Convert LD90 from Dimmer to Hot Power

[IRW]

Hey folks,I've recently become head of lighting and sound for a major receiving house, and part of my (long term) brief is to work towards improving the facilities with regards to hosting incoming companies, as many elements that I am responsible for have had very little investment since the venue's last major refit in the late 90s.

 

Ultimately, I'm going to be putting together a phased update/upgrade plan, but there are a few areas in which I'm going to have to apply a couple of sticking plasters in the short to medium term using what I've got available to me. (i.e. non safety-critical stuff, but stuff that would be really, really useful on a day-to-day basis.)

So, here's the situation: I've got LD90's galore in the dimmer room, with 100 channels that are available to a hard patch field in the dimmer room which services all of the FOH areas. That'd be great if I was ever expecting to put 100 dimmable lamps up FOH, but as it is, I'm already being inconvenienced by only having 8 independent outlets available to the patch field (via an LD90 'independents' rack), so it'd be great to be able to repurpose a bank of one of the dimmer racks to provide more hotpower on the patch field. Obviously I'm aware there will be control implications, but this isn't insurmountable.

 

I've actually got plenty of spare dimmer blocks that have kindly been passed down to us from another local venue, so if it's a case of just bypassing the dimming circuitry on the module so that the MCB directly feeds the outputs, then I could set one of these up, and then simply replace the last block of dimmers that appear on the patch field- is this something that anyone else has done in the past, or does anyone know of any LD90-specific reason why this may be a bad idea? I've not actually gone to the trouble yet of opening up our independent rack to see how that's wired, but I can't imagine there's anything particularly clever in there? If I recall correctly from last time I had my head in one of these (about 10 years ago), they're pretty bare bones?

 

Obviously all replies will be read with the usual 'danger danger, high voltage!' caveats in mind! There shouldn't be any mixed-phasing related problems, as everything that comes out on the patch field is all on red (yeah, I know, but that's the colour of the little tag above each channel...!).

 

Thanks,Ian

 

[adamharman]

just bypassing the dimming circuitry on the module so that the MCB directly feeds the outputs

 

If you don't need contol, thats very easy to do.

Assuming all LD90s were the same, they're double pole MCBs so you can just disconnect the wires that fed the dimmer PCB and put the load wiring straight in.

All the zero point sensing etc is done by the processor module which has it's own feed from the incoming supply. The ribbon cable from the processor to the dimmer blocks just feeds into the opto isolators on the dimmer PCB, there is no direct connection to the mains side. I'm fairly certain that even if you unplugged the ribbon, the processor won't notice the difference.

[IRW]

just bypassing the dimming circuitry on the module so that the MCB directly feeds the outputs

 

If you don't need contol, thats very easy to do.

Assuming all LD90s were the same, they're double pole MCBs so you can just disconnect the wires that fed the dimmer PCB and put the load wiring straight in.

All the zero point sensing etc is done by the processor module which has it's own feed from the incoming supply. The ribbon cable from the processor to the dimmer blocks just feeds into the opto isolators on the dimmer PCB, there is no direct connection to the mains side. I'm fairly certain that even if you unplugged the ribbon, the processor won't notice the difference.

 

Thanks Adam, I kind of realised just after I hit post that doing it this way would remove all control...It would be grand to be able to keep some form of remote on/off control, even if it's a case of having to park channels on on the desk, rather than through our handy control panel in prompt corner.I realise this may complicate matters a little!

Edited August 21, 2021 by IRW

[DrV]

1) For individual channels, without any control you could just remove the triacs and put a link across where the MT1 and MT2 terminals of the triac were. Nice and easy as the triacs are on barrier strips.

2) If you can cope with the hard power still being triac controlled then of course you can set them as switched channels on the processor.

3) If you want to go a bit sophisticated and have individual relays controlling the channels from the desk then it might be possible to modify the power module to drive relays. The trigger signal on the LD90 is a steady DC current which switches on at the appropriate point after the zero cross and then stays on steady until the end of the cycle. This means that when a channel is at 100% you get a DC level out of the processor. It might be possible to drive a relay directly from this via an additional switching transistor but I'd have to check how much power is available. Might not even need the transistor but I'd have to check. I'm away until tomorrow night but I'll investigate possibilities if you like.

My test jig just has a load of LEDs on it but I can't remember how much current each channel draws - probably no more than the optos.

 

just bypassing the dimming circuitry on the module so that the MCB directly feeds the outputs

 

If you don't need contol, thats very easy to do.

Assuming all LD90s were the same, they're double pole MCBs so you can just disconnect the wires that fed the dimmer PCB and put the load wiring straight in.

All the zero point sensing etc is done by the processor module which has it's own feed from the incoming supply. The ribbon cable from the processor to the dimmer blocks just feeds into the opto isolators on the dimmer PCB, there is no direct connection to the mains side. I'm fairly certain that even if you unplugged the ribbon, the processor won't notice the difference.

The power board has links which tell the processor what type of module is plugged in so it would notice if you unplugged the ribbon cable.

[IRW]

1) For individual channels, without any control you could just remove the triacs and put a link across where the MT1 and MT2 terminals of the triac were. Nice and easy as the triacs are on barrier strips.

