Any interest in IV-9 numitron clock kit?

[Moses]

Hi Everybody,

I seem to have enough IV-9 tubes to where I can produce a batch of 100 or so 6 digit numitron clocks. Thinking ESP32 powered like my last one (Thank you to all those who gave suggestions, the kit came out great!). I'm thinking maybe try for a full through-hole solder kit with some shift registers.

I know nixies are more popular.. but anyone interested in a numitron clock kit?

Oh, one question I do have, anyone have long term experience (ie, years?) running this particular tube? How do they fare?

Thank you!

Regards,

-Moses

[Richard Scales]

I have built several numitron kits from nixieclock.biz and am a big fan of numitrons, I have only had them running far a couple of years and then for only half a day at most. I have heard that their tube life is 'long' but have not seen any actual numbers. I have also re-worked the design for other numitrons like IV-19 and DTF104B.

- Richard

[Michail Wilson]

I have been making numitron projects.  Namely Bitcoin tickers that run a live stream of the price (tubes constantly changing digits since the prices changes 24 hours a day by the second).   So, well over 2000 tubes out there and I offer free repair/warranty meaning people are apt to contact me if there was a problem.  They run 24 hours a day.

 

I would say that I have replaced about 15 power supplies (USB).   And, over that time, I have had to replace 4 tubes (3 customers) as a segment or two have stopped working.   That said, I have probably replaced a total of about 20 tubes, but those other ~16 were damaged (destroyed) in shipping.

 

Lastly, I run no cross-fading or ramp up/down technique.  All tubes are direct drive of either on or off switching.

 

Hope that helps.

 

Have a link and/or pics to your clock kit?

 

Michail

--
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/f461a854-20da-49f2-bf35-b8d5f50922cen%40googlegroups.com.

[Dylan Distasio]

Hi there.   I'd be potentially interested in a numitron kit!

--

[Mark Moulding]

I've built quite a few devices using the IV-9 tubes, and found them to be extremely durable. (I probably wouldn't be interested in a clock kit, as I have about 3500 tubes in stock...)  If operated below their maximum operating parameters (dimmed a little, and soft-started filaments), they'll run essentially forever.  The DTF-104Bs are even better, and are really fantastic-looking tubes.  They're also extremely rugged mechanically.  I've built Burning Man Mutant Vehicle dashboards using these, and the rather extreme conditions of that application didn't faze them at all.  I've also had clocks running for nearly a decade with exactly zero failure due to the tube burning out.  Impact with flying objects is another matter, but in fact a tube survived for a minute or two even with its outer glass envelope completely shattered.  Again, if operated within their specified environment, their lifespan is essentially unlimited.

~~

Mark Moulding

[Ian]

I’m in

On 17 Nov 2022, at 04:28, Richard Scales <ric...@scalesweb.co.uk> wrote:

I have built several numitron kits from nixieclock.biz and am a big fan of numitrons, I have only had them running far a couple of years and then for only half a day at most. I have heard that their tube life is 'long' but have not seen any actual numbers. I have also re-worked the design for other numitrons like IV-19 and DTF104B.

<RCS_5954.jpg>

- Richard

On Wednesday, 16 November 2022 at 22:26:36 UTC mo...@neonixie.com wrote:

Hi Everybody,

I seem to have enough IV-9 tubes to where I can produce a batch of 100 or so 6 digit numitron clocks. Thinking ESP32 powered like my last one (Thank you to all those who gave suggestions, the kit came out great!). I'm thinking maybe try for a full through-hole solder kit with some shift registers.

I know nixies are more popular.. but anyone interested in a numitron clock kit?

Oh, one question I do have, anyone have long term experience (ie, years?) running this particular tube? How do they fare?

Thank you!

Regards,

-Moses

--

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/421b1cfd-f165-49ad-a63b-1afae649b1ebn%40googlegroups.com.

<RCS_5954.jpg>

[Peter Hall]

If Richard is in-So am I 😊

      Peter..insidiousnixies:Utube                                    

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6975019D-E714-4B91-94DC-878BE396B531%40yahoo.co.uk.

