8453/Z550M erratic

[peter bunge]

Using the schematic from the data sheet

This works with the count moving around but other numbers flash erratically, especially close to the number that is supposed to be lit.

When I moved  the bottom of C1 to the other side of R1 it works perfectly.  It is rock steady and does not care about line voltage or the 5v supply (shown above as 8v but used at 5v)

If this is an error it is continued through all the documentation and is consistent. Changing the value of C1 up and down by 10 had little effect but a smaller C1 helps a bit. My Rst are all directly on the socket pins and the wires are all about 5 inches long.

I have varied the line voltage with little effect. 

Any suggestions???

[chuckrr]

That schematic is so wrong in so many ways.   You need high voltage transistors operating the tube cathodes.

You need high resistance drving the transistor bases.   You need a buffer such as 4049 or 4050 driving the

resistor, which in turn drives the transistor base.  Only then would I dare to use the 4028....to operate the buffer, which in turn operates the

transistor base via appropriate high resistance.  That is the only sure fire way I know of to attain noise-free performance from CMOS logic

driving cold cathode tubes.   That schematic there is a noisy deal.

--
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 view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/b996ef0b-edbb-4c77-8ec1-57255dd58a82n%40googlegroups.com.

[peter bunge]

I think you are wrong. The literature explains this tube is designed to be driven by 5v logic and it does work. The steering electrodes are close to the Anode voltage which is grounded and only 5v pulses (square waves) are needed. Note all the cathodes are connected together internally.

I got to thinking about the C1 position. The data sheet shows it per my schematic and I can't believe they made that mistake several times including the hand drawn notes of the designer. I think maybe I needed to have C1 directly on the socket pins and will try that tomorrow.

Thanks for your interest though but you are thinking Nixie Tubes, this is a special tube with steering electrodes..

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/202102101842701.SM05295%40ADT-IM1.

[gregebert]

I suspect the differential voltage between lit vs non-lit numerals is too low. The CMOS device is basically driving grids to determine which cathode will be illuminated. I've seen similar behavior with an A-101 dekatron. What voltage are you using for VDD ? Is it 8V as indicated on the schematic, or a more-conventional 5V ?

[peter bunge]

I tried varying it with absolutely no effect. It runs perfectly at 5v with the capacitor moved to the other side of the resistor. However I was clocking slowly. It may not run at spec speed. I seem to remember reading of using 5 to 8 volts for the logic.

The data sheet shows the following which I'm inclined to believe. I will try Ck directly on the pins tomorrow. Someone must have tried this circuit.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/9543e23e-ebb0-46e7-8624-ddcf4a843667n%40googlegroups.com.

[gregebert]

If you have an isolation transformer, can you put a scope on the cathode line to verify the ripple is not excessive ? Rk and Ck create an RC filter around 77Hz, and the line is 50-60Hz, so there could be some ripple.

[Dekatron42]

Have you tried the circuit in the J.B Dance book below? Here they use a center tapped transformer and also a much lower value capacitor, only 33nF versus 250nF in your circuit - I've only evere seen 33nF used in real instruments using these indicators.

Quite a few of the Z550M/ZM1050 are broken internally, I have a box of them, and that seems to be due to the welding of the internal parts coming loose when shaken or hit hard (the same problem exists with the B9012/NL9012 tubes - I have a few broken ones of those too with internal pieces that have come loose) - sometimes you can hear these loose pieces if you shake the tube very carefully close to your ear. What usually happens is that a rather large round center piece comes loose and in the worst case shortens some of the electrodes but usually only makes it hard or impossible to get all digits to light up poperly and some of them trigger easier than others due to the distances between electrodes are differing, shaking the tube a little moves that effect around so other digits will start to work and vice versa.

/Martin

[peter bunge]

There is ripple but the steering elements are clean and the display should be off when these are grounded. I tried biasing the anode positive by 5v but it did not help.

Moving the Rk and Ck directly on the tube socket did not help.

Note that the different capacitor values are used depending on half wave (0.25uF) or full wave rectification (0.033uF).

I have not tried full wave rectification and will do so this afternoon. 

See my answer to gregebert for more details.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/9281e612-84bd-45bf-8cad-a239bed5c9b1n%40googlegroups.com.

[peter bunge]

I meant dekatron in my last reply to gregebert.

I have 3 of these tubes, all apparently new (NOS).

Per your suggestion I tried the other two. One is almost perfect but flashes the 2 a bit when 8 is selected. The third is perfect unless I hold the tube in my fingers by the glass where I get lots of random flashing.

Remember that the really bad one works perfectly with the capacitor moved as I mentioned. I don't know if touching the glass affects it.

