Hi all,
I was trying my Heathkit curve tracer on a gas discharge transient suppressor
(striking at about 200V) and I've found the following behaviour:

https://www.youtube.com/watch?v=DeKPkG5T1xc

Basicly it seems to heavily ring on the turn-off curve?
This kind of devices should just show a two lines curve in theory.
My curve tracer appears to work fine on other less exotic devices (BJT,
zeners).
Can someone explain what is happening?
Thanks

Frank IZ8DWF

On Mon, 3 Jul 2017 09:57:35 -0000 (UTC), frank
wrote:

Hi all,
I was trying my Heathkit curve tracer on a gas discharge transient suppressor
(striking at about 200V) and I've found the following behaviour:

https://www.youtube.com/watch?v=DeKPkG5T1xc

Basicly it seems to heavily ring on the turn-off curve?
This kind of devices should just show a two lines curve in theory.
My curve tracer appears to work fine on other less exotic devices (BJT,
zeners).
Can someone explain what is happening?
Thanks

Frank IZ8DWF

My guess(tm) is that the unspecified maker and model gas discharge
tube might be filled with a gas that exhibits a negative resistance
characteristic, similar to a neon lamp, and therefore may be
oscillating. Neon, argon, and krypton are common gasses. Some gas
discharge tubes include something radioactive to lower the threshold
voltage.
https://www.google.com/search?q=neon+lamp+relaxation+oscillator
You might try substituting a common NE-2 neon lamp for the suppressor.
If the mess on the screen looks similar, then oscillation is the
likely culprit.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Monday, July 3, 2017 at 3:01:28 AM UTC-7, frank wrote:
Hi all,
I was trying my Heathkit curve tracer on a gas discharge transient suppressor
(striking at about 200V) and I've found the following behaviour:

https://www.youtube.com/watch?v=DeKPkG5T1xc

Basicly it seems to heavily ring on the turn-off curve?

The impedance of a gas discharge, especially at low current
after a discharge event, is negative. So, it can oscillate.

Geiger and proportional counter tubes have a gas fill that includes
quench gas (CO, or methane, or somesuch) that is intended to
reduce the negative-resistance behavior. Those gasses may be
omitted from your device, or they may have been used up (depends on
how much 'test' discharge the tube has seen). So, this behavior
might not happen on an as-delivered device.

On 07/03/2017 05:57 AM, frank wrote:
Hi all,
I was trying my Heathkit curve tracer on a gas discharge transient suppressor
(striking at about 200V) and I've found the following behaviour:

https://www.youtube.com/watch?v=DeKPkG5T1xc

Basicly it seems to heavily ring on the turn-off curve?
This kind of devices should just show a two lines curve in theory.
My curve tracer appears to work fine on other less exotic devices (BJT,
zeners).
Can someone explain what is happening?
Thanks

Frank IZ8DWF


Maybe Barkhausen oscillation (aka the Barkhausen-Kurtz effect). This
used to show up in old-fashioned incandescent light bulbs that were
evacuated instead of argon-filled.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

Phil Hobbs wrote:


Maybe Barkhausen oscillation (aka the Barkhausen-Kurtz effect). This
used to show up in old-fashioned incandescent light bulbs that were
evacuated instead of argon-filled.

according to a wikipedia article, the Barkhausen-Kurz oscillator used
triodes. So I'm still puzzled.
I'll try to make a test setup with constant current on a GDT and try
to measure the frequency of the oscillation.

Frank

On Thursday, July 6, 2017 at 3:19:34 AM UTC-4, frank wrote:
Phil Hobbs wrote:


Maybe Barkhausen oscillation (aka the Barkhausen-Kurtz effect). This
used to show up in old-fashioned incandescent light bulbs that were
evacuated instead of argon-filled.

according to a wikipedia article, the Barkhausen-Kurz oscillator used
triodes. So I'm still puzzled.
I'll try to make a test setup with constant current on a GDT and try
to measure the frequency of the oscillation.

