There is four main ways to make our pj brighter.

1) A more transmissive LCD. New polarizer may help a lot on this as more are found and tried. I would think there will be limit to the contrast v brightness but some improvement would be welcome. I think part of the reason some of the pj look so good is the lower transmissive polarizer they have give vary high contrast.

2) A better reflector system. A small arc condenser system will yield more LUX than a long arc non condenser system. Short arc lamps tend to have a short life but work vary well with reflectors. A parabolic or ellipses reflector could be even brighter than a condenser system if one can be found that will work with longer arc lamps.

3) A larger triplet. The larger the triplet the larger the lamp image can be. This is not a small thing. If the LL standard triplet had a 120mm diameter then the lamp image could be twice as large. You could then use a longer arc lamp or a larger condenser lens [shorter fl] or double the diameter of the reflector system. This would make for a much brighter pj all we need is a 400mm fl x 125mm triplet or doublet? {I have an idea but it has to wait. Would it be twice as bright?]

4) A brighter light source. The 1000 watt system is here but there are some problems. The 80mm arc tube makes an arc image too large for most triplets even the 128mm diameter Buhl lens can't fit all the light in. Any one with the standard or pro lens will only be able to use about 45mm of the arc's 80mm output if they use a 330mm rear fresnel.

SAFTY FIRST

If you always leave your pj on then this might not be the best thing to do. Your lamp life will be shorter and the fire hazard will increase.
I recommend for anyone doing this to make an all steel light box with forced air cooling and safety glass incase of lamp rupture!

A regular lamp will most likely just get dirty inside and the shell will darken to the point that the lamp just doesn’t make enough light to be usable. The new failure mode of an overdriven lamp will be arc tube rupture. I would guess that all the over driven lamp will suffer arc tube failure. All the lamps we use have an extra safety shell but if that shell is over 450deg it too can weaken and break. This could then let the lamp parts arc togather and drop hot melted metal in your box and start a fire.

It gets better if a ballast has a larger capacitor on it will now be able to over heat and smoke too. If your lamp internally shorts out from arc tube failure a normal ballast can operate into a dead short. A ballast that has a larger capacitor will make more current and more heat. A fuse should be added inside the flame proof box for the ballast.

One more thing, the kind of electricity that ballast makes is just the kind that will kill you, High voltage and high current is not what you want to be hit with. If you want to live don't touch the wires when hot.

It is possible to make a standard metal halide lamp brighter by increasing the current to the lamp. The high performance projector lamps have high intensity arc and no extra lamp shell to get in the way of the reflector and condenser. We don't want the high price pj lamps and the short life but we don't really need a 20,000 hour lamp life either so a comprise might be best.

The plan is to overdrive [overclock] the lamps by using a larger ballast or by increasing the power output of the existing ballast by increasing the capacitor value. I don't have an e-ballast and don’t plan on modifying one. If you want to add your ballast to the test feel free to do so.{ Brian said this"The only concern for me is that we won't honor any warranty for a ballast that has been modified or a lamp that has been overdriven.} If you burn out your lamp then don't try to return it.

All the ballats That I will be using will be of the coil and core types. Right now I have a 400 watt M59 CWA ballast and 1500 watt M48 ballast. I will try to find S52 1000 watt HPS ballast.

The capacitors that they use are rated in voltage and microfarads [uf] All 400 watt ballast use a 400 vac cap this is different than 400v or 400 if you see a cap that does not say 400 vac don't use it. In fact unless you know exactly what to look for I would not try any unknown cap. The 1000 watt and 1500 watt ballast use 525vac caps.

I have two GE 400 watt lamps to start testing with. The model # is MVR400/U. This is an ED37 lamp shell which is the big one 4.5" x 11" I got them on eBay for $15.00 delivered so if they don't make it's no great loss. They look to have a 40mm long arc tube by about 33mm wide. This looks just like the 1k lamps but shorter.

Click to view attachment

The 40mm arc fits inside a standard triplet and will fit the pro lens also.

The full test of each lamp will include the current of the lamp the voltage the LUX output and the heat of the lamp shell and ballast. I have an old oscilloscope so we can see the waveforms too. I will do some long term testing on mabey one lamp.
My target lifespan for the lamps is 2000 to 4000 hours. Most people will want longer life so should not run there lamps so hard.

More later.

Some test has been done already with the 400 watt GE lamp and look better than I expected.
[Edit]

I have changed lamps to a 400 watt ed ed28 lamp from 1000 bulbs.com

for the 400w setup are you just changing the capacitor to a larger one or are you gonna purchase a whole new ballast

What about running a 400 watt bulb on a 575, 750, 775 or 875 watt ballast?

http://www.venturelighting.com/BallastHTML...llast.html#750W

And btw, how do you know when its safe to overclock a bulb? (something tells me, you don't) is there only one way to find out? How do you know you won't blow up your projector when you fire it up?

We already know that it is risky and can be unsafe. A small amount of overdriving might be OK though. It will definately decrease the life of both the lamp and ballast. Here's a guy that went too far...

http://www.lumenlab.com/forums/index.php?s...&hl=capacitance

I was waiting for you to post your favourite thread.

DJ

I can't believe you found that thread. I read it a long time ago. He ran a 48uf cap that is double the stock 24uf cap this could deliver almost twice the current to the lamp and he did not add any cooling. He did not say what lamp he had but it might have been one of the S400dd types.

