RZ568m learnings

[gregebert]

I recently bought my first 3 RZ568m (Dalibor's replacement for the expensive and rare z568m) and wanted to share what I've found out, and see if others have additional info to contribute.

1. Construction.  The most significant things I noticed are

• Pins are not mounted thru the glass. Instead, there is an actual PCB in the base of the tube with pins that connect to bondwires into the tube. The big advantage here is that inserting/removing the RZ568m tube will not put any stress onto the critical glass-metal seal at the base of the tube. I think the best way to insert/remove the tube is to grasp the base, NOT the glass.
• The crimp-style seal should reduce gas leakage and increase the tube lifetime because it's a longer barrier thru glass compared to other tubes/nixies.
• Fill nib at the top of the tube.  None of my other nixies have this; common for VFD and traditional amplifier tubes. Just be careful not to bump it.
• Big !!  Even though the symbol-height of 50mm seems like a small increase above a 40mm IN-18, the width is also substantially wider and visually it's much bigger tube than you would expect just by a mental extrapolation. I have a clock with fourteen IN-18's and I thought that was a big-enough tube.
• Consistency. I have 1 tube from a much earlier batch than the other 2, and the internal design is the same as far as I can tell and except for the serial numbers the tubes look identical. Even spot welds are the same number and location. 
• Mercury pellet / getter. I see 2 places where a small plate of material is located; I'm assuming it's getter material that absorbs hostile impurities from the tube..
• I strongly suggest you watch the video of how the tubes are built
  
• In my own opinion, this is a high-quality and solidly built tube. I saw no hint of workmanship or quality issues. It literally looks like it came from the assembly line at Burroughs or Valvo, etc.
• Each tube is shipped in a "collector's edition" style box with excellent foam padding, and then enclosed inside another shipping box. You probably could drop and kick it many times without breaking the glass.

2. Electrical. I plotted I-V curves for 2 tubes and see expected behavior.  I will post my plots in a later update.

• Current always increases when voltage is increased; this is important to ensure stable operation.
• Once ionized, the voltage drop varies minimally over the specified operating current (5-6.5mA in datasheet, 5-7mA on website). Both tubes were within a few volts of eachother. Around 130V for normal operation. Thus, it's probably not necessary to use a circuit based current-limiter; a simple anode resistor should be fine as long as you plan for tolerance, tube-aging, tube-to-tube variations, and power-supply variation.
• I did not measure the ionization voltage accurately, but it's below the datasheet spec of 170V. Stated another way, make sure your anode supply voltage is an absolute minimum of 170V under worst-case conditions. When I have more time, I will measure this for each numeral. I'll try to measure it and post the findings. What I expect is that the ionization voltage will not vary significantly across cathodes due to the anode mesh construction.
• Numerals are fully illuminated at a small-fraction of the operating current. At very low currents (less than 1mA), numerals are only partially glowing. In my opinion this is important to ensure uniform "wear" on the cathodes.
• The anode mesh fully surrounds all of the digits. I think this will prevent any sort of metal-whisker growth between cathodes, such as what I have documented on IN-1 tubes.
• I did not notice a large variation in brightness when the current was varied from 5-7 mA (the recommended operating range). Based on this, I will operate my tubes at 6.0mA, using direct-drive.
• You can buy a set of 3 PC boards for socketing this tube from oshpark.com for about 15USD total, including shipping. I made a small nixie tester using this and a thumbwheel switch.

3. Next steps

• I'll measure the ionization voltages, then post along with the IV curves.
• I will design my own set of PC boards for a RasPi-based clock in the future, but I dont have a timeframe due to so many other projects in the queue. It's more than a year from now.
• Over the next few months I'll acquire the additional 4 tubes I need for a clock, and collect data on them.

[Bill van Dijk]

Great summary!

 

When you do the ionization tests, could you do those in light and dark conditions? If not, at least document lighting condition under test, i.e. natural light, incandescent light, fluorescent light etc. I am curious if that has a significant effect.

--
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+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/4e8e9183-a463-4137-a4a7-99ea812b5453%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jeff Walton]

I’ve been running Dalibor’s RZ568 tubes for more than 3 years and they are certainly beautiful tubes/works of art. 

 

There have been some significant (visible) changes to the product since the tubes went into production and sales.  I’ve seen tubes with serial numbers in the hundreds (xxx); one thousands (1xxx) and my latest are in the seven thousands (7xxx).  I have tubes from the three series above and obviously, it is a small sample set but offers some insight into the  changes in the process and materials.
  

·         Beginning with tubes after about 1000, Dalibor added some very nice packaging that is currently used today.  As Greg described it, they are shipped in a collector style box with a custom fit foam liner.  Recently, the Millclock ZN18 has replicated the same style and color boxes that Dalibor is using.  Individual tubes are shipped in single tube boxes and full sets are shipped in a longer box that has foam cutouts for each tube.  It is very rugged packaging and much safer for shipping than the early hand wrapped packing material method.

·         Tubes after 1xxx have a metalized tube identification marking/branding on the back of the glass envelope which is both durable and much more professional than a painted-on marking.

·         Each tube carries a unique serial number on the inside of the tube which provides traceability for Dalibor to identify process and material changes.  For those that have followed the progress that Dalibor has shared, traceability is something that is taken seriously and has made it easier to identify processes that have been refined.  The serial numbers are hand written on the back of the Anode frame.

