Pulse Width Modulation

[Richard CSL]

OK let me start by saying what I am trying to do is very unconventional.

 

I have a 6ch lightfactor dimmer pack. this has internal 6 optoisolators. MOC 3020 or similar.

 

I have 14 cheapo soundlab scanners. to which I have fitted optoisolators then triacs to the lamp circuit to allow dimming.

 

4 channels are piggy backed to 2 scanners each.

 

and 2 channels are piggy backed to 3 scanners.

 

the 2 channels work ok unless I use all of them in which caser I need to dim one channel to about 60 percent to get all heads to illuminate. the 3 channels only work if I disconnect the last 2 heads otherwise all heads turn off. I have measured voltages and ge about 3 volts on each channel when activated at full. As each head is plugged in the voltage drops very slightly.

 

I thought of a solution using a LED parcan remove the led panel and use the PWM of this to drive the moc isolators. this does not work as I get a flicker at all levels. seems to be ramping for no reason I can see monitored this on an oscilloscope and ne evidence of ramping , but there it is on the output. I know this is very unconventional, but I could really do with some help here as the show starts soon.

[KevinE]

Is the drive waveform to the opto-triacs synchronised to the mains frequency?

[oilyspark]

This sounds exactly like the sort of gubbins I get called upon to rig up a few hours before it's needed! I can't see it working from the LED PARs as the PWM will be unlocked from the mains as Kevin suggests. The intensity ramps are beats between PWM clock and mains. Presumably you have a dimmer pack and all heads run from a single phase. That ought to do, at least to get reasonable control.

 

The pattern-sensitive behaviour suggests to me that you have accidentally cut off a power rail to whatever drives the optotriac LEDs. E.g. it can now only see the rail through any un-energised optos, which is impossible when all are on. This occurs with CMOS logic, you can leave a power rail off and so long as one logic pin is near the rail it will attempt to power through that but will lack drive for its outputs. Or perhaps you have sent something other than the proper rail out as the common for your optos.

 

Also, have you disconnected the internal opto LEDs and /or have you provided yours with separate series resistors. If not then you have two LEDs in parallel which will not current share and marginal operation or overload of the driving device could occur. Are your optos of the same type as the internals? So many variables when we can't see the circuit! Dare I even ask why you can't send dimmed AC to the scanners separately?

 

Lucien

[Richard CSL]

Kevin the power supply for each is the same mains power so the wave forms should be locked. but I understand the pwm of the par can may generate its clock pulse internally , hence the flicker.

 

Lucien the scanners can not be run off a dimmer as the internal transformer supplies all the electronics too. The system works for 12 scanners perfectly, well almost as there is a small amount of dimming goes on when all lights are on full. It is where I have 3 scanners (optos) parallel linked that one always goes off when the third one is connected.

[Brian]

...It is where I have 3 scanners (optos) parallel linked that one always goes off when the third one is connected.

There's your problem. You can't parallel optos unless they each have their own series resistor. So you'd need to feed raw DC out of the dimmer to each fixture, then to the opto, then to a series R then out of the fixture and then back to the dimmer.

 

 

E2A...

 

[Tomo]

Lucien the scanners can not be run off a dimmer as the internal transformer supplies all the electronics too.

So break out the supply for the lamp to a separate plug, and plug that into the dimmer.

The electronics stay fed from the internal transformer, only the lamp is disconnected.

- I'm making the assumption that the lamp is mains voltage tungsten, which appears to match with your description.

 

All you need to do this is access to the lamp socket.

 

This is how external dimmer moving lights like the MAC TW1 and VL1000T function.

 

- I am confused as to why you're bothering though. Surely it would be cheaper and more reliable to use scanners that have their own dimming system?

[KevinE]

Where are you getting the PWM from? I think Im confused what you're actually doing. As I understand it, you've fitted a triac driven by a random-switching opto-triac (MOC3020) inside your Soundlabs, controlling the secondary 24V halogen lamp supply as it exits the mains transformer. PWM on AC ideally works by switching whole mains cycles on a variable mark-space at the zero-crossing point, rather than the older partial-mains cycle principle used by phase-angle control which comes in part-way through a cycle and commutates at the zero point. Either way, they're best driven by a synchronised waveform aren't they? What generates it? The dimmer pack drive electronics?

If so, you need to arrange for a constant and equal current through each opto-triac emitter; this is usually done if driven by a common single source by connecting the emitters all in series and adjusting the single series resistor to suit. Parallel-connected LEDs / opto emitters dont readily current-share due to production spread and ageing...I think this is essentially how Brian sees it in his post. The other way of course is one emitter, one resistor, common voltage source.

[Richard CSL]

Tomo we have these scanners and they hav an ELC 24 volt lamp internally. they do not have internal dimming, that is why I have altered them to do the job, the dimming is essential as the group is a Pink Floyd tribute, if you have seen their pulse concert videos and know their music you will understand the dimming is essential.

 

The optoisolators are from an old set of dimmers and do already have a 1K resistor in series. I think the problem is the voltage 3Volts over a 30' cable run (mic cable used) and the voltage drop over each opto isolator drags the voltage down to just below the threshold of the isolator I believe. The measured vltage drops with each head added, even if it is on another channel.

 

A concurrent post has been automatically merged from this point on.

 

Where are you getting the PWM from? I think Im confused what you're actually doing. As I understand it, you've fitted a triac driven by a random-switching opto-triac (MOC3020) inside your Soundlabs, controlling the secondary 24V halogen lamp supply as it exits the mains transformer. PWM on AC ideally works by switching whole mains cycles on a variable mark-space at the zero-crossing point, rather than the older partial-mains cycle principle used by phase-angle control which comes in part-way through a cycle and commutates at the zero point. Either way, they're best driven by a synchronised waveform aren't they? What generates it? The dimmer pack drive electronics?

If so, you need to arrange for a constant and equal current through each opto-triac emitter; this is usually done if driven by a common single source by connecting the emitters all in series and adjusting the single series resistor to suit. Parallel-connected LEDs / opto emitters dont readily current-share due to production spread and ageing...I think this is essentially how Brian sees it in his post. The other way of course is one emitter, one resistor, common voltage source.

 

 

The pwm is coming form a 6ch DMX dimmer pack, I am using the feed for the internal optoisolators (disconected) to feed individual regulator boards fitted inside the Scanners. because of the amount of scanners I have paralleled the feed to 2/3 heads on each channel.

[KevinE]

Can the dimmers actually power more than one opto per channel?

 

Maybe if you buffer the signal first. I doubt the dimmer circuits really appreciate their circuitry being extended for 30m outside their circuit boards!?

Most circuitry that is designed for driving long cable runs will have a raft of anti-rf and other circuitry to buffer and isolate any drive waveform before it leaves the board; cables can act as effective aerials for vunerable solid state circuitry.