Nixie driver problem when going over 170 VDC

[nix]

Hello I just build my own Nixie clock with input from Arduinix.com and a power supply from ledsale (Build a nixie power supply - LEDsales)

Attached find the circuit and the pcb from fritzing. The clock works fine but when I adjust the high voltage with R36 the ds1307 module from Twig

http://shop.boxtec.ch/twig-ds1307-p-40355.html starts to flicker because on the serial interface it shows strange numbers like 165:165:85. After a while it does not return the time over the i2c interface at all. A restart of the nixie clock starts showing the time again. With an oscilloscope I can see that after a certain position on R36 the SDA puls disapear and it shows a flat line. I think it is at high level. Not so shure how to read the oscilloscope. 

Any Idea why? I don't see a direct link from the high voltage power supply to the 2 wire interface. (see attached circuit)

Kind regards

Sergio

[nix]

[Spencer W]

Do you have decoupling capacitors on all ic's?

Sent from my iPhone

--
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 an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msg/neonixie-l/-/jgtKRhvfY7IJ.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[nix]

not sure. There is a capacitor on the highvoltage power supply. is this a decoupling capacitor?

[Spencer W]

You want a small capacitor, usually 0.1uF on each power supply of the ic as close as possible to the pins.

So on the ds1307, you want it on pin 8 as close as possible.

This image shows what I mean, 

This goes a little deeper into why its needed, http://hackaday.com/2008/09/29/parts-01uf-decoupling-capacitors/

Sent from my iPhone

To view this discussion on the web, visit https://groups.google.com/d/msg/neonixie-l/-/TaAzQP23KfEJ.

[dr pepper]

What spence says, and keep your wiring short esp to the nixies.
A 1000u cap right at the power input wouldnt be bad either.
If the issue occurs at a certain ht voltage then maybe your reaching
the breakdown voltage of the switching fet/transistor, I've had this
before.
Another thing is the inverter high voltage supply inductor could be
saturating, this will happen suddenly and when it does it'll pull lots
more current and possibly deck the supply to your logic causing the
rtcic to freak out, if the inductor is warm then thats a sign this is
happening.

On 13 May, 15:17, Spencer W <upnxwoo...@yahoo.com> wrote:
> You want a small capacitor, usually 0.1uF on each power supply of the ic as close as possible to the pins.
>
> So on the ds1307, you want it on pin 8 as close as possible.
>
> This image shows what I mean,
>

> This goes a little deeper into why its needed,http://hackaday.com/2008/09/29/parts-01uf-decoupling-capacitors/
>
> Sent from my iPhone

>
> On May 13, 2013, at 6:49 AM, nix <daniels.ser...@gmail.com> wrote:
>
>
>
> > not sure. There is a capacitor on the highvoltage power supply. is this a decoupling capacitor?
>
> > Am Montag, 13. Mai 2013 13:45:46 UTC+2 schrieb Spencer:
>
> >> Do you have decoupling capacitors on all ic's?
>
> >> Sent from my iPhone
>
> >> On May 13, 2013, at 6:19 AM, nix <daniels...@gmail.com> wrote:
>
> >>> Hello I just build my own Nixie clock with input from Arduinix.com and a power supply from ledsale (Build a nixie power supply - LEDsales)
>
> >>> Attached find the circuit and the pcb from fritzing. The clock works fine but when I adjust the high voltage with R36 the ds1307 module from Twig

> >>>http://shop.boxtec.ch/twig-ds1307-p-40355.htmlstarts to flicker because on the serial interface it shows strange numbers like 165:165:85. After a while it does not return the time over the i2c interface at all. A restart of the nixie clock starts showing the time again. With an oscilloscope I can see that after a certain position on R36 the SDA puls disapear and it shows a flat line. I think it is at high level. Not so shure how to read the oscilloscope.

>
> >>> Any Idea why? I don't see a direct link from the high voltage power supply to the 2 wire interface. (see attached circuit)
>
> >>> Kind regards
>
> >>> Sergio
>
> >>> --
> >>> 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 an email to neoni...@googlegroups.com.
> >>> To view this discussion on the web, visithttps://groups.google.com/d/msg/neonixie-l/-/jgtKRhvfY7IJ.

> >>> For more options, visithttps://groups.google.com/groups/opt_out.

> > --
> > 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 an email to neoni...@googlegroups.com.
> > To view this discussion on the web, visithttps://groups.google.com/d/msg/neonixie-l/-/TaAzQP23KfEJ.

> > For more options, visithttps://groups.google.com/groups/opt_out.- Hide quoted text -
>
> - Show quoted text -

[nix]

Hi Spencer,

thank you for your replay. Since the RTC is made with an smd component, I was not able to solder the capacitor directly to the pin of the ds1307. But i solderd it quite near to the power pin.

