Author Topic: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595  (Read 3825 times)

0 Members and 1 Guest are viewing this topic.

Offline 2shyTopic starter

  • Newbie
  • Posts: 8
  • Country: pl
IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« on: February 02, 2020, 11:46:29 am »
Hi, inspired by Dave's series of videos I designed schematic and PCB of a simple Nixie clock and I hope that someone will be able to explain what's happening.
The problem is that when I'm driving the cathodes with TPIC6B595 shift register, some of the digits are not readable/fuzzy. It's worth noticing that on different tubes different digits (in the same driving circuit)  get fuzzy, eg. cathode 0 and 7, the rest is okay.

I'm attaching a photo of a blurry/fuzzy digit and a well displayed digit, as well as part of the schematic. HV is 180 V DC, series resistors are 2x10k, I'm switching on the anode voltage using a pair of MPSA42 and 92. I tried to lower down the anode voltage to 165 V, but it seems that the problem remained.

What I've done:
I checked individual tubes with a resistor and HV power supply - they're all ok.
I checked the TPIC6B595 outputs with a LED and series resistor - they're all ok, I'm driving one cathode at a time, one of them is always on, I'm filling the registers first, then turning on the anode supply, never the other way round.

The displayed sequence is 0-3-4-5-6-7-8-9-0-..., because cathodes 1 and 2 are connected to the next register, which is always filled with 0's. This is for test purpose only, as I have only two shift registers soldered to the PCB for now (the nixies can be just put into PCB for now, they form a tight fit)

I also recorded a video:
https://youtu.be/ibOdviAl_b4
Digit 7 is not displayed correctly. If I check this tube directly with a resistor and HV power supply, digit 7 lights up correctly. Replacing tubes result in a different digit being blurry.

Here is the test code (STM32):

Code: [Select]
void test_load_and_latch_two_registers(uint8_t output_number)
{
uint8_t value_to_load, j;
if(output_number > 7) return;
value_to_load = (1 << output_number);

// second register fill with 0
for(j = 0; j < 8; j++)
{
    set_shift_register_clock(GPIO_PIN_RESET);
    HAL_Delay(1);
    set_data(GPIO_PIN_RESET);
    HAL_Delay(1);
    set_shift_register_clock(GPIO_PIN_SET);
    HAL_Delay(1);
}

set_shift_register_clock(GPIO_PIN_RESET);

// first register fill with data
for(j = 0; j < 8; j++)
{
    set_shift_register_clock(GPIO_PIN_RESET);
    HAL_Delay(1);
    if(value_to_load & (1 << j)) set_data(GPIO_PIN_SET);
    else set_data(GPIO_PIN_RESET);
    HAL_Delay(1);
    set_shift_register_clock(GPIO_PIN_SET);
    HAL_Delay(1);
}

set_shift_register_clock(GPIO_PIN_RESET);
latch_shift_register();
HAL_Delay(1000);
}

I suppose there's a problem with floating cathodes voltage, but have no idea what is happening in this circuit.
I measured voltages when a single digit is displayed in two cases.

Digit 5 displayed correctly:
A - 135 V
C0 - 123 V
C1 - 66 V
C2 - 124 V
C3 - 54 V
C4 - 102 V
C5 - 0 V
C6 - 54 V
C7 - 103 V
C8 - 54 V
C9 - 54 V

Shouldn't all off cathodes be around 50-55 V?

Digit 0 displayed incorrectly:
A - 169 V (way too high?)
C0 - 0 V
C1 - 58 V
C2 - 82 V
C3 - 54 V
C4 - 54 V
C5 - 54 V
C6 - 53 V
C7 - 53 V
C8 - 54 V
C9 - 54 V

As far as I understand, the off cathodes voltages in ideal case should be equal around anode voltage, but the internal clamping of TPIC6B595 should drop this floating voltage to around 50-55 V. But it seems that it doesn't work like that.

I would be grateful for any hints on this, guys.
 

Offline MasterT

  • Frequent Contributor
  • **
  • !
  • Posts: 851
  • Country: ca
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #1 on: February 02, 2020, 01:52:21 pm »
Leakage collector current of T9, probably? I'd try to lower R24 down to 2-3k
 
The following users thanked this post: 2shy

Offline 2shyTopic starter

  • Newbie
  • Posts: 8
  • Country: pl
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #2 on: February 03, 2020, 06:25:55 am »
I lowered R24 to 3k3, it didn't help, actually, two more digits become fuzzy.
I lowered also 220k resistors to 2x100k and 2x47k, it didn't help neither.
Then I increased 220k resistors to 470k, still no change.

