Sure, I now have one more wavetek dm27 my twin brother has a twin meter Are you thinking about the input voltage and current to the 600 watt ballast?

Rox Can we all agree that what we have hear with the ballast is a series current pass filter? It works just like a 6db band bass filter except it is resonate.

I can now graph the first ramp of the curve. The graph will be lamp voltage V power up to 95 volts of lamp voltage, then lamp will make about 1100 watts. If the lamp ages and the voltage across the lamp increases to 105v what will be the new wattage?

I am thinking of making a dummy load to find a few more points above 95v by adding the load in series with the lamp and calling the new combined load the new power level.

yes, with the 600W ballast and the 1200W lamp.

mmm, I tried to work out the transference function on the variac +series L + seriesC + series lamp system... but havent have success yet... (dam, those things are much easier for europeans )

I have missing some data... the coil is one or two? I mean, there is any "known" magnetic relation between the variac and the series inductor (L) ?

have to go... see you tomorrow.

There are probably 2 coils.

The problem is, that Inductance measurement, we are unceartain that it is just the one coil for lag purposes, or is it a parrallel of both coils.

I'm not using the variac in the circuit.

I think the transformer (transformer with an inductor built in=ballast) can be reduces to a voltage source with a resistor value. Even thought the inductor is on the primary side of the transformer.( I believe they move the inductor to the primary side because the voltage is higher on that side now that the lamp is only 110v) The transfer function is almost 100%(95% of the power of the secondary is at the primary) so for all practical purpose except for the voltage difference of the secondary to the primary we can call it a voltage source with a R value, The inductor value will be in series with the rest of the circuit regardless of which side of the transformer it is on.

So the circuit is still the same I think maximum power will occur when X (L) = X©?


Click to view attachment

thats called correcting the power factor.

thats why suggested tryin to add inductance, because when you change the factory Xc, you then change the power factor.

When Xc = X(L) you have a perfect power factor of 1.0. You probably arent gonna be able to get there easily, I doubt even in "Stock" conditions that Xc = X(L). Since X(L) = 2(pi)FL, then the low frequency of 60Hz, means a large L value is needed.

no ed28 i believe. mini football

yes I know you are not using the variac with the 600W setup, I was just refering to the first stage on the 600W ballast... this is a "voltage increaser" what I was refering as a variac... and then there are 3 more elements in series L+C+Lamp.

here in europe we do not need more voltages... we just have them on the plug (220VAC) just need a choque (L) in series with the lamp and then a paralelled condenser to the whole system for the factor correction...

is is posible for you to remove eveything from the ballast and meassure ONLY output of the "variac" ? (remove L wires...)

We must be thinking the same thing I wanted to look at the voltage and wave form from with out the cap and inductor. The inductor is wound on top of the four primary taps and then feed out to the lamp.

I did find a problem. All of the output and input wires are 15 gauge, not 14 gauge like the 400 watt or 12 like the 1200 watt ballast. All the wires on the output side get fairly hot. I measured 110deg on one of the hotter spots. It looks like some of the plastic has started to melt in spots but it is so hard to tell because they have gunk all over from the dipping of the core.

Click to view attachment
I removed one layer of the plastic wrapping so I could get a better look. On the secondary side the output is made of two 14 gauge (about) winding paired together then it has a 15 gauge wire soldered onto it.

I can't get a good look at the inductor side without removing more layers. It looks like a single winding but I can't see it yet. I can't tell the wire gauge either. The inductor could prove to be the weak link.

Click to view attachment

I might be able to take a reading of just the inductor if it is soldered to the final tap on the out side like most of the taps. Rox is that what you were looking for?

I'm kind of stuck on this right now. I don't want to hack my ballast up if I can’t change the output wires I really think the two ballast should work if paired together and there would be no heat problems.
I could just leave it and put a fan on it. I am not going to be using it for thousands of hours any how.

My twin brother's twin meter. What’s funny is we never told each other what type we had

Click to view attachment

I took a reading and a scope shot of the unloaded ballast. I did remove the cap but that choke must work real well because I got no voltage on the output.

The RMS voltage was 230vac It made a nice triangle waveform. My 40 year old scope’s 1000v range doesn’t work any more (yes I know I could just make a voltage divider) So you cant see the whole waveform. There is a slight hook on the top and a dip on the bottom.

I ran a 32uf cal and a 55uf cap and got the same voltage with no load.