It seems to be a little bit better but still has the error.

With the oscilloscope, when I look at the SCL, I see a 100kHz square wave. the rise is about 700 ns.

When I increase the voltage on the DC/DC step up circuit, all of a sudden i see instead the SCL square wave just a flat line at 5 Volt level.

The high voltage from the DC/DC step up converter showes at the upper end where the issue exist bigger ripples.

Sergio

[nix]

The wires to the nixies are 10 to 20 cm long. The output of pin 3 on the 555 looks like this in the highest voltage where the problem exists.

[nix]

With the decoupling capacitor on the ds1307, it is possible to go a little bit higher then the 170 Volts which is enough to illuminate all the digits. The error still occurs when going over the 180 Volts, but the clock seems to work :-)

[threeneurons]

Is there some hand wired perfboard circuitry involved ? If so, can you provide a photo of both sides. Back in school (when dinosaurs roamed the Earth), I noticed a lot of my fellow students, used thin wire for everything. Power and ground should be heavier. Those switching supplies generate a bunch of interference. This makes "physically" routing signals, especially power and ground, even more important. Everybody should get a copy of Don Lancaster's TTL Cookbook, if just for the construction guide lines. For actual circuit ideas, you can get his CMOS Cookbook, too. Both are old books, form the 70's, but many of the guidelines still apply.

[Jens Grabner]

I have simulate the scematic.

You need a resistor between NE555 out and Gate Q1. I think 150 Ohm is good. The peek current is verry high, about the input capacity of the gate from the mosfet.

Jens

Am 14.05.2013 02:38, schrieb threeneurons:

Is there some hand wired perfboard circuitry involved ? If so, can you provide a photo of both sides. Back in school (when dinosaurs roamed the Earth), I noticed a lot of my fellow students, used thin wire for everything. Power and ground should be heavier. Those switching supplies generate a bunch of interferene. This makes "physically" routing signals, especially power and ground, even more important. Everybody should get a copy of Don Lancaster's TTL Cookbook, if just for the construction guide lines. For actual circuit ideas, you can get his CMOS Cookbook, too. Both are old books, form the 70's, but many of the guidelines still apply.

On Monday, May 13, 2013 4:19:05 AM UTC-7, nix wrote:

Hello I just build my own Nixie clock with input from Arduinix.com and a power supply from ledsale (Build a nixie power supply - LEDsales)

Attached find the circuit and the pcb from fritzing. The clock works fine but when I adjust the high voltage with R36 the ds1307 module from Twig

http://shop.boxtec.ch/twig-ds1307-p-40355.html starts to flicker because on the serial interface it shows strange numbers like 165:165:85. After a while it does not return the time over the i2c interface at all. A restart of the nixie clock starts showing the time again. With an oscilloscope I can see that after a certain position on R36 the SDA puls disapear and it shows a flat line. I think it is at high level. Not so shure how to read the oscilloscope. 

Any Idea why? I don't see a direct link from the high voltage power supply to the 2 wire interface. (see attached circuit)

Kind regards

Sergio

--
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 an email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/8f9bd2a6-1578-4dfe-b029-bc81159848d4%40googlegroups.com?hl=en-GB.

[nix]

Hello Jens,

thank you for your help. How did you simulate it?

Kind regards

Sergio

[Per Jensen]

Hand wired perfboard ? it's worse than that...

Look at the board layout in the corner regarding the SMPS, it's a luck that it actually works :-O

https://lh4.googleusercontent.com/-PsF8r0jvHEQ/UZDN7eiidtI/AAAAAAAABH4/x_mkJRi4IPU/s1600/nixieclock_v1.1_Leiterplatte.jpg

That layout is _flawed_

// Per.

[Jens Grabner]

The most scematic i simulate with LTspice
http://www.linear.com/designtools/software/#LTspice. It have no cost and is with no limits.

Good for begiining ar this documents from a german teacher.
http://www.gunthard-kraus.de/LTSwitcherCAD/SwitcherCAD-Tutorial_English/pdf-File/
http://www.gunthard-kraus.de/LTSwitcherCAD/index_LTSwitcherCAD.html

Jens

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/f3ed76b0-e2c2-4276-a9b7-7cf1aaa5fd86%40googlegroups.com?hl=en-GB.

[nix]

Hi Per,

I found some errors myself, which I would like to correct in version 2 of the pcb. 

Your comments below unfortunately do not help at all. What exactly is wrong with the layout of the SMPS

Thank you for your input.

Sergio

[Per Jensen]

Hi Sergio. 