The S400dd type lamps have a smaller arc tube than many 400 watt lamps I think the arc tube is thinner too. I look at them as over clocked to start with only a 10,000 hour life they have half the life of the GE or many other 400 watt lamps. He may have been fine with a 26uf or a 28uf.

There is one more post that I found and the guy has good luck by adding a small amount of capacitance but he didn't say how much he added.

The small t15 and s400DD lamp will not have much room for over driving, with cooling I would think 500 watts or so max. I am not going to try any of the small ones because there is no real reason unless you plan on adding a condenser lens then the smaller arc will pay off more.

Well I need to get setup for some ballast and lamp testing.

Man I look like a tweaker or something.

Click to view attachment

That better.

Click to view attachment

Some test equipment I have three multi meters and a laser temp gauge and a scope. I need to start with some baseline readings for the 400 watt ballast and lamp of current, temp and what the wave form looks like.

Hear is a screen shot of the waveform[edit runing the 400 watt lamp at full power] after warm up.

Click to view attachment

Never buy a GE ballast this thing is the loudest thing I have ever herd It is way louder than the venture 1500 ballast.


I installed a switch so I could jump in and read the lamp current but the lamp went out. The curent readings are key to determining the true wattage because the resistance of the lamp does change alot as it gets hotter. I need at least to find the missing test leads for one of my meters. It’s a start

I have two ballast a 400watt and a 1500 watt. I will be using both on the 400 watt lamp to see how well the lamp takes it. I have already hooked the 400 watt lamp up to the 1500 watt ballast with the smaller cap[26uf or about 1000 watts] and was ok for the short term.

The caps that I have right now are a 24uf 400vac, 26uf 525vac, 32uf 525vac, three 2uf 660vac. I could use a 8uf or 10uf just to cover different ranges.

Click to view attachment

All can be used on the 400 watt balast but the 1500 watt ballast needs the 525 volt caps.

The ballast the 1500 watt is even bigger than it looks

Click to view attachment

.

I think it will be quite difficult to know exactly how long the lamp will work overdriven.
It might do it for 50 hours or 2000. the only way to find out will be using till it meets its maker, isnt it?

Edit: BTW Arizona, I really hope you wont meet yours by doing this.

AV, Ive been pondering the thought of this very exact same thing, I just havnt had the time, nor do I have the equipment to do this at home (No Scope, not that its exactly needed).

I have access to a bench at work, that has good analog watt meters on it, and I was gonna ask if I gun use it to run tests, but never really got around to it.

Do you have access to Capacitance/Inductance meter. Id love to know the value of inductance the ballast has. If I were to know this I could probably mathematically calculate the capacitance needed to reach a certain wattage.

Typical M59 ballast will draw 450W, and with a power factor of .95 they deliver ~400W to the bulb. If were to completly correct the power factor (setting it to 1.0) then theorectically you would deliver a full 450W to the lamp.

Its important to know that as you add capacitance you are going to DECREASE the total IMPEDANCE of the circuit. Yes capacitors have Reactive impedance, but we know that the caps impedance(1/2(pi)fc is a lower value than the inductive impedance(2(pi)fl). As Im sure you know Capacitance and Inductance offset each other, so as you raise capacitance you are just gonna begin nullifing the inductance. So as the power to the bulb goes up, you will supply slightly more voltage and more current. (one might think that one component of power would be the primary increase)

Im also very intriqued by the wave form, With a capacitor that big, the time constant must be pretty low (what resistance on the bulb was present, when you took that waveform). It becomes more resembling of a squarewave than a sine wave. If you add capacitance, this wave will probably look even more square. This will also result in more effecient power to the bulb.

We will be able to make a guess of lamp life by knowing about how much a given lamp is over driven and the temps. Some lamp lamp makers may be better than others. Are the eye lamps are supper tuff? Can Venture lamps can take the heat?

Right now we know vary little about how hard the lamps can be driven. Finding the thermal limits is fairly easy, if it melts or explodes then that’s a clue but for long term lamp life there is only one way and that is run the lamp for many hours. I am thinking of making a fixture for long term testing. I can use a light for out side. This still will not let us know the true life because the life of an over driven lamp will be determined by the thermal stress it is exposed to. So running for 12 hours a day may be better than just turning on for a 1 hour movie each day.

I think for any of the 40mm or so arc tubes an extra 100 watts could be had with almost no shorting of life just by adding active cooling. Remer most of these lamps run in steel can with no cooling at all in 100deg heat.

I look at it this way a 27mm arc tube[sd400dd] at 400 watts is 14.8 watts per mm of lamp arc so a 40mm lamp may be able to run at 592 watts with only shorting the life to 10,000 hours. With some cooling mabey 700 watts and a 5000 hour life? Who knows lets find out.

I like the direction you're taking now, well, except for possible explosions.... anyway, this may be a tad OT but what about the opposite of this thread...can a 250watt ballast drive a 400watt bulb?
not worried about the brightness, just would allow me to test some things and still be able to use my projector .

and back on topis...is it a fairly linear progression as youoverdrive a lamp? ie. 10% more watts, 10% more lumens?

Thats a good question Mikyd, I wonder if the lumens per watt ratio is maintained.