Frank

Not knowing Barkhausen-Kurz from Berkinstocks, I think Phil means it
has to do with the time delay. Enough gain and a time delay leads to
oscillations. (hmm just like in thermal control loops.)

George H.

On 07/06/2017 03:15 AM, frank wrote:
Phil Hobbs wrote:


Maybe Barkhausen oscillation (aka the Barkhausen-Kurtz effect). This
used to show up in old-fashioned incandescent light bulbs that were
evacuated instead of argon-filled.

according to a wikipedia article, the Barkhausen-Kurz oscillator used
triodes. So I'm still puzzled.
I'll try to make a test setup with constant current on a GDT and try
to measure the frequency of the oscillation.

Frank


You can get Barkhausen oscillations in light bulbs, as I said. If you
have a spectrum analyzer, you can find out pretty fast.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

On 07/06/2017 08:35 AM, wrote:
On Thursday, July 6, 2017 at 3:19:34 AM UTC-4, frank wrote:
Phil Hobbs wrote:


Maybe Barkhausen oscillation (aka the Barkhausen-Kurtz effect). This
used to show up in old-fashioned incandescent light bulbs that were
evacuated instead of argon-filled.

according to a wikipedia article, the Barkhausen-Kurz oscillator used
triodes. So I'm still puzzled.
I'll try to make a test setup with constant current on a GDT and try
to measure the frequency of the oscillation.

Frank

Not knowing Barkhausen-Kurz from Berkinstocks, I think Phil means it
has to do with the time delay. Enough gain and a time delay leads to
oscillations. (hmm just like in thermal control loops.)

George H.


No, it's a coupled plasma/surface effect. It's really cool. It peaks
down around a few tenths of a millitorr, but it's still appreciable at
higher pressures.

The light bulb thing was discovered by somebody turning on an
incandescent lamp and wiping out his reception.

I don't know if the pressure in the transient suppressor is low enough
for Barkhausen--if not, it's probably the neon bulb oscillation as
others have said.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

Phil Hobbs wrote:

You can get Barkhausen oscillations in light bulbs, as I said. If you
have a spectrum analyzer, you can find out pretty fast.

I do have one (of course, I'd say). My radio reception (hamradio bands) has
never been disturbed by incandescent lights though (and it's about
the only electric thing that doesn't make any RF nowadays).
In the next days (as I find time) I'll try to run the transient suppressor
with a small dc current and see if it oscillates (and at what frequency).

Thanks
Frank IZ8DWF

On Thu, 6 Jul 2017 07:15:39 -0000 (UTC), frank
wrote:

I'll try to make a test setup with constant current on a GDT and try
to measure the frequency of the oscillation.
Frank

Could I trouble you to disclose the maker and model of the gas
discharge transient suppressor that you're testing, and the model
number of the Heathkit curve tracer? I don't have a curve tracer but
I do have a high voltage variable power supply that I can manually
sweep to see if there are any negative resistance regions.

The manufacturers of gas discharge tubes should have something on the
gas discharge tubes and negative resistance.
http://www.littelfuse.com/~/media/electronics/product_catalogs/littelfuse_gdt_catalog.pdf.pdf
Yep, the graph looks much like a neon lamp. The tube triggers at 75
to 600VDC, and rapidly drops to 15VDC.

I have a different guess(tm). When the gas discharge tube conducts,
it effectively shorts the terminals of the device. I don't think your
Heathkit curve tracer will not like a short circuit. It should have
some kind of short circuit or overcurrent protection circuit which I
would guess(tm) is doing the oscillating. Insert a large value
resistor in series with the gas discharge tube and try again.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Thu, 6 Jul 2017 09:51:44 -0400, Phil Hobbs
wrote:

No, it's a coupled plasma/surface effect. It's really cool. It peaks
down around a few tenths of a millitorr, but it's still appreciable at
higher pressures.