The unloaded wave form from the 600 watt hps ballast

Click to view attachment

Click to view attachment

Rox I was able to go around the Inductor and run the transformer with no cap or coil. They are all wired separately so I was able to take readings on everything.

First is the voltage on the secondary with no load or cap or inductor. Input is 118vac

164.9 vac RMS or 128.3vac normal (non rms meter)
The DCR of the secondary is .7ohms
The inductance is .285H

The primary tap has a DCR of .3ohms
Inductance of 9.7mh

The series inductor has a value of 26.6mH and a DCR of .6ohms

I can use the igniter taps for the out put if I wanted too and if I really wanted to I could just not use the internal inductor it is a separate winding.

That still leaves the cap wire which is single so to get to that I will have to cut away the covering on both windings. I'm not sure if I need to do that right now. A little air can go a long way to cool the small wires

The waveform with no cap or coil.

Click to view attachment

Click to view attachment

mmm, lot of work last night

mmm, a "poor man's ampmeter" reading at input for 600W+1200Wlamp is still something I would like to know yet (yes, I know I ask too much things that you just can´t reply me all my requests.. )

by now let me play with your numbers;

1063W at lamp, 52W "poor man´s...", 0.6 ohms inductor @11,3A=76W
1191W at secondary by now...

plus 0.7 ohms secondary @ 11.3 A=90W we are already at 1281 USED wattage or active WATAGE.

assume 0.8 power factor (mmm, >0.9 is calimed by the stated system at 600W with no overdrive... so it will decrease I bet...)

mmm 1600W (aparent) inputing to the system... at 117 Vac this means somewhere 13,7A at primary.... that is 0.3 ohms... 56W disipating...

well, the lamp wattage was 1063W but my rough calcs are showing 1600W input... this is 66% efficiency on the whole system... also I am omiting any SATURATION at the core... this is going to increase the input wattage too... so who knows the true efficiency... (50%?)

conclusion; a very hot ballast

I can take an input current reading now at the 1100 watt setting. We do have a fairly good out put current reading. I didn't know you were looking for system efficiently. You should have just asked.

I'm beting on 85% to 90% efficiently.

Rox: Have you ever tried to pick up a stock ballast after a few hours of running? My GE runes at 200deg F The 600 HPS at 1100 watts runs around 180deg.

I think calling it a power factor or a current filter is basically the same thing. In both cases as you move away from the center point the power factor will be lower so less power will be made. Or if you look at it as a current filter if the load moves from the center point of the filter the power will be less.

It is probably more accrete to call the circuit a resonate trap circuit. It has a high "Q" value of 1.8 I suspect that there is 50% less stored energy in the system with the higher power level so the "Q" might be 1.4 now.

I made a movie of the power ramp up watching the two meters one with the current and one with the lamp voltage. The current was surprising flat.I will make a graph of the readings I am going to add some points with some added resistance. I want to find out what happens to the power if the lamp voltage increases some but I don’t want to do twenty lamp cycles. In fact I don’t want to put a hundreds lamp cycles on the lamp just for testing. It's my baby

I do have other lamps that I can do anything too one cheep 400 watter and two 1000 watt ones.

Soon I want to set up my condenser lens and do some lux reading if my lux meter ever gets back.

I have three projects coming up. Changing some of the carpet in the house, fixing up my old 1973 blazer to sell. And the PJ If I can get the last part I may just blow through the PJ fairly fast just to get a picture up and running and do more tuning later.

yes I have ran 250W and 150W lamps with their ballasts and they get hot... I don´t think as much as yours though (maybe 160 Farenheit, never measured further than my fingers accuracy ;D, I remenebr I could keep my hand on them no problem, with no air forced cooling of course.

beting on 85 to 90% eff... mmm I am taking it (just joking )
lets see what your PMA says.

Not wanting to ever waste any power I did a quick test of the input voltage and current of the 600 watt hps ballast running at 1056 watts (using .402 ohms for the resistor vaule)

The input voltage was 110.2 RMS and 110.5 non RMS( the two readings were real far apart during during startup because the wave form was vary distorted. It started to look fairly good at full power.)

The voltage across a .402ohm load was 4.4222vac RMS and 4.24 vac non rms.

Some quick math and the input power was 1212.2 RMS and 1164.7 non rms.