The term power factor is generally used to express the ability to transfer current into an inductive or capacitive load. The lamp should be almost purely restive so I think they are referring just to the transfer from the primary to secondary. If we change the cap we will make the power factor less and the ballast less efficient. That means more heat. The primary windings will be trying to sink a load into a slightly capacitive load. Because of this I would rather use larger ballast and reduce the cap rather than the other way around. The 1500 watt ballast doesn't even get warm at 1000 watts.

The secondary has three taps and a separate inductor winding. [I found this out with the bad ballast that I had. It had a nick in the inductor] the cap and inductor are mainly to control the current and to provide the high voltage for start up. We can only do so much with just the cap but that is all we can do. I do have a good supply of inductors on hand but I think that would just lower the current.

My Waveteck does cap/ind so I can check the inductance.

Reading the current is more of a problem because I don't have a real amp gauge just the 10 amp max that is in the meters.
I may make a 1 ohm load and measure the voltage drop to determine the current.

The power factor in this specific circuit is telling us alot.

A power factor of .90 is telling us that 90% of the power is being dissipated in a Resistive Load (THE BULB) and that 10% is being dissipated in Reactive load (Inductive/capactive impedance losses)

If we raise that power factor to say, .95 then that will up the Power being dissipated on Resistive loads by 5%. In this case, our bulb will see a 5% increase.

As you know, That Capacitive Reactance and Inductive Reactance fight each other. Since the capacitor included with a magnetic ballast is often COINED a POWER FACTOR correction cap. Then you can assume that you begin with a GREATER inductive reactance than a capactive reactance. Thus, if we increase capacitance, more inductance will be offset (circuit will still have a inductive characteristic untill, a Power factor of 1.0 is achieved)

In this case, that is why the effeciency on a M59 Magnetic ballast and the Power Factor are both the same value.

http://www.venturelighting.com/VLPS/Ballas...MH/V90D6412.pdf

you will see that the ballast draws 450W, with 400 of those watts ideally going to the bulb

400/45 = ~.90, in which 0.90 is the power factor of that particular ballast, adding capacitance will increase the power to the bulb. Thats why the guy in the link that deathray gave you blew up the lamp. He increased the Power factor so much,(and probably raised overall total circuit resistance, which sunk more current) that it blew the lamp. He went too far by adding a total of 3 caps, I believe. (which would be ~72uF)

It is possible that if you improve the power factor, that less wattage would be dissipated on the coil, thus perhaps a 400W running cooler.

If the ballast draws 450W with a .9 power factor. Then ~50W is lost on the coil. If we increase that power factor, while MAINTAINING a 450W input. Then you can divert a larger portion of the 450W to the bulb. Perhaps we can put 420W on the bulb, and only 30W on the ballast.

Power factor correction is a big buisness, Companies pay Engineers that are good at power factor correction big big dollars if they can save them alot on their electricity bill. Thats why many plants have large capacitor banks with automated switching. As large motors in the factor change speeds (or a plant only running at 50% production ect), then the power factor can consistently be corrected, meaning the POWER is delivered where you want it to.

Get me the inductance measurements, and I can draw a simulation is Pspice or something, so that we can see the effects over a Broad range of different capacitor values.

Also, if you can give me the resistance values of the lamp. Im sure there is a change when its cold, and when the lamp has reached operating conditions. Im more concerned with its operating conditions measurement.

lastly, I dont expect the frequency to be changed, is it still at approximatly 60Hz with that waveform you showed in the pic above.

For scubasteve

The inductance is .54 henrys for the secondary.


My amp meter is working now and I have a full set or readings. My main concern is not the lamp at this point; it will take a lot more power, but the ballast overheating.

For reference the stock 400 watt lamp and ballast readings with the stock 24uf cap
The stock ballast at startup

Click to view attachment Click to view attachment

Lamp voltage 135v amps= 2.91 Ballast secondary temp= 214deg LUX at 1m= 343 lamp shell temp =340deg

[edit] Brain fart removed P=VxI so 135v * 2.91= 392 watts the lamp Resistance is E/I so 135/2.19=47ohms

Now with a 26uf cap

Lamp =139v amps= 3.13 Ballast temp secondary = 225deg LUX at 1m=450 lamp shell temp=384

Power is 435 watts the R= 44.4 ohms

And with a 32uf cap

Lamp volts= 146v amps= 4.14 ballast temp=240deg LUX at 1M=690 lamp shell temp=406

Power = 604 watts R= 35.3 ohms

I think This lamp could run at 600 watts for a varry long time . The ballast I'm not so sure.

the 32uf cap wave form

Click to view attachment Click to view attachment Click to view attachment

Some observations.

The lamp is suppose to run at 135v at 3.2 amps with the stock 400 watt ballast and cap. Looking at the wave form it has a nice spike so I bet that the 135 volt reading from my meter is high. None of my meters are true rms so 131 to 133v is more likely true.

The amp readings are low. The stock setup The curent should be at least 3.2 amps instead of 2.91 I don't think the ballast is that far off. The meter that I am using for current is crap but my waveteck I burned out the amp function and the other doesn't read a/c current.

The LUX readings are at 10x again. The room has lots of reflected light so all the readings are good for is to see the % of gain. Going from 343 to 690 is more of a jump than I thing is right. I may do a LUX reading outside for a better reading.

