The AM Forum

Does anyone remember Nixie Tubes? I'm going to start a new HB tube receiver project sometime later this summer and I want to use Nixie tubes for the frequency readout instead of an analog dial. That way it will still be vintage ;D

Has anyone used these in a while? I need to come up with a way to somehow track from the L.O. or something and generate the correct readout.

Anyone have any ideas?


(http://www.amforever.com/amphone/nixie_tube.gif)

You can build in the equivalent of a Heathkit SB-650 :) 

I set up one of the Heath stations like the 1973 catalog (I think I have the year correct) with the SB-303/401/200 accompanied by the SB-600/610/620/630/650.  The display has a nice glow.

As I recall Heath used presets for the BFO to avoid actually counting all frequencies.  PM if you need more info.

I have seen some old frequency counters that used nixie tubes. They are usually very large and cheap.

All of the older HP series use Nixie tubes.  I have a couple of the 524D series that came as part of a vintage gear deal.  They are bigger and weigh a lot more than most vintage transmitters.  Most of the 5245 series were also Nixie. 

[quote   They are bigger and weigh a lot more than most vintage transmitters.  [/quote]

They also had a habit of burning out segments.

Yep, I have a 50 MHz 5245L which has Nixies. Always liked the look of them too.
It is quite heavy but also very accurate, and was $20 at Dayton  8)
Shows (1)46.940008 from my 2m HT.

(For a modest additional expenditure you can get the frequency converter plugins and measure into the GHz range if you are so inclined).

Segments? Each digit is a separate cathode... or did you mean a failure of that driver?  ???

I've got a frequency counter that predates even nixie tubes - it uses individual lamps behind a shadow mask for each digit.  You'll find some of the very first digital instruments did the same thing using neon lamps, and I suspect the nixie grew out of an attempt to combine all 10 digits in one envelope.

The counter I have is model 602A made by Computer Measurements Company, mid-60s vintage.  All discrete transistors, before ICs were available, very nicely built.

How about digitrons?

I bet you could find some of those in junk calculators.

http://foxtango.org/FR-101%20FL-101/FR101%20Antiflicker%20Mod.htm

Better link with more info

http://foxtango.org/ft101/Foxtango%20YC601%20page.htm

There's this guy for parts too.

http://www.sphere.bc.ca/test/nixies.html

The Russian NIXIE tubes are very cheap on ebay and since they are light shipping is also inexpensive.  I bought a few varieties earlier this year for construction and possible retrofit for some of the U.S. units.  The display tubes used in many of the Heathkit clock and weather products have gotten very expensive making field substitution a consideration.

If I were going to build a new unit and didn't already have the tubes I would definitely take advantage of the Soviet era surplus.

Replace the 7490s in a Nixie counter that has them with 7493 presettable dividers (not a pin for pin compatibility here) and you can automagically adjust the count for the IF.

The 7441 or 74141 are the proper drivers for these tubes.  Hard to find, except the Russians sell big lots of them on eBay.

I like nixies!!!!
i have an external frequency counter and syntetizer for my Drake B line that uses nixies. I have also a very nice receiver from the seventies that uses nixies, the Plessy PR-1553 a very good performer.

While developing the four-tx antenna array for TIROS 1, the World's first weather satellite (1959), the freq counter I used was an HP with nixies. Don't remember the HP number of the unit, but it was large and really heavy.

Walt

Walt,

I bet that was a 524C.  Just after my wife and I were married we traveled to see the Fall color change in NC/TN and picked up a Hammarlund SP-210LX as part of the trip and the Hammarlund came with the 524C counter and a couple of the accessory modules.  I didn't feel like having it in the back of the pickup while we were sightseeing so it stayed in the motel room.  I am sure the maids really wondered what was up.

My original 524C didn't come with a manual and a few years ago I found a manual at a hamfest for $5 IF I also took another 524C with it.  So now I have a cabinet type and a rack mount 524C.  Both work (although the timebase thermometer on one is way off) and the fan is loud enough it sounds like a small jet idling before takeoff.  If I power up one of the 524 counters and my Tektronix 555 simultaneously it would probably bring me a visit from the "green police".