I guess I have to use the higher # for efficenticy P out =1056/1212.2 Pin Efficency=87.1%

Power wasted as heat is around 156 watts. That's Smoking

It might be a good idea to try to increase the power factor by increasing the inductor and the cap. What that value would be I don't know. My question is you can increase the power factor by adding an inductor and the core will become more efficient and cool down and make more power. But my real problem is the small wires from the cap and inductor, will they run cooler with a better power factor?

The ballast does weigh 17LB so it does have a large surface area to cool

The one thing that was interesrting was the wave form across the input resistor. It shows a trace of the current wave form of the input so it should be the same as the current at the output. There is no flatting of the wave form from core saturation in fact it shows almost perfect current transfer. The core could make more power

Click to view attachment

The 60 Hz wave form now has some rounding suggesting the power factor (there I said it) is off and is distorting the wave form.

Click to view attachment

I did find a web site that listed a ballast for different HMI lamps and they listed different caps for different size lamps the hmi575 took a 75uf cap(a 75uf and a hmi575 sounds about right for this ballast) and the hmi1200 watt took a 145uf cap They must have added a large inductor to use that size of a cap.

I have to say I am really suprised about your measuremts...

they show numbers that I would have never thought being that "efficient".

actually I don´t believe you can improve the power FACTOR since It it very close to 1 as I can show you;

1212 inputW--- 1056W lamp+76W inductor+36Wprimary=1168W mmm (I am not considering the PMA wasted W)...

I worked out the each W since we already know the current on every point and the R of each stage... I am having a W input (aparent) = W ACTIVE (or close, 96,6%) value so the power facttor needs to be very close to 1, you could hardly impove this...

just a last idea...
mmm, you have two twin rms ampmeters... they should have very similar internal R when ampmeter function... what about paralleling them at 10A so you could meassure up to 20A RMS (they should have close to half current on each...) stupid idea?

I just wanna be sure the PMA readings are belivable...

4uF

You can tell, but it isnt mind blowing. I would do it again if I had to.

I only have one RMS meter both of the DM27 are non RMS and the amp function in one of the wavetek is burnt out in. It has never been the same the "bad day" My brother saw mine and told me not to use the amp function at all.

I don't think 85% power transfer is the greatest but we are really asking a lot from the ballast. Everything gets warm in this setup. The power cord is small at 16gauge. All of the feed wires on the ballast. The small piece of 14 gauge on the extension cord. The normal ac voltage at my house is 118vac. so to be down to 110vac at the ballast means I could get more power just by replacing the power cord and plugging the ballast into the 30amp outlet with a heavy cord.

I think this setup could be made to work perfect with a HMY575 lamp with no heat or cooling problems at all you might not have to use any fan on the ballast at all.
If some one wants to try the hmi1200 watt long term inside ( Mine will be used 2 to 4 hours a week outside with me standing next to it so if bad things happen I just unplug it and grab the garden hose ) using two ballast may still be a OK idea. The heat will be half the current on each ballast will be the same as stock. The real problem is the cost. A $100 lamp and two $100 ballast is not a cheep solution. The only problem I see is the start up. The igniter will see both ballast and their internal resistance. It might be necessary to separate the two ballast during start up and just switch the other in.

If I am unhappy with the brightness of the 1,200 hmi I might look at the 2500 watt HMI But they are always $220 and have even shorter life.


I can use the windings of my 1500 watt ballast for an inductor that will take 10 amps and not saturate. The secondary windings have a .4h value the primaries could be used as a mulitap inductor too.( not at the same time) This would be just for fun because I would not want to use a $90.00 inductor. I only have about 50uf more of cap though. I just may try to find a high power inductor to try. A lot of mig welders use them as filters I just need to find a dead one.
The primary windings on a lot of transformers might work too.

I have a new Lamp I don't know why really it just was a good deal on eBay It is the HQI 400 watt OSRAM Powerstar HQI-TS 400W/D.

This is a great lamp It is the right size and package and color temp for our PJ's but it is not sold here in the USA execpt on ebay and the ballast is a not standard either.

The purpose here is to see how much power some of the ballast I have will make with this lamp.

It is rated at around 430 watts

GadgetSmith also bought a few of these and I hope this helps

I wanted to try my 600 watt HBS ballast(S-106) to see how it would run the lamp. It started fine and after worm up it made 672 watts with the stock cap. The lamp looked fine and the color temp was still slightly blue. This power level is higher than most people would want to run the lamp so a smaller cap will be needed.

