" Content-Transfer-Encoding: 7bit Content-Description: Card for Christopher Eddy, PE Content-Disposition: attachment; filename="vcard.vcf" Attachment converted: wonderland:vcard.vcf (TEXT/CSOm) (00004842)

On Fri, 17 Apr 1998 15:32:40 -0400 "Montaigne, Mike" <montaignemKILLspamAECL.CA> writes: > I wanted to build a PIC controlled zero crossover lighting >dimmer, for our church and briefly discussed this a few months ago. I >was using a 555 timer triggered on the zero crossover to get my delay. >The problem was I wanted 0-10 volt remote control so I could locate my >controls in the balcony and the power unit under the stage. (The >other >problem is, we have no money). A standard analog method of generating phase control is to generate a negative going ramp that is reset to +10 volts at each AC line zerocrossing. You can then use analog comparators to compare the control voltage with the ramp. If the control voltage is greater than or equal to the ramp voltage, turn on the triac. This technique is used in lots of theatrical dimmers. They use a separate ramp for each power line phase and have a comparator for each dimmer output. {Quote hidden}> > The standard way of dimming lights is to turn on a triac some >time after the zero crossover of the ac waveform. The problem with >this >is, the triac turning on causes rf noise (which you then have to >filter >with an inductor and capacitor). By the way, the TEA1007 chip does a >real nice job for this. See the Vellem K8003 DC controlled dimmer >kit, >about $25 (change triac to 25 amp & add larger current inductor). Yes, the turn-on can cause singing of the lamp filaments. The rise-time can be delayed either by adding a series inductor or by using a slow turn on device. Some manufacturers are using a FET with a slow turn on, but this is substantially less efficient than using a triac (which has a lower voltage drop than a pair of FETs and has either high voltage across it with no current or low voltage across it with high current, both resulting in relatively low dissipation). {Quote hidden}> > I then thought, using a PIC I could hit the light with (only >one) 1/2 sine wave, say once per second, and increase the number of >1/2 >sine waves every second until the light was full on. I could then >send >(serial 422 from a pc) a light number and a count of 0-256 to tell the >PIC how dim or bright. > > I tried this with a zero crossover triac and a signal >generator >that I could vary the pulse period and the frequency and found that >even >when I had 1/2 cycle every say 10 seconds, you could still see the >flash >of the filament on a 100 watt bulb. I know that a larger bulb, say >1000 >watt would not respond as quickly, but I am trying to dim 1500 watts >of >xmas tree bulbs. Your're right. The flicker is going to be horrendous. I think you're going to have to stick with phase control. > > (Other than a variac) is there any (cheap) way of zero >crossover >dimming for this project - am I missing something very basic? I don't >mind using the TEA1007 phase shift solution, but I thought industry >lighting control for theater, used zero crossover, and I was trying to >understand how it worked. The lighting industry (at least where I work http://www.dovesystems.com) uses phase controlled triacs (or solid state relays) with a series choke to limit rise time. Again, some use FETs or IGBTs, still switching at 60 (or 50) Hz, but with a slower rise time, making the choke unnecessary, but suffering low efficiency. I've looked at the possibility of high frequency switching of the AC to give a "solid state variac", but costs are high (four FETs and associated drivers, snubbers, and output filters per channel) and efficeincy is lower than the same circuit with a triac (due to required diode drops, FET or IGBT saturation voltage, etc.). Besides the analog comparator technique discussed above, we also receive DMX512 and generate phase control all in a single 16c74a (see the Shoebox dimmer at dovesystems). Harold _____________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com Or call Juno at (800) 654-JUNO [654-5866]

I'm in the midst of a similar light dimmer project (110V, 500W, Phase delay triac). I'm in a quandry over the value of the series choke. They're either too big and hog board space, too small (seem to have no effect), or too expensive. Best so far is a home wound 35 turn on a surplus torroid core (From http://www.bgmicro.com). I have no specs on the core and 35 turns is downright painful on the fingers. Anyone have any rules of thumb on how much filtering is enough??? Some vendors spec by the milliHenry, others publish current/rise time specs only. Any suggested sources for parts? Joe (Atlanta) Big Snip---------- lighting industry (at least where I work http://www.dovesystems.com) uses phase controlled triacs (or solid state relays) with a >>>>>>>>>>>series choke<<<<<<<<< to limit rise time. Again, some use FETs or IGBTs, still switching at 60 (or 50) Hz, but with a slower rise time, making the choke unnecessary, but suffering low efficiency.