The ballast with the 32uf cap gets too hot to touch and you can smell some burning smell. I did not have any fan or a box just watched it get smoking hot. Even with a fan I think a 32uf cap is a little too much for this ballast. I might make a box and put a 120mm fan on it to see how much it helps. It is new so the smell may be it breaking in. The secondary windings get way hotter than the primaries do so I would put the fan to blow on the secondary first. [The secondary are the big ones]

The lamp was fine, just brighter and hotter I have hit it with a lot more so I think 4.2 amps are fine.

I took more pictures but I lost track of which ones were which

At this point a 28 to 30 cap might be best for the M59 ballast. I have 3-2uf caps so I can edge up to 30uf to see the ballast temps.

I think this is the formla for power. P=V x I Resantance = V/I

Do you mean these formulas?

P=V^2/R=I^2*R=I*V

David.

AV

was all those measurements taken at full brightness, I assume it has.



The formula for REAL power is V*I

Other variations for power that include resistance are: V^2/R and I^2 * R

If you wanna count for total circuit power, then you will have to calculate total circuit impedance. This is called Apparant Power.

the resistance of the coil, cap, and lamp combined, over the full secondary voltage and current.

In that case its V*I again, but this could be measured on the primary. Input system voltage and current will tell you how much power is drawn.

Real Power / Apparent Power is the Power Factor. Thats why PF is important when it comes to this.....

What is the resistance on the lamp? and the resistance on the secondary of the transformer.

Moving on to your readings.

The 24uF reading,

135V * 2.91A = 392.85W

The 26uF reading,

139V *3.13A = 435W

The 32uF Reading,

146V * 4.14 = 604.44W

What is happening is that the added capacitance, its changing the power factor like I predicted, but UNFORTUNATLY it is also changing circuit resistance. Meaning that more CURRENT is drawn. As you can see approximatly the same voltage is present in all 3 readings, A 10V window isnt very much, when compared to more than a 33% increase in current. The jump from 2.91A to 4.14A is telling me that the added capactive impedance is playing a huge role.

The effective Impedance of the 3 caps should be:

The 24uF,

1/2(pi)FC = 1/(2)(3.14)(60Hz)(.000024) = 110.62 Ohms

The 26uF,

1/2(pi)FC = 1/(2)(3.14)(60Hz)(.000026) = 102 Ohms

The 32uF,

1/2(pi)FC = 1/(2)(3.14)(60Hz)(.000032) = 82.9 Ohms

The effective impedance of the secondary Coil:

2(pi)(F)(L) = (2)(3.14)(60Hz)(.54) = 203.472 Ohms

So what is happening here is this:

As you increase capacitance, you are increasing the current. This is why the ballast is getting hotter. The resistance is going down. A consistent voltage is applied, so current is rising.

Thats why the manufacturer has the capacitor of 24uF. They could further correct the .9 Power factor, but that does no good if the current goes UP to much. The best solution would be to better the power factor, without DECREASING circuit impedance. That way more of the 450W is diverted to the bulb. What you have above A.V. is more power going to the bulb, but you also have the ballast drawing more power. This means more power is dissipated on the coil as well.

I think the limiting component of this is the BULB. There is nothing we can do to change the resistance of the bulb, hence I predict at best, you could only get away with the 26uF and a shortened lamp life.

Zona,

It's probably impossible to tell, but did you happen to notice the lamp color temp changing when testing
larger cap values?

I was wondering what happens to the lamp resistance when it is hit with more current. It lowers. We had 47,44,35 ohms so the more current the lamp sees the lower resistance.

Decreasing resistance is what causes thermal runaway. I think the lowering resistance is only due to the increasing current not the temp.

I can do a simple test this. Take a voltage reading with a fan on the lamp with no reflector then remove the fan and add the reflector centered on the lamp. This will cause the lamp to heat up greatly. If the resistance increases [which it should] then every thing is fine. if it decreases then you have part of the condition for thermal runaway.
The ballast will limit current but if less power is dissipated in the lamp then more power will be dissipated in the ballast causing it to heat up more..

To scuba

The power factor can only be correct for a set load.
In normal use the only value that changes is the lamp resistance. Right after start up the voltage across the lamp is 25v so the lamp resistance is 8.6 ohms. The ballast always makes the same current [that why they cal them CCW constant current wattage] the relationship of the inductor and cap is only to control the current in the secondary for a changing load.
Power factor only comes in play if your faze angle is highly inductive or capacitive and Looking at the start up, the wave is almost flat with little cap lag or inductive overshoot, at full power there is more inductive overshootwith though.
By moving away from the standard L/C values wee will lower the total power factor. This is also what I expected. The ballast will waste more power. This is why they have different ballast for different power levels.
I wasn't trying to make the ballast more efficient I was trying to get more power to the lamp.

By looking at the wave form at full power, adding some capacitive would make the impedance closer to zero, but this is a new lamp, a old one may vary well have a lower resistance at the lamps half life and the wave form would be closer to non reactive one.

To paladin

The lamps color temp does look to increase some but I don't have a good way to measure the change.

EDIT the color temp goes down with more power.

I was thinking of shining the light on a screen with two or three other known lights sources. A 3200k 3000k and 2800k lamps and just looking at the difference. Maybe the photo shop experts could come up with some RGB values.

