General > General high-speed discussion

driving low pressure sodium lamp without 60Hz flicker

(1/2) > >>

Hi all,
I'm hoping to tap into the collective brain for a problem I'm working on.  I've been using my lab's Chronos for recording turbulence patterns in my soapfilm tunnel.  Here are some test videos I took:

The pressure distribution on the flowing soap film causes light and dark interference fringes when illuminated with a monochromatic light source.  Currently, I use a 55 watt Philips Low Pressure Sodium lamp (  To drive the lamp, I'm using a Fulham workhorse 5 (, as I saw many people had had success using this ballast with LPS lamps.  When I bought it, I thought I would be free from 60Hz flicker because it is advertised as an "electronic" ballast.  Unfortunately (as you can see in the videos), the 60 Hz flicker is alive and well...

I scoped the output to get a sense of what was going on, and I saw the nice ~100kHz waveform I had expected (TEK00110.PNG attached -- don't worry about the voltage scale; it's measured through a capacitive pickup, so the absolute scale doesn't mean anything).  When I zoom out in time I see the amplitude of this high frequency signal is modulated at 60Hz (TEK00111.PNG attached), explaining the flicker.  So, my ballast is indeed an electronic type, but it is passing through the 60Hz frequency content from the mains. 

Initially, I had assumed the circuit in the ballast was something like the block diagram here:
I thought the flicker could be coming from not having enough capacitance after the rectification to sufficiently smooth the signal.  So, I opened up the ballast and de-potted the pcb from the disgusting black tar... and found what looked like the diagram above (image attached).  My thought was just to replace the blue electrolytic capacitors with larger ones.  But, when I (carefully) probed the circuit, I didn't see what I had expected after the rectification stage.  Instead of a near-DC signal, I saw the sine wave clipped into positive and negative pulses before getting sent into the chopper.  So, I didn't know how to proceed...

So, I'm looking for some advice.

1. Does anyone in this forum has experience with LPS lights and their ballasts?

2. Anybody know other electronic ballasts that work with LPS that might have less flicker?  I tried this one (, but I don't think the striking voltage was high enough, as the arc never formed (just a halo around the electrodes).

3. Is there an external power filter module I can put in series with the lamp to filter the 60Hz (and avoid any depotting...)?

4. DIY no-flicker LPS driver? 

5. any other ideas?

The ballast seems to use a valley-fill rectifier to improve the power factor without using an active pfc circuit:

You could modify the circuit by replacing both capacitors with ones having a higher voltage rating (200V), short out D1+D2 and remove D3+D1.
This should reduce the ripple and thus the flicker significantly.
But this will also increase the average dc voltage and therefore the output power slightly.

Another solution may be to simply rectify and filter the AC mains to DC and feed it into the ballast. I would probably use a variac for control, and adjust it so that the lamp is operating at the proper power. Just plugging the circuit into mains directly would likely result in the lamp being overdriven.

I had a similar Problem with my LED-Lights.
Most of them are Pretty Cheap ones, and most of them Flicker a lot.
So i built a litle Box to Plug in between my Lamps and Mains Outlet.
Just a Simple rectifier and Electrolytic Capacitors and Ceramic Capacitors to Filter out all Riple.
Use Big enough Capacitors to get rid of all Voltage Ripple, Size depending on what kind of Power your Lamp has.
I would also recomend adding a Fuse, and a 100K Resistor to Discharge your Capacitors to Avoid Harm from Loaded Caps when unpluged.
The Diagram pictured Below is what i use for LED, donīt Know if that would work with your electronic ballast and your type of Lamps.
If the Ballast unit is not too expensive, and easily replaceable, maybe just try it or Work your Way up with the Voltage with a variac, like Tesla500 said.
Should Work for any Device or Lamp, that Uses DC voltage internaly anyways after a (cheap or poorly designd) rectifier stage.
For me this Works great, and i never had any Problems again with any Flicker on my Slow-Motion-Clips.
Use this Circut only, if you know what you are doing and on your own risk.

Hi all,
Thanks for the input.  I've got a nice system set up now, I'll describe below.

bkt, Thanks for the link to the valley fill circuit.  I had been wondering why there were 7 diodes on the pcb, and it's nice to know the source of the 60 Hz that's making it through the ballast. :)

Tesla500, nice idea -- very simple but I would have never thought of it.  I must say, I'm still a little confused how the DC is making it through the input filter on the ballast.  I thought there would be an isolation transformer that wouldn't let any DC pass... ?  Clearly not...

Nikon1, thanks for the schematic.  What is the transformer in this circuit doing?

Since I'd love to avoid depotting any more ballasts, I opted for Tesla500's solution, feeding an appropriate DC voltage into the ballast.  I didn't have a variac, but I borrowed a high voltage DC supply to test.  Sure enough, with ~100V DC, the ballast produces a constant amplitude 100kHz signal (TEK00115.PNG and TEK00113.PNG, attached) and the lamp lights without any flicker.  For a permanent solution, I ended up buying this power supply instead of building my own rectification circuit:  The controllability and current display is nice, and I'll definitely use it for other projects as well.  I think the appropriate voltage is around 105-110V -- after the lamp has warmed up this puts the power draw near 55 watts.

Here's a soap film video (3kfps) without any flicker:


[0] Message Index

[#] Next page

Go to full version