2) If you can cope with the hard power still being triac controlled then of course you can set them as switched channels on the processor.

3) If you want to go a bit sophisticated and have individual relays controlling the channels from the desk then it might be possible to modify the power module to drive relays. The trigger signal on the LD90 is a steady DC current which switches on at the appropriate point after the zero cross and then stays on steady until the end of the cycle. This means that when a channel is at 100% you get a DC level out of the processor. It might be possible to drive a relay directly from this via an additional switching transistor but I'd have to check how much power is available. Might not even need the transistor but I'd have to check. I'm away until tomorrow night but I'll investigate possibilities if you like.

My test jig just has a load of LEDs on it but I can't remember how much current each channel draws - probably no more than the optos.

 

Thanks for the comprehensive reply Dave- the idea here is to be able to provide hot power for touring movers/LED kit in remote FOH locations- does that make option 2 unwise, or would that sort of thing be reliably happy on the end of an LD90 'switched' channel?

[DrV]

Option 2 would work for most devices needing hot power but it would be subject to the distortion that a triac produces i.e. there is still a small off period just after each zero crossing. My humble opinion is that the distortion this produces would not affect most loads but my experience is limited in this respect. Others may have other opinions and I would be only too happy to learn if I am wrong. The way the trigger works on the LD90 is the least worst for using a triac to control hot power because the trigger signal is already there even before the mains cycle begins so there is no delay after zero crossing. This is in contrast to e.g. old fashioned thyristor dimmers (STM, PIP etc) which used a train of pulses through a transformer which didn't start until just after the zero crossing.

 

But I would genuinely be interested in counter arguments if anyone has any?

 

Edit to add: I'll set up some kit to record some waveforms off a real processor and a real power module when I get home

Edited August 21, 2021 by DrV

[Whiskers]

" 1) For individual channels, without any control you could just remove the triacs and put a link across where the MT1 and MT2 terminals of the triac were. Nice and easy as the triacs are on barrier strips."

Are the noise reduction inductors still in circuit by this method?

 

[sunray]

I know its a completely different beast but the principle can't be too different... I added changeover switches to an Anytronic installation which simply bypassed the triac, it was a while ago so I can't remember the exact detail

 

My first inkling was to swicth the outgoing wire between the MCB and the O/P terminal but in the end I did ensure the filters stayed in place.

 

The switches were fitted in an adaptable box under the dimmer rather than the main enclosure.

 

 

 

[dje]

It would be grand to be able to keep some form of remote on/off control, even if it's a case of having to park channels on on the desk, rather than through our handy control panel in prompt corner.I realise this may complicate matters a little!

 

Just as a thought, that with my touring hat on I would always much prefer to use venue hot power if it is not remote controlled and literally switched from a breaker somewhere.

Of course, tours tend to carry dimmers/distro as much as possible but there are always occasions where venue dims/hot power can save hassle... difficult positions ie booms, circle bar, advance truss etc.

But I am always reluctant where these are controlled by house DMX. One erroneous plugging in of something into house data, or recalling a show file, or somebody playing with the house console... and we have unpredictable power loss and a troubleshooting nightmare.

 

So for me personally, total bypass is preferable, even if that's at the expense of losing an electrician for 10 minutes during load in to go up to the top floor of the building an energise the relevant channels.

[DrV]

" 1) For individual channels, without any control you could just remove the triacs and put a link across where the MT1 and MT2 terminals of the triac were. Nice and easy as the triacs are on barrier strips."

Are the noise reduction inductors still in circuit by this method?

Yes

 

It would be grand to be able to keep some form of remote on/off control, even if it's a case of having to park channels on on the desk, rather than through our handy control panel in prompt corner.I realise this may complicate matters a little!

 

Just as a thought, that with my touring hat on I would always much prefer to use venue hot power if it is not remote controlled and literally switched from a breaker somewhere.

Of course, tours tend to carry dimmers/distro as much as possible but there are always occasions where venue dims/hot power can save hassle... difficult positions ie booms, circle bar, advance truss etc.

But I am always reluctant where these are controlled by house DMX. One erroneous plugging in of something into house data, or recalling a show file, or somebody playing with the house console... and we have unpredictable power loss and a troubleshooting nightmare.

 

So for me personally, total bypass is preferable, even if that's at the expense of losing an electrician for 10 minutes during load in to go up to the top floor of the building an energise the relevant channels.

Thinking aloud...

A solution for this case would be a panel of single pole switches, preferably next to the dimmer rack, which short out the triacs. There's no harm in shorting them out and the processor won't notice. However, when the switches were off, the wires (antennas!) would have the high frequency edges from the triacs on them so EMI might be an issue if they were too long.

Edited August 22, 2021 by DrV

[indyld]

Thinking aloud...

A solution for this case would be a panel of single pole switches, preferably next to the dimmer rack, which short out the triacs. There's no harm in shorting them out and the processor won't notice. However, when the switches were off, the wires (antennas!) would have the high frequency edges from the triacs on them so EMI might be an issue if they were too long.

