Author Topic: some experiments  (Read 3420 times)
hannahs lights
Member
*****
Offline

Gender: Female
View Posts
View Gallery

some experiments « on: April 22, 2016, 01:59:00 PM » Author: hannahs lights
I've been given an old circular lamp fitting and tube to play with it works well so I tried replacing the choke with a 15 watt bulb (the tube is 22 watt) and the starter tried its best but wouldn't strike the lamp that way so I tried increasing the voltage to 265 volts from my variac and it still wouldn't strike our normal mains here is around 245 volts so my question if I put a 60 watt normal bulb in series with the tube will it destroy it? I don't want to kill the tube
Logged
funkybulb
Member
*****
Offline

Gender: Male
View Posts
View Gallery


Re: some experiments « Reply #1 on: April 22, 2016, 02:47:36 PM » Author: funkybulb
One is u dealing with 60 arc volt lamp for fluorescent
And the 60 watt incandescent ballast u are using is under 250 mA and cant properly heat the cathode
100 watt incandescent is 410 mA a closer match.
The 15 watt lamp is 61 mA no wonder it wont strike.
« Last Edit: April 22, 2016, 02:51:04 PM by funkybulb » Logged

No LED gadgets, spins too slowly.  Gotta  love preheat and MV. let the lights keep my meter spinning.

hannahs lights
Member
*****
Offline

Gender: Female
View Posts
View Gallery

Re: some experiments « Reply #2 on: April 22, 2016, 06:30:58 PM » Author: hannahs lights
Thanks funkybulb il try it with a 100 watt bulb and see what happens both at normal volts and with increased voltage il let you know what happens
Logged
Ash
Member
*****
Offline

View Posts
View Gallery


Re: some experiments « Reply #3 on: April 22, 2016, 08:16:42 PM » Author: Ash
The ballast not only limits the current, it provides an arc striking voltage and discharges some initial kick of energy into the newly created breakdown path, to make sufficient ionization there so it can keep going on below mains voltage

If the lamp is not flashing at all, there is no breakdown. The size of Incandescent have no effect on the available voltage, its still on the order of 240V, or whatever the glow in the starter clamps it down to

If the lamp is flashing but not "catching on", there is breakdown and some little cold discharge at times where the starter is not shorting, but no sufficient ionization to make the discharge grow into a "full" arc. Those cold discharges are usually at very small currents, so there is no voltage drop on the Incandescent and its full 240V across the FL lamp anyway

Logged
Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: some experiments « Reply #4 on: April 23, 2016, 01:07:03 AM » Author: Medved
@Ash: It is true the resistive ballast gives no kick, but the T9 circlines (I understood it is FC8T9) have no problem starting with just the 230V or so (even 2 lamps in series do ignite, but probably only with manual starting switch), but only with heated cathodes. And there the ballast bulb wattage is, what matter.

The 22W T9 circline is a 0.36A lamp, so filament preheating current around 0.5A or so. That means you will need a 100W incandescent to really work correctly. Of course with that the power consumption of the system will be about 75W or so. With two fluorescents in series the same setup will light both with the same input power...
Logged

No more selfballasted c***

hannahs lights
Member
*****
Offline

Gender: Female
View Posts
View Gallery

Re: some experiments « Reply #5 on: April 24, 2016, 01:43:26 PM » Author: hannahs lights
On Tuesday I will try my lamp with a 100 watt filament lamp as ballast I know it won't get the voltage kick it gets from a choke ballast so if it won't start on 245 volts mains I will put it on at 265 volts and try again
Logged
Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: some experiments « Reply #6 on: April 25, 2016, 02:19:31 AM » Author: Medved
What helps a lot is having the ballast on the Neutral side and a grounded metal (reflector,...) close to the lamp. The trick is to have as high as possible voltage between the filaments and the external electrode to help to form some ionization around the electrodes, this then lowers the required ignition voltage quite a bit.
Otherwise the T9 circlines do ignite quite easy...
Logged