[Peter Hall]

Or am I mistaken—Its Ian that’s in…sorry..I’m looking at Emails,that go everywhere.Gotto go

Too Neonixe site,next time.

        Peter😉

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/MEYP282MB40755A707FB42B98D0B61076A6069%40MEYP282MB4075.AUSP282.PROD.OUTLOOK.COM.

[Ian Sparkes]

You really should try to make the drivers constant current drivers: I used to do this with TI parts, but they've been out of stock for over a year.

The TLC5916N drivers are shift registers with constant current drivers attached - great for this sort of thing. The Numitron clocks I made are also NTP/WiFi enabled and based on an ESP8266, with linear digit dimming and a whole bunch of other things (motion sensor, NeoPixel backlights etc).

The ESP32 is a great platform. Absolutely love working with it, but it's a bit overkill for this.

[gregebert]

Constant-current is an excellent approach for incandescent devices because it eliminates the inrush current-surge that shortens the life of the filament. Recall that a cold filament has lower resistance versus when hot; current-limiting also allows the filament to heat-up slower, which reduces mechanical shock.

I always hated it when I walked into a dark room, flipped-on the light switch, and POOF went the light bulb.

[Mac Doktor]

On Nov 17, 2022, at 1:48 PM, gregebert <greg...@hotmail.com> wrote:

Constant-current is an excellent approach for incandescent devices because it eliminates the inrush current-surge that shortens the life of the filament. Recall that a cold filament has lower resistance versus when hot; current-limiting also allows the filament to heat-up slower, which reduces mechanical shock.

PTC thermistor. I didn't understand this simple principle for a long time.

I always hated it when I walked into a dark room, flipped-on the light switch, and POOF went the light bulb.

One of my other hobbies is old Christmas lights. The inrush current is a huge deal because most of what I have is rated 15V and wired with 8 bulbs in series. You can't mix different bulbs together unless you use a dimmer or better yet a variac to slowly ramp the voltage up because the curves are different. The bulb with the lowest resistance will always heat up more rapidly and will be the first to blow.

What's needed are miniature base festoons with the sockets in parallel. These are pretty rare as the only thing wired this way were some sets with 6~8V blinker bulbs and a transformer that were only made for a short time in the mid-'50s. I managed to get three of them with 20 sockets apiece back when they were only $25 on eBay. I use variacs and 12V transformers intended for low voltage outdoor lighting. Most of them will bubble reliably at 12V which derates them 20%, further increasing lifetime.

Bonus feature: if a bulb blows all the others stay lit. And despite my best efforts it still happens occasionally but then the bulbs may have many hours on them.

Just as many here were stocking up on Nixies decades ago, I was chasing Christmas lights on eBay, especially bubble lights. I was that noob guy who outbid everyone by going fifty cents over the average price per unit. Today they can't be had for three times what I paid in 2007. Stupid noob, huh? I've never taken an exact count but I have well over 100 bubble lights in my collection. At least a dozen are purple, the rarest color. They're not for sale.

Terry Bowman, KA4HJH
"The Mac Doctor"

https://www.astarcloseup.com

"If only you could see what I've seen with your eyes."—Roy Batty, Blade Runner

[Moses]

Michail,

Thank you for your feedback! I was talking with a friend of mine and we seem to come to the same conclusion about reliability. The specs seem to under rate the drive on these tubes, may be one reason they last so long.

Here is the last clock design: http://neonixie.com/Z57XM6DV2/

A few groups members here have one.

Regards,

-Moses

[Moses]

Trying to reply to the rest of the posts.. but something is deleting my replies as soon as I post them :/

Will try again later.

[Moses]

Mark,

Going to try these posts again, not sure why they are getting deleted. Hope this makes it through.

Good to hear on the IV-9 durability.. I've never ran them other then briefly on the bench. The specs on the IV9 show 3.15v and 20ma or so for the segments, which I believe may be underdriven already just looking at the color. They look pretty good at at 3.3v. What do you run yours at?

Burning Man is awesome! My first burn was this year! Fun times!

