Your prediction of 8 dim levels is without basis or merit.
Assuming 60hz. Typical phase control the triac must be triggered 120 times per second at a TBD offset after the zero crossing. If the triac is instead triggered exactly twice as often, it will function exactly the same -- the output A.C. waveform will be identical. Same if the triac is triggered exactly 4x as often. Or 8x as often. 8x is 960/sec. If I instead trigger at 1000/sec, that is a max of 4% error when it happens at random instead of synchronized to zero crossings. This error has imperceptible effect (to my eye when using incandescent lighting).
I was wondering why you took down your famed "FAQ". I have a copy of it sucked with Teleport Pro in HTML and the .zip file in RoboHelp format taken from your /images directory you left unprotected.
But they do. I and many others see the flicker from LEDs run on unfiltered 60hz A.C. simply by turning ones head causing eyes to sweep past the LEDs. Many people complain of this with LED christmas lights and night lights.
Fluorescent lighting phosphers are a bit slower, and the 120/sec flashing is much faster, but they also produce easily discernable flicker that bothers many people and is proven to exhaust eyes even for those people who are not consciously aware of flicker.
And I am building dimmers, thus I need to turn the light on and off many times per second and DMX is irrelevent (unless you can offer DMX dimmers in the $5 to $10 per channel range when buying 4 channel blocks).
For the most part. And their slow speed is why a 60hz A.C. dimmer works. But just because the output cannot reproduce the frequency, that is no reason not to have the frequency. E.g. CD audio was (and remains) a useless waste a fidelity for most purchasers of CD content.
The DMX dimmer switches the current on 120 x per second. The only thing that changes is how long it waits after zero crossing to switch it on. It can't switch the current off again in mid-phase. It can't turn the load on and off more times because it has to wait for zero crossover for the current to turn off.
Perhaps not but he and I both know how triac dimmers work and I know you can't switch them on and off more than 120 x per second.
It's not about the software. There are limitations of the hardware itself. If you want to use a triac you're not going to switch it on/off more than 120 times a second, regardless what software you use.
In article , ZsdbUse1+noZs snipped-for-privacy@Zbigfoot.Zcom.invalid (sylvan butler) writes: | On Mon, 23 Oct 2006 23:42:22 -0400, Marc_F_Hult wrote: | > No I am not ;-) What you describe is a very low-resolution way ( max 8 dim | > levels at 1000 hz) to dim AC power using TRIACS or back-to-back SCRs. | | Your prediction of 8 dim levels is without basis or merit. | | Assuming 60hz. Typical phase control the triac must be triggered 120 | times per second at a TBD offset after the zero crossing. If the triac | is instead triggered exactly twice as often, it will function exactly | the same -- the output A.C. waveform will be identical. Same if the | triac is triggered exactly 4x as often. Or 8x as often. 8x is 960/sec. | If I instead trigger at 1000/sec, that is a max of 4% error when it | happens at random instead of synchronized to zero crossings. This error | has imperceptible effect (to my eye when using incandescent lighting).
What (not synchronized to the zero crossing) triac trigger waveform will you synthesize to obtain an approximately 50% dim level?
Now ya dunnit, Dan ;-) This may be the last we hear from sylvan on the issue.
As you know, the independent variable controlling output in a TRIAC-based dimmer is phase or time offset with respect to the zero crossing. But sylvan does not (knowingly) synchronize the trigger with the AC waveform. So sylvan has no adequate answer to your question.
There are at least three possible explanations for his claim that he can control TRIAC dimming with a 1000hz signal not synchronized to the zero crossing but none that I can divine that support the claim of actually deterministically providing many different dim levels without flicker.
1) His actual trigger (control) signal has a significant component of 60hz AC noise ("hum") and so is in fact inadvertently synchronized to the AC line
2) His dimmer circuit skips entire half-cycles. This would happen if his trigger repeated at less than 120 hz. One could, for example, make an approximately 50% dimmed by repeated cycles of a 8.2 millisecond burst of
1khz square wave followed by a low signal 8.4 millisecond in duration, thus only triggering on (approximately) alternate 1/2 cycles. This is why I pressed him on the flicker issue. (We have been assured that his TRIAC-based dimmer system meets rigorous theoretical and empirical tests for flicker.)
3) The dimmer is in his head and hasn't been instantiated.
We disagree, and you also disagree with the physics of the matter IMO.
Right. And because you are not synchronized with the power line, you are not synchronized with the zero crossing and so the TBD (To Be Determined) offset from the zero crossing is not known and *cannot* Be Determined. be
Whatever periodic waveform you use, the dimming value on the first cycle depends on _when_ you start that periodic waveform, right?
And since you are not synchronized, by definition, the dimming level on the first cycle is unknown. Right? And is unknown on the second half cycle, etc.
In other words, your 'approach' does not use the between dim level and delay in sec, degrees or radians.
Sure. As I wrote previously, what happens to the control voltage _after_ the threshold is reached in any given half cycle has no effect on operation.
The dim level is determined completely by the "To Be Determined" offset as you've stipulated. But your system has no way of determining the offset, so it has no (intended ) way of dimming in a deterministic way except by skipping entire half-cycles.
??? Where does 4% come from?
0.4% (not 4%) relates to how close 960 is to 1000 but that's not relevant to calculation of the dimming error.