In the evening I'll remove T3 and connect the supply directly to 180 V, to eliminate the anode driving circuit.
« Last Edit: February 03, 2020, 04:37:49 pm by 2shy »
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #3 on: February 03, 2020, 12:37:35 pm »
"Shouldn't all off cathodes be around 50-55 V?"

Shouldn't they be higher. :)

Are the TPIC6B595 outputs only 50V.
https://www.ti.com/lit/ds/symlink/tpic6b595.pdf
.  That took much longer than I thought it would.
 
The following users thanked this post: 2shy

Offline Prehistoricman

  • Regular Contributor
  • *
  • Posts: 216
  • Country: gb
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #4 on: February 03, 2020, 12:53:33 pm »
Unless I'm completely misunderstanding this circuit, I don't see why the issue would have anything to do with T9, T3, or any of the plate driving circuit. You could eliminate the possibility by checking the plate voltage for both the faulty and correct states.

I think your fault is much more likely to be in the shift registers. Try keeping it on the fuzzy digit and poke around at the circuit to see if you can fix it. Does shorting an 'off' digit to ground change anything? See how much current each digit is consuming (if you can be bothered :)).
 
The following users thanked this post: 2shy

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #5 on: February 03, 2020, 01:16:34 pm »
The TPIC6B595 isn't suitable for nixies. You need a driver which is rated for the supply voltage of the nixie tubes. You'll probably end up needing transistors rated for >180V. BTW you can multiplex nixie tubes so you don't need that many drivers.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 
The following users thanked this post: 2shy

Offline DBecker

  • Frequent Contributor
  • **
  • Posts: 326
  • Country: us
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #6 on: February 03, 2020, 01:42:42 pm »
The TPIC6B595 isn't suitable for nixies. You need a driver which is rated for the supply voltage of the nixie tubes. You'll probably end up needing transistors rated for >180V. BTW you can multiplex nixie tubes so you don't need that many drivers.

From the datasheet, the TPIC6B595 has clamps on the output drivers.
The part is rated for 50V, but the clamping voltage isn't specified.  It's likely much higher but still under 180V, and all of the outputs are turning on slightly to draw enough current to drag the voltage down to the clamping voltage.
 
The following users thanked this post: 2shy

Offline NivagSwerdna

  • Super Contributor
  • ***
  • Posts: 2507
  • Country: gb
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #7 on: February 03, 2020, 01:48:17 pm »
It looks a bit like something I have seen before... not sure of the details... but R24 10k, R52 47k, R64 remove?
 
The following users thanked this post: 2shy

Offline 2shyTopic starter

  • Newbie
  • Posts: 8
  • Country: pl
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #8 on: February 03, 2020, 04:36:48 pm »
A little bit of update:
1. Removed T3 and started the app with tube put into board, then applied 180 V directly to T3 collector pad - the result is the same, some digits are okay, some fuzzy. So the anode driving circuit seems to be fine. I guess the problem is in TPIC6B595, as there are not many options left.
2. To be sure, I checked the test tube again using 180 V and 2x10k series resistors - all digits light up just fine (one cathode tied up to ground, the others are floating)
3. Replaced TPIC6B595 with a new part directly from tape - still the same problem.

I'm going to fiddle around the circuit to see if I can make the fuzzy digit light correctly. The measured cathode and anode voltages are specified in the first post of this thread.

I'm also starting to wonder if I have genuine Texas parts. I got them from a local electronics store, but the epoxy case and the labels look quite weird. Please see attached photo.

Edit:
I lowered down voltages back to 160 V. Now the fuzzy digits turned off.

https://youtu.be/ntrPKsrAp8I
« Last Edit: February 03, 2020, 04:59:44 pm by 2shy »
 

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #9 on: February 04, 2020, 03:56:59 pm »
"Replaced TPIC6B595 with a new part directly from tape - still the same problem."

The problem is they're 50V ICs used on 180V.