Regards,

-Moses

[Mark Moulding]

I usually drive the filaments non-multiplexed from 74HC595 shift registers (technically slightly out of spec, but have always worked fine), and use the output enable line to PWM all of the displays at once for brightness control.  I run a separate PWM on each filament that's off, keeping it warm to just below the level of visibility.

I use a six entry look-up table for the PWM values, to create subjectively similar steps in brightness.  Most people seem to run my clocks at level 3 or 4 (10/32 or 18/32 duty cycle, respectively). at level 6 (100% duty cycle), the tubes see about 4.3 volts due to the drop across the -595 outputs.

I was a regular at Burning Man for quite a few years, but last went in 2017.  Now that it's happening again, I think we'll go back next year.  Glad you had a great time!

~~
Mark Moulding

[Moses]

Mark,

Interesting info, thank you! 

What is your voltage and filament current?

Regards,

-Moses

[Nick Andrews]

I might be interested.   I have a small supply of iv-22 tubes and some of the iv-4/17 alphanumeric ones too.

Terry, I remember those bubble lights at grandparents farm.  I had started to look into buying some many years ago,  but never did...

--

[Mike Mitchell]

I built two clocks using these displays:

They're not Numitrons, but rather use fourteen grain-of-wheat bulbs per dual display.  I've had them running constantly for over four years.

I gave one of the clocks to a friend.  Recently his one clock had a few blown segments.  I opened up the display and replaced the bulbs, after which a few more segments failed to light.  It took several tries to get all segments working again.  I did check the bulbs, they really were blown and not just a poor connection, and I had checked the replacement bulb before re-assembly.  I'm guessing "rough" handling was enough to damage the small bulb, even though I was careful in the assembly/disassembly.

I use the TLC5947 to drive the displays.  I never turn a segment fully off, but dim enough so I can't tell.  The clock has a brightness control and I run my clock at about 3/4 brightness.  Mine has not had a problem.

[Adrian Godwin]

Are they these bubble lights ?

https://www.youtube.com/watch?v=qTDtyFaAZy8

[Tom Nolan]

Hi

I've been designing an IV-9 alarm clock using all 4xxx CMOS. With little IV-9 experience ( I only have a one digit IV-9 clock that is working ) I'm adding a motion sensor and PWM brightness in the hopes it will extend the life of the filaments.

I'd be interested in one of your clocks.

All good things,

Tom

[theold...@gmail.com]

 

I made an IV-9 based clock some years ago which was featured in the Nuts & Volts magazine September 2016. I used 4511 BCD to seven segment decoders with latch function to drive the tubes directly. An HC164 shift register is clocked by a mpu to make sure the data gets to the correct tube. The tubes therefore are not multiplexed, and the whole thing can drive as many tubes as you want with only one 8 bit port. There are a few hundred of this clock in circulation, and after 6 years I am starting to see the odd tube failure aside from a couple of “infant mortalities”. Below is an excerpt from the schematic showing the part related to the tubes. JP1 is simply a way to lower the voltage of the tubes by 0.7V if so desired.

 

[Nick Andrews]

Yeah.   Apparently they still make them,  too.

--

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/CALiMYruxaMt2iefrSUC9ynAXQ0kdh1OGrdL_FG5nLQW_DDjaHw%40mail.gmail.com.

[Moses]

Bill,

I just read the article on nuts and volts, looks nice!

I must have misread the IV-9 datasheet initially.. I didn't see it listed values up to 4.5v. The 3.15v it refers to may just be a mid point nominal value or something.

Looks like you are driving the tubes at 5v. The optional jumper with the diode is a nice idea.

Regards,

-Moses

[Michail Wilson]

Bubble lights.

 

I remember when I was trying to get them and finally found them cheap on christmascentral.com

They have them for Halloween as well   (diff colors).

 

Example:

https://www.christmascentral.com/pack-of-3-multi-color-c7-retro-bubble-light-replacement-christmas-bulbs/

$7 for 3 pack.

 

Michail

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAJD4P-hyb8aR9Qeasf2Z78xXxcQcKLNJrQFSF0zrqnfU_tWwkw%40mail.gmail.com.