You've already agreed that dimming is determined by the offset after zero crossing (units = seconds, radians, or degrees), so the calculation of the dimming error by definition looks like:
Percent Error = (TO-AO)/TO *100%
where TO = Theoretical Actual Offset after zero crossing , in seconds, radians, or degrees AO = Actual Offset after zero crossing, in seconds, radians, or degrees What values do/would you use in this calculation ?
What does your TRIAC-based dimmer output look like on a scope?
A dimmer curve for TRIAC dimmers shows the relationship between dim level and offset (in seconds, radians or degrees).
See my spreadsheet at
formatting link
Please describe the relationship you use between your 'waveform' and the resulting _constant_ , non-flickering dim level.
This is not how nor the conditions under which the human flicker fusion frequency is conventionally determined. The fff not a measure of fatigue. Please read what I wrote.
But 'exhausted eyes" are not a measure of the human flicker fusion frequency. You are confounding a formally defined measure (at least it was when I took psychology 35 years ago) with other vision-related phenomena.
But all this is silly. The point here was to assure than if we discovered that your dimmer kinda sorta works by skipping 1/2-cycles, you wouldn't claim that to be acceptable performance.
At first I thought sylvan might just not be explaining himself well, but it now seems more likely that he doesn't understand the fundamentals.
For folks interested in _actually_ creating dimmers using microcontrollers or analog techniques, I created a link to a spreadsheet on my web site that I prepared about five years ago:
formatting link
This spreadsheet solves the equations that provide the numerical values to program the delay-after-zero-crossing for a TRIAC-based dimmer either by bit-banging with a microcontroller or other processor including a PC or with an analog comparator.
It provides tables and graphs of:
0-10vdc control signal, Vdc DMX512 0-255 value, also sequence number for dimmer steps in dimmer curves Fraction of AC cycle Phase angle, degrees Phase angle, radians Delay after zero-crossing, milliseconds Output voltage, Vrms Relative luminous intensity for a typical tungsten lamp, %
Email me if you want the .xls version (a .pdf is posted).
'Course neither the equations nor the data has much meaning for sylvan who claims to create high-resolution, flicker-free control of TRIAC dimmers _without_ synchronization to the powerline with a 1000hz fixed-frequency trigger!
One purpose for my recent posts on this topic has simply been to provide an antidote to the mis-information posted so that others don't waste their time.
But controlling 110/220 volt AC also needs to be done SAFELY and in compliance with applicable building codes. Folks that might be mislead into homebrewing sylvan's (apparently) hypothetical nonsense might also create potentially unsafe hardware which could be disastrous -- especially with holiday lighting which is typically installed outdoors and in a temporary fashion.
So ... *Please* be careful and know your limitations.
That said, folks that are interested in _actually_ dimming TRIACS, back-to-back SCRs or SSRs based on them may be interested in a spreadsheet that I created about 5 years ago.
formatting link
This spreadsheet solves the equations that provide the numerical values to program the delay-after-zero-crossing for a TRIAC-based dimmer either by bit-banging with a microcontroller or other processor including a PC or using analog techniques.
It provides tables and graphs of:
0-10vdc control signal, Vdc DMX512 0-255 value, also sequence number for dimmer steps in dimmer curves Fraction of AC cycle Phase angle, degrees Phase angle, radians Delay after zero-crossing, milliseconds Output voltage, Vrms Relative luminous intensity for a typical tungsten lamp, %
Email me if you want the .xls version (a .pdf is posted).
I don't know up front, that is a software calibration issue related to the lamp being controlled and visual perception.
If you meant 50% power level rather than 50% light level, then I would synth a 50% on-time, and know that with a 4% error limit I would be close enough. But again, that is power, not light.
Depends on dim level. For 100% bright the output voltage waveform is essentially complete (a bit of noise near 0-crossing). For 0% bright there is no output voltage. At 50% duty cycle the light appears significantly dimmed and the voltage applied to the light shows apparently random parts of the complete sine form with a rare complete half-cycle.
For typical triac dimmers the output waveform shows a relationship between voltage (or current) and time. This does not have any linear or easily calculable relationship to dim level.
(1000-960)/1000 gives 4% error in frequency. Or I thought it did...
It does? Hmm.
Oh, I should clarify that right now I'm not frequency limiting to 1khz.
1khz is my lower bound on the design, not to be confused with "will operate at 1khz" and definitely not an "up to 1khz" target. The only limits I have on output waveform right now are duty cycle of the output and how often the processor can repeat the synth loop.
In article , ZsdbUse1+noZs snipped-for-privacy@Zbigfoot.Zcom.invalid (sylvan butler) writes: | On 28 Oct 2006 20:33:23 GMT, Dan Lanciani wrote: | > In article , ZsdbUse1+noZs snipped-for-privacy@Zbigfoot.Zcom.invalid (sylvan butler) writes: | >| If I instead trigger at 1000/sec, that is a max of 4% error when it | >| happens at random instead of synchronized to zero crossings. This error | >| has imperceptible effect (to my eye when using incandescent lighting). | >
| > What (not synchronized to the zero crossing) triac trigger waveform | > will you synthesize to obtain an approximately 50% dim level? | | I don't know up front, that is a software calibration issue related to | the lamp being controlled and visual perception. | | If you meant 50% power level rather than 50% light level,
I didn't, but indeed the 50% power level question is more appropriate for discussion purposes as it does not require knowledge of the light output to power mapping.
| then I would | synth a 50% on-time, and know that with a 4% error limit I would be | close enough. But again, that is power, not light.
Power at 50% is fine, but can you describe the actual waveform that you would synthesize to trigger the triac?
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.