So I can run a 26 watt lamp properly on a WH5?

It looks like it. I'll run a 26w SOX-E again tonight to measure the light output with magnetic vs the WH5.

It looks like it. I'll run a 26w SOX-E again tonight to measure the light output with magnetic vs the WH5.

Cool thanks George. What's the purpose of the tape wrapped around the base of the lamp in your photo?

awesome! This will run 55 watt lamps without life shortening consequences?

I'd suspect that the WH5 ballast will die first. The 2 red wires together are just enough to run these lamps without overdriving. I won't try 3 red wires together. Someone here used 3 red for the 90W SOX. 90w lamps need .95 Amps lamp current. He also said the lamp (90w) looked underdriven.

The tape is a wrap made of fiberglass. Running the lamp for a few hours gets the electrode area very hot preventing sodium condensing at the end. I also use this method to dislodge any condensed sodium. After turning the lamp off, I tap it hard to drop any molten sodium down.

I'd suspect that the WH5 ballast will die first. The 2 red wires together are just enough to run these lamps without overdriving. I won't try 3 red wires together. Someone here used 3 red for the 90W SOX. 90w lamps need .95 Amps lamp current. He also said the lamp (90w) looked underdriven.

The tape is a wrap made of fiberglass. Running the lamp for a few hours gets the electrode area very hot preventing sodium condensing at the end. I also use this method to dislodge any condensed sodium. After turning the lamp off, I tap it hard to drop any molten sodium down.

Would doing that to one of my lamps work to get the sodium near the electrodes to move away from them? The sodium isn't in droplets or puddles it's more like a layer of sodium mist on the inside of the tube. I'd like to get all the sodium to move away from the electrodes and combine into droplets or puddles closer to the top of the lamp. Unless it doesn't matter.

What is that tape called if I need it and I'll look it up online?

Running the lamps horizontally for an hour or so will clear the arc-tube area of a layer of sodium film because it gets hot close to the electrodes. I would take care to fire up a lamp that has a thick layer of solidified sodium. Typically I'll examine the lamp prior to firing it up. Then I'd position the thinnest pool above, and the thickest pool below the electrodes.

You can get such rolls, they're called fiberglass insulation, in places where they sell fireplace/woodstove accesories.

Running the lamps horizontally for an hour or so will clear the arc-tube area of a layer of sodium film because it gets hot close to the electrodes. I would take care to fire up a lamp that has a thick layer of solidified sodium. Typically I'll examine the lamp prior to firing it up. Then I'd position the thinnest pool above, and the thickest pool below the electrodes.

You can get such rolls, they're called fiberglass insulation, in places where they sell fireplace/woodstove accesories.

I've had that lamp running for several hours before and the sodium film/specks doesn't seem to go away. Right near the electrodes the glass is free of sodium. But below the electrodes in the bottom of the tube the sodium specks stay there.

I searched LG for that issue. Look at all threads below. I ended up not worrying too much because my lamps are mostly made by Philips and from the millenium on-ward.

2006 post https://www.lighting-gallery.net/index.php?topic=123.0
2010 post https://www.lighting-gallery.net/index.php?topic=1837.0
2011 post https://www.lighting-gallery.net/index.php?topic=2183.0

Cool very good information. I read through all of those topics. Seems like mine should be ok. I might try aluminium foil though anyways just to see if it does anything.

I updated the file info. I compared the light output for 3 different watt lamps.

I would try and find another means than a WH5 to light a 55w SOX. 3red outputs wired together however might be ok.

I have no hesitation to light any 26w SOX-E lamp with the WH5 using 2 reds together. It's also alright for a 35w SOX. A used 35w lamp might be even better.

I updated the file info. I compared the light output for 3 different watt lamps.

I would try and find another means than a WH5 to light a 55w SOX. 3red outputs wired together however might be ok.

I have no hesitation to light any 26w SOX-E lamp with the WH5 using 2 reds together. It's also alright for a 35w SOX. A used 35w lamp might be even better.

Would you say it's ok to use a WH5 to run a 26 watt lamp in an outdoor dusk to dawn fixture like a bucket or a roadway fixture?

I have no problem doing so. I intend to do just that shortly. I forgot to mention that the workhorse5 barely warmed up in all the tests. Lamps were running for 45 minutes to get the light stable.

Edit: I now think that a WH3 (2 reds together, 1 yellow on the other side) might be better for a SOX-E 26watt lamp.

Cool so at least we know we have other options for ballasting SOX lamps. Since SOX ballasts are so expensive anymore.