According to Wikipedia:
https://en.wikipedia.org/wiki/Incandescent_light_bulb
the gas pressure in a common light bulb is about 70 kPa or 525 torr.
That's quite a bit higher than a few tenths of a millitorr.

The light bulb thing was discovered by somebody turning on an
incandescent lamp and wiping out his reception.

I don't quite believe it. In the early daze of light bulb research,
it was wrongly assumed that a better vacuum produced a better light
bulb. So, early light bulbs had a fairly high vacuum, which might
explain the RF interference, except that radio hadn't really become
common at the time. Eventually, someone figure out that it was the
water in the glass that was killing the filaments. Once the water was
baked out of the glass, subsequent light bulbs had a much lower
vacuum.

I don't know if the pressure in the transient suppressor is low enough
for Barkhausen--if not, it's probably the neon bulb oscillation as
others have said.

I don't think that the ceramic gas discharge tubes have a vacuum. More
like they are under pressure in order to lower the conduction current.
However, I'm guessing and don't have time to look it up right now.

Cheers
Phil Hobbs

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann wrote:
On Thu, 6 Jul 2017 07:15:39 -0000 (UTC), frank
wrote:

I'll try to make a test setup with constant current on a GDT and try
to measure the frequency of the oscillation.
Frank

Could I trouble you to disclose the maker and model of the gas
discharge transient suppressor that you're testing, and the model

I don't know, I've pulled them out of a scrap PCB.
They are marked as 201M TM4M (on two separate lines).
201 might be the discharge voltage which is in good agreement on what I see
on the curve tracer.

number of the Heathkit curve tracer? I don't have a curve tracer but
I do have a high voltage variable power supply that I can manually
sweep to see if there are any negative resistance regions.

it's IT-3121, connected to an HP-1801A/180AR as display.


The manufacturers of gas discharge tubes should have something on the
gas discharge tubes and negative resistance.
http://www.littelfuse.com/~/media/electronics/product_catalogs/littelfuse_gdt_catalog.pdf.pdf
Yep, the graph looks much like a neon lamp. The tube triggers at 75
to 600VDC, and rapidly drops to 15VDC.

I have a different guess(tm). When the gas discharge tube conducts,
it effectively shorts the terminals of the device. I don't think your
Heathkit curve tracer will not like a short circuit. It should have
some kind of short circuit or overcurrent protection circuit which I
would guess(tm) is doing the oscillating. Insert a large value
resistor in series with the gas discharge tube and try again.

series resistor was already selected to limit the current peak value, it's
a knob on the curve tracer.
I've tried going from 5K to 50K and oscillation remains, of course peak
current varies. Using 100K as series resistor doesn't make the tube ignite
at all. Max voltage value on this curve tracer is around 240V.
Do you mean I should add an external resistor? I believe it would just go
in series with the internal one.
Indeed these tubes are meant to adsorb transients, so once fired, their
impedance could become very low, that would explain what I'm seeing.
Frank

On 07/06/2017 01:00 PM, Jeff Liebermann wrote:
On Thu, 6 Jul 2017 09:51:44 -0400, Phil Hobbs
wrote:

No, it's a coupled plasma/surface effect. It's really cool. It
peaks down around a few tenths of a millitorr, but it's still
appreciable at higher pressures.

According to Wikipedia:
https://en.wikipedia.org/wiki/Incandescent_light_bulb the gas
pressure in a common light bulb is about 70 kPa or 525 torr. That's
quite a bit higher than a few tenths of a millitorr.


The light bulb thing was discovered by somebody turning on an
incandescent lamp and wiping out his reception.

I don't quite believe it. In the early daze of light bulb research,
it was wrongly assumed that a better vacuum produced a better light
bulb. So, early light bulbs had a fairly high vacuum, which might
explain the RF interference,

How exactly, if not Barkhausen? Hot wires don't produce a lot of RF IME.

except that radio hadn't really become
common at the time. Eventually, someone figure out that it was the
water in the glass that was killing the filaments. Once the water
was baked out of the glass, subsequent light bulbs had a much lower
vacuum.

