b7971 segment current

[gregebert]

According to the b7971 datasheet, the maximum allowable current is 21mA. Yet when you add-up the currents for each cathode to produce the  '8' or '*' characters, they add up to 40mA, which is way over the spec limit.

For those of you who design their clocks, did you scale-back all segment currents so they never exceed 21mA ?

I'll probably current-limit each cathode individually, and put another current-limit on the anode. Even though that adds up to a lot of transistors for a direct-drive clock, these tubes are getting pretty scarce and expensive so I'm not taking any chances.

[jf...@my-deja.com]

I did nothing heroic.  I used the cheapest transistors I could find (PolyPaks, John Meshna, etc.) and carbon comp resistors.  I used the resistor values in the old "73 Magazine" article "Build a Giant Nixie Clock" from the mid-1970s, and increased the HV power supply voltage until it was "bright enough" (about 175V on my old Lafayette VOM).   The resistor values may have changed with time, but I think they were initially 1-2ma per segment.

I built three clocks with a total of 18 tubes.  So far, after 40+ years, there have been zero tube failures.  The $20 I spent on spare tubes (at $1 each) has been a waste of money.

The tubes were multiplexed.  I found that the biggest problem is that the tubes sing at the mux rate, and the volume increased with the drive current. (Someone once suggested that the sound was coming from the power supply, but this was way back in the old days and I only built linear power supplies operating at the line frequency of 60 cps.)

[gregebert]

> I built three clocks with a total of 18 tubes.  So far, after 40+ years, there have been zero tube failures. 

40 years ? How many hours per day are you running your tubes ?

[jf...@my-deja.com]

On Tuesday, August 23, 2016 at 10:25:09 PM UTC-7, gregebert wrote:

40 years ? How many hours per day are you running your tubes ?

Mine has been running 24x7x365.25, with brief exceptions for things like moving to a new home, power outages due to earthquakes (California!) and weather, etc.  The other two were gifts, but I have not needed to replace any tubes in those (and I get to check on them occasionally).   It was a basic design typical of the 1970s: linear power supply and a CT7001 clock chip.  It did not have the additional circuitry that would have been required to dim the bulbs or turn them off.  At $1/tube, there was no incentive to complicate the circuitry in an attempt to "extend" the life of the tubes, and as the experiment has demonstrated, it would have been a waste of time and effort.

[Jeff Walton]

Greg – my 1971 clock with Burroughs CK8754 nixie tubes ran 27/7/365 for over 35 years with no tube or digit failures.  It used 74141 drivers driven at ~170 volts by a rectified line voltage doubler circuit using recommended anode resistors from the example circuits out of the Texas Instruments data book.  During the life of the clock, there were a couple failures of the caps in the voltage doubler but no failures in the tubes or chips.  The clock was basic with no steps to prevent cathode poisoning and the tubes were actually soldered with wrapped wire around the pins.  The entire clock was point to point wiring.

 

While I can’t attest that the brightness was the same as in the beginning, the display tubes were still perfectly viewable with full digit coverage and easily visible in a room with windows.  In the early life of the tubes, I did note that there were occasional blue arcs that occurred with reducing frequency as the tubes “broke in”.  I think the arcing stopped sometime within the first year.

 

I came to the conclusion that a well designed nixie, which had good seal integrity, the right gases/doping and was operated within the specs could last a good long time.

--
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/265d4eac-fec6-491b-90a2-818055107d7f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[gregebert]

On Wednesday, August 24, 2016 at 12:21:56 AM UTC-7, Jeff Walton wrote:

>During the life of the clock, there were a couple failures of the caps in the voltage doubler 

 When your cap(s) failed, was it catastrophic ?  I've only had 1 electrolytic fail in recent history, and it was a low-voltage cap that dried-out and shorted at medium-resistance in a Heathkit device (not a clock). No smoke, etc.

I've tried to prevent/mitigate cap failures in my designs by using the smallest possible fuse, keeping the caps away from any heat, staying well below the rms/ripple current spec, and using a higher voltage rating  than necessary (eg 450v cap running at 340VDC).

Recently, I found caps designed for solar-energy applications (TDK Epcos) that boast 85C operation for 10,000 hours, so I use them now. Most electrolytics are rated for 2000 hours. That doesn't mean the caps will fail (ie, explode) in 2000 hours; they just wont be within spec (capacitance out-of-spec, but otherwise functional).

Electrolytics are a strange beast compared to other components.

[JohnK]

I haven't looked at electro specs closely since the mid '90s. I was involved with a product that used a conventional aluminium electro in an apparently undemanding application. The value was 10uF and variously two types were used- a 10VDC and a 25VDC [ or a bit higher - I forget]. The DC voltage across the cap was constant at a bit under 2 volts DC and was on one of the inputs to a comparator.  There were issues with the product but early in the troubleshooting the input circuitry came under scrutiny and it was noticed that the capacitors were being run significantly under their 'working voltage'.  The capacitor manufacturers [one Euro] were both asked for comment and both advised against the use of that style of electro under 75 to 80% of the working voltage. The reasons from both related to 'forming'. It wasn't clear whether they meant initial forming or continued forming though. To my mind it doesn't explain the seemingly satisfactory operation of electros as coupling capacitors in early transistor radios for instance. Hard facts concerning time-frames and numerical values for degradations weren't forthcoming.