All SOX lamps equal and below 55w run at 600mA and the bigger ones c90w and above, run at 900mA. So you can run any SOX on either a 600 or 900mA constant current source. The arc voltages they develop will be dependent on the lamp (arc tube length).


George, if you open that white box and draw the circuit of what is inside, then you should get some inkling of just how it does what it does.

If this is a rectifier-crest factor corrector-oscillator-impedance match type circuit, then the output will be a few tens of KHz AC. To measure the arc current, wind a small current teansformer around a ferrite toroid, say 10:1 or 20:1...load its secondary with a known resistor, about 10-20 Ohms and place a precision rectifier in parallel with the load resistor (build with a Op-Amp and, resistors and two diodes)....that will give you the peak arc current. You can safely place the CRO across the load resistor to see the current waveform and thus estimate its crest factor and hence estimate the RMS arc current.

Actually, with the CRO you would not need the precision rectifier, you could calculate the arc current from the anplitude of the waveform appearing on the CRO screen.

Say the peak amplitude is 1v, then 1v across 20 Ohms is 50mA. Multiply by the transformer ratio to get (20, 20 turns sec, one turn pri) to get 1A primary current. so with a 20:1 transformer and a 20 Ohm sec load you will get a 1:1 reading...i.e 1A arc current gives 1v on CRO screen.


By measuring the arc current you will know if you are over or under driving the lamp..and this would be a starting point to modifying the impedance match part of the control gear to match it to the lamp perfectly.

The max current output is 1.15 Amps for the WH5. I've assumed that two red wires together would be 1.15/2 = 0.575 Amps.

Not necessarily true: 1.15a is the line current. There's a good chance that it in no way relates to lamp current.

Also 1.15a would be 4 of whatever fluorescent lamp the ballast was designed for. If you took 2 of the same lamp and ran them in 2x overdrive mode (2 reds) You'd have something less than 1.15a (brightness & power used would be somewhere between 1.5x & 1.75x)

----
Your later posted results look pretty good with brightness on the SOX being close to magnetic though..

If this is a rectifier-crest factor corrector-oscillator-impedance match type circuit, then the output will be a few tens of KHz AC. To measure the arc current, wind a small current teansformer around a ferrite toroid, say 10:1 or 20:1...load its secondary with a known resistor, about 10-20 Ohms and place a precision rectifier in parallel with the load resistor (build with a Op-Amp and, resistors and two diodes)....that will give you the peak arc current. You can safely place the CRO across the load resistor to see the current waveform and thus estimate its crest factor and hence estimate the RMS arc current.

Thanks for this (you've described this procedure for me before), but since I have no time nor the technical kmow-how to get all the correct parts and making them work, I'd rather find something off the shelf instead of making something myself to measure HF current going through lamps.

For me, the easiest thing to pursue has always been to compare the light output. If the lamp was overdriven, even slightly, I'd spot it. This method is quick, letting the lamps power up and allowing me other stuff, like eating, taking a shower, etc. When I retire, I'm sure I'll tackle your procedure out of extreme curiosity.

As for the draw current from the WH-5 ballast not telling me anything about the lamp current, I looked at the maximum 1.15A and divided by 4. Then took the 2 out of 4 reds together thinking that would be safe for these lamps. But now, since the SOX-E 26watt lamp requires 0.45 Amps, then the WH-3 ballast might also do the trick. Max current for WH3 is at 0.56A. More testing I guess.

Making these type of assumptions from the primary side current is only a vague way of determining what is going on.

Opening that white box and drawing the circuit of what is inside is the logical next step toward determining what it does and how it works.

Pairing up "red wires" could end in disaster, (or not) depending upon just what those red wires are connected to inside the box and how.


Working in "black box", (or is that white box in this case) mode is like trying to navigate across a map from Calgary to Montreal with Toronto and its environs covered with a black piece of cardboard you dare not remove.

To understand...or head towards understanding what went through the head of the bloke who designed whatever is inside that white box one needs to open it and work out what is in there and why....sort of like lifting the black card from the map of Canada when you knew nothing of Eastern Canada before....Geez, there's a whole city of five million people under there.

If you don't feel you are up to it...take really good, clear pictures of each side of the board inside the white box and post them up here and I will attempt to remotely draw the circuit from here and ascertain the mode of function from that. If there are big things on the board...like transformers and chokes, they may need to be removed so I can see underneath. Use solder wick and a solder sucker to get them off carefully then replace them after the pictures have been taken.

I have a CT (Current Teansformer) diagram up on LG...been there for ages...

https://www.lighting-gallery.net/gallery/displayimage.php?album=2488&pos=71&pid=118491

In your case, you'd scrap the bridge rectifier as the diode bias drops (0.6v x 2), would muck up your readings. Also scrap the capacitor in parallel with the load resistor. Keep the secondary turns under 20 at high AF (15-50KHz) where I suspect this thing operates as the indictive reactance (of the secondary) will become problematic if you put too many.

Bear in mind that the (lamp) current flowing in the primary has to be equal to the current in all the parallel lumped turns of the secondary...so if the secondary passes through the ring 20 times, its current has to be 1/20th of the primary current. It is essential to provide the secondary a parh for that current for it all to work properly, so that is what the secondary load resistor is for. (That's thinking in current).

If you use "voltage thinking" then the secondary voltage will be N times the primary voltage drop, (which you don't know in this case, but it will be a volt or two), if the primary has one turn and the secondary, N turns.

Using the transformer works sort of like this...it starts with an estimation of the primary current...in this case about 600mA for a SOX 26....if the turns ratio is 1:20 (Pri-Sec) then the secondary current will be 1/20th of the primary or about 30mA. So using a 30 Ohm load resistor will produce a voltage drop across it of about 900mV...(0.9 = 0.03 x 30, Ohm's Law) so 1v measured here will represent 1.5A in the primary.

You will need the Oscilloscope (CRO) to look at the waveform appearing across the secondary load resistor as a multimeter won't be able to react properly to the c30KHz. If you have no CRO and cannot beg, borrow or pinch one, you will need to build a precision rectifier circuit to turn the 30KHz into DC you CAN then measure with a nuilimeter!

Using a current transformer allows you to make measurements ISOLATED from the mains, so you can ground reference one side of the secondary safely. Use well insulated wire on the primary though....it only has to pass through the hole in the core once to be considered a turn. Get the toriod core out of a stuffed CFLi, use its steering transformer's core...this is the correct type of ferrite for this frequency range.

Our eyes are NATORIOUSLY non-linear. I have seen your web site with all the light pollution stuff... (Great Photos by the way). I too, have done Astronomy and had to deal with Light Pollution...so from the dark observations you would know just how non-linear the eye is. Looking at a SOX-26 withing ten feet of it will saturate and overload your eyes completely, particularly some of the cone cells.

Here's a little "experiement" to try to demonstrate this saturation. Fire up the SOX 26 in a room with all the blinds down and door closed. Look at it for a few minutes. Then get up, and go into another room where there is standard 3000 or 4000K lighting.....notice anything odd for the first minute or so?

Interesting. I have never had a problem with sodium getting anywhere near the electrodes. The only thing I can think of that would cause it is running the lamp base down, and nobody is daft enough to do that, right..?

Would doing that to one of my lamps work to get the sodium near the electrodes to move away from them? The sodium isn't in droplets or puddles it's more like a layer of sodium mist on the inside of the tube. I'd like to get all the sodium to move away from the electrodes and combine into droplets or puddles closer to the top of the lamp. Unless it doesn't matter.

It matters. Sodium too close to the electrodes will cause premature lamp failure

What I find difficult to understand is how sodium gets too close to the electrode end of the lamp to start with (see my post above).

For sodium getting behind the electrodes, it might happen if under-driving the lamp or possibly not running it long enough horizontally. Running a lamp vertically with base on top should clear any sodium there.

For sodium getting behind the electrodes, it might happen if under-driving the lamp or possibly not running it long enough horizontally. Running a lamp vertically with base on top should clear any sodium there.

I was just about to suggest that, you beat me to it. Not for too long though, unless it's 55W or below.

I think some of the latest lamps have had sodium around the electrodes straight from the factory
I've got a 2017 35w SOX-Plus which had loads of sodium behind the electrodes, but running it base up for a few hours has cleared most of it.

I was thinking the same: I received a number of SOX lamps with sodium behind the electrodes. Eventually I didn't worry too much about the issue.

What i do for sodium behind electrodes and this works every time is to run the lamp from cold to the point where the lamp is still cool, but the sodium has turned from a solid to a liquid. Gently tap the glass end of the lamp on a soft surface a cushion, bed, sofa, etc this will tap the sodium down the arc tube. Make sure as much of the sodium has gone down the tube as possible. The less sodium left up there the better. And then continue the lamps run for the rest of the day. The area should then clear ( like mercury clearing on a tube)

If you dont tap sodium down it will spread around the arc tube where the electrodes are and it will crack the arc tube. Stopping your lamp working,

Never run the lamp so its fully run up and then do this. This ca cause thermal shock as the sodium is hotter than the glass and it will move onto a colder bit of glass shocking it and potentially cracking it. Always run it up so it is cool but warm enough that the sodium is a liquid

Some might be left in that area but it will be no more than just condensation