Years ago I had a digital voltmeter that used these displays, and used
stepper relays for the "logic". I turned it into a clock.
The displays were OK, I prefer nixies. Having LEDs, with the ability to
control the brightness of individual digits would be a big improvement.
There rearmost digits
Thought I would post a follow-up regarding a RasPi as a nixie controller.
So far, using a RasPi Zero W for controlling my latest nixie clock (8-tube
b7971) has far-exceeded my expectations:
- No 'helper' FPGA is needed; there is far more than enough horsepower
in the Pi to handle all of
Was hoping for CD47's..
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to neonixie-l+unsubscr...@googlegroups.com.
To post to this group, send an email to
Be careful with 5031's.I've bought a few in the past from different
sources, and half of them had defects, most likely gas leaks.
They did not respond to depoisoning attempts at high current. Oh, they got
really warm. Someday I hope to melt one. Did that before with a 0A2 and it
was
On Thursday, June 15, 2017 at 6:05:42 PM UTC-7, Terry S wrote:
>
> 8 watts is a lot, you should easily be able to find that with your finger.
> It'll be hot.
>
>
>>
I re-measured, and after some allowances for the RasPi, logic,
LEDs/optoisolators, I have about 5W that's leftover and presumably
Have any of you MOD_6 owners measured the power consumption of the entire
clock ?
When I have 6 tubes running on my setup, it's around 25 watts
Well, you could make a really goofy single-digit clock with tube #10. Since
digits 5 & 6 are out, you could have a mechanism to turn the tube upside
down.
A mechanical version of what the Soviets did with some of their tubes, like
the IN-1.
--
You received this message because you are
Thanks for the warning. Generally, I leave everything plugged-in 'forever',
which is probably why I havn't had problems yet.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send
I've seen prices spike in the last few seconds; I was bidding on a set of 4
NIMO tubes a few years ago and the final price jumped 10X to over $1500 US
in less than a minute. You really have to know when to call it quits, or
enjoy your addiction.
--
You received this message because you are
I got my RasPi-based clock to display time last night, so proof-of-concept
is done.
Right now, I'm using seven b7971 tubes. All of the compute-related stuff is
done via a 'C' program running on the RasPi. I had planned on adding an
FPGA to offload the RasPi, but so far I have not seen a need
Does the TO-220 device get hot during normal operation ? If so, I would
suggest mounting it vertically and adding a heatsink.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send
One of the photos appears to show cracked glass at the edge, which almost
certainly would make the tube leak. Still, it's a great museum piece.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving
I will add that:
1. Not all nixies will respond to depoisoning attempts. I have a few
Burroughs 5031's (very early nixie) that are hopelessly gone.
2. It takes weeks to depoison at normal current.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l"
A101 ? How fast are you running it ? I'm amazed how well 3 distinct dots
show up.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
For those of us who dont need to shell-out $800 US for a soldering station,
but still want decent SMT capability, try a Weller WESD-51 . I've had mine
for over 2 years now and I have no regrets. It has enough power to solder
larger objects, such as panel switches.
--
You received this message
It's been 25-30 years since I used a metcal; it has the finest tip I've
ever seen and is perfect for SMT work. Even the finest tip for my Weller
(it's a decent one, not a cheapo) isn't as sharp as a Metcal.
Not sure how small you plan to go, but this one will easily handle 0.5mm
pitch ICs and
If you like segmented displays, and want to do some experimenting, this is
a chance to get some at reasonable price. I've seen these displays listed
for sale a few times. There's a good chance the equipment has a nice
transformer for producing nixie and logic supply voltages.
I've used a
Lots of high-quality free PCB tools out there, and plenty of inexpensive
good-quality PCB manufacturers on Ebay.
I stopped breadboarding years ago (and will never go back..) because PCB
costs are on-par with breadboarding, and much higher quality.
If you plan to make 2 or more boards, PCBs will
LE does the data-transfer
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to neonixie-l+unsubscr...@googlegroups.com.