  More later, I will be away this morning.  

Peter.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/7085da7d-27dd-46f3-9bba-9e2d369dcc4fn%40googlegroups.com.

[Dekatron42]

Touching the glass means that you get a capacitive coupling which will upset them somewhat, there was a special socket made for this by Philips which has a partial metal screen surrounding the tube. The socket is nicknamed "der Kuss" , "The kiss", due to its form. I couldn't find a photo on the Internet now but I know it exists as I have some in my storage and in an instrument that uses them. That instrument is called PW4261 Timer, some photos of the externals can be found on the Internet. On this socket all of the resistors are mounted flush to the pins to minimize the distance, but the capacitors and power supply is mounted some 40cm from the tubes themselves, likewise there are long wires to the drivers.

You can also have a look at the manual for the PW4231 which I scanned that can be downloaded from here: https://frank.pocnet.net/other/sos/Philips_PW4232.pdf if that can help you with the voltages for the drivers.

/Martin

[peter bunge]

  I take it that you never got one of these working?

Peter

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/ad3deb14-52ce-4b11-b048-dd6dea83bdban%40googlegroups.com.

[peter bunge]

I have solved the problem but I don't understand why.

Having tried all the suggestions except this, I went to a full wave rectifier. It is not quite the same as the data sheet but does the same thing.

It also has the capacitor connected per the data sheet, the same as my half wave circuit that did not work.

This is rock solid with no indication of any other flickering. I am using the first (worst) tube. Touching the glass has no effect.

Apparently the tube does not like half wave. Perhaps it was my 1N4007 diode? They can switch too fast and maybe it needed a snubber.

I also tried DC from a variable regulated power supply (HP 6448B up to 600 volt@1.5 amp, and no, I did not go that far). That locked the display on one digit. The data sheet says it needs rectified line, not DC.

I am running my tests from an HP supply for the 5v and an isolation transformer driven by a Variac for the AC so I have full control of all voltages. 

This is my final schematic. It is driven by a PIC with parallel connectors for each display. I used a PIC to generate the BCD because CMOS would not drive the three 74141 for the Nixies. The PIC counts up and down at varying speeds which would have been more difficult with CMOS.

On Thu, Feb 11, 2021 at 10:33 AM Dekatron42 <martin....@gmail.com> wrote:

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/ad3deb14-52ce-4b11-b048-dd6dea83bdban%40googlegroups.com.

[Dekatron42]

Nice that it works!

My understanding is that they need rectified AC, with just a little smoothing so as not to make it a pure DC voltage, as they extinguish during the portion of the voltage when it is to low to maintain the glow, and they are re-ignited when the voltage rise again as long as a trigger electrode has the correct voltage to initiate a glow, much like a thyratron.

Maybe you can find the reason for it working now compared to earlier failures if you check the voltage across the Anode-Cathode-Ignition electrode with an oscilloscope?

I've also had some problems with CMOS ic's when I have had poor, or no, connection to VCC/GND on the power pins - the circuit worked quite well until I touched some pins.

/Martin

[peter bunge]

Hi Martin, I changed the circuit back to halfwave using a single diode from the bridge rectifier and changing the capacitor. This was absolute minimum change without moving anything else. Back came the erratic flashing. So it's not the diode. It just does not like half wave.

My circuits are direct copies from the data sheet.

Peter

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/dcd99e19-3b9c-406c-8057-d69f1547ce03n%40googlegroups.com.

[Bill van Dijk]

Just a guess here…….

 

Where are you located, i.e. what is your net frequency? A 50Hz net frequency with a single diode will produce a 50Hz ripple. A bride on 50Hz will produce a 100Hz ripple. In areas where there is a 60Hz net, the numbers will be 60, and 120Hz. Could it be that the specifications are designed for a 60Hz net frequency?

 

Is it possible the frequency is too low to stabilize the tube?

 

Bill

This works with the count moving around but other numbers flash erratically, especially close to the number that is supposed to be lit.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAFxqQo%2ByAodg_gx-%2BWjbSKBed8Oax4%3DXLEwWaMovY17kcR3CMA%40mail.gmail.com.

[Bill van Dijk]

Further to my last thought…..

 

Are you driving the control grid with a DC or pulsed signal? If not DC, could there be a strobe effect between your high voltage and the control signal?

 

Bill

 

From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of peter bunge
Sent: Friday, February 12, 2021 8:14 AM
To: neoni...@googlegroups.com
Subject: Re: [neonixie-l] 8453/Z550M erratic

 

Hi Martin, I changed the circuit back to halfwave using a single diode from the bridge rectifier and changing the capacitor. This was absolute minimum change without moving anything else. Back came the erratic flashing. So it's not the diode. It just does not like half wave.