The Drake MSR and DSR receivers used nixies, as did the Sylvania R-1414, some Racals, the HRO-600, one of the R-14xx "Waterboy" receivers, and some CEI/WJ receivers come immediately to mind.

Suggestion - if you're gonna add a nixie display, you'll want to completely enclose it in a metal case.  Too many digital signals that'll radiate harmonics.

Also, there are many old sk00l freq counters around with Nixies.  Scour ebay for one of the small ones and use that as your base.  That way you already have the time base, conditioning, and gating circuits.  Replace the 7490 decade counters in each digit with a 7493 so you can pre-set the IF (ie, for a 455 KHZ IF you would preset the counter to 99,545.0, or 455.0 if you're using low side injection.  As I said theyr're not a pin for pin replacement.

This is way too complex to attempt using tubes.

No, the nixies don't radiate much.  But consider the fact that your clock will be divided down, as well as your input signal.  So you can have all kinds of fun mixing products and spurs received with 1 MHz, 100 KHz, 10 Khz, etc floating around being switched at TTL levels.    Well shielded with a cutout for the nixies seems to work fine.  Don't bypass the nixies, TTL really hates highly capacitive loads.

Also the presettable decade counter is a 74160, NOT a 7493.  Dunno what i was thinking there, and I posted the same error twice.

Nothing at all wrong with TTL.  We went to the moon with TTL.  And you'll have a hard time finding a CMOS nixie driver IC  ;D  It's all there already in TTL with the 7441/74141.  The other ICs in the counter, you can use 74LS/ALS/ACT or whatever so long has they have enough scrote to drive the '141s.

For a 6 digit display * 10 digits per display that's a LOT of FETs, a lot of work, and a lot of PCB real estate just to be parochial about TTL...  Like I said, it'd be a very easy conversion to find an old small nixie counter like the Heath IB-series and repackage it.

Brian, you're way overthinking this.  I've had several Nixie tube receivers and none had problems with noise coming from the nixies.  It's the time base/counting boards you gotta worry about.  The tubes aren't multiplexed so at the very worst the 100 hertz digit will be switching at a 100 hz rate.  The harmonics of that drop off very quickly.  The time base, OTOH takes 10 (or 5, or 1, or whatever) MHz and divides it down.  That's where you can get some radiation.  Just keep the leads short and use normal shielding practices.

You might find FET arrays in ICs but I don't know if any are available that will switch 200V.  Too, I doubt substrated FETs would appreciate a completely unnecessary .1 uF load on 'em either.  Neither Watkins Johnson nor HP found it necessary to bypass nor shield the nixies in their counters, and I've never seen a piece of equipment that had nixies bypassed.

I have a complete (except for the time base) counter board that came from a Watkins-Johnson receiver.  5 digits, it's about 5"X9".  It dispenses with the 7475 latches so the digits are counting 10% of the time.  The nixies are soldered right onto the PC board.  No reason you couldn't breadboard something equivalent in size, as that board isn't at all crowded.

I agree with your comments on shielding the TTL counters/dividers. It sounds like you're implying that the Nixies will only radiate 100 Hz and harmonics thereof.
Consider that the switching time might be (for nice round numbers) 0.1 us... that will be a strong 3.5 MHz signal (with 100 Hz modulation, a nice raw buzz).

Brian, I would be careful adding bypass caps to the nixie leads.  Because the nixie is just a group of neon bulbs, you could end up with a relaxation oscillator and blinking digits. 

HV positive DC is applied to a large electrode in the back of the nixie.  The nixie digit that is illuminated is pulled down to a low DC voltage by the driver IC.  The dark digits are just floating high.  The length of the nixie leads should not be an issue, as no high frequency signals ride on this part of the circuit.  As JN says, shield the count-down chain and other digital logic.  The 74141 drivers should be rather quiet.

Years ago I built a digital clock using nixies for the day of the month, 24 hour time, minutes, seconds, and tenths, clocked from a 10mhz crystal, with 7400s, 7490s, 7493s, etc.  Open construction on perf board, no enclosure, running on the desk near the receiver, absolutely no interference anywhere on HF.  I still have this clock, but have not used it for some time.

Between your counter chips and the 74141 nixie drivers, you may want to use 7475 latches to store each count so that you can enjoy a stable frequency display between counts.  Any circuitry from the four-bit output of the latches and the nixie decoder/driver should not be a source of EMI for your receiver.

You're adding 120 parts where you only need 6.

It depends on how you build the counter.  There are any number of ways.  Usually for 100 hz readout and smooth operation you'll want a .1 sec timebase.  Configure the gating and reset circuits such that the counters count for .1 sec then the data from them is loaded into the 7475 latches.  The latches then drive the 74141 decoder/driver.  While the count is displaying the counters are reset, and the process starts over again.  So the display is being updated 10 times/sec, as is the count.  All displays are updated simultaneously from the latches.

Another method is to do it the way WJ did it in that board I mentioned.  It dispenses with the 7475 latches and does a count/hold.  The counters count for .1 sec (or .01 sec, or whatever your timebase is set to), then the count is held for.09 sec (or .9 sec).  The disadvantage of this is the display isn't as "defined", for alack of a better word, since you'll see it counting for 1/10 the time.

Remember if you insist on using FET drivers instead of the 7441/74141 you *still* need to decode the 4 bit binary to  1 of 10 for the nixies.  I think the 7442 will work for this.  You can probably do the whole thing in a PIC microcontroller, that's how the nixie clock boys do it.  You can also use CMOS TTL equivalents for the counters, time base, and gating/reset circuits - ie 74ACTxxx ICs if you're against using TTL.  However I'm unaware of any LS, or ACT equivalents for the 7441/74141 and 7475s.  And remember your fanout calculations - I'm pretty sure the ACTs will drive a 74141, for a 5V supply voltage, but check to be sure.

Look at the Drake MSR/DSR skizmatic.  My DSR-2 has absolutely NO receiveable spurs or noise from the display, nor does my Sylvania R-1414.  Neither have any bypassing gynmastics on the nixie tube elements.  Remember also that the 74141 grounds each element to turn it on, so switching transients are extremely low, compared to if the IC simply turned on the 180V to each element.

With a .1 second refresh rate, you won't notice a lag.  With a 1 second time base, you certainly will.

You have complete control of the count time duration and interval.  The interval you set between counts determines how quickly the display is updated when you tune your oscillator.  Use a longer count to get higher resolution - ie: low-order digits.  Simply move the decimal point as you change your time base selection. 

I found a rather interesting article in QST for April '74.  On page 34 is part II of an article by Wes Hayward, W7ZOI, entitled "A Competition-Grade CW Receiver".  He installed a HB counter, using 7-segment LEDs, not nixies, inside the receiver.  Most of the receiver sections are shielded, but the counter logic is not.  He designed it using a 2.0 MHz clock so the spurs would not lie in the ham bands of interest.  Then the novel approach he used was to gate signal from the crystal oscillator to the divider chain, such that the chain was inactive when no count is in progress.  The oscillator is the only logic running continuously.  This virtually eliminated the interference problem.

No reason the Heath counter won't work below 3 MHz.   In its intended application it doesn't need to though; there are no oscillators in the Heath SB-series under 3 MHz.

Use at least 74LS ICs in the first two dividers if you need to go to 55 MHz.  Regular 74xx TTL craps out between 15-20 MHz, depending on the chip.

I used to have a HP counter that used Nixies that wasn't big. It was about the size of the Heathkit 650 seen here. I don't remember the model, but it was a 10 MHz top end counter I believe. I gave it to a friend few years ago.

Why do all that?  The driver ICs will drive the nixies directly.  There's absolutely no reason whatsoever to add FETs - the ICs are much smaller than 10 FETs will be.  You can build the whole counter on one board, nixies included.

Why not just run each 4 bit line from the counters to the decoder/driver mounted at the base of each nixie?  That way you're only sending 5V thru those cables instead of 180V.

"Better" is the enemy of "good enough" and it's certainly true here.  I can't fathom what you have against using the 74141s to drive the nixies directly - it's been done in hundreds of designs very successfully.