Need some advice...UPB Lights pulsing
- Thread starter tasandkrs
- Start date
Desert_AIP
Senior Member
Those 40V spikes occur in the transition just prior to the zero volt regions chopped by the triac, not the zero crossing point of the sine wave.
Desert_AIP
Senior Member
Page 10 and 11 describe the UPB pulse positions.
http://www.pulseworx.com/downloads/upb/UPBDescriptionv1.4.pdf
http://www.pulseworx.com/downloads/upb/UPBDescriptionv1.4.pdf
After spending 10's of $1000 on UPB, this really is killing the WAF. I wish there was a solution to fix this. It's in such a large scale in my home. My entire house is LED, so it happens with every light. The simple solution is not dimming a light, but with so many lights in my house it's too bright to have them at 100%.
Yeah, the only solution would be a coil or cap after each switch. Just there isn't much room in switch boxes, so the first light is prob the best bet. Not sure what's going to work better. The cap should soak up the additional voltage when dimmed, and spread it out. But still might see a bump in light. Just smooth do it may not bother you. The reactor, should do similar, in that the extra shot would be soaked up in increasing the magnetic field then put back in the circuit. As the volts go down. Try both. Should be able to track down a cheap cap. I don't know the size you would need. Not very large. Just ensure its 120vac. Reactor shouldn't matter.
Another solution would be a 1:1 isolation transformer. That should soak it up aswell. But where do you install that? Start to run into box issues. As it would be the same size as a doorbell transformer.
Have you. Contacted your switch provider to see what they think?
Another solution would be a 1:1 isolation transformer. That should soak it up aswell. But where do you install that? Start to run into box issues. As it would be the same size as a doorbell transformer.
Have you. Contacted your switch provider to see what they think?
What puzzles me about this problem is that LEDs are not dimmed in the same way as incandescent bulbs are, LEDs can be dimmed in two ways - through pulse width modulation or through reductions in the current to the LED. The LED bulbs contain a small power supply circuit that translates the incoming AC waveform into a signal that will dim or brighten the LED. So even though the UPB signal uses 40V pulses on the AC waveform, this should have no effect on the dimming of the LED.
It really seems to me that the power supply circuit in these particular LEDs is a marginal design that isn't doing a proper job of controlling the dim level of the LEDs.
It really seems to me that the power supply circuit in these particular LEDs is a marginal design that isn't doing a proper job of controlling the dim level of the LEDs.
It just depends on how well their rectifier works. If its pulse width modulation, and increase in line voltage, will directly translate to an increase in output. Even though it's pulses, bigger pulses, with same rate, there will still be more going through. They don't have any form or regulation. That's the issue. So you gotta filter in.
What about something like this downstream of the switch
http://www.hometech.com/hts/products/automation/upb/conditioning/po-ilc.html
http://www.hometech.com/hts/products/automation/upb/conditioning/po-ilc.html
I swapped out one of the HAI CFL/LED switches with a SAI 1140 switch yesterday, and the pulsing is gone!!!! What puzzles me is I went with HAI because they were made for LED lights and the SAI wasn't. Not only does the SAI NOT pulse, it also handles the dimming curve MUCH better.
A related white paper from Lutron on the subject. They implemented a analog filter along with the standard PLL to accurately track the zero crossing: http://www.lutron.com/TechnicalDocumentLibrary/Power_quality.pdf
As a side note (not the problem in this case) - do not add a capacitor to the output side of a leading phase control dimmer. The leading edge turnon of the triac produces high dV/dT into the capacitor and will generate large line spikes.
The following shows the phase controlled output of a dimmer at 100% level feeding a dimmable CFL (red = AC into lamp). The yellow is the AC input to the switch with the 60Hz filtered out. Line in duced spikes due to the CFL capacitance are 23.2 Vp-p.
As a side note (not the problem in this case) - do not add a capacitor to the output side of a leading phase control dimmer. The leading edge turnon of the triac produces high dV/dT into the capacitor and will generate large line spikes.
The following shows the phase controlled output of a dimmer at 100% level feeding a dimmable CFL (red = AC into lamp). The yellow is the AC input to the switch with the 60Hz filtered out. Line in duced spikes due to the CFL capacitance are 23.2 Vp-p.