I have an e-ballast and a smaller cap to try a 32uf should be close.

A few pictures.

Click to view attachment

Click to view attachment

One Note is the lamp arc tube is about 38mm long but the electrodes stick way in too make the gap 28mm.

Kind of like my HMI lamp.

When are you going to post HMI lamp results. They are a bit scattered amongst forums. I would greatly appreciate it .

Thanks again

The first test with the MSI 1,200 watt lamp I could not fill the entire fresnel. The smaller arc made a tighter light pattern it would be vary close for a 17" LCD. So I need a longer fl condenser lens. I got one yesterday but it is the same 170mm fl.
I am also waiting on the Altman reflector to return and a new single ended HMI1200 watt lamp and a new color tran spherical reflector.

I don't want to put the lamp through 200 cycles if I don’t have to.

I still think that for most people the HMI575 watt will be perfect. What wattage are you thinking of? The right cap could be found fact If I had a lamp. A S51 ballast should work too.

I made a thread just for the hmi It had one post and no questions. I'm not sure the best place to put the stuff.

A supper light engine? Found a HMI ballast? Overclocking lamps?

A/V,

I believe this is the bulb that you have correct? ... Osram HQI-TS 400W/D from ebay. They are only $17 ea. What is the lumen rating for this bulb?

Click to view attachment

However, I'd like to see some screen shots, if you can provide some please. I'm interested in how well the color red appears (is it red or is it orange).

Cheers!

Fulcrum

I can't believe how cheap they are! I paid around AU$100 for mine




I use this lamp and I can say that the reds look just as good as the greens and blues. Have a look at a couple of screen shots in my plog. I’m running it using NAV control gear so it’s around 37000 lumens, 5200k here’s a link to the spec sheet that AV posted in my thread.

DJ

edit:
Now that you Americans have seen these lamps I bet the price is going to go up. :angry:
So this lamp is crap its colours are all wrong and it’s not even good enough to use as a flash light. j/k

For a long arc lamp It is the best I have see yet. 90 CRI and a vary small lamp shell. Mine is still blue but it has less than an hour on it. Who said it had crap for colors? I don't know how long the guy on ebay will have then for $17 I just bought it to do some test with may sell it.

We need to find a OK ballast for cheep. The venture ps 430 watt is looking good so is the S51 a e-ballast will work too. I have one and will hook it up soon.

mmm, whats the problem with the lamp? is it not standard 400W Metal halide compatible? need special ballast?

heheheheh now the light has been seen,,, spectrally speaking,,, too blue does not matter,,, RGB, thats where it is

Arizona,

I've been looking at using the Osram HQI-TS for some time. I've got the lumenlab eballast but not sure how well this bulb will run on it. I believe another member Tenderheart has tried it with the eballast and reported it to light but wasn't sure if it was as bright as it should be. Exclusiv-online sells this bulb in a kit with a ballast that sounds like it will run the bulb at higher power for extra lumens. If I can use my current eballast with this bulb and save some money I will be happy.

I've openend up my eballast but havet looked at it properly to see if I can up the power to the bulb, any ideas if this is possible. Would have thought a little increase in power would be fine with adequate cooling, just need to figure out the circuit as on first look there seems to be more to it than a simple switch mode.

Looking forward to seeing you light the HQI-TS.

Mike.

my4 already did it last year with HQI osram,, it fires the lamp ok,, (eballast that is)

I knew this was coming. The e-ballast should make 400 watts with any lamp that it can run It just does the voltage x current to make 400 watts. I think Exclusive-online may be selling what they call NAV gear which will run the lamp at 430 watts. I'm sure the lamp can take more power than that and still have good color temp.

I could hook up a e-ballast ballast for a bit and test the power level. I don’t know the frequency that they run at but I'm sure it's over 2kH so my RMS meter wont work but the scope should show me the wave form and if it is a square wave I can make a good guess of the power level.

I was thinking of a way to cheat to find the power level that most people could do with out a true RMS meter. One way would be to use a wire of known size and measure the temperature rise with the lamp current passing through it any good the meter will be all that’s needed.

And of course there is the question of can we get an e-ballast to make more power? I sure we can if some one want to help me find the part to change I do have a decade box.

Exclusive-online price is fairly good for what you get.

Has any one found some one with a socket for the HQI?