You might be able get away with some of it in software by turning the anodes off, changing the shift registers, then turning the anodes back on, hopefully the cathode that is 50V ahead of the others would be the only one to start glowing.
.  That took much longer than I thought it would.
 
The following users thanked this post: nctnico, 2shy

Offline NivagSwerdna

  • Super Contributor
  • ***
  • Posts: 2507
  • Country: gb
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #10 on: February 04, 2020, 04:03:44 pm »
If you decide to re-think.... re-think in a HV5812 direction.
 
The following users thanked this post: 2shy, Ian.M

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3382
  • Country: gb
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #11 on: February 04, 2020, 06:07:18 pm »
From the datasheet, the TPIC6B595 has clamps on the output drivers.
The part is rated for 50V, but the clamping voltage isn't specified.  It's likely much higher but still under 180V, and all of the outputs are turning on slightly to draw enough current to drag the voltage down to the clamping voltage.

The headline in the datasheet says output clamp voltage is 50v, the absolute maximum Vds is specified as 50v and Vds breakdown is specified as 50v minimum @ 1mA.  To me this strongly suggests a clamp voltage that is somewhat close to 50v.
 
The following users thanked this post: 2shy

Offline DBecker

  • Frequent Contributor
  • **
  • Posts: 326
  • Country: us
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #12 on: February 05, 2020, 02:48:02 am »
From the datasheet, the TPIC6B595 has clamps on the output drivers.
The part is rated for 50V, but the clamping voltage isn't specified.  It's likely much higher but still under 180V, and all of the outputs are turning on slightly to draw enough current to drag the voltage down to the clamping voltage.

The headline in the datasheet says output clamp voltage is 50v, the absolute maximum Vds is specified as 50v and Vds breakdown is specified as 50v minimum @ 1mA.  To me this strongly suggests a clamp voltage that is somewhat close to 50v.

I think that we agree that the maximum voltage is listed as 50V, and that any voltage over 50V exceeds the part's rating.  I the statement that the clamp voltage is 50V, but I expect that actually means that the clamp voltage is above 50V but below the breakdown voltage of the output transistor.

Anyway, 50V.
 
The following users thanked this post: 2shy

Offline 2shyTopic starter

  • Newbie
  • Posts: 8
  • Country: pl
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #13 on: February 05, 2020, 07:03:53 am »


As I understand this output stage schematic of TPIC6B595, the drain voltages above 50 V should cause the internal clamping Zener diode to conduct, turning on the (previously turned off) output FET just a little bit to drop the drain voltage to around 50 V. In this case the drain-source channel should not break, right?
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 3382
  • Country: gb
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #14 on: February 05, 2020, 01:40:05 pm »
As I understand this output stage schematic of TPIC6B595, the drain voltages above 50 V should cause the internal clamping Zener diode to conduct, turning on the (previously turned off) output FET just a little bit to drop the drain voltage to around 50 V. In this case the drain-source channel should not break, right?

The zener clamps to protect the FET, so the FET should not be damaged provided excessive current isn't being dumped into the device.
 
The following users thanked this post: 2shy

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2850
  • Country: se
  • Country: Broken Britain
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #15 on: February 05, 2020, 03:55:47 pm »
So what's the final fuzzy digit solution(s) ?
.  That took much longer than I thought it would.
 
The following users thanked this post: 2shy

Offline schmitt trigger

  • Super Contributor
  • ***
  • Posts: 2431
  • Country: mx
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #16 on: February 05, 2020, 04:10:50 pm »
How fast are you multiplexing the tubes?

There was a Nixie app note, from Burroughs if I recall correctly, which mentioned the fastest multiplexing was in the whereabouts of 140 Hz.

Not saying this is your case, but as an experiment could you slow the multiplexing to a very slow rate, such that you can see the individual Nixie being lit up?

EDIT: if you have an oscilloscope, what does the anode voltage waveform looks like?
« Last Edit: February 05, 2020, 04:13:27 pm by schmitt trigger »
 
The following users thanked this post: 2shy

Offline 2shyTopic starter

  • Newbie
  • Posts: 8
  • Country: pl
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #17 on: February 05, 2020, 04:25:45 pm »
How fast are you multiplexing the tubes?

There was a Nixie app note, from Burroughs if I recall correctly, which mentioned the fastest multiplexing was in the whereabouts of 140 Hz.