If you read my original post, I pointed out that it was only the old
fashioned evacuated bulbs that showed the effect, not the modern argon ones.

I don't know if the pressure in the transient suppressor is low
enough for Barkhausen--if not, it's probably the neon bulb
oscillation as others have said.

I don't think that the ceramic gas discharge tubes have a vacuum.
More like they are under pressure in order to lower the conduction
current. However, I'm guessing and don't have time to look it up
right now.

You don't want to reduce the conduction current in a gas suppressor
tube, though, do you? My guess is that they'd be a torr or two, to get
lower breakover voltage, which is on the high side for Barkhausen, for sure.

Cheers

Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

On 07/06/2017 12:12 PM, frank wrote:
Phil Hobbs wrote:

You can get Barkhausen oscillations in light bulbs, as I said. If you
have a spectrum analyzer, you can find out pretty fast.

I do have one (of course, I'd say). My radio reception (hamradio bands) has
never been disturbed by incandescent lights though (and it's about
the only electric thing that doesn't make any RF nowadays).

Right, the argon-filled ones don't show the effect--the gas pressure is
much too high. If the transient suppressor has that much gas in it, it
definitely won't be Barkhausen oscillation.

In the next days (as I find time) I'll try to run the transient suppressor
with a small dc current and see if it oscillates (and at what frequency).

I'll be interested in the results.

Cheers

Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

On Thursday, July 6, 2017 at 9:51:54 AM UTC-4, Phil Hobbs wrote:
On 07/06/2017 08:35 AM, wrote:
On Thursday, July 6, 2017 at 3:19:34 AM UTC-4, frank wrote:
Phil Hobbs wrote:


Maybe Barkhausen oscillation (aka the Barkhausen-Kurtz effect). This
used to show up in old-fashioned incandescent light bulbs that were
evacuated instead of argon-filled.

according to a wikipedia article, the Barkhausen-Kurz oscillator used
triodes. So I'm still puzzled.
I'll try to make a test setup with constant current on a GDT and try
to measure the frequency of the oscillation.

Frank

Not knowing Barkhausen-Kurz from Berkinstocks, I think Phil means it
has to do with the time delay. Enough gain and a time delay leads to
oscillations. (hmm just like in thermal control loops.)

George H.


No, it's a coupled plasma/surface effect. It's really cool. It peaks
down around a few tenths of a millitorr, but it's still appreciable at
higher pressures.

The light bulb thing was discovered by somebody turning on an
incandescent lamp and wiping out his reception.

Huh, OK... Get any sort of link?
B-K tube,
https://en.wikipedia.org/wiki/Barkha...80%93Kurz_tube
sounds like a delay... like a reflex klystron.

Is it a surface plasmon thing?
https://en.wikipedia.org/wiki/Surface_plasmon_resonance

Back to Frank, Can you do an I-V (at DC) to see if there is
negative resistance. (i've never looked at a negative
resistance so I don't really know if I'm asking the right
question.)

George H.



I don't know if the pressure in the transient suppressor is low enough
for Barkhausen--if not, it's probably the neon bulb oscillation as
others have said.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

On Thu, 6 Jul 2017 15:21:48 -0400, Phil Hobbs
wrote:

On 07/06/2017 01:00 PM, Jeff Liebermann wrote:
On Thu, 6 Jul 2017 09:51:44 -0400, Phil Hobbs
wrote:

No, it's a coupled plasma/surface effect. It's really cool. It
peaks down around a few tenths of a millitorr, but it's still
appreciable at higher pressures.

According to Wikipedia:
https://en.wikipedia.org/wiki/Incandescent_light_bulb the gas
pressure in a common light bulb is about 70 kPa or 525 torr. That's
quite a bit higher than a few tenths of a millitorr.