The SMPS-part of your Circuit should not be in close contact with the rest of the Circuit. The parts are all over the place on your board. They should be together in an area of less than 1in. square. Remember to keep the gate lead short and fat, and supply tracks short and fat too. The feedback path should be as far away from the inductor as possible. 

// Per. 

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/ba7116fa-9f7e-4e24-945b-f38fa28113a5%40googlegroups.com?hl=en-GB.

[JohnK]

Sergio, I am not sure if you think that Per is having a go at you. He won't be. If you look back at the topics in this group at both Google Groups and at Yahoo Groups you will find a lot of discussion of power supply merits and pitfalls.

Maybe you have not been in the group for very long; those that have will have an automatic reaction to topics that have been discussed to death.

There is a wealth of info in the 'archives'.

 

John K.

----- Original Message -----

From: nix

To: neoni...@googlegroups.com

Sent: Tuesday, May 14, 2013 5:49 PM

Subject: Re: [neonixie-l] Nixie driver problem when going over 170 VDC

Hi Per,

I found some errors myself, which I would like to correct in version 2 of the pcb. 

Your comments below unfortunately do not help at all. What exactly is wrong with the layout of the SMPS

Thank you for your input.

Sergio

...clip...

[Dekatron42]

Here's an example of a NE555 Nixie PSU with layout that shows the fat tracks on the PCB: http://bleuchez.wordpress.com/2011/09/05/a-low-cost-nixie-tube-power-supply/ (it could be made even tighter but it works). At the end of the text there are also references to a design by Nick De Smith : http://www.desmith.net/NMdS/Electronics/NixiePSU.html but that one is an SMD design and it also uses a different switching ic, but it shows the need of the fat tracks.

 

/Martin

[AlexTsekenis]

Hi nix,

I have read this thread with interest; first let me say that I really like your clock's case graphics!

The issue you are having with the RTC module is, to summarise what others said, due to very poor power and signal distribution. This is caused by improper/inadequate circuit board layout. To quote a friend: 'PCB design is not just going from A to B'. It is not just the RTC that is 'suffering', your arduino and those driver chips are too. In your next driver board revision you can study more and improve upon (search terms in parentheses):

1) Related components being close together (PCB component aggregation)
2) Traces to be thick enough to support the current they are carrying and generally as thick as possible (PCB trace width)
3) Clearances (gaps) between traces and pads to be big enough for these high voltages. (PCB trace creepage distance)

4) General trace routing guidelines (an introduction here)

To show off your clock as quickly as possible I suggest you stop trying to fix it. You see, changing/adding one component here and there might get the clock to work eventually, but only marginally. You will find that after a month it fails again and even the arduino might get damaged.

I was going to suggest you buy an arduinix but then I saw their schematic and layout. So I suggest you buy a small HV nixie power supply like the ones linked which will give you a robust HV power supply. Then use your own (temporary) board, removing the HV supply components, connecting the PSU module to your '+170V' point and the grounds from the two boards together. That will give you a robust clock and the time to design a new board (pun intended).

As a side-note, if you are recently starting in electronics and/or CAD design, Fritzing is a not a good choice as it has a very small community.

Alex.

[David Forbes]

Sergio,

I see the problem.

As others have said, your PC board layout may be correct for your net
list, but it is completely wrong for a switching power supply.

The Nixie power supply design by Nick DeSmith is a good one. Copy it
exactly.

When doing a PC board layout, the first thing to do is to move around
all the parts so that the connecting lines are as short as possible.
This takes a lot of time. I will typically spend a day moving the parts
around to get a nice placement. After I do that, then the wiring part of
the job only takes an hour.

Also, be sure to use heavy traces for power. Those bypass capacitors
next to the ICs are essential too.

A PC board layout that looks pretty is more likely to work well than one
with the traces spread nilly-willy all over the board.


> <https://lh4.googleusercontent.com/-PsF8r0jvHEQ/UZDN7eiidtI/AAAAAAAABH4/x_mkJRi4IPU/s1600/nixieclock_v1.1_Leiterplatte.jpg>
>


--
David Forbes, Tucson AZ

[nix]

Hi All,

thanks all for this valuable comments. I see now that I should have studiedt more before creating the PCB. I really thought its just wiring the things together!

I see that the arduinix layout is much cleaner, but does it fullfill all the arguments from Alex?

I will take all this input in my version 2.

Many thanks again

Sergio

[Jens Grabner]

I have an other scematic,

simulatet and it works in hardware.
http://threeneurons.wordpress.com/nixie-power-supply/

Jens

--
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 an email to neoni...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/c8312724-907c-4c6e-a066-f0b25ddac3a1%40googlegroups.com?hl=en-GB.