To post to this group, send an email to
Based on the time-delays, the scope trace is showing 2 different signals at
the level-shifter, not input & output of the same bit.
Double-check the control bits; it's very easy to make a mistake with POL,
BL, and LE.
--
You received this message because you are subscribed to the Google Groups
I hope to find that out this weekend. My hunch is that segments with the
highest voltage-drop will be impacted the most. I dont have my IV curves
handy at the moment, so I cant say which particular segments would be
affected. I recall seeing 3 clusters of IV curves for each tube.
--
You
As far as "what" the socket pins mentioned in the original post are, I'm
fairly certain they are pins for PCB-mounted connectors. Normally, you
would buy the entire connector, pins included. These particular pins are
split at the end (good photo in Paul's link to the Ebay listing), and are
In addition to current-control per-segment, I also have an anode-current
limiter. It's probably overkill but it does cover the case where I have a
few characters that exceed to max total current.
I'm trying to fire this up for the first time this weekend. I've never used
a RasPi before, so I
Greatest thing since sliced bread..I first used them with my 14-tube
IN-18 clock, and also use them in my current b7971 clock.
I think one of our neonixie members sells them.
The pins themselves have low insertion-force onto common nixies (5092,
6091, IN-18, b7971 from my own experience)
The problem is that if the segment currents stated in the Burroughs
datasheet are summed-up for various characters, many characters exceed the
max-spec for anode current. However, the average character-current is
on-par with the max anode-current spec. Unfortunately, we'll never know
what
Luka - What segment currents are you using for the 7971 ?
If you turn all segments on at the rated datasheet current, there are
several characters, such as 8 and Q, that will cause the total current to
significantly exceed the max rating of 22mA. I analyzed current-draw for
most characters,
Assuming you have high-enough HV power-supply headroom for the nixies (180V
or more), the choice for the resistor begins with the isolated power
supply; I chose 12V because there are a lot of small, efficient DC-DC
converters that provide 12V. You can use other values, but be careful to
stay
Nobody in this forum is dumb; some of us just havn't been bitten as hard as
othersI found out the hard way you have to read and understand
every spec item in the datasheet, or you will get into trouble. I've
refused to use quite a few parts because a spec value wasn't provided;
other
A 5K-ohm load with 2mA thru it only drops 10V; with a 200V supply, you
would have 190V across the mosfet. At 2mA load current, that translates to
380mW dissipation in the mosfet. From the datasheet, the max Theta ja I saw
listed was 170 C/W. That would translate to a 65C rise, which is fine for
Can you determine (or measure) the current consumed by the nixes ? If it's
a direct-drive clock and the voltages you provided are measured across the
anode resistor (10K ?) the current is on the low-side, roughly 3.3 to
3.9mA. I believe the IN-18 is spec'd for 4-6mA.
--
You received this
One of my 6091's had darkened in the front, presumably from sputtering.
After running for several hours in a different position, and at 3X rated
current, the substance that was deposited on the inside front of the glass
has dispersed.
I dont know if it's cathode material, or mercury
On Friday, May 5, 2017 at 3:17:53 PM UTC-7, charles wrote:
>
> [... I belive Richarson bought up Very large stocks of any tubes thay
> would get their
> hands on in "warehouse" condition mostly, and Box and Brand them
> "National Electronics" ...]
>
Does that mean NL nixie tubes were
Nice design; I'm sure you spent many, many hours on the case. The nixies
almost appear to be 'wireless' in the photo.
BTW, my google icon is a picture of my big clock. The numbers are b6091's,
so you get a rough idea how big the clock is (22" across). The hands are
individual NE-2H bulbs
Terry- I wouldn't be surprised if they were re-dated. 1985 seemed too
recent, as LEDs & VFDs were churned-out like jelly beans at that time. Tons
of nixies from 1960's, some from 1970's. Mine tested-out fine, no signs of
darkening, no blue dots, etc. Definitely new-in-box, but questionable if
These are noticeably larger than the more-common 0.6" b5092s.
Not my auction, but I bought a lot of 6 (new-in box/mint condition).
Apparently the seller has a stockpile.
Never knew about the extra '8'.where was it located in the vertical
stack ?
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
I think the reason why pulsed operation makes IN-9's work better is because
when the tube goes wacky, such as lighting from the other end, etc, the
tube is turned-off in the next cycle, then restarts.
It's a lot like Microsoft windoze: Crashes often, for no apparent reason,
then is fine for
The only way I got my IN-9 to behave rationally was to use a
current-controlled power supply.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
My $0.02 here
"Tomorrowland" was the inspiration for my first-and-only IN-18 clock I
created. It even has a pushbutton to turn on the clock made from a
tomorrowland souvenir pin.
My wife instantly recognized the nixie clock in the movie and gave me
strict orders to make on for her.
--
I suggest you try another experiment: Try pulling the plug to turn it
on/off. Does that mimic the way the sign turns off with the smart outlet ?
Once you have that info, we can speculate about what the power switch in
the sign does, and also the smart outlet. It might not be a simple
However, for nixies you wont encounter enough load-current that would
require paralleling of devices.
Here's a tech paper, albeit old, that describes paralleling of MOSFETs in
detail. It was written by my college prof back in the 1980's. At that time,
there were far fewer choices for devices.
Clockulator - Functions as a calculator but times-out to be a clock if no
keys are pressed after a few minutes ?
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
I'd suggest you start with an IN-18 first, preferrably one that is
electrically and mechanically interchangeable.
If you can get the pin-to-glass seals to pass a helium leak-test, I'd say
you have cleared a major production hurdle.
--
You received this message because you are subscribed to the
It's done in FPGA code (verilog). I created a crude A-to-D converter using
a resistor tree into 4 FPGA pins. The resistors are ratioed to give the
FPGA indication when the anode voltage is 140,160, 180, or 200V. From
there, the FPGA adjusts the duty-cycle of the DC-DC converter to change the
My wristwatch uses a 'boost' approach to ionize the display above 180V for
25msec, then throttles back between 140 to 160V after the display is
stable. The saved energy is significant. It's 3-1/2 digits, direct-drive,
and uses NPN current-regulators for each segment (24 total).
My bench
I ran a few more tests on serializing a 64-bit stream using C-language and
the wiringPi library.
Typically, I'm getting around 32usec (measured on a scope) to send the
64-bit burst, which is decent for software-controlled GPIO's.
There are timers available in wiringPi, but I found they are not
I encountered problems with light reflecting off the nixies, and it also
washed-out the glow, so I thought the solution was to photograph in
low-light. Attempting low-light photography led to annoying reflections of
light from adjacent tubes. The picture I use for my icon from my big clock*
I have an FPGA as my 'Plan B', in case the there isn't enough
predictability in the Pi to display time down to the second. It wouldn't be
much work to modify the Verilog code I did for my previous clock to allow
the Pi to handle daylight savings, insert text messages, and possibly
correct for
I'm just using C, and so far I have not chosen a development tool. The info
I found on sparkfun mentions Geany (and it's included in the current
raspian image), so I will try that first.
There are sample programs included in the wiringPi distro for I2C and SPI.
Have not tried either yet, but
My concern with current limiters that rely heavily upon the datasheet specs
(Vgs for Depletion-mode regulator; Vbe for current-mirror) is that
variations due to process & temperature will have significant impact on the
actual current. Using a slightly more complex+costly design will mitigate
I bought a pair of 100x probes for Nixie/dekatron work. Rather inexpensive
on Ebay, maybe 20-25 USD for set of 2.
Fairly generic, and will work with typical scope that has a BNC connector,
and 1meg/20pf input impedance.
Although I'm using lab-grade equipment from the early 1990's, the features
I'm not able to reason-out how the last circuit acts as a regulator.
Basically, the path from the 200V supply to Rload (output) is 2 diode-drops
(base-emitter junction) in series with 100K of resistance. Assuming the
nixie tube takes about 150V to ionize, a current of 0.5mA thru R1 causes a
Dont use the current mirror without emitter resistors; it will likely be
inaccurate with discrete devices.