Harold Hallikainen wrote, quoting Mike Montaigne: >> I tried this ... and found that even when I had 1/2 cycle every say >> 10 seconds, you could still see the flash of the filament > Your're right. The flicker is going to be horrendous. I think you're > going to have to stick with phase control. This is exactly what might be predicted. Obviously you (Mike) have never tried experimenting with transmitting sound using a common torch bulb. Quite surprisingly, it really does work, and up to quite a few hundred hertz, perhaps even a thousand or so. OTOH, you are going to be amazed how short you can make your lamp life your way (pulsing them at full voltage from cold) due to inrush current. One of the most useful aspects of phase control is the ability to set a minimum or "start up" brightness level where lamps are pulsed with the last part of the cycle only, say 10 volts, which limits the current *even* with a cold filament to no more than the nominal bulb rating. At this power level you can barely see the filaments glow, yet they will have raised their resistance to most of its full value. Of course, using a PIC controller, you may implement either a minimum power level, an obligatory "warm-up" time during which the phase is restricted to this level, or a "smart" algorithm which performs both; i.e., keeps bulbs warm for a preset period since last use during which time they may be brought to (near- or) full- brightness immediately, after which they will be shut off and when next commanded on, will be limited to the warm-up level for the pre-determined period. >> is there any (cheap) way of zero crossover dimming for this project - >> am I missing something very basic? In a sentence; Yes, you do *not* use zero-crossing for any device that doesn't like being repeatedly turned on and off. > The lighting industry (at least where I work http://www.dovesystems.com) uses phase controlled triacs with a series choke to limit rise time. I'd pay heed to Harold; to those that have done it before. They would be using the cheaper way already *if it worked*! The use of series inductors is most elegant, they should be selected for a rise time of a millisecond or so in combination with the load resistance to be controlled. There should also be a snubber capacitor, *rated for the load current* across the combination of Triac and choke inductor where it interrupts the load circuit. Note that 1} there must *not* be a capacitor directly across the Triac. 2} The inductor actually protects the Triac from surges (such as that due to the snubber capacitor), keeping it within its current ratings. 3} As the thyristor switches itself off at near- zero current, negligible "kickback" is generated by the inductor. If a much smaller capacitor is provided to absorb this, it has a series resistor to limit surges. 4} Most interference generated by the Triacs, if they, the inductors and snubber caps are in a shielded enclosure, will be at audio frequencies. It is picked up by microphone cables and coupling transformers by mutual induction from the mains wiring, and is more audible simply because the higher frequencies are coupled more efficiently. It may also be picked up through un-shielded power transformers in the audio equipment. Most importantly, it can *NOT* be removed by bigger choke inductors in the dimmer, but only by a well- designed and installed audio system. Do not be told otherwise! Cheers, Paul B.

It's kinda tricky. Ideally the choke would be rated in mH with a maximum current limit (to avoid core saturation). However, core saturation doesn't seem to be a real concern here. The chokes for our dimmers were selected before I got there, so I have not seen the design criteria. The ones we are using definitely could not be board mounted. They're too big. Ideally, you could do a simple time constant calculation (T=L/R), but this assumes constant L (which actually varies as the current varies due to core nonlinearity). A higher power lamp actually requires less inductance since the load resistance is less. However, you need a larger wire size and sitll have so many ampere-turns to deal with in terms of core saturation. You might want to call Lyle Nighwonger at Clearwater Technology magnetics division in San Luis Obispoi, CA (area code 805). Harold On Mon, 20 Apr 1998 11:42:00 -0500 Joe Little <EraseMEJoe.Littlespam_OUTTakeThisOuTSCIATL.COM> writes: {Quote hidden}> I'm in the midst of a similar light dimmer project (110V, 500W, >Phase > delay triac). I'm in a quandry over the value of the series >choke. > They're either too big and hog board space, too small (seem to >have no > effect), or too expensive. Best so far is a home wound 35 turn >on a > surplus torroid core (From http://www.bgmicro.com). I have no specs on >the > core and 35 turns is downright painful on the fingers. Anyone >have any > rules of thumb on how much filtering is enough??? Some vendors >spec > by the milliHenry, others publish current/rise time specs only. >Any > suggested sources for parts? > Joe (Atlanta) > > > > Big Snip---------- > lighting industry (at least where I work > http://www.dovesystems.com) uses phase controlled triacs (or >solid > state relays) with a >>>>>>>>>>>series choke<<<<<<<<< to >limit > rise time. Again, some use FETs or IGBTs, still switching at 60 >(or > 50) Hz, but with a slower rise time, making the choke >unnecessary, but > suffering low efficiency. > _____________________________________________________________________ You don't need to buy Internet access to use free Internet e-mail. Get completely free e-mail from Juno at http://www.juno.com Or call Juno at (800) 654-JUNO [654-5866]