[BTW, the product problem was actually related to the specs for a triac being considerably improved by the manufacturer; the old snubber values were now causing the problem. ]

 

 

John K

 

 

--

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/66df0632-699e-4d44-866b-a9c52e1fc701%40googlegroups.com.

[Dekatron42]

I ran across similar capacitor problem when using surface mount tantalum capacitors and MLCC capacitors, without the proper working voltage across them they behave poorly. I had to read a lot of design notes and datasheets to realise that they did not fit my design, MLCC capacitors has to be reformed by heating over time (high capacitance types at least) and they change value when stored and soldered. I didn't know about this before I was using them but had to change to polypropylene capacitors in the end. You can check the measurement specifications for MLCC capacitors and smd tantalums from several manufacturers as well as from the instrument manufacturers to see that these capacitors behave in manners that most people wouldn't think of. Just measuring them to get correct readings isn't possible with simple capacitor testers as the bias voltage when measuring is usually to low - learned this while trying to measure them!

/Martin

[rmp]

To the folks who are still running the "Giant Nixie Clock". From the early 1970's:
I built one of these way back when. Unfortunately, it is long gone, but as I recall it treated the tubes as 7-segment devices, and so the 2 middle vertical and the 4 diagonal segments will NEVER have been lit. Am I correct? It would be an interesting exercise to make a test jig that can illuminate all the segments and see how much, if any, the unused segments are brighter than the used segments.
Just food for thought.

[jf...@my-deja.com]

On Saturday, August 27, 2016 at 1:38:28 PM UTC-7, rmp wrote:

To the folks who are still running the "Giant Nixie Clock". From the early 1970's:

... as I recall it treated the tubes as 7-segment devices...Am I correct?

 Yes.  It was based on the MM5314 from National Semiconductor.  

[A.J. Franzman]

You got that backwards. Unused digits/segments in a nixie become "poisoned" by stuff given off from the active segments. Trying to light those middle vertical and diagonal segments in a tube from a well-used "Giant Nixie Clock" from the early 1970's now, might reveal that those segments don't light at all. They may also light dimly, partially, or unevenly. They might be shorted to the back substrate and possibly even each other via the substrate. Or they might work just fine; those are large tubes and any given point on one segment is on average much farther away from the nearest point on each other segment than typical cathodes of small conventional nixies. But I can't think of any reason why they would ever be brighter than the segments which have been in use.

On Saturday, August 27, 2016 at 1:38:28 PM UTC-7, rmp wrote:

[jf...@my-deja.com]

On Monday, August 29, 2016 at 9:31:39 PM UTC-7, A.J. Franzman wrote:

... They might be shorted to the back substrate and possibly even each other via the substrate...

 I think this is improbable.  On the one tube that I dissected*, the back substrate is an insulating white ceramic with an insulating  black ceramic overcoat.

BTW, this is one of the reasons I think there are no NOS tubes.  It would be a lot of unnecessary additional work to deliberately put the shadow pattern into the black ceramic overcoat.

* My one catastrophic failure because the glass envelope could not withstand being struck by the floor.

[Jeff Walton]

The only NOS tubes that might have been around would be branded as Burroughs and not Ultronics.  I actually had a couple Burroughs tubes NIB that I purchased back in 1970 or 1971 before I found the Ultronics tubes in the back of magazines for sale by the surplus houses.  I have kicked myself for not keeping the new Burroughs tubes and the boxes as a pristine reference for 45 years later, but who would have thought that we would be talking about this so long after the fact.  Even in 1971, the Ultronics tubes were all pulls from systems that had been in operation and bought up by liquidators.

 

The shadow on the substrate may even exist on a NOS tube depending on the burn-in process that was used and the doping mix of materials in the gas that were used for stabilizing and sealing preceded the burn-in.

 

 

From: 'jf...@my-deja.com' via neonixie-l [mailto:neoni...@googlegroups.com]
Sent: Tuesday, August 30, 2016 3:28 PM
To: neonixie-l
Subject: Re: [neonixie-l] Re: b7971 segment current

 

On Monday, August 29, 2016 at 9:31:39 PM UTC-7, A.J. Franzman wrote:

--

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/38ea757e-b836-486c-bcf8-90b6f9c1f0e3%40googlegroups.com.

[jf...@my-deja.com]

I would believe it was NOS if it was accompanied by a pristine box, the lettering on the tube was all sharp and crisp, and the pins were all straight (with no tool marks from straightening) and unscratched (or maybe one scratch from burn-in).

[A.J. Franzman]

If you sputter enough metal onto any insulator, you can make it conduct.

[jf...@my-deja.com]

I will believe it when I see it.

[A.J. Franzman]

Got any plastic model kits with "chrome" parts? Those are made by vacuum metal deposition, essentially the same process by which the glass inside those 7032 tubes we've been commenting on has become silvered. Put an ohmmeter onto one of those model parts and see what kind of a reading you get.