The current mirror as shown uses the operating point of Q2 (which is very
controllable and stable) to set the current through Q1. The problem is that
in this configuration, Q1's collector
There are lots of inexpensive isolated DC-DC converters. I've been using
products from Cui and Recom. Here's a 12V unit I use:
http://www.digikey.com/product-detail/en/VIBLSD1-S5-S12-SIP/102-1432-ND/989887
They are available in all sorts of input and output voltages. Be careful
that some
I do mine in 2-steps:
- Low-insertion socket pins soldered to PC bpard
- 3D-printed collar that sets over the pins. It's mostly for aesthetic
reasons, because the pins dont fit snugly into the collar. The collar
certainly helps guide the tube into the socket. However, you should
A word of caution...the next article in the allaboutcircuits link in Paul's
posting discusses current mirrors. You *can* use that technique for
multiple anodes with a few caveats
1. If you use the R1//R2 voltage divider, you will need to account for
base-current of additional regulators.
I'm certainly not expecting 40+ years of continuous operation from any of
my tubes, but if that's what you've getting I may as well give it a try and
run my 7971's 24/7 . I assumed that bigger tube + more pins + more
interior surface area & materials would lead to significantly shorter
My favorite: Burroughs 5092. Small (0.6" numbers), but very reliable. I've
never had one fail on me, and I've never encountered a defective one via
Ebay.
I have clocks with other tubes (6091, IN-18, and coming soon, b7971), but I
can tell they dont have the same ruggedness.
--
You received
To elaborate a bit more about what happens to the b7971 when the anode
current-limiter kicks-in, I will need to study the I-V plots I generated
from tube measurements. What will happen is that the anode current-limiter
will effectively vary the voltage at the anode to maintain the total
Timely subjectmy 7971 PC boards just cleared US customs and are on the
plane to Oregon
I have independent current-regulators on each cathode. 4 tubes * 15
cathodes = 60 current regulators. The driver transistor is a dual NPN in a
surface-mount package to save area. The cathode
I dont see anything fundamentally wrong; I just have paranoia about running
devices in an environment where this is potential for overvoltage. The
LM317 is rated for ~35V, and the HV supply is around 180V, so you are
relying on the voltage drop across the nixie tube. There are all kinds of
Some other things I forgot to mention...
1. Generally, you would use a current-regulator on the anode side for
non-segmented tubes (0-9), where all cathodes use the same current.
2. Segmented displays (b7971) have different currents for various segments,
so you will need cathode-side
That's basically what I use in my designs. I'll highlight the differences:
- I use a PMOS instead of PNP, mainly because it requires no
drive-current.
- R1 & R2 are replaced with a pot to make the current adjustable.
- The above pot can driven from a small DC-DC converter (my
IN-9's are 2-terminal devices, so they are not quite the same. However,
there could be a similar mechanism that impairs their performance.
My limited experience with IN-9's is that they can behave erratically,
such as not fully illuminating at rated max-current (even at 2X, I only got
about
I've seen this on 5031's; I was unable to cleanse them even with very high
current (tube got very warm...)
If you cant get the 6844's to clear-up, switch to 5092's. They are
trouble-free; I have not had a single failure out of 24 tubes I've been
running for the past 6 years.
--
You received
Just guessing here, but the supply of IN-18's continues to dwindle.
Meanwhile, Dalibor continues to crank-out more tubes, so my crystal ball
says you will see a kit with some flavor of new-production nixies.
--
You received this message because you are subscribed to the Google Groups
I think a lot of kickstarter campaigns are started by people who have a
good idea, but dont have first-hand experience taking a concept into
production. It's NOT easy, and it's not cheap.
Every time I finish another clock, my wife asks me why I dont sell them. My
usual answer is that I make
I chose the RasPi zero-W (the newest Pi, for $10 US) because it has
built-in WiFi to do periodic time-sync, and it will allow me to update my
clock software remotely, rather than taking the clock apart each time to
update FPGA code.
