SCR driven lights get dimmer as more lights come on

[Wes]

Hi - I'm hoping someone can help me here. The basic problem is
that I have a circuit consisting of a bunch of #47 mini-lamps
that get turned on and off via a processor, and everything works
great EXCEPT that the lights get noticeably dimmer as more total
lights are on at the same time. The lights are controlled in
groups of 4, and each group of 4 that turns on has a relatively
dramatic effect on the brightness of all the bulbs that are
currently on. I would like the bulbs to stay close to their
designed brightness (relatively constant brightness regardless
of how many lights are on).

There are 16 lights total (4 groups of 4). Each group of 4 is
controlled by a shift register (595) output. The 595 output
drives a pair of MCR-106 SCR gates, and each SCR drives 2
bulbs. I have a 1K resistor connected between each 595 output
and the pair of SCR gates it drives.

For lamp power, I'm using a 12.6VCT 4A xformer. I've fed the
center tap and an outer lead to the AC inputs of a bridge
rectifier, so I get the 6.3VDC I need to drive the bulbs -
this works fine. The lamps are 6.3V / .15A, so with 4 amps
available I would think I should be able to drive 16 lamps OK.

I connected a sort of "status bulb" directly across the 6.3V
lamp power supply. I thought I would see it dim along with
the SCR-driven lights, but this one bulb stays relatively
nice and bright.

So, if just 1 group of 4 lights is on, they look nice and
bright. Calling that 100% brightness, when the next group
of 4 comes on, then all 8 lights drop to, oh, 85% bright.
Then when the 3rd group of 4 comes on, then all 12 lights
are about 70% bright. Finally, 16 lights on are all at
about 55% or so.

Anyone have any ideas as to how I can keep all of the lights
near 100% brightness? Thanks!! -Wes

 

[Dana Raymond]

Hi Wes. Thats an interesting design you have there... Why are you using
SCRs?

Anyway, use a scope or meter to see where the voltage drop is occuring.
Also, make sure that your wiring and/or PCB traces are heavy enough to
handle the combined current.

Since the bulb power is 120Hz half wave, the SCR's trigger point may be
affected by small IR voltage drops between the SCR cathode and TTL ground.
Higher SCR cathode voltage will probably result in later turn on of the SCR
(if the trigger level is marginal to start with) and result in lower
brightness.

Using FETs instead of SCRs would lower parts count and eliminate problems
like this.

Dana Frank Raymond

 

[Wes]

Hi Wes. Thats an interesting design you have there... Why are you using
SCRs?

Hi Dana - thanks for the reply! I decided on SCRs simply because I've
become pretty familiar with older Bally pinball machines, and it uses
SCRs to control all of the feature lights.

Anyway, use a scope or meter to see where the voltage drop is occuring.
Also, make sure that your wiring and/or PCB traces are heavy enough to
handle the combined current.

Well, I bypassed the electronics completely and just hooked up all the
lights directly to the transformer in parallel. Same problem - the
more I hook up, the dimmer the bulbs are. I'm using AWG22 solid, and
I'm thinking maybe I should try about AWG16 or so. Any thoughts on
that?

Since the bulb power is 120Hz half wave, the SCR's trigger point may be
affected by small IR voltage drops between the SCR cathode and TTL ground.
Higher SCR cathode voltage will probably result in later turn on of the SCR
(if the trigger level is marginal to start with) and result in lower
brightness.

Well now I am using a full wave bridge rectifier, and I'm not
attempting any phase control with ZCD or anthing like that. Just
straight DC from the + and - outs of the BR.

Using FETs instead of SCRs would lower parts count and eliminate problems
like this.

Worth looking into - thanks again!

 

[Robert Monsen]

Hi - I'm hoping someone can help me here. The basic problem is
that I have a circuit consisting of a bunch of #47 mini-lamps
that get turned on and off via a processor, and everything works
great EXCEPT that the lights get noticeably dimmer as more total
lights are on at the same time. The lights are controlled in
groups of 4, and each group of 4 that turns on has a relatively
dramatic effect on the brightness of all the bulbs that are
currently on. I would like the bulbs to stay close to their
designed brightness (relatively constant brightness regardless
of how many lights are on).

There are 16 lights total (4 groups of 4). Each group of 4 is
controlled by a shift register (595) output. The 595 output
drives a pair of MCR-106 SCR gates, and each SCR drives 2
bulbs. I have a 1K resistor connected between each 595 output
and the pair of SCR gates it drives.

What is the shift register for? Are the lights getting turned on in
sequence? This sounds like a problem with 'duty cycle' rather than voltage
to me. The more groups, the less total time they are on.

 

[Dana Raymond]

OK, so you hooked up all of the bulbs directly to the transformer and they
get dimmer the more you add. Exactly what does that look like? Just bulbs,
sockets, wires, and a transformer? You need to be much more specific so that
we can help you, OK?

Did you make any measurements? Did you know that a transformer rated X Volts
@ Y Amps does not have constant voltage as the load increased from 0 to y
Amps. You need to measure the voltage at the point where all of the lamps
are connected together in common (at the transformer or bridge or whatever)
to see if that is your problem.

You also said that you connected a status bulb across the power supply and
it did not dim. But you say above that you connected all of the bulbs to the
power supply and they did dim. Thats contradictory to me.

Get your meter out and make some measurements, OK?

Also, try the AWG16 instead. The current of 16 bulbs may have a significant
voltage drop if its carried by a single AWG22 wire or wires.

BTW I used to fix pinball and video games for a small arcade chain 20 years
ago.

Good luck, and tell us what you found.