No more selfballasted c***

hannahs lights
Member
*****
Offline

Gender: Female
View Posts
View Gallery

Re: some experiments « Reply #7 on: April 26, 2016, 11:27:40 AM » Author: hannahs lights
As promised I tried it and with a 100 watt filament lamp as ballast I initially tried it on the variac with 250 volts applied and amazingly it started a bit slower than normal but it lit! Then I tried it on straight mains thinking that maybe the variac transformer was providing some sort of inductive kick. I was pleased to see that it lit again just on the mains sequence as follows switch on faint glow in tube flickering a bit bulb not lit then starter closes cathodes glow bulb lit brightly then a couple of clicks from starter and tube lights and bulb glows a bit dimmer. Current measurement show 105 watts and around 300 ma current at the mains measured on a kil a watt type device
Logged
Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: some experiments « Reply #8 on: April 27, 2016, 02:49:27 PM » Author: Medved
I was just curious, so tried to extend the experiment: Via a 105W halogen I managed to light a series combination of the 32W and 22W T9 circlines.
But it needed a bit of fiddling:
- Short both "starter switches" (well, in my case these were just two jump wires...)
- Let both lamps preheat for about 5..10 seconds
- Start the 32W one (remove the preheating jumper wire)
- Then start the 22W one

The overall current is around 0.32A, total arc voltage is around 140V, the real power into the fluorescents is around 31W (all values indicated by a power meter connected to the lamps).
Logged

No more selfballasted c***

hannahs lights
Member
*****
Offline

Gender: Female
View Posts
View Gallery

Re: some experiments « Reply #9 on: April 29, 2016, 02:04:08 PM » Author: hannahs lights
I tried my experiment again and got some proper measurements with the choke in line as follows mains volts 236 current 345 ma voltage out of choke76.4 volts wattage measured 30.8 watts
With 100 watt lamp input 240 volts 73 volts after applied to tube 84.4 watts 356 ma  current and wattages measured at mains socket on kil a watt device ( British equivalent of)  voltages measured at convenient points in circuit referenced to earth
Logged
Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: some experiments « Reply #10 on: April 29, 2016, 06:56:35 PM » Author: Medved
I had the measurements directly across the fluorescent, because I have a setup, where I have one E27 socket, normally with 300W incandescent in it, in series with the input to a test connection with the meter (combined panel meter from e-bay). So I just replaced the 300W bulb for a 105W halogen and connected the fluorescents directly to the output of that test jig, so the meter's volatage input then became connected directly across the fluorescent, so I do know the input power there... It is well possible, the meter's supply circuit (a capacitive dropper supplied from the voltage input) may have interfered with the lamps (discharges do not like capacitors) and maybe get stressed a bit more than originally intended (due to the voltage waveform shape)...

With that I was just curious, if both fluorescents in series will light with just a resistive ballast...
Logged

No more selfballasted c***

hannahs lights
Member
*****
Offline

Gender: Female
View Posts
View Gallery

Re: some experiments « Reply #11 on: May 01, 2016, 02:14:49 PM » Author: hannahs lights
One thing I am a bit concerned about is that the lamp current and power consumed seem a bit high after all it's 22 watt tube and with a 100 watt bulb as ballast  it getting around 80 odd watts am I in danger of destroying the tube? Finally the tube is rectifying a to some degree a distinct flicker is visible but it's not visible on the ballast lamp
Logged
Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: some experiments « Reply #12 on: May 04, 2016, 02:05:10 AM » Author: Medved
The overall setup may consume 76W or so, but it is just because of the ballast inefficiency. With just one 22W tube, the tube gets barely it's 22W (even when the current is close to spec in rms value, what does "convert" to lamp power is the "rectified average" and due to the distorted current shape that is lower, I would guess 20W or so). The remaining 56W is then power dissipated on the ballast mainly as the IR heat (and some as the light as well). This low efficiency (barely 30%), so the high system input power is the main reason why the filament bulb ballast have disappeared so long time ago, it gave no overall system efficacy benefit than a regular incandescent (about 20lm/W include the light from the ballast bulb, compare to about 14lm/W for a way simpler and cheaper incandescent only light).
With the "proper" ballast choke the lamp power is 22W, the remaining 9W is then wasted as a heat from the ballast losses, so the overall ballast efficiency is about 65%. That means about 43lm/W, way greater than the 16lm/W of the filament bulb ballast.

