With preheat ballasts for F40T12 lamps still readily available, why waste time converting a ballast? Advance has the L-140F-TP ballast and it is in stock on 1000bulbs.com and other places.

I looked at it, and I find it interesting that the product description doesn't mention that it works with F40 lamps. Probably a result of the new efficiency regulations. Unfortunately there are no magnetic ballasts for two F40 lamps.

Quote from: jrmcferren on July 09, 2013, 06:33:05 PM

With preheat ballasts for F40T12 lamps still readily available, why waste time converting a ballast? Advance has the L-140F-TP ballast and it is in stock on 1000bulbs.com and other places.

I looked at it, and I find it interesting that the product description doesn't mention that it works with F40 lamps. Probably a result of the new efficiency regulations. Unfortunately there are no magnetic ballasts for two F40 lamps.

No it does not state that it is for F40T12, but some Advance documentation states the same part number.

Then is there any with model L-240- ? ^_^

And nothing really prevents from using 2 1-lamp ballasts for the 2 lamps, this is ultimately how it is done here in 240v

It seems the two lamp model is RSHM-240-TP. These were made until the mid 2000s when they were banned by new efficiency regulations along with other HPF preheat F40 ballasts (single lamp HPF F40 and three lamp HPF lead-lag F40). In case you're wondering the corresponding model from Universal is 205-C-TCP. New old stock 2 lamp F40 preheat ballasts is common on eBay if you need to use one. 

Using two F40 ballasts also works too, in fact most of the F20 fixtures are wired that way and even some (non lead-lag) of the two lamp F20 ballasts were just two chokes with a common feed inside a case.

LPF ballasts were exempt from efficiency regulations and so are still made.

The description on 1000bulbs is missing F40T12 but if you click on the spec sheet link on that page, you'll see that F40T12 is also listed

I realize that it's easy to use two 1-lamp ballasts for a 2-lamp application, but I still wish the 2-lamp models were made. I wish HPF magnetic ballasts were still around; they're my favorite kind and also the most efficient magnetic ballasts, correct?

Capacitor+choke in series are the most efficient, but do they exist for 40w on 120v ?

Capacitor+choke in series are the most efficient, but do they exist for 40w on 120v ?

Such configuration need the mains to be at least 10% higher than the lamp arc voltage for steady operation, so on 120V it would not have any room for mains tolerance and dips.

And even when you would tolerate such sensitivity for the mains voltage, you would have problems to start the lamp.
And for the starting, it would need current (thermal,...), or voltage waveform shape (I had some discussion about how to implement this, but never seen anything like this commercially available) controlled starter, as there is not enough voltage room to fit a classic voltage controlled starter.

Glowbottle starters need the mains voltage to be at least twice the lamp arc voltage, electronic would suffice with a bit lower margin, but still they are designed along the specs of the glowbottles.

Five F15T8(55 volts at .3 amps) ballasts in parallel should overdrive an F40T12 (99 volts at .43 amps) slightly (if the lamp can start). You MUST use a manual start switch as there isn't enough voltage overhead for a glowbottle starter. The Five ballasts in parallel should give you an impedance of 42 ohms (correct drive would require 44 Ohms impedance). You have about 19 volts to drop in the ballast at 118 volts for an F40 lamp.

Let's expand this to the ever popular (on lighting gallery) F90T17 (62 volts at 1.55 amps). To drive this with a choke you need an impedance of 36 ohms in series. With 6 F15T8 ballasts in series you have an impedance of 35 ohms. This will work if the chokes put out enough of a kick to start the lamp. Since there is no glow-bottle starter designed for this application (voltage to low for the FS-6 Starter.

@jmcferren:
Calculating the ballast in that way (the impedance only) does not work. Beside the correct current, the ballast have to provide sufficient voltage for the arc restrike after each current zero cross. For a series choke ballast this impose a limitation to the arc voltage to not exceed about half of the mains voltage. With a series choke you have only the mains peak voltage for the lamp to reignite.
And you should not forget, than the voltage across the tube is nearly rectangular, so the simple "vector math" does not work anymore, when the arc voltage become close to the mains.

On top of that the ballast should maintain the current even over some mains voltage, as well as real lamp arc voltage variation, with so little voltage margin and no stabilizing element the spread would be simply too large.


What help (although I'm not sure, if the nearly 100V lamp is not too much even for that) is the use of a series LC, operated off-tune on the capacitive side: For first harmonics the capacitor's and inductor's impedances subtract from each other, forming the desired 44Ohm. But what makes it way different from the simple series choke and what is very important with discharges are two aspects:
First, when the current crosses zero, the capacitor is charged to it's peak voltage (corresponding to the designed impedances) and when the arc extinguish during the current zero cross, so the current stop flowing, the capacitor voltage add up to the mains, so the lamp have way higher voltage available for reignition (practically you can reach at least double the mains peak voltage).
Second aspect is related to the fact, than typical ballast coil lower their inductance at higher currents (part of the core saturate). Now if such coil is part of an series LC ballast operated on it's capacitive side, this reduction of the inductance at higher currents cause the LC to detune even further, so increase it's impedance. And higher impedance mean lower current, so the core saturation in fact mean the current stay pretty constant over quite wide voltage range, a feature you need when the lamp voltage approaches the mains, just to compensate for variations

But the manual start or a special electronic starter is still necessary, as all the effects before start to play their role only AFTER the current start to flow.

By the way this method is used in commercial ballasts, e.g. in this ballast , but only till F30T8...

@jmcferren:
Calculating the ballast in that way (the impedance only) does not work. Beside the correct current, the ballast have to provide sufficient voltage for the arc restrike after each current zero cross. For a series choke ballast this impose a limitation to the arc voltage to not exceed about half of the mains voltage. With a series choke you have only the mains peak voltage for the lamp to reignite.
And you should not forget, than the voltage across the tube is nearly rectangular, so the simple "vector math" does not work anymore, when the arc voltage become close to the mains.

On top of that the ballast should maintain the current even over some mains voltage, as well as real lamp arc voltage variation, with so little voltage margin and no stabilizing element the spread would be simply too large.


What help (although I'm not sure, if the nearly 100V lamp is not too much even for that) is the use of a series LC, operated off-tune on the capacitive side: For first harmonics the capacitor's and inductor's impedances subtract from each other, forming the desired 44Ohm. But what makes it way different from the simple series choke and what is very important with discharges are two aspects:
First, when the current crosses zero, the capacitor is charged to it's peak voltage (corresponding to the designed impedances) and when the arc extinguish during the current zero cross, so the current stop flowing, the capacitor voltage add up to the mains, so the lamp have way higher voltage available for reignition (practically you can reach at least double the mains peak voltage).
Second aspect is related to the fact, than typical ballast coil lower their inductance at higher currents (part of the core saturate). Now if such coil is part of an series LC ballast operated on it's capacitive side, this reduction of the inductance at higher currents cause the LC to detune even further, so increase it's impedance. And higher impedance mean lower current, so the core saturation in fact mean the current stay pretty constant over quite wide voltage range, a feature you need when the lamp voltage approaches the mains, just to compensate for variations

But the manual start or a special electronic starter is still necessary, as all the effects before start to play their role only AFTER the current start to flow.

By the way this method is used in commercial ballasts, e.g. in this ballast , but only till F30T8...

Wow, I'm so used to dealing with LEDs it is crazy. I just realized that I for forgot to calculate that those lamp voltage numbers were based probably at the maximum voltage across the lamp.

Fortunately, there are a few other countries that still seem to have some preheat fluorescent HX autotransformer ballasts that are designed for 4 foot 40W T10 and T12 fluorescent tubes that also should run on 120V 60Hz supplies no problem.

In Brazil, you can easily find a single lamp 40W preheat fluorescent HX autotransformer ballast designed for 127V 60Hz mains.

In Taiwan, there are still plenty of single lamp 40W preheat fluorescent HX autotransformer ballasts designed for 115V 60Hz mains.

Lastly, you can also consider the Eastern Japanese 100V 50Hz FL40S preheat fluorescent tube HX autotransformer ballasts due to the fact that they should have similar inductive properties on 120V 60Hz supplies.

Here is the math that proves this:

https://www.lighting-gallery.net/gallery/displayimage.php?album=6304&pos=0&pid=239122

Those places should be good sources for acquiring 4 foot 40W preheat fluorescent tube HX autotransformer ballasts for 120V 60Hz supplies if you can afford high shipping costs in some cases.