If some one has coil S51 ballast I'm sure it would work just fine but it may need a cap change.

"Looking forward to seeing you light the HQI-TS."

I really wanted to spend time on the MHI lamps and maybe get back to the HQI some day.

ok, this could tell you the current (previously done temp Vs current table). But what about the volts?

Sorry AV, not trying to intrude on your MHI tests.

I'll prolly just buy the HQI-TS and try it on the eballast first, just not sure if i wanna go all out and buy the whole kit with ballast from exclusiv or not. Would be nice to know if the eballast can be easily changed for a little more juice.

Elken, how'd the HQI-TS look on your eballast?

Thanks.

Well I sold my 1500 watt ballast and EYE lamp to some one on LL. This left me without the 32uf cap I was using to make my 600 watt HPS ballast run my MHI 1,200 watt lamp What to do? Well I don't have the right combo of caps to add to 87uf so I could order a new $30.00 cap from venture,… Or I could try to use two ballast hooked in parallel to see if that works.

I used my 400 watt standard MH ballast with a 26uf cap and hooked the common to the other 600 watt common output leads and wired in a switch for the hot wire.

This worked fine!! I had to increase the cap on the 600 watt hps ballast to keep the lamp running during startup so I added a 24uf cap.

It made a lot of power and did not get hot or make a lot of noise.

The power levels with just the 600 watt ballast with the 55uf+ 24uf cap was, 902.4 watts and with the extra ballast switched in it was 1190.8 watts, perfect!

An only did two scope shots and it’s hard to see any difference in the waveform

Lower power

Click to view attachment

And high power

Click to view attachment


I could have made a mistake on the low power photos I can't tell which one it is. Both the low and high power waveforms looked almost the same.

If this worked with ballast with different powers(400 watt +600 watt) I see no reasons why two of the same could not be paired up if you needed to. If two can be paired up so could three?

So how does she look!


Im looking to purchase a HMI575W soon, and need your go ahead on which cap to use with which ballast. ( i know im the annoying guy asking the same questions 500 times)


with love

You could use either a S51 400 watt or a S106 600 watt ballast for the 575 HMI
I don't have one to test so I can only guess as to the right cap needed. If you go with the 600 watt it may make a little less power with the 55uf cap if this is trure then you will need to add one. It'is going to vary close to 575 watts at 95v but we will have to see. Ebay had a good price on the HPS 600 watt balast fron hightec garden.

The MHI 1200 looks great at full power and a little blue at 900 watts. I really like the high low switch idea so I can run the lamp at lower power and let it cool down slowly. The full power 1200 watts is a little whiter that the 11000 setting that I had before. I may add another switch for a 500 watt seting for a cool down before shut off setting. You may want to add a fan on these lamp while running but I would not cool the lamp after the power to the lamp is turned off. the slower they cool the better.

I did do some LUX test but The fl of the condenser is a little too short to cover a 19" LCD I may need to do some more to see if it will cover a 17" with the 165mm fl condenser if it does then that condenser lens will work good for the 575 and a 17" LCD. I have three.

What setup are you thinking of? The HMI lamps even with a condenser will work fine with all the triplets and be just as bright.

i was going with just the


proreflector hmi575 220mm 550mm 450mmbuhltriplet


I dont know anything at all about condeners. Is it worth the effort?

Thanks again

hey mikelish..you have a luxmeter don't you? if so can you take measurements without the lcd (using a cutout the size of your lcd(9x12 for 15" lcd))? we're trying to get some "build" measurements(non-lcd) of different combinations since lcd % vary so much
AV can answer you about the brightness, but I can say in my experience that using a precondenser really evens out the brightness if nothing else(vignetting)

just wondering if that evenes is achieveble by just placing the lamp a bit further (longer than 220mm from fresnell). Because one of the effects of introducing a rpecondensor lens is that it virtually extends the lens to lamp distance. (makes the lamp look further..)

good question Rox, never tried placing it further away than the focal length, I'll try that in the next couple of days...

My bet is that evenes will improve but at expense of lumens. This should be less significant with precondenser lens (evenes improved but not at expense of lumens...)

Rox, yes, this is what happens, vignetting % will increase (meaning more uniform lighting), but the overall brigthness drops pretty significantly. I'm assuming that this is why a precondenser is essential. It's location near the bulb does not change much, so the overall brigthness does not suffer while increasing corner lux readings.... which can lead to higher calculated lumens ? ... (not sure on this point, I would think the center point lux reading doesn't change much with the precondenser, but i've got to go back and look at some of the numbers people have listed...)