This works with the count moving around but other numbers flash erratically, especially close to the number that is supposed to be lit.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAFxqQo%2ByAodg_gx-%2BWjbSKBed8Oax4%3DXLEwWaMovY17kcR3CMA%40mail.gmail.com.

[Dekatron42]

Which datasheet do you use? I only have information with the full wave rectifier.

What happens if you use a 470nF or 1uF capacitor with the half wave rectifier? I made some quick LTSpice simulations which shows that a higher capacitor might work as the voltage seems to fall low enough between each period, but only real tests can show if that works properly.

/Martin

[peter bunge]

Martin; See attached for data sheet. I have tried different capacitors. They make no difference for either full wave or half wave - full wave always works with no indication of problems and half wave flickers all over the place and never works (unless I move the capacitor connection).

Bill van Dijk: I am in Canada, 60 Hz. Control grid?? See data sheet attached.

I was hoping someone else would have tried these circuits. I have used a 'scope to check the various signals and they look fine.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/93f9a715-47dd-4f8b-8fd4-89e03e0bb06bn%40googlegroups.com.

[peter bunge]

Looking at the data sheet and the last graph: Maybe the problem is my resistor is too low and should be 12k or higher.

I will try it.  used 8.2k because that is what they suggested..

On Fri, Feb 12, 2021 at 9:06 AM Dekatron42 <martin....@gmail.com> wrote:

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/93f9a715-47dd-4f8b-8fd4-89e03e0bb06bn%40googlegroups.com.

[Dekatron42]

Thanks, I've never looked at that particular datasheet previously, only the Philips ones even though I have a few Amperex stamped Z550M's.

In this document: https://www.dos4ever.com/Z550M/Z550M2.pdf (a somewhat clearer PDF can be downloaded here: http://www.tube-tester.com/sites/nixie/dat_arch/Z550M_2.pdf) there is a paragraph on "Supply voltage" on page 120 where they mention a minimum repetition frequency of 80 pulses per second as a minmum for proper operation - perhaps that is the reason why it doesn't work as well on just half wave rectification. Maybe Amperex had some specially made to be able to drive them at 117 VAC @ 60Hz?

/Martin

[peter bunge]

I tried Rk at 12k and 22k and it is still erratic.

Next I went back to full wave and connected it to my display chassis that runs from a PIC that counts at various speeds. I see the circuit only counts slowly but at 10/sec it fails. No flickering, it just does not count.

I changed the capacitor to the other side of Rk and it works perfectly.

I can't believe it is a typo carried through three documents but companies have been known to intentionally place errors to cause grief to anyone trying to steal their designs.

This circuit works.

[Alex]

Just a thought, maybe the diodes originally used had a much different reverse leakage which, in half wave rectified setup, caused the polarity in the tube to flip slightly - at low leakage current levels - and thus extinguish it more cleanly between ripples. Modern(!) silicon (verses germanium) may be too sharp for it? I think your full wave design is probably best anyhow, the half wave was likely a cost compremise!

- Alex

[Dekatron42]

What happens if you remove the capacitor alltogether?

Ronald Dekker who wrote about the manufacturing of these showed that they can be driven with a CMOS 4017 in an article in Elektro Magazine, it is listed in the other thread by you: https://groups.google.com/g/neonixie-l/c/gNwRMF9iq1s/m/eUH7Qx-jBQAJ - there he doesn't even use a capacitor after the rectifying diode (I didn't think about that previously). Maybe his circuit works somewhat better if at least one output is active high whereas the 4028 can have all outputs low if it is driven with binary 10-15 on its inputs.

I only know that the 33nF capacitors are there in my PW4261 counter but I haven't actually checked how they are connected. I'll see if I can do that it the future as the PW4261 is in storage.

/Martin

[peter bunge]

Alex; you are missing the point. Neither of the circuits shown in various data sheets work properly!

The circuit that does work and the one I'm using has the capacitor connected to the other side of the resistor.

A quick question so it does not get lost below: Can a Dekatron be driven like a Nixie from a 74141 instead of using the steering electrodes?

Each tube has its own small circuit board that connects to a PIC on my power supply by a 6 wire (+5v, Gnd, 4 BCD) harness. and by a separate two wire harness if High Voltage is needed (the 2 pin 0.2" spacing header on the perfboard above). My wires are fed through a hole and loop down to be soldered to provide strain relief. All boards mount on two spacers. Some mount directly on the back of the tube if space allows. Any board can be removed in less than a minute, usually 2 screws but 4 screws where the board is not attached to the tube, then pull one or two connectors. 