My concern is the limited GPIO speed of the Pi, so I will
Based on your info, gas-leaking theory sounds reasonable to me.
I have a fourteen-tube clock with IN-18's, and they are similar date-code
(1991), and similar age (started service Feb 2016) so it's a good baseline
for comparing. I start the clock around 7AM in he morning and it runs all
day. At
Let's hope that LTspice doesn't get torpedoed now that Analog Devices is
in-charge. It's really handy for quick circuit analysis.
I had a similar fear when Intel bought Altera; so far those free tools are
still free.
--
You received this message because you are subscribed to the Google Groups
Someone mentioned coilcraft as a supplier in another thread. I'll give that
a +1. Good selection on their website, some design tools, and they
direct-sell in single-unit quantity.
Make sure you have a good scope, and NEVER use the ground-clip from your
probe. Instead, tie directly to the
Anyone know how the radiation "hazard" of nixies compares to Xrays produced
by color televisions/computer monitors that used CRTs ?
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it,
My $0.02 here
I assume this is for a nixie supply; what you probably want is a flyback
converter; instead of a single inductor, there is a transformer. The
operation is similar.
Simulations will help somewhat, but what I've found is that modeling of
non-ideal transformers (aka, coupled
I no longer breadboard projects, because PC boards are so inexpensive and
their quality/neatness far exceeds anything I could do on a perfboard. I
recently did a small HV supply for nixies; for $21 US I got 3 high-quality
PCBs from OSH Park. They charge $5 US per square-inch, which is great for
*If everything is properly connected*, even an internal short in the 74141
that leaked high-voltage back to the Arduino, which seems unlikely, would
be limited to a few mA so it's unlikely to cause damage.
Your biggest risk is a hookup error, or possibly ESD damage to the Arduino
from
1 tube may produce 0.4uSv/Hr, but a clock with many more tubes will produce
more radiation. I have a total of 47 nixie tubes operating in my home, plus
another 6 at work. However, most of my tubes are at least 50 years old, so
I'm not too concerned.
--
You received this message because you
Considering the prices of IN-18's these days, if you can get an assembled
6-tube clock for 500 USD, it sounds reasonable to me.
I wont confess how much I've spent completing 4 nixie clock designs the
past 6 years, but to give you a rough idea each run for a PCB is at least
$100 ($300 for a
Take a close look at the signal amplitudes (best to use a scope); if the
resistor values are causing the logic gates to be overloaded [either due to
wrong IC, wrong logic family, wrong resistor value, solder short/open] ,
you will see incorrect levels. Generally, logic-0 will be 200mV or less
Clocking problem. I've been looking at the peculiar buffering on the clock
signals (7417 open-collector buffer), and cant come up with a good reason
why R11, R13, and R15 are present. I'd be especially concerned about
unstable operation, as this circuit has positive feedback.
Are you using
While on the subject of VFD's, anyone know of a particular VFD that has a
long lifespan ?
I'm reluctant to build a clock with them unless I'm certain they have a
long lifetime, as I've seen the VFD on my kitchen stove is fading near the
filaments (I suspect electron bombardment on the cathodes
OT ??? Nope; every nixie clock needs a case and lots of us use wood. Thanks
for posting.
Originally I thought this would be more like a "3D anti-printer", which
would move the router for you in XYZ direction. Interesting idea, but it
still requires the operator to do the work, rather than
I see no reason why it wouldn't work; the voltage-drop across the diode
will vary as needed to limit the current. This is the typical behavior of a
current source; even a resistor behaves somewhat in this manner. The
datasheet states the device can handle 600mW, so if you are running at 6mA
If your tubes are socketed (assuming they are, because they re IN-18), it's
probably best to run them from a separate supply. If your supply has
current-limiting, this is easy: Just set the voltage higher than the
ionization voltage of the tube (around 160-180), then increase the current
to
Nice find!!! I've never heard of these devices before. Should work
exactly as you mentioned. I dont see them at DigiKey, but I do see a nice
selection at Mouser.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group
I cant help wondering if the sheer size of that tube makes it more
difficult to manufacture with a high-quality seal. They weren't produced
anywhere near the volume of other nixies, such as the 5092, so I doubt
there was the same level of know-how at Okaya that there was at, say,
Burroughs.
For the battery-powered clock, I'd suggest 2 separate batteries. A
coin-cell for the timekeeping chip, such as DS3231, and another battery
that handles everything else. There are some DC-DC converters on Ebay that
might work for the nixies; no idea about quality or efficiency but for $7
US
On Thursday, February 23, 2017 at 3:20:46 PM UTC-8, NeonJohn wrote:
>
> You could use the modern version of the old edge-lit displays. A stack
>
I pondered that awhile back, but I remember my first-ever digital clock I
built used those same displays (along with stepper relays,
One of the (many) projects in the back of my head is a 'beer bottle nixie
tube', basically cutting-off the top of a brown beer bottle and inserting a
PC board with orange LED's arranged like nixie numerals. The tinted glass
would obscure most of the PC board, and blur the LEDs so it would have
I quick-glanced the HV5812 datasheet, and it's a push-pull driver, meaning
that it drives-out approx 0 volts to turn the tube on, and 80V to turn the
tube off. Even though the anode supply is significantly higher than the 80V
rating of the HV5812 device, it's OK to do this because any leakage
That's the driver topology I use, except I omit the zener diode (good idea,
though, for NMOS) and the base resistor because I dont see a hazard without
them. If there is a power-on transient that occurs when the HV supply is
energized (very unlikely, because the HV DC filter cap would need to
Current-regulators are very simple to implement: NPN transistor with
emitter resistor. Though you can use NMOS devices, they have a Vgs(on) of
several volts, so they need to be driven from a higher voltage, around
+10V, otherwise minor variations in Vgs (it's temperature-dependent, and
varies
I took a quick look at the HV5532 datasheet, and the 15V output-low voltage
is at 100mA, which is far more current than you will need for a nixie. The
outputs will tolerate an anode supply of +220V, so that gives plenty of
margin.
So, if you have an anode supply of +200V, you can calculate a
Are you driving segmented tubes (like the 7971) ? If so, I would advise
against the HV-series drivers because you will need several different
segment-currents.
If it's a traditional 10-cathode nixie, where each numeral uses the same
current, the HV devices work great. There's a variety of
Read-up on Paschen's Law, which can explain why a leaky tube is harder to
ionize after it's pressure starts to increase.
I wonder if we should store our nixies in a hyperbaric chamber with
neon/argon.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l"
Meh, you can buy gobs of USSR nixies today for less than that. No time
machine required.
However, they wont have nearly the same quality or reliability as the
Burroughs tubes.
I hate to think that someday in the future, 7-segment LED displays might be
highly coveted and command high
Most people will think it's a bunch of randomly flashing lights, so I can
see why it wasn't a success. I can guarantee that if I had one of these,
everyone at home would insist I got rid of it.
Even though most people wont understand (or appreciate) the innards of a
nixie/VFD/Numitron clock,
How many displays do you have ?
You could build a box with all of the displays , controlled by a
WiFi-equipped RasPi , then send it messages from your computer
(time-of-day, temperature, stock prices, etc...)
--
You received this message because you are subscribed to the Google Groups
Special emphasis here on the constant-current source mentioned in the
review, which is an excellent way to lengthen incandescent tube life. You
can actually see the warm-up latency from the video in the ones-second
digit; it looks similar to cross-fading.
--
You received this message because
2.5mA sounds about right; must be the gas mixture/mercury that causes the
color.
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
I bought 4 of these last year on a whim because they are SO unusual
Problem is, electroluminescent displays have a rather short lifespan, on
the order of a few hundred hours. Not a good choice for a 24/7 clock.
They operate on low-frequency AC, on the order of a few hundred Hz.
Basically, the
901 - 1000 of 1373 matches
Mail list logo