The light bulb thing was discovered by somebody turning on an
incandescent lamp and wiping out his reception.

I don't quite believe it. In the early daze of light bulb research,
it was wrongly assumed that a better vacuum produced a better light
bulb. So, early light bulbs had a fairly high vacuum, which might
explain the RF interference,

How exactly, if not Barkhausen? Hot wires don't produce a lot of RF IME.

All that I know is that someone allegedly turned on a light bulb which
produced enough RF to wipe out their radio reception. My point was
that until 1916, when Irving Langmuir invented the gas filled light
bulb at GE, light bulbs were made with a sufficiently high vacuum to
produce Barkhausen noises. The light would also need a tungsten
filament, which was introduced in 1908. So, if the bulb really did
produce noise, it would have to have been manufactured between 1908
and 1916. The Barkhausen effect was discovered in 1919, so the person
flipping the light switch would have needed a pre-1916 light bulb, in
order to call the noise Barkhausen Effect by running the test after
1919.

Meanwhile the radio receivers of that era were crude at best
consisting of coheres, magnetic detectors, cat whiskers, liquid
diodes, and early triodes. Regular entertainment broadcasting didn't
begin until 1922. Therefore, whatever was being broadcast between
1908 and 1916 would probable have been amateur radio or military
signals. The receive technology lurched forward rapidly, but I
suspect there was nothing worth listening to until the 1930's by which
time all the light bulbs would probably have been gas filled.

If not Barkhousen noise, the light switch could have caused an arc
somewhere in the light circuit. It could have turned on something
else in the circuit, such as a carbon arc lamp, which generates enough
RF noise to clobber even the most deaf receivers of the day.

I'm also having problems believing that the weak paramagnetism of
tungsten, can generate enough RF to be heard on the presumably AM LF
(low frequency) receiver of the day. Also, I don't see why a high
vacuum is required since there are several YouTube videos
demonstrating the effect at 1 atm.
https://www.youtube.com/watch?v=H7nJi5episc
https://www.youtube.com/watch?v=YLycGnOCqLc

So, do I wrap some wire around an incandescent light bulb and feed it
to my spectrum analyzer? Or, is this a waste of time with modern gas
filled light bulbs?

except that radio hadn't really become
common at the time. Eventually, someone figure out that it was the
water in the glass that was killing the filaments. Once the water
was baked out of the glass, subsequent light bulbs had a much lower
vacuum.

If you read my original post, I pointed out that it was only the old
fashioned evacuated bulbs that showed the effect, not the modern argon ones.

I read that, which is why I'm questioning the timing. At the time
when there were radio receivers sensitive enough to hear Barkhousen
Effect noises, the light bulbs were all gas filled. Well, I guess the
person flipping the switch might have been using an antique light
bulb:
http://www.centennialbulb.org

I don't know if the pressure in the transient suppressor is low
enough for Barkhausen--if not, it's probably the neon bulb
oscillation as others have said.

I don't think that the ceramic gas discharge tubes have a vacuum.
More like they are under pressure in order to lower the conduction
current. However, I'm guessing and don't have time to look it up
right now.

You don't want to reduce the conduction current in a gas suppressor
tube, though, do you? My guess is that they'd be a torr or two, to get
lower breakover voltage, which is on the high side for Barkhausen, for sure.

I wish I could xray the ceramic GDT and measure the spark gap spacing.
My guess(tm) is that it's fairly wide at perhaps 1 mm, something
similar to a common NE-2 neon lamp. I wouldn't need to have a hard
vacuum in order to get a lower breakdown voltage when neon breaks down
so easily. So, if I don't need a vacuum, and want plenty of ionized
neon atoms to provide conduction, I would pressurize the tube with as
much neon gas as I could cram into the package.

Cheers
Phil Hobbs


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Thu, 6 Jul 2017 18:25:31 -0000 (UTC), frank
wrote:

Jeff Liebermann wrote:
On Thu, 6 Jul 2017 07:15:39 -0000 (UTC), frank
wrote:

I'll try to make a test setup with constant current on a GDT and try
to measure the frequency of the oscillation.
Frank

Could I trouble you to disclose the maker and model of the gas
discharge transient suppressor that you're testing, and the model

I don't know, I've pulled them out of a scrap PCB.
They are marked as 201M TM4M (on two separate lines).
201 might be the discharge voltage which is in good agreement on what I see
on the curve tracer.

I tried to identify the manufactory from the part numbers supplied.
Nothing. I would agree that 201 might mean 200V discharge voltage.

I found one in my junk box inscribed Joslyn M-3 2204-44. I also
couldn't find any specs on that device. I was going to try it on my
hi-v power supply, but ran into 2 problem. The power supply is dead
for no obvious reason. My workbench is currently being used to test
my latest Harbor Freight purchase, a 1x30" belt sander suitable for
sharpening all my dull kitchen knives. That's a priority project.

number of the Heathkit curve tracer? I don't have a curve tracer but
I do have a high voltage variable power supply that I can manually
sweep to see if there are any negative resistance regions.

it's IT-3121, connected to an HP-1801A/180AR as display.

Schematic and operating instructions at:
http://www.vintage-radio.info/download.php?id=225

On Pg 62 of the operations manual, it shows various display anomalies.
Some are loops which look something like your photo. They suggest
using the LOOP control to reduce the effect or lower the power supply
voltage. You've probably already tried this, but I thought I would
mention it anyway.

series resistor was already selected to limit the current peak value, it's
a knob on the curve tracer.

Yep. Resistor values from 0 to 1M. Aim towards the higher values
here.

I've tried going from 5K to 50K and oscillation remains, of course peak
current varies. Using 100K as series resistor doesn't make the tube ignite
at all. Max voltage value on this curve tracer is around 240V.

You might be too close to the conduction threshold on the tube. Hard
to tell without specifications. If the device appears to fail to fire
at high series resistances, it might have fired, dropped the voltage
to about 15v across the tube, and stayed there. Grab a DVM and
measure the voltage across the device. My guess(tm) is that it's
about 15v and holding.

For fun, put a capacitor (300v or high) across the tube with a high
series resistance and see if you can get it to oscillate as a
relaxation oscillator.

For additional entertainment, replace the tube with a common NE-2 neon
lamp and see if it does the same thing.

Do you mean I should add an external resistor? I believe it would just go
in series with the internal one.

No. I didn't realize that there was a series resistor in the curve
tracer. This switchable resistor should be sufficient.

Indeed these tubes are meant to adsorb transients, so once fired, their
impedance could become very low, that would explain what I'm seeing.

Yep. Good luck.

--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

wrote:

Back to Frank, Can you do an I-V (at DC) to see if there is
negative resistance. (i've never looked at a negative
resistance so I don't really know if I'm asking the right
question.)

well yes, I could do that probably. I have a 300V DC source but it's not
variable. I could try with resistor dividers and measure a few points.

Frank

Jeff Liebermann wrote:

it's IT-3121, connected to an HP-1801A/180AR as display.

Schematic and operating instructions at:
http://www.vintage-radio.info/download.php?id=225

On Pg 62 of the operations manual, it shows various display anomalies.
Some are loops which look something like your photo. They suggest
using the LOOP control to reduce the effect or lower the power supply
voltage. You've probably already tried this, but I thought I would
mention it anyway.

Yes, I've tried the loop control but it doesn't change much of anything on
that display.


I've tried going from 5K to 50K and oscillation remains, of course peak
current varies. Using 100K as series resistor doesn't make the tube ignite
at all. Max voltage value on this curve tracer is around 240V.

You might be too close to the conduction threshold on the tube. Hard
to tell without specifications. If the device appears to fail to fire
at high series resistances, it might have fired, dropped the voltage
to about 15v across the tube, and stayed there. Grab a DVM and
measure the voltage across the device. My guess(tm) is that it's
about 15v and holding.

the test voltage is a half waveform of the 50 Hz (in this
part of the world) coming from a secondary winding and some transistor
buffers. The voltage waveform is sent to the test probe and the X axis
of the oscilloscope. So if the trace on the scope goes all the way to 240V,
I think it's not firing. With 100K series resistor I can't see any
vertical (current axis) movement even at 1mA division.


For fun, put a capacitor (300v or high) across the tube with a high
series resistance and see if you can get it to oscillate as a
relaxation oscillator.

another thing to test, ok.


Frank

On Friday, 7 July 2017 05:51:38 UTC+1, Jeff Liebermann wrote:
On Thu, 6 Jul 2017 15:21:48 -0400, Phil Hobbs
wrote:
On 07/06/2017 01:00 PM, Jeff Liebermann wrote:
On Thu, 6 Jul 2017 09:51:44 -0400, Phil Hobbs
wrote:

No, it's a coupled plasma/surface effect. It's really cool. It
peaks down around a few tenths of a millitorr, but it's still
appreciable at higher pressures.

According to Wikipedia:
https://en.wikipedia.org/wiki/Incandescent_light_bulb the gas
pressure in a common light bulb is about 70 kPa or 525 torr. That's
quite a bit higher than a few tenths of a millitorr.


The light bulb thing was discovered by somebody turning on an
incandescent lamp and wiping out his reception.

I don't quite believe it. In the early daze of light bulb research,
it was wrongly assumed that a better vacuum produced a better light
bulb. So, early light bulbs had a fairly high vacuum, which might
explain the RF interference,

How exactly, if not Barkhausen? Hot wires don't produce a lot of RF IME..

All that I know is that someone allegedly turned on a light bulb which
produced enough RF to wipe out their radio reception. My point was
that until 1916, when Irving Langmuir invented the gas filled light
bulb at GE, light bulbs were made with a sufficiently high vacuum to
produce Barkhausen noises. The light would also need a tungsten
filament, which was introduced in 1908. So, if the bulb really did
produce noise, it would have to have been manufactured between 1908
and 1916. The Barkhausen effect was discovered in 1919, so the person
flipping the light switch would have needed a pre-1916 light bulb, in
order to call the noise Barkhausen Effect by running the test after
1919.

Meanwhile the radio receivers of that era were crude at best
consisting of coheres, magnetic detectors, cat whiskers, liquid
diodes, and early triodes. Regular entertainment broadcasting didn't
begin until 1922. Therefore, whatever was being broadcast between
1908 and 1916 would probable have been amateur radio or military
signals. The receive technology lurched forward rapidly, but I
suspect there was nothing worth listening to until the 1930's by which
time all the light bulbs would probably have been gas filled.

If not Barkhousen noise, the light switch could have caused an arc
somewhere in the light circuit. It could have turned on something
else in the circuit, such as a carbon arc lamp, which generates enough
RF noise to clobber even the most deaf receivers of the day.

I'm also having problems believing that the weak paramagnetism of
tungsten, can generate enough RF to be heard on the presumably AM LF
(low frequency) receiver of the day. Also, I don't see why a high
vacuum is required since there are several YouTube videos
demonstrating the effect at 1 atm.
https://www.youtube.com/watch?v=H7nJi5episc
https://www.youtube.com/watch?v=YLycGnOCqLc

So, do I wrap some wire around an incandescent light bulb and feed it
to my spectrum analyzer? Or, is this a waste of time with modern gas
filled light bulbs?

except that radio hadn't really become
common at the time. Eventually, someone figure out that it was the
water in the glass that was killing the filaments. Once the water
was baked out of the glass, subsequent light bulbs had a much lower
vacuum.