The E1T tube does not count in sync with the others because it is clocked, not fed by the common BCD bus. It takes time for the filaments to warm up so it needs a reset every decade to get it in sync. However it cannot count backwards without a lot of effort I'm not about to expend so it stays like this. The hole below is for LED displays.

You received this message because you are subscribed to a topic in the Google Groups "neonixie-l" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neonixie-l/uVu3BIr_9Uc/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neonixie-l+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/06111dd3-86b7-454a-b34d-5b0488b3ec09n%40googlegroups.com.

[Jon]

>Can a Dekatron be driven like a Nixie from a 74141 instead of using the steering electrodes?

I've not actually done the experiment in the way that you say, but the answer should be 'yes'. You'll need a selector dekatron though where all the main cathodes are brought out to their own individual pins. So GS10C / 6476 / A101 for example rather than GC10B / 6802 / OG4.

There is actually a dekatron-lookalike, the GR10A, which is designed to be used like this - it's essentially a GS10C without the guide electrodes (different base, but that's a detail).

Jon.

[gregebert]

A 74141 might work properly, because it's a decoder that only allows 1 output to be driven low at any given time. To properly drive a dekatron, such as an A-101, you should drive it with an 'overlapping' drive pattern, which has 6 phases:

1- Drive cathode 'N'

2-Drive cathode 'N' and the first steering cathode ring

3-Drive first steering ring

4-Drive both steering rings

5-Drive second steering ring

6-Drive second steering ring and cathode 'N+1'

(Repeat)

I did experiment with a non-overlapping pattern several years ago, and it was flaky. The 6-phase pattern is rock solid.

[Jon]

Actually, musing on it further, you might have a bit of a problem with the 74141 and dekatrons. I suspect the chip output transistors won't have a high enough Vceo to cope if you try to use the typical dekatron anode voltages. Likely you're either going to need to play around to run the tube at lower anode voltages (not sure that'll work very well) or switch to a mixture of logic and discrete components - something along the lines of a 7442 driving MPSA42 (you'll need an inverter between them).

The GR10A is the best option here I think - that should work fine with 74141.

Jon.

[Dekatron42]

If you can get your hands on a GSA10G you could even let it drive a Nixie at the same time, or you can use one of the A-201 Polyatrons with a Nixie.

You can use a set/reset circuit for each cathode on the GSA10G (or any selector dekatron) as well as on the auxilliary anodes of the A-201, it will however require a few components for each of the cathodes/auxilliary anodes to be able to set/reset them to each position, an extra transistor will let you drive a Nixie too on an ordinary selector dekatron. You can find the circuit for the set/reset capability for the A-201 Polyatron here on the forum as I've published it here a few years ago. You should let the guide electrodes at their bias voltage so that they don't float and disturb the dekatron/polyatron from working properly.

For the selector dekatron you can use the circuit below (taken from an Advance Rate Meter Counter), it will drive a Nixie at the same time as having the capability to set/reset the dekatron to each cathode - you'll of course need 10 of these circiuts for a GS10C dekatron. The circuit was designed for use with a GS10H dekatron together with a GR10K Nixie. Don't leave out either D1 or D2 from the circuit as they are there for protecting from positive/negative pulses that would upset the circuit and blow Q3. If you use a GSA10G you can skip Q3 as the GSA10G will drive the Nixie directly. You will probably have to adjust R4 & R7 depending on what dekatron you use but the rest of the components should be fine with any dekatron, the reason for adjusting these two resistors is to accomodate for the correct cathode current to pass to common ground both during normal counting (R7) and during set/reset (R4). R2 should be a 1W resistor as it will have just above 0.6W power or use two resistors in series for enough power rating. The circuit is designed to make a short enough pulse for set/reset and not for a continous voltage on the cathode of the dekatron, the glow will stay on the set/reset cathode due to current flowing through D2 & R7 to common ground (current also flows through R4 & D1 to common ground but the main part of the current flows through D2 & R7). I've used this circuit in many experiments and it works very well.

/Martin

[peter bunge]

Thanks Jon; I am starting a new thread for this Dekatron project.

I will post tomorrow after some tests.

Peter

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/19dedc5a-4589-4f33-ace2-370a4e448af3n%40googlegroups.com.

[peter bunge]

  Thanks Martin; I am starting a new thread for this Dekatron project.

I will post tomorrow after some tests.

Peter

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6285e13c-a2ff-4f4a-82f3-af3d871bf74en%40googlegroups.com.