Not saying this is your case, but as an experiment could you slow the multiplexing to a very slow rate, such that you can see the individual Nixie being lit up?

I'm not multiplexing, they're supposed to be driven statically.
Actually, the first nixie clock I made was multiplexed, I used a single 74141 to multiplex 6 tubes @ 80 Hz, it has a cathode depoisoning sequence when HH=MM=SS. I also implemented dimming the tubes at low ambient light. It's working flawlessly at my home three years now. You can see it working here (tube flickering on the video is not seen in real life):

and see the simplified build process:
https://goo.gl/photos/6rs2Qc8VyK6q7ynN8
Here's another version of the same clock, but with different tubes LC-516 (Polish ripoffs from Russian IN-1, I did a double footprint) - I personally like this one much better.
https://imgur.com/a/NXRS8

I made another version with IN-1 tubes and wanted them to be statically driven, so I chose to use TPIC6B595. And now it seems it wasn't the best solution :/
I'll try to look at the anode voltage.

So what's the final fuzzy digit solution(s) ?

Just as your nick... StillTrying to figure it out.
I didn't have much time to tinker with the circuit in the past two days.
I'll update the thread as soon as fuzzy digits start to light properly.
« Last Edit: February 05, 2020, 04:30:29 pm by 2shy »
 

Offline 2shyTopic starter

  • Newbie
  • Posts: 8
  • Country: pl
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #18 on: February 06, 2020, 04:25:53 pm »
Okay, I finally had some time to look at the PCB and fortunately was able to solve the problem.
It's really embarassing, but I have to confess - it turned out that some nixie pins didn't make any contact with PCB pads  :palm:
I noticed it while measuring cathode voltages - the pins and the pads had different voltages. Then I turned everything off and tested the 'press fit' connection using continuity tester - two or three cathodes were unconnected. I soldered the pins to PCB and now everything works as expected.

I'm glad that the problem's solved, but ashamed of the reason.
Thank you all for your comments, I learned the lesson. Next time I'll probably replace TPIC6B595 with something more suitable, as using this IC for driving nixie tubes is not the best engineering practice.
 

Offline schmitt trigger

  • Super Contributor
  • ***
  • Posts: 2431
  • Country: mx
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #19 on: February 06, 2020, 09:23:47 pm »
No need to be embarrassed. Everyone , and I mean everyone has made a foolish mistake...at least once.

For me, it was assuming that the band on a tantalum cap was also the negative terminal, like most aluminum caps are.................
The results were spectacular.  ^-^
 
The following users thanked this post: 2shy

Offline intabits

  • Frequent Contributor
  • **
  • Posts: 334
  • Country: au
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #20 on: February 07, 2020, 03:00:04 am »
... it turned out that some nixie pins didn't make any contact with PCB pads 
...Next time I'll probably replace TPIC6B595 with something more suitable, as using this IC for driving nixie tubes is not the best engineering practice.

I only soldered one or 2 pins per tube on mine, but they are a tight fit into plated through holes, and seem to make good contact.

I'm using TPIC6B595 chips with IN-1A tubes and it works pretty well:-
https://youtu.be/rQEzPThgjuU

The tubes count at ever increasing speeds, with the neon bulbs between them also showing the count in binary, after some time doing that, it switches to clock mode showing MM DD;HH:MM:SS.

In case you find it useful, my arduino code for the demo in the video above.
https://pastebin.com/RjTw21z9

 
The following users thanked this post: 2shy

Offline 2shyTopic starter

  • Newbie
  • Posts: 8
  • Country: pl
Re: IN-1 nixie tube - fuzzy digits when driving with TPIC6B595
« Reply #21 on: February 07, 2020, 06:25:19 am »
Quote
I only soldered one or 2 pins per tube on mine, but they are a tight fit into plated through holes, and seem to make good contact.

I'm using TPIC6B595 chips with IN-1A tubes and it works pretty well:-
https://youtu.be/rQEzPThgjuU

The tubes count at ever increasing speeds, with the neon bulbs between them also showing the count in binary, after some time doing that, it switches to clock mode showing MM DD;HH:MM:SS.

In case you find it useful, my arduino code for the demo in the video above.
https://pastebin.com/RjTw21z9

I like the modularity of your design, I think I've seen it before on YouTube.
Thanks for the code, I'll look at it, though in this particular case I write for STM32.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf