Waveform of a 750W Mercury Vapor Bulb

[SG-1]

Here is the waveform of a 750W Mercury Vapor Bulb with built in ballast.

A Fluke 196C Scope was used.

The last picture was taken through a solar filter like you would use to view a solar eclipse.

Power was through a 1KVA distribution transformer wired 120:240. It was in turn supplied by a variable auto transformer plugged into a standard 15A 120V receptacle.

 

[gar]

181028-2359 RDY

SG-1:

Interesting. The ballast is simply hot filaments, and therefore resistors. The mercury vapor bulb is a non-linear somewhat constant voltage load, but from the waveforms more than that.

Interesting that on the negative slope voltage crossing that the current goes to zero at the voltage zero. That I would expect. Not so on the positive slope. I don't know the internal construction of the mercury vapor bulb. Clearly there is electron emission going on at the voltage positive zero crossing. There are hot mercury and electrodes in the bulb. How are they configured?

.

 

[SG-1]

Here is a sketch of the inside. The screw shell is at the bottom.

 

[mbrooke]

Thank you! Analyzing everything here.

 

[gar]

181029-1706 EDT

SG-1:

I don't see the circuit as drawn working.

Your drawing shows the arc shunted across one resistor, the left one, and the other resistor as the current limiter. But both filaments, resistors, in the bulb appear of equal brightness.

If everything was in series, then what starts conduction? Is it just enough peak voltage?

Where is there any pool of mercury?

.

 

[mbrooke]

181029-1706 EDT

SG-1:

I don't see the circuit as drawn working.

Your drawing shows the arc shunted across one resistor, the left one, and the other resistor as the current limiter. But both filaments, resistors, in the bulb appear of equal brightness.

If everything was in series, then what starts conduction? Is it just enough peak voltage?

Where is there any pool of mercury?

.

X2- I don't think thats it- but then again that looks like a slightly different bulb then the ones I am familiar with. Perhaps a close up of the bulb?

 

[Besoeker]

Here is the waveform of a 750W Mercury Vapor Bulb with built in ballast.

A Fluke 196C Scope was used.

The last picture was taken through a solar filter like you would use to view a solar eclipse.

Power was through a 1KVA distribution transformer wired 120:240. It was in turn supplied by a variable auto transformer plugged into a standard 15A 120V receptacle.

Bit of third harmonic but not too bad.

 

[SG-1]

181029-1706 EDT

SG-1:

I don't see the circuit as drawn working.

Your drawing shows the arc shunted across one resistor, the left one, and the other resistor as the current limiter. But both filaments, resistors, in the bulb appear of equal brightness.

If everything was in series, then what starts conduction? Is it just enough peak voltage?

Where is there any pool of mercury?

.

I have made a new sketch. I did find an error inside the arc tube, which is corrected in this version. I see the same thing you do, except I now see the second (shorter) path for the right hand filament is through the aux electrode. Notice the carbon resistors in series with each aux electrode. I also see both filaments in series with the main electrodes.

There is no pool of liquid mercury. I do see lots of tiny spots of condensed mercury on the inside of the arc tube. I did find this web site with some information.

http://edisontechcenter.org/MercuryVaporLamps.html#inventors

 

[SG-1]

X2- I don't think thats it- but then again that looks like a slightly different bulb then the ones I am familiar with. Perhaps a close up of the bulb?

I work on getting you some pictures. The issue with that is the restraint the forum places on file size.

Are you familiar with 120V or 220V models ?


This bulb was most likely manufactured circa early 1960s.

 

[SG-1]

One other thing I failed to mention above is the aux electrode appears to have the diameter of a hair. I would suspect this is a very low current path.

 

[winnie]

Interesting that on the negative slope voltage crossing that the current goes to zero at the voltage zero. That I would expect. Not so on the positive slope.
.

Gar,

I don't think there is as much asymmetry on the positive and negative slopes as that.

The way the scope was set up, the voltage zero cross was half way between the two hash lines. I think that the zero cross of the current leads the zero cross of the voltage on both the positive and negative slope, by about the same amount.

Interesting that there is a bit of positive current when the voltage is negative, and a bit of negative current while the voltage is positive.

-Jon

 

[gar]

181030-0911 EST

SG-1:

Several comments.

The current waveform is not symmetrical.

I would expect the construction of the mercury bulb portion and related current limiting resistors to be roughly symmetrical from a mechanical and wiring perspective. Your circuit diagram does not show that.

In a gas discharge device at an instant of time one electrode is an electron emitter, a cathode, and the other electrode an electron receiver, anode. In a thyratron a hot cathode is the electron emitter, and the thyratron is a one way device, a controlled diode. Only the hot cathode is hot enough to emit electrons, and when polarity reverses the plate being much colder does not emit and thus no conduction. An ignitron is also a one way controlled diode. Here a mercury pool is the electron emitter with an ignitor rod to start conduction.

In a mercury vapor bulb both electrodes are both an electron emitter and receiver because both are hot and can switch roles from 1/2 cycle to the next.

Mercury vapor bulbs contain some excitable gas, like argon, that is used to get conduction going at a moderately low voltage. When the bulb is cold the mercury is condensed and pressure is low. The mercury is not in an excited state. The excited gas conducts at some moderate voltage defined by the pressure and warms the mercury. The warm mercury vaporizes, conducts, and temperature further rises as does pressure from the heat. Voltage goes up as pressure rises. More electrical power input from the higher voltage drop by about the same current set by input current limiting, more power dissipation in the conducting path, and more light output.

If your external circuit is unchanged, including the scope connection, and you rotate the bulb orientation 180 degrees in the vertical direction, then does the waveform change? Does the coincident zero crossing change from the negative slope to the positive slope?

.

 

[gar]

181030-0957 EDT

winnie:

You are correct. SG-1 did not have his zero base on a major grid line, but halfway between two. So we do have a symmetrical waveform and that makes more sense.

The current is lagging voltage probably from the external series inductance of the transformer and Variac.

.

 

[mbrooke]

I work on getting you some pictures. The issue with that is the restraint the forum places on file size.

Are you familiar with 120V or 220V models ?


This bulb was most likely manufactured circa early 1960s.

Both. 230 volt models are exactly like a regular mercury lamp, there is just a filament in series with the top electrode:

120 volt versions are a bit more complex, they have filaments in the arc tube (usually at both ends) that branch off from the main electrodes, and are shorted by a bimetal switch that opens when the arc tube heats up. This is to account for the lower starting voltage and work more like a florescent lamp when first switched on.

Your version simply has starting probes at both ends, which threw me off.

 

[SG-1]

181030-0911 EST

SG-1:

Several comments.

The current waveform is not symmetrical.

I would expect the construction of the mercury bulb portion and related current limiting resistors to be roughly symmetrical from a mechanical and wiring perspective. Your circuit diagram does not show that. I thought it would have symmetry also. I have re-checked the upper aux electrode is connected between the two filament resistors. The lower is connected directly to the screw shell.

In a gas discharge device at an instant of time one electrode is an electron emitter, a cathode, and the other electrode an electron receiver, anode. In a thyratron a hot cathode is the electron emitter, and the thyratron is a one way device, a controlled diode. Only the hot cathode is hot enough to emit electrons, and when polarity reverses the plate being much colder does not emit and thus no conduction. An ignitron is also a one way controlled diode. Here a mercury pool is the electron emitter with an ignitor rod to start conduction.

In a mercury vapor bulb both electrodes are both an electron emitter and receiver because both are hot and can switch roles from 1/2 cycle to the next.

Mercury vapor bulbs contain some excitable gas, like argon, that is used to get conduction going at a moderately low voltage. When the bulb is cold the mercury is condensed and pressure is low. The mercury is not in an excited state. The excited gas conducts at some moderate voltage defined by the pressure and warms the mercury. The warm mercury vaporizes, conducts, and temperature further rises as does pressure from the heat. Voltage goes up as pressure rises. More electrical power input from the higher voltage drop by about the same current set by input current limiting, more power dissipation in the conducting path, and more light output.

If your external circuit is unchanged, including the scope connection, and you rotate the bulb orientation 180 degrees in the vertical direction, then does the waveform change? Does the coincident zero crossing change from the negative slope to the positive slope?

.

The Plan: I will have time Friday to do this.


Operate the bulb in the vertical position base down.

Operate the bulb in the vertical position base up.

Operate the bulb horizontal.

Follow the circuit through the bulb in pictures.

If you think of something else to do, post it before Friday morning.

 

[mbrooke]

The Plan: I will have time Friday to do this.


Operate the bulb in the vertical position base down.

Operate the bulb in the vertical position base up.

Operate the bulb horizontal.

Follow the circuit through the bulb in pictures.

If you think of something else to do, post it before Friday morning.

Do you by chance have a regular Merc bulb? You can ballast it with a few incandescents.

 

[gar]

181030-2532 EDT

SG-1:

There is no need to do the orientation test. I made an erroneous assumption of where your zero base line was. Winnie pointed this out. Your waveform does seem to be symmetrical.

Try the experiment that mbrooke is suggesting.

.

 

[gar]

181031-1157 EDT

SG-1:

Where was voltage measured? Were the scope channels on AC or DC coupling? What means was used to measure current?

.

 

[peter d]

A merc lamp that uses halogen lamps as a ballast? I guess it was state of the art at the time.

 

[SG-1]

181031-1157 EDT

SG-1:

Where was voltage measured? Were the scope channels on AC or DC coupling? What means was used to measure current?

.

CH1: 10X probe attached to 12 gauge conductor between H4 & H1 about 3ft from the bulb.

Scope channels A & B set on DC coupling.

CH2: Fluke i200s Clamp set on the 20A scale around H4.