 

I think from a touring LX point of view above, hard bypassing them ON is only half the equation. If the concern is over remote control, then I'd imagine that concern works both ways albeit to perhaps a lesser extent.

Edited August 22, 2021 by indyld

[adamharman]

Another option, if you want control without using the triac to directly switch the output might be to add some contactors. Set the channels to switched on the processor, and use the outputs to feed the contactor coil.

 

I know it's possible to fit additional MCBs onto the power blocks in the gap between the existing ones so rail mounted contactors should fit. The plastic covers are all identical so the centre one has the mcb positions ready to cut out,

[J Pearce]

Given my experience of LD90 processor cards I'd divert the desired non-dim circuits out of the LD90 into a proper distribution board fed by a contactor or multiple contactors. A low voltage control line could be run to somewhere convenient for a keyswitch to power up/down.

 

Worth noting that Zero88 make a dimmer that does pretty much what has been talked about above. The Chilli bypass acts as a dimmer until you push in the 'bypass' button and then that circuit gets hot power - unfortunately with no control from DMX.

[IRW]

Just as a thought, that with my touring hat on I would always much prefer to use venue hot power if it is not remote controlled and literally switched from a breaker somewhere.

Of course, tours tend to carry dimmers/distro as much as possible but there are always occasions where venue dims/hot power can save hassle... difficult positions ie booms, circle bar, advance truss etc.

But I am always reluctant where these are controlled by house DMX. One erroneous plugging in of something into house data, or recalling a show file, or somebody playing with the house console... and we have unpredictable power loss and a troubleshooting nightmare.

 

So for me personally, total bypass is preferable, even if that's at the expense of losing an electrician for 10 minutes during load in to go up to the top floor of the building an energise the relevant channels.

 

Oh yes, I'm totally with you there, I always prefer a switch on the control box wall to desk-controlled non dims. If however the choice is desk control vs. no control, I would accept desk control, and have encountered many venues on my tours where this has been the case.

 

Another option, if you want control without using the triac to directly switch the output might be to add some contactors. Set the channels to switched on the processor, and use the outputs to feed the contactor coil.

I know it's possible to fit additional MCBs onto the power blocks in the gap between the existing ones so rail mounted contactors should fit. The plastic covers are all identical so the centre one has the mcb positions ready to cut out,

 

Given my experience of LD90 processor cards I'd divert the desired non-dim circuits out of the LD90 into a proper distribution board fed by a contactor or multiple contactors. A low voltage control line could be run to somewhere convenient for a keyswitch to power up/down.

Worth noting that Zero88 make a dimmer that does pretty much what has been talked about above. The Chilli bypass acts as a dimmer until you push in the 'bypass' button and then that circuit gets hot power - unfortunately with no control from DMX.

Both good points, but exactly the sort of thing that I need to avoid as my 'little-to-no-cost-short-term' solution. I don't want to get into heavily modding these racks, as long term, I'll be looking to spec a non-LD90 solution, be that simple hot power distro, or something cleverer. Right now, I need workarounds such as DrV's Triac short/bypass method.

For operational reasons, I'd prefer to not have to go to the dimmer room to power things on/off on a daily basis, but if saves the time and effort of installing/stripping out ridiculously long toured looms, it's an idea I'm willing to entertain as a short to medium term solution.

 

 

 

 

[DrV]

If anyone's interested I connected up an LD90 module on the bench this morning and grabbed some waveforms. The screenshots are here

In each case the yellow trace is the drive to the opto, i.e. the trigger, and the green trace is across the load which was a 1k Par can. The opto drive looks a bit weird because of where I've probed it but the important thing is the timing of the leading (low going) edge which is what actually turns on the triac (actually an alternistor but that's a whole other conversation!).

They are as follows:-

ld90_12.png This is just the mains input for reference and to prove that the flat top is nothing to do with the dimmer! Awful isn't it.

ld90_18.png Output set for 10%

ld90_19.png Output set for 50%

ld90_20.png Output set for 75%

ld90_21.png Output set for 100% The trigger is still off for a visible time after the zero crossing so there is a small glitch on both halves of the output waveform

ld90_22.png Output configured as a non-dim. The trigger is now on almost all the time. The off time is only 5uS. However there is still a small (smaller) glitch at the start of the -ve half cycle

ld90_24.png This is zoomed in on the worse of the two glitches when set as a non-dim.

 

I don't think this distortion of the waveform is significant enough to cause a problem with pretty much any load but I'm just a bloke on t'internet.

 

Other points to note - the LD90 has a max output voltage setting which can be configured on a channel by channel basis (or globally). I set this to 250V to get all of these waveforms so that the processor would not be limiting the 100% level. However I was surprised to see that non-dim channels are also subject to this max voltage setting. In other words if your mains input is significantly higher than your max setting, a non-dim channel will still be dimmed slightly. Obviously the solution is to set the max voltage higher than you ever see on the input.

 

All of this is pretty academic now as the dear old things are steadily being retired , especially with the demise of tungsten lighting, but I suspect that there will be a few more having their outputs set to non-dim in the interim. And it kept me off the streets for a few hours! And it gives me an excuse to mention that I still repair the processors...

 

Dave