Did any one notice that we made 50% more power with the 32uf cap and the 400 watt ballast? and the lamp seamed fine? This will give you a 50% increases in LUX and lumens.

you have an inductor? Most M59 ballasts that I have seen, only have an outboard Cap. Plenty inductance is present in the coil already, no additional inductance should be needed.

I know you werent tryin to make the ballast more effecient, and that your goal was to get more power to the lamp, but it may be better to accomplish that by changing the effiency. However, that may be moot as it may be impossible.

I really do think the Power Factor to be a more important aspect than you are willing to credit it. It comes into play more.

http://www.venturelighting.com/TechCenter/Capacitors.html

When you had the 32uF cap, you had 600W on the lamp. Who knows how much was lost on the coil.

The goal would be to lower resistance, while MAINTAINING Power factor. I know you dont think the power factor means alot, but I think its the key to Overdriving these lamps more controllably, especially since we can regulate power dissipation on the coil.

with 200ohms of inductive impedance, and and 80-110 ohms of capactive impedance the circuit should remain primarliy inductive regardless of which of the 3 capacitors you used. The stock configuration has a 200:110 ohm inductive/capactive ratio. I think the best way to approach this would be LOWER resistance, while mainting that reactive ratio.

Do they make an 27, or 28uF caps in this flavor.

I think 32uf is a bit too much, but I think we can do a little more than 26uf.

If we could lower the capactive impedance to 95 ohms. Then we would need to change the inductive impedance to 175 Ohm, to maintain the same ratio. We could add an inductor in parallel to the coil to do this. that inductor value would be:

200 * ZL / 200 + ZL = 175

200ZL = 35,000 + 175ZL

25ZL = 35,000

ZL = 1400

2(3.14)(60Hz)(L) = 1400

L = 3.72 Henrys.

I think im going to far with this ..... 3.72 Henry inductor probably very hard to find, so mathematically it would work, but realistically probably too hard to track down the components. theory and real world collide again, LMFAO.

What kind of numbers are you getting with the 1000W and small Cap. Im interested in the input power as well.

are you talking about underclocking a 1000 watt ballast to 600?

I think im going to far with this ..... 3.72 Henry inductor probably very hard to find, so mathematically it would work, but realistically probably too hard to track down the components. theory and real world collide again, LMFAO.

Ya finding a large 3.72 inductor might take some time. Where would you find stuf like that

Click to view attachment

oh on my desk.

Seting the power factor back to 90% might reduce the ballast temp some it might be worth a try.

I don't think 3.73 H is right thats more than the total .55 H that is internal to the ballast. How about .373 H?

That's quite a selection you have there!

You part owner of PartsExpress or something?

Hmmm, chokes for speaker xovers?

The 3.73 Henry is statistically correct. you can check my math above.

we start with 200 Ohms of inductive impedance and would want to lower that to 175 ohms. only way to do that is to parallel the 200 ohms with 1400 ohms.

the result is one big ass inductor, that probably doesnt exist.

The best approach is possibly using larger ballast, with lower than stock capacitors. I do think 600W will prove unstable for the lamp, but only time will tell.

If I go to home depot and have them wrap 125' of 12 gauge on a nice small bobin?

A 1000 watt ballast is $70.00 and won't get hot.

Tonight I will run the M59 ballast at about 500 wats with 28uf of cap. Then the big one.

I wanted to try using a switch in series with the extra cap’s to act as a high low power setting for the lamp. Switch on would be high and the extra caps would be engaged and off would disconnect them.
For this I used 3 -2uf caps[660vac] hooked up in parallel with the main 24uf cap. I am still using the 400 watt ballast. This will switch from 400 watts to 600 watts. The setup looks like this.

Click to view attachment

This worked great you can flip from high to low with the lamp on with no problem. The ballast still gets real hot at 600 watts but I like the switch idea.

Next I hooked up the 26uf and a 2uf cap for a total of 28uf. The readings were
Lamp voltage=144 amps=3.51 wattage= 505

Then I used 30uf cap. The readings were
Lamp voltage=144 amps=3.93 wattage= 565

I didn’t take ballast temp readings right now.

Next is the 1500 watt ballast with a smaller cap of 26uf instead of the 32uf that came with it.. The readings were
Lamp voltage= 153 amps= 4.8 wattage= 734 watts Lamp shell temp 420deg LUX=920

The wave form is much more inductive with the light load and smaller cap. A picture if the start up wave form and at full power.

Click to view attachment Click to view attachment

The lamp arc is a little shaky. It wobbels a little with this setup and the scope readings and the LUX readings go up and down some about 5% or so. I saw the same thing with the 1000 watt lamp with the same ballast and cap.

Click to view attachment

I want to get the power testing done and then do a long term test at a certain power level with the lamp. I just haven’t decided what power level? I think the lamp will be fine at 600 watts but the ballast might not cut it. 735 watts is more than most people would want? Any suggestions?

The next test will be the lamp hooked up to the 1500 watt ballast with the stock 32uf cap. This will bring the lamp close to 1000 watts. edit that was just a little to much power.

I tried to test if there is a color temp change by setting the camera to manual and taking a picture at 400 and then 600 watt. We now know that the color temp goes down with more power some times a lot.

Have you thought about attaching an aluminium heat sink to the ballast core?
I can’t seem to find your base line lux measurements. I may have just overlooked them some ware though.

DJ

DJ, post 16. Baseline is 400W ballast with 24uF cap using the GE 400W bulb. Lux @ 1m = 343 (x10)

gs

Not too much interest in the over clocking of the lamps so far?