What I do not believe is the lamp starting rectifying n the series resistor ballast. I would rather guess the 50Hz flicker around the tube ends gets more pronounced, just because the light is concentrated to narrower pulses compare to the series choke. Mainly the overall 100Hz flicker will be "stronger" due to the same effect (when the light is concentrated to narrower pulses with darker gaps, the flicker appear as way stronger to the eye than with wider pulses and smoother edges, even when the frequency is exactly the same)
Logged

No more selfballasted c***

hannahs lights
Member
*****
Offline

Gender: Female
View Posts
View Gallery

Re: some experiments « Reply #13 on: May 13, 2016, 02:31:20 PM » Author: hannahs lights
Hi again this time I got a 1982 vintage 2 foot flourescents tube and fixture to play with the tube is one of the old style one not thin like new ones also fitting is not a twist in one I'm sure you know what I mean.so I tried it with its choke ballast it flashed several times and started OK then I tried with a 100 watt bulb ballast it started the first time but second time it struggled and wouldn't go so after a while I tried again and it did start then it wouldn't start again after that seems It likes a delay between starts on the bulb but doesn't have any trouble on the choke. Next I looked at the current waveforms I noticed on the bulb there was a slight delay before it started conducting after each zero cross on the choke it had a very sharp turn over at the peaks of the wave. When I looked at the waveform of just the filament lamp I can see that our mains has the peaks of the waveform crushed by other loads these measurement were done with a 100 watt lamp straight off the mains. I was given a current transformer by my mum just recently that's why I'm able to do this measurement now
Logged
Medved
Member
*****
Offline

Gender: Male
View Posts
View Gallery

Re: some experiments « Reply #14 on: May 13, 2016, 04:38:00 PM » Author: Medved
It is not that much about if the mains voltage peak is one volt higher or lower, but it is about the time the lamp has to spent without any current around the zero cross. During this time the ionization is not supported by the current, so what happens is the ions only recombine, so the ionization decays down. More it decays, longer it takes to reestablish the conductivity again after the voltage is restored in the opposite polarity.
On the other hand with a resistive ballast the current disappears when the sinewave drops below the arc voltage and can not flow, till it reaches enough voltage in the opposite direction. That takes some milliseconds, what is quite a long time, so the ionization level has plenty of time to get reduced. Consequently it takes way longer for the ionization to reappear the next half cycle and it may even happen it does not get restored at all.

The lamps differ a lot in this, even those with similar arc voltages. And I won§t be surprised that much, if there would be significant differences among different makers of the same lamp designation (lamps are rated for either inductive and/or HF ballast, there all work OK, but for a resistive ballast they are not rated, so the functionality can not be guaranteed)

And another effect plays it's role:
With the inductor as a ballast, the current zero cross happens when the mains voltage is already quite far in the next halfwave, so once the current disappear, there is already enough voltage in the opposite polarity to start the discharge again, so the zero current gap lasts for barely few 10§s of us. So the arc gets reignited very soon, so it has no time to decay that much. Plus the ballast capacitance may lead the voltage to even swing way above. When the arc disappears, the voltage across the ballast is already the arc voltage plus the mains in the opposite direction. This is then the initial voltage of the LC circuit (ballast inductance plus whatever capacitances collect there) starts to oscillate, so after half of the resonance period the voltage across the ballast becomes the same as before, but opposite polarity. And that adds up to the mains voltage and so gives plenty of extra voltage (arc voltage plus twice the mains voltage at the moment of current zero cross; with a 100V lamp it means up to 100V+2*280V=660V in a hypothetical loss-less scenario; with even plenty of losses in a real circuit the peak may go way above the normal mains peak voltage) available to reignite the lamp. By the way the capacitor in the starter has quite significant effect on boosting the Q of that LC circuit, so makes this swing stronger.

With the resistor, there is nothing like that available for the reignition, just the mains voltage itself...
So no wonder the lamps work way better on a series choke frequently have troubles working on a resistive ballast (usually it requires at least manual preheat and frequently a some fiddling).

Logged

No more selfballasted c***

Print 
© 2005-2024 Lighting-Gallery.net | SMF 2.0.19 | SMF © 2021, Simple Machines | Terms and Policies