If you read my original post, I pointed out that it was only the old
fashioned evacuated bulbs that showed the effect, not the modern argon ones.

I read that, which is why I'm questioning the timing. At the time
when there were radio receivers sensitive enough to hear Barkhousen
Effect noises, the light bulbs were all gas filled. Well, I guess the
person flipping the switch might have been using an antique light
bulb:
http://www.centennialbulb.org

I don't know if the pressure in the transient suppressor is low
enough for Barkhausen--if not, it's probably the neon bulb
oscillation as others have said.

I don't think that the ceramic gas discharge tubes have a vacuum.
More like they are under pressure in order to lower the conduction
current. However, I'm guessing and don't have time to look it up
right now.

You don't want to reduce the conduction current in a gas suppressor
tube, though, do you? My guess is that they'd be a torr or two, to get
lower breakover voltage, which is on the high side for Barkhausen, for sure.

I wish I could xray the ceramic GDT and measure the spark gap spacing.
My guess(tm) is that it's fairly wide at perhaps 1 mm, something
similar to a common NE-2 neon lamp. I wouldn't need to have a hard
vacuum in order to get a lower breakdown voltage when neon breaks down
so easily. So, if I don't need a vacuum, and want plenty of ionized
neon atoms to provide conduction, I would pressurize the tube with as
much neon gas as I could cram into the package.

People do use old bulbs. A place I stayed in the 80s was still using 200w carbon filament bulbs in the bathrooms to stop them freezing. Back in 1900ish nothing would have been thrown away. Even here I recently used a lamp that look like it was from the 60s or earlier.

ISTR being told 15w lamps stayed vacuum a very long time after 60 & 100w went gas filled, I don't know if that's correct or not.


NT

Just a little update, I've tried regular neon indicator bulbs (most of them
start conduction at around 75V) and only one showed oscillations and
only at low max current ratings.
https://www.dropbox.com/s/padumhxip1...-neon.zip?dl=0

Here're a few pictures of the V/I traces of regular small lamps.
Tilted traces are lamps with internal resistor, so the resistor is
tilting the trace.
Frank

On Sun, 9 Jul 2017 10:48:58 -0000 (UTC), frank
wrote:

Just a little update, I've tried regular neon indicator bulbs (most of them
start conduction at around 75V) and only one showed oscillations and
only at low max current ratings.
https://www.dropbox.com/s/padumhxip1...-neon.zip?dl=0

Here're a few pictures of the V/I traces of regular small lamps.
Tilted traces are lamps with internal resistor, so the resistor is
tilting the trace.
Frank

The one that oscillated (IMG_1950.JPG) looks very much like your
original video.
https://www.youtube.com/watch?v=DeKPkG5T1xc
It would be helpful if you increased both the vertical and horizontal
gain of your photos so some detail could be seen:

If you happen to have a few identical NE-2 lamps, put three in series
to get something close to the 200v as in the gas discharge tube. Then,
compare the displays.

Also, smaller JPG's please. This test does not require 12 megapixel
images.


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann wrote:
The one that oscillated (IMG_1950.JPG) looks very much like your
original video.
https://www.youtube.com/watch?v=DeKPkG5T1xc
It would be helpful if you increased both the vertical and horizontal
gain of your photos so some detail could be seen:

that's the max vertical "gain" that I can have on the IT-3121, 0.5mA/div.


If you happen to have a few identical NE-2 lamps, put three in series
to get something close to the 200v as in the gas discharge tube. Then,
compare the displays.

I've just tried that, no oscillations.

Also, smaller JPG's please. This test does not require 12 megapixel
images.

Sorry for that, I've not plaied with the camera settings.

Here's a video of the oscillating neon lamp, the higher current is obtained
with 100k resistor, the lower current (and larger oscillations) is with
500k limiting resistor:

https://www.youtube.com/watch?v=cgObvN8xiWM