·         The color and appearance of the anode frame is improved after 1xxx.  The back surface of the anode is now a bright metal as opposed to the previous dull annealed color.  The inner face of the anode is now darker and appears to be a coated surface that is dark and provides less reflection and better contrast to the cathode digits stacked in front of it.  It is a noticeable improvement.

·         The shape of the anode cage changed after 1xxx.  Earlier tubes had a back frame of the anode that was solid about half way up the stack of cathodes.  This made side viewing much more limited, especially for the back digits in the stack.  The “0” was particularly impacted and required nearly head on viewing, or the outer left and right portion of the “0” would disappear as you moved off of front viewing to viewing from either side.  The newer tubes have the honeycomb portion of the anode cage extending nearly to the back of the anode plate and the side viewing is much improved.

·         The shape of the top of the glass envelope has become more domed in shape as opposed to the earlier tubes which were somewhat flatter than now.  It likely is a reflection of who is doing the glass firing and sealing work.

·         The internal pinch seal is a bit larger on the later tubes than the early xxx serial numbers.
 

·         For base lighting of the tubes, consideration needs to be given to the fact that there are insulated wires that are welded to the wires coming from the pinch seal down to the soldered connection of the PCB.  A wide angle LED would be preferable to a focused LED for light dispersion into the glass.  The connection wires coming to the PCB can be gently adjusted to provide a nice opening into the base of the tube for clear passage of light.  Obviously, in terms of light transmission, a pinch seal tube construction is no match for the pin/wire welded glass base that the original Z568 tubes and similar construction tubes from the “good old days”.

 

In my experience with Dalibor over the past few years, he has provided much, much more detail regarding each of the above topics along with specifics about the changes in construction.  I had a couple early leakage failures and Dalibor  has been excellent in resolving the problems.  He stands behind his work and I think he is very conscientious about making sure things are right before they ever leave his shop.

 

Jeff

--

[gregebert]

I did some basic testing of the ionization voltage, with the tube inside a closed box, and found that it would consistently ionize around 122-123 volts, and then the normal glowing would kick-in with the tube operating normally, though at a low current (about 500uA). I was expecting a much higher ionization voltage; regular NE2 bulbs have a much larger difference between the ionization and extinguishing voltage, on the order of 30 volts. With the tube exposed to ambient light, the ionization voltage was about 1-2 volts lower, which I did expect.

Though I did not test every cathode, the 3-4 that I did test were in the same ballpark. This suggests that if you have a 160V anode supply, you should have plenty of margin with these tubes.

Once I get a clock built, I will collect data again and then see if it changes after a few years of operation.

[Dalibor]

Hello!

Really good thread, thank you for all the feedback.

- Regarding consistency of the look - the final appearance (anode cage opening, digit-base distance, color of the backplate dark coating, logo on the back) settled down around S/N #1000. Since then we've made only minor changes so all tubes manufactured after #1000 can be replaced/mixed with new tubes. Consistency of the appearance, quality of the build and overal craftsmanship is my #2 priority.

- #1 priority is the gas purity and consistency of the gas composition inside. To ensure maximum lifespan, the tube should operate at lowest possible maintaining voltage (lowest voltage on tube = least intensity of the field = lowest velocity of the Ne ions bombarding the cathodes = lowest wear rate).The maintaining voltage is determined by the gas composition and even very low amount of impurities will influence the electrical parameters and lifespan of the tube (my first-hand experience). So, the maintaining voltage is usually within 1-2 volts and is around 130V.

- Power supply voltage: I decided to use min 170V for power supply to keep some margin because in 10-20 years the striking voltage may increase (gases evolving from inside the parts, glass..). Also, the higher supply voltage, the higher anode resistor you need to use and the closer you are getting to constant current power supply (and changes in tube has lower impact on current). The tube would work with 140V DC power supply with only small anode resistor, but any small change in characteristics of the tube would have serious impact on current flowing through it (decreasing it) and than the tube might stop lighting up or the glow would not cover whole digit uniformly. It is like a resistor divider - if the tube is only a small resistor compared to the anode resistor, then change in tube will have small impact on overal current flowing through the circuit.

When you look into datasheets of RFT, NL or Burroughs, all the jumbo tubes (50mm+ cathodes) need 200V power supply. This is to make sure the power supply has this constant current behaviour in case something changes inside the tube - at the time these jumbo tubes were produced, the manufacturers probably didn't have enough data to evaluate lifespan and possible changes, so they went for much higher margin. I made a risk and decided to go for 170VDC as it is easier to implement and safer for customer. Comparing to production of Tesla ZM104X tubes (reliable tubes), we achieve higher vacuum and cleaner pumping/filling process (I spoke to man who used to be their process engineer) - so the risk of gas evolution with time should be lower than in case of old jumbo tubes. Thats a theory, time will show the truth..

To ensure very consistent electrical parameters, we started development of a combined burn-in/tester (picture attached) which will measure the electrical characteristics during burn-in phase. Data will be send to server where it will be analysed and all tubes whose parameters differs (possible sign of contamination, leak etc..) from reference will be marked and rejected. The biggest issue with quality control is that there are no visible/measurable differences between contaminated tube and good tube. The issue will appear long after shipping (e.g. because of saturated getter), this tester should give us new possibility how to see into the tube during the burn-in process where we actually found some differences..

Have a nice day,

Dalibor

[Bill Notfaded]

Wow the new test STE is really nice! Thanks for the updates and picture. It's neat comparing the originals to the newly manufactured tubes. It seems the new tubes have tight tolerances.

Bill