OK dont run the HQI on Eballast,,,, it does not suit

although I ran it for a couple of months, it did settle,, but tonight when imade another real Hi-Tech lamp mount,, it would not even try to strike.. at ll

so i dug up the mag coil and set it up,, it fired straight off,, very blue start then settled to pure whit again,, altho bluish around edges, being it is 6500k,,

my ushio had failed last thurs night too on the Eballast,, it struck straight up too on the mag,,, will reburn both of these on mag then try eballast again,,

I am not sure, and still have not measured the voltage out on Eballast,, but seem for some reason to think they only run at about 90 volts or thereabouts,, because it runs the HQI very dirty,,,

these lamps are 120 volt

its funny but so far I have not found the volt spec for the ushio

another spec is that some lamps only need 700 volts to ignite whilst others need 2kv to 3kv5

oh and its the voltage that sets the colour temp,,,, there needs to be more attention the proper matching of ballasts to lamps,,,,, you need the lamp specs and the mag specs too

i have seen lamps with a spec of 4.1 amps 120 volts being supplied with ballasts that are 100 v 3.1 amp,, and thats the supposed proper match,,, well guess what not in my books, and this might be why so many are getting many varied results etc

even if you supply a lower voltage at the right current,, the plasma is not going to be discharging correctly,,, its not just the current that is important,,,,,, its the voltage that maintains the correct excitement of the electrons, that raise up to the K shell where photon's are emitted,, or otherwise the electrons wont be stable,, also hence the dancing arc's etc. etc

votage = electron excitement
current = power


edit;;;; should I say that the krypton gas mixture inside the inner chamber and its density are what determines the voltage required to maintain a (whats called) a steady state of electron excitement,, and thats why I say that the voltage is very important to the best running of a lamp and its correct colour temperature

I have a question about capacitors. I don't know much but as I wander through the digikey/mouser catalog there are quite a number of types of capacitors. In the past I have used X2 rated caps for general AC line filtering and large electrolytic caps for DC line filtering and ceramics for the high frequency stuff. The application of general line filtering is not the same as starting or power factor correction or whatever application the cap in our lighting circuit is called. I have some 1uf 220vac tautalum caps lying around. Will these work or some of my X2 rated caps. I don't want them to explode because they can't disipate the heat or whatever other stresses they may see.

rearden

The voltage of a lamp will be a function of the current in the lamp and thus the temp of the lamp. We don't really get to set the voltage the lamp for a given current or temp. I can change the voltage of a lamp just by adding a reflector to the lamp and raising the temp. The lamp voltage will increase and the lamp will make more power, especially if the ballast does not limit current.

This is an effect I don't want to happen. Lamp gets hotter the voltage goes up I x V =P goes up and up.

The lamp clamps the voltage peak by its own internal resistance. It may be the cast that a highly peaked wave form may excite the metal a little different and thus change the color temp but we are not really doing this

A high voltage AC pulse might give a different reaction from the metals in the lamp. If the pulse was riding on top of the current wave. But we are not doing that either.

The e-ballast by making a pulsing AC square wave will have a lower peak ac voltage and may look slightly different but I have not heard a lot of people saying the color temp changed a lot for the same power level.



"I have seen lamps with a spec of 4.1 amps 120 volts being supplied with ballasts that are 100 v 3.1 amp,, and that’s the supposed proper match,,, well guess what not in my books, and this might be why so many are getting many varied results etc"



I was searching for HMI ballast one day and found a ballast manufacture that had a nice big ballast for 400, 575, 700, 1200, and 2500 watt lamp same ballast all they said is tell us what lamp you want to run and we can pre install a 70uf, 75uf, 80uf, 85uf or 145uf cap for the 1200 watt lamp all on the same ballast. I should have bookmarked it but it was not too hard to find under HMI ballast. They didn't say if the output was multi tap
Or if the input was but I think ballast has a fairly large range it can run over.

They say for ballast they like the running voltage to be about 60% of the open circuit voltage. If you look at the way they rate ballast they take open circuit voltage and shorted current. This is a important spec because it tell us the amount of current the ballast will make for the value of L/C if the ballast maker gets cheep and does not make the inductor large enough the ballast will act more like a transformer making a huge amount of current into the lamp at startup when lithe lamp has a vary low internal resistance.
This is more of a problem than the lamp voltage.

A ballast making 100v at 3.1 amps will run a HQI at around 320 watts. You would have to make the cap larger and then it will be "proper” I have not seen any one run a HQI on non standard ballast besides me. When my new caps get here I will try to make the HQI run at 430 watts like the spec sheet says and I’m sure it will have the same color temp as any other HQI running at 430 watts.

If I'm right the Tantalum are Electrolytic so are only good for DC not AC If they have a + or (-) marking then you can't use them.


I havent herd of X2 caps before. Most of the caps are still oil filled just like they were 20 years ago, now they want no PCB in them. The 24uf and even my new 55uf are dry caps. (film?)

The 24uf cap I have for the standard 400 watt MH ballast is 440vac The 600 watt HPS balast that runs a lower voltage lamp is 240vac.

I bought some of these

http://cgi.ebay.com/Used-2uF-660-VAC-Capac...itemZ3861982633

660v is way more voltage rating than you need but they are cheep and he shipps fast.

These are a little smaler and should be OK too. He didn't list shipping. There film so no oil if they pop

http://cgi.ebay.com/ASC-2uf-800V-HIGH-VOLT...emZ250000264559

800v dc is good for about 450vac.

What setup do you have now?

Another person was asking about using 600 watt electronic ballast for a HMI 575 lamp there is a chance it might work but it would help to look at how an e-ballast puts out current.

I used the MHI 1200 watt lamp to see if the ballast would try to put out its max current into the load. That should be 4.1A at 110v

I took some scope shots of the wave form but it was real messy. I was able to trigger at two different frequencies so I think there may be two wave forms. In other words it a two phases ballast.(edit elkin now says the e-ballast runs at 50kh so I'm set at half that so that explanes the double wave form)

My scope is tubes made in 1960 or so it's just work like crud at higher frequencies well most frequencies . But here’s a look.

Click to view attachment

Click to view attachment

I was expecting a square wave.

The ballast would only make about 3.1 amps into the MHI lamp with a 25v lamp voltage for about 80 watts output.( edit my RMS meter is not exact at this frq so don't bet on the numbers)

I did run it on a standard 400 watt MH lamp that I have and you could hear the high frequencies ring inside the lamp arc tube. I didn't notice this with the smaller sd400 I had.

One of the wave forms.

Click to view attachment

None of my meters are worth a crap at this frequency but I compaired the reading from the e-ballast to a reg ballast and the meter was .905 high so I guesed that would hold up for all the readings.

I don't know what to say about using a 600 watt electronic ballast for a HMI 575 lamp. It might be a big risk without having the ballast in hand.

I tried to mesure the input current on the e-ballast with two different lamps first was the standard 400 watt MH lamp after warm up it made 406 watts drawing 3.48 amps at 116.9vac line voltage.
The ramp up 2.9A to a peak of 3.48A

My 400 watt lamp rings most of the time with thr epballast some times real loud you can see it clear as day in the top of the peaks in the input current waveform.
Click to view attachment

It did settle down some times and then you can see on the 60hz input waveforl a nice high fq curent riding on the input wave form

Click to view attachment

Click to view attachment

Next was the HMI1200 watt lamp the ballast does not like this lamp but will start it but it never gets vary bright

Current was 2.51 amps at 116.9vac so input power was 293 watts.

The wave form now had noise on both hald of the wave form.

Click to view attachment

The RMS meter was working on the standard 400 watt lamp but with the MHI it has a lot of gosting.

I wached the current ramp up with the 400 watt lamp expecting it to peak when the lamp had around 110v on it but this did not happen.

yea this is the kind of shape I was expecting... some kind of envolvent of pure shine 50Hz and a hi frecuency subtone under that sine. Would you say the hi frecuency is identical to output frecuency? (25Khz / 50Khz?)

I would say that only a temp based poor man's current meter will tell us the "RMS value".

Thankyou AV

Rox: How did you like the lamp ringing? You could hear it from across the room. I would love to see what that would do to an LCD on the same power line. Every day I like the e-ballast less and less

I cant tell the FQ of the ring.

I never turned on my lamps at electronic gears, just magyes.
but 50Khz shouldn´t be heard (maybe your dog would hear it...)