I think trying to get a large amount of power out of the 400 watt ballast is probably a bad idea [edit a fan fixed this]. Adding 4uf with a hi/low switch is about all I would do for the safety minded people. That’s 505 watts this will not hurt the lamp and the ballast will get hotter but any fan will keep it cool. I will do a two or three day test with this setup. [edit 550 watts]


Now for the more fun stuff.

I hooked up the 400 watt GE lamp into the 1500 watt ballast with the full sized cap[32uf] This does not make the lamp run at 1500 watts because the lamp runs at as lower voltage than a 15000 watt lamp.
It does run at 166 volts at 6.3 amps. That is 1045 watts. WOW now that’s cooking.
The wave form is fairly reactive with this setup some more cap is needed for the lower voltage lamp [A 1K mh lamp runs at 250v] but if you add a cap the current will be way too high. So what to do?

Use the 208 volt tap. This reduces the voltage in the primary so it is also lower the voltage in the secondary now I can add a cap to make the wave form flatter so I added a 26uf cap for a total of 58uf. The new power level is now 946 watts and the lamp is stable and the arc does not wonder. The wave form now looks good.

I don’t think the lamp will last too long at this power level. The arc tube looks exactly like my 1000 watt lamp. The 1500 watt venture lamp is rated at 3000 hours and is 85mm long. I think it is the exact same arc tube with a slightly different pressure and gas mix. If this true then that lamp has a power output of 17.65 watts per mm of arc tube. So a 40mm arc tube at the same power per/ mm of arc tube should last 3000 hours. So 17.65 x 40mm= 705 watts. Well that is not 946 watts. Power per mm of arc tube for the 946 watt setting is 23.65 watts per mm. I can count on a short life at this power level. A fan sounds like a good idea.

For a 3000 hour life we need a 700 watt setting I got this with the 208 volt tap on the primaries with the 32uf cap. The lamp voltage was 146v and the current was 4.6 amps. The wave form is not the best here and some more caps are needed. I will redo this test with 36uf and this should make 705 watts and have a better wave form. I really do think you can run this lamp at 700 watts and get 3000 hour life.

It would be better for over-driving this lamp to have 1k ballast or a 600 watt hps, instead of a 1500 watt. [edit the 400 watt ballast is ok with a 120mm fan at 550 watts] The stock power output without changing the cap for the 1k ballast should be around 500 watts if you use the 208v primary tap. You should increase the cap to around 32uf to 36uf total this should put you around 700 watts and the wave form should look nice.

Tonight I am going to do some LUX reading with the test box and the small 80mm triplet at 1000 watts

I am very interested. Just waiting for results of lux and longevity tests.

David.

Well I ran the lamp at 950 watts on the bench and it was OK and I had the reflector on it [I always have it place to block the light and well you never know what may happen] but this time I made sure the reflector was centered like it was suppose to be [did some screen test with a 220mm and 330mm fresnel] well this makes the lamp a lot hotter and it melted!

Click to view attachment

Click to view attachment

Click to view attachment

Well I was running a 400 watt lamp at 1000 watts what would you expect. The only bummer is with two lamps I could swich setups and not have to wait for the lamp to cool for restart.

It may be that even if we had a stock 1500 watt sports lamp if we added a spherical reflector it might overheat.

I am looking for a great spherical reflector if any one knows of one.

so what are your conclusions as to the 400 watt ballast? 28uf? I didn't see alumens reading for that one... I know you said the 32 was too much?

and for those of us electronically challenged, do you have a link to a suitable 28uf cap?

I wanted to get through the entire power test to see what we had and now to see what the failure modes were first.

Adding 4uf does look to be a good amount of capacitance to add. The lamps can take the extra power and the ballast should not get too hot.

I will do a complete test at that level for the long term test with 400 watt ballast.[edit going 550 watts using 30uf] The lamp should live for 5 to 10,000 hours at 550 watts.

The idea of using a switch to engage the extra capacitance I like more and more. By adding capacitance the ballast makes more current all the time. Well start up is hard on the lamp this is when it has to melt all that metal and the glass has to heat up to thousands of degrees. With a switch it can do this slowly so the lamp will get a lot less stress on the arc tube after full warm up it then can be turned to high. The switch will arc inside a fair amount so a hevy duty switch should be used.[15 amp 250 volt]

I need to talk about lamp stability. It looks like the lamps are more stable with the 400 watt ballast with a 2uf cap added or with a total cap of 26uf. This may be just because my lamps are new. They will change a lot in it's 20,000 hour life. I still don’t know what changes as a lamp ages. Do the current carriers end up on the lamp shell and therefore the lamp arc draws less current? Or do the deposits on the lamp shell make a path for arc to fallow?
Some people have had a flickering problems with some lamps. If you watch the lamps with an arc welding helmet on ,if the ballast is too inductive[the 1500 watt ballast ] the arc will dance or spin around in the arc tube.[ I think the arc tuching the side of the arc tube caused the lamp failure] If you use a lux meter or look at the scope the power levels changes as the arc spins around. When the ballast is right the arc makes a nice straight line in the tube. You can see this in the final light output on the screen too.

you don't want the arc wobbling around. The light variations in LUX are high enough to be seen and there is a color shift that takes place too , making this something you don't want. More heat may make the arc settle in better so don't over cool the arc.
The temp of the lamp affects this too, for some of the power levels the arc will dance around then if I add the spherical reflector the lamp heats up more and stabilizes.

Most of us won’t be using a lamp that is 10,000 hours old. At that point the lamp is making a lot less light so adding 2uf might be a good thing for just about all users. And it is worth 50 watts of power too.

This where I got my 2uf 660vac caps they are used and cheep $1.50.

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewI...item=3861982633

Or the seller

http://stores.ebay.com/Cascade-Surplus-Electronics

These are the exact right kind for a 4uf cap just cost a little more.

http://cgi.ebay.com/2-Mars-440v-4uf-A-C-Mo...1QQcmdZViewItem


A five pack for your friends too.

http://cgi.ebay.com/5pcs-General-Electric-...1QQcmdZViewItem


It is a lot cheaper to add a small cap than to buy a new large $30.00 one. A 28uf 400vac is hard to find. The 2uf cap does NOT have an internal resistor like the 24uf does so it can remain charged to around 330 volts if you disconnect it by the switch when its charged [pj is on] it will stay charged. It will discharge if it is connected to the other cap but if you have a switch hooked up and turn it off before you turn off the ballast then it will remain charged, waiting for you to change the wire connection on top of it and teach you about how caps store electricity.

You should always assume a cap is charged and short out the leads on top of it before touching it. Or only buy "protected" caps They will discharge by themselfs.

cool, thanks for the info and safety tips... if you can get a 30%(waiting to see the results of your tests of course) increase relatively safely and a "high/low" switch for the price of a couple of caps and switches, it seems like a good deal ...

AV, at this point you may want to explain how you added the capacitors.

If you cascaded the additional caps in series (current flows through one cap, then the next) than the 2 cap values will add like resistors in parallel. In this case 24*4/24+4 = 3.43uF

If you wired the caps in parallel to each other, then the caps add like resistors in series. 24+4 = 28uF.

Im certain the latter of the two is the way you wired it, but for those that dont know about caps, they may assume that caps add in series like resistors. They dont.

or to help even more, a wiring diagram.

You guys will be sorry that you asked for a drawing You must not have seen me draw before!

Always wire the caps in PARALALL like this.

Click to view attachment

A note of caution:

Capacitors, especially high voltage ones, can be lethal. They hold a charge for an indefinite amount of time after the power is removed. Be sure the caps have bleeder resistors across the posts to help drain the charge when the power is off.

There are a lot of different types of caps if you read the side of you cap from your ballast it will say 400vac EDIT It must say if it has a resistor vaule if it does then it will discharge the cap when not in use so it can’t shock you with 330 volts if you touch it. You can use non protected caps but they will stay charged when not in use. But like I said before they will discharge if hooked to the other cap. The small one's [2uf] won't kill you just teach you.

I had to take a closer look at the melted lamp so I cracked the shell. The inside arc tube broke at the same time. It made a nice loud bang. I don't know if the outside or inside made the bang.

Click to view attachment Click to view attachment Click to view attachment

Now I can measure the arc tube and test the internal parts.

It is really 38mm and thinner but it still looks just like the 1000 watt tube just shorter

Click to view attachment

Click to view attachment

Click to view attachment

There was no damage to the electrodes

I might put all the reading of power, caps, and temps on one post. It will make it easer to see the trends.

I ran the 400 watt lamp all day first stock then with a 28uf cap and then with a 30uf cap I wanted to get a better baseline reading of the ballast temp before over clocking it. I have been taking the readings at two points on the iron core and the copper windings of the sender coils.

After 3 hours the stock 400 watt ballast was at 205deg secondary and 179.
I then added a cap to make the value 28uf or 496 watts after an hour or so the temp was

Core 195deg secondary 240deg

Then I went to 30uf or 545 watts after an hour or so I got

Core= 202deg secondary 242deg

That is too hot to touch. Well I have 25 fans maybe I should add one? So I added a low CFM 120mm fan. No box just let it blow on it. I measures points that the air was not directly blowing on in fact where the air was blowing on the secondary the were like 90deg or so. This at 545 watts after a hour or so.

Click to view attachment


Core=136deg secondary = 170deg

Well that is a lot lower than I expected. I do have a nice box that the ballast came in maybe I should cut a hole in it for a fan.

The cooling worked so well that I may change my plans some.

Not tryin to steal your glory any, just though a slight cleaner diagram for those who arent electronically/schematically inclined.

Caps in parallel to each other, and those in series with lamp.

I will be doing this when I return from Europe. Ive been wanting to do this test for awhile and now that AV has done them for me, I get to plug and play ...... Thanks

I really hope that you broke open that lamp outside where you were not exposed to the poisonous gasses contain in the arc vessel. And you used gloves when handling the pieces.

DJ

I do have one thing to discuss, those that may know more about active cooling.

One could go the path of thermal paste and a heatsink, but I dont think that is required, as these ballast can take a chunk of heat before the give up.

I have my ballast in an enclosure I built already as seen in the following images:





As you can see, I already thought about the holes when I made the box. I can just mount a fan on the outside, my only thought is would it be better to blow onto the ballast, or suck air out of the enclosure.

I'm crushed Really

Well the heat from the ballast won't be a problem so what wattage do we run the lamp at?

This a small list of lamps and wattage per mm of arc tube length and life. just to see a trend.

Arc tube.. Power...... Life...... Total power.

43mm.... 10.45 ... 20,000... 450 watt
38mm.... 10.5... 20,000... 400 watt
80mm..... 12.5... 20,000... 1000 watt
25mm .... 12.4... 20,000... 310 watt
27mm..... 14.8... 10,000... 400 watt
51mm.... 14.7... 10,000... 750 watt
85mm.... 17.7... 3000... 1500 watt


It looks like 14.7 or so is commom for the more high performance lamps so 38mm x 14.7= 558 watts.
looking at the list there is one lamp that lookes like it has a lower power to watts raito.The venture 450watt PS lamp with a 43mm tube 43 x 14.7=632 watts. A 43mm arc image will still fit in a stock triplet.


550 watts or 600 sounds good.

I did an outside test with the 400 watt lamp at 550 watts.

[Edit by placing the lamp on a tripod at 1m in the air and having the lux meter on the ground there is almost no reflected light from walls or roof so the reading should be close to correct]

The LUX meter showed 45,100 LUX[edit lumens] at 1m. This seams a little low for the wattage but shouldn't be too far off.

[Edit Most of the universal lamp from venture make 36,000 when new. I haven't found the GE numbers yet ]

I orderd a spare 400 watt lamp today from 1000 bulbs.com It has a 4200k color temp and the arc tube is unkown. $17.00
It may well be that most lamps male less light in the horizontal position. This might explain the low lux readings.

I have decided to do some long term testing at 550 watts. How long I go I don't know right now.
If everything goes well I might go to 600 watts.

EDIT

I wanted to make a list of the power output and the cap and temps. the real thing to look at is the current.

Stock 400 watt ballast and GE 400 watt lamp with a 38mm arc tube.

Power watts...Lamp voltage.........Current amps...............Lamp shell temp....Balast temp......... Cap

392............135........................2.91..............................343...........
......... 214..................... 24uf..................................................

435..........139......................... 3.13..............................384......................225....................26uf....
...........................................

489......... 137........................ 3.57....................................................................................28
uf...............................................

562......... 142.........................3.96..............................401....................170[ with 120mm fan].30uf...........................................

604..........146.........................4.14..............................406............
.........240..........................32uf..........................................

I did the input current for three power levels

392 watts was 3.56 amps at 114vac .. Input watts =405
562 watts was 5.1 amps at 113vac . . Input watts = 576
604 watts was 5.68 amps at 114vac . Input watts = 648

The input waveform is perfect even with a 604 watt load.

Click to view attachment

At 604 watts only wasting 44 watts as heat is a lot better than I thought. It should be easy to cool the ballast with any fan.

AZ
Would the same principals apply to a 400 hps ballast? I am almost thru with build 2 and would like to overclock my S400DD lamp. I am doing a verticle build which means I will lose some lumens with the bulb being horizonal and the FS mirror. This build also uses a precondenser.Also have a lux meter on the way so it will be interesting to see the increase in total output in a hps ballast.

My cap says MMP Capacitor 300vac 55uf +-3%.

The sd400 has a smaller arc at 27mm [I think] so it already runes with a higher power density.

The hps ballast make more current so the L/C ratio is larger they run a 240vac cap without testing I would not go over 4uf this should give you 50 more watts. Any more and I would count on a vary short lamp life.

It would help to know if the arc is running across the top of the arc tube. If it is this may shorter lamp life even more. You will need some welding glasses to really look at the arc even then your face will still see the up. Keep the exposure time down to less than 30 seconds if you do this.

A before and after lux reading will give a fair idea of the power increase too.

The 400 watt lamp runes at 14.8 watts per/mm at 450 watts that becomes 16.67 watts. The highest power normal lamp is the sports lamp at 17.7 watts per/mm of lamp arc[3000 hour life]. You might only get 4000-5000 hour lamp life at 450 watts.

You could be the first to find out Hey what was that sound?

So if you add a 4uf 240vac cap it should be OK. [any ac voltage larger than 240vac is ok but the cap will get larger with voltage rating pm me if you find one] I would use a switch and only hit it with the extra power after warm up and when the movie starts.

The lamp will cool much better outside than in a box with poor air flow. I would not over cool the lamp it may make them unstable but the hot air does need to be removed from the box. You dont want the reflector to cover the whole lamp everyone with the large bowl with the notch cut out should just cut the whole bowl at the notch. The rest of the reflector is not used any how but the light and heat does heat up the lamp. [yes the light system does wast all that light.]


I have started the long term testing of the GE lamp at 550 watts. I expect this to be about much fun as watching paint dry.


It is outside on a steel covered table with the 120mm fan blowing on the ballast. I am using the cooper reflector that acts like a spherical reflector so this should heat up the lamp some. The air does blow towards the lamp reflector but I did no direct cooling of the lamp. I may chane this to let it get hotter.

Click to view attachment Click to view attachment

I don’t know how long I will let it run. After some 15 hour days I may let it go for some 24 hour runs. After a few hundred hours I may go to 600 watts. I do have another lamp coming and I will want to try it too. I will take some temp readings along the way.

hey AV,

will you tell me the input current, and line voltage (~120V) going into the ballast.

I wanna see how much power the ballast is drawing, versus how much power the lamp is recieving.

poisonous gas??!!!