[Nick Andrews]

Yes, I noticed they make new ones, but I believe Terry is playing with the antique ones.  I don't know if there are size differences in the new ones or if they are better or worse quality-wise.  I do see several different varieties in the glass envelope shape.  Maybe Terry can elaborate on the differences.

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/MW2PR0102MB3435897E567A800244391D03820A9%40MW2PR0102MB3435.prod.exchangelabs.com.

[Michail Wilson]

I use TLC5916 for all my numitron projects.

Stock not available until next year and I only have about 50 left.  I have had 500 on order for just over a year now.  ☹

 

Michail

 

From: neoni...@googlegroups.com <neoni...@googlegroups.com> On Behalf Of Moses
Sent: Friday, November 18, 2022 11:12 AM
To: neonixie-l <neoni...@googlegroups.com>
Subject: Re: [neonixie-l] Re: Any interest in IV-9 numitron clock kit?

 

Ian,

 

Yes, I've been looking at the TLC5916N.. and no stock anywhere! I did find TLC5917.. same ship but with short detection I believe. There is sporadic stock available here and there. I'll order some for testing. Do you know if there is a pin equivalent from another manufacturer? Would be nice to have an alternate.

 

The ESP32 worked really well for my last clock, and I've got a good deal of time invested in the code. To be able to reuse it without too much fuss on another clock is a bonus.

 

Regards,

-Moses

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/37c7fc94-5c36-45ce-bdaa-8b8e1222ce09n%40googlegroups.com.

[Nick Andrews]

Like these or a different variant?

https://www.ebay.com/itm/314093421980?var=0&mkevt=1&mkcid=1&mkrid=711-53200-19255-0&campid=5338590836&toolid=10044&customid=2c0d2461f1b21bc7b1c4471ca2c06a78

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/MW2PR0102MB343595A29AA3E0456E4CB63A820D9%40MW2PR0102MB3435.prod.exchangelabs.com.

[Michail Wilson]

Well, yes.   That is the TLC5916IN which is the DIP version.    I use the SMT version.

 

That price is robbery; however, if you need it, then you need it.

 

If you only need a couple and need SMT, then I wouldn’t mind sending to you (as a member of this group).

The great thing about this chip is you only need a single resistor for the chip instead of one per segment.

 

This is what I use them for:   https://www.voltagegoat.com/product/bitcoin-rocket-ticker/

 

 

Michail

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/CAJD4P-j6HhfQ66DbT3Td2GJLPX9HPUgu5CQFw-B0GOM_V0MR%3DQ%40mail.gmail.com.

[Nick Andrews]

Cool project!  I don't need any currently, but thank you for the offer.  I am swamped with sorting the more recent auction junk and parts, trying to clean up the shed so I can safely use the lathe and milling machine.  Stacked up the 6 or 7 pieces of old test gear recently acquired that use nixie for the displays.  Big old Sangamo mica transmitter caps, huge HV high power resistors and rectifiers, RCA SCRs, all kinds of fun things!

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/MW2PR0102MB3435AE6159C5E20B1A3544D3820C9%40MW2PR0102MB3435.prod.exchangelabs.com.

[Mac Doktor]

On Nov 20, 2022, at 8:14 AM, Adrian Godwin <artg...@gmail.com> wrote:

Are they these bubble lights ?

https://www.youtube.com/watch?v=qTDtyFaAZy8

Yes. I'll make a detailed post about them later today. They're quite interesting and have a history to match.

Terry Bowman, KA4HJH
"The Mac Doctor"

https://www.astarcloseup.com

“...the book said something astonishing, a very big thought. The stars, it said, were suns but very far away. The Sun was a star but close up.”—Carl Sagan, "The Backbone Of Night", Cosmos, 1980

[Nick Andrews]

I have a few of both types.  One string of the Silvestri 70s/80s-produced smaller ones that run on a typical smaller-socket light string (box looks like old K-Mart style printing).  You can find them on ebay sometimes relatively inexpensively.  These are a spherical holder that the glass vial slides into, and an opening that slides over the smaller incandescent bulbs on a typical light string.

Then I have 9 of the older Noma style screw-in replacement bulbs made in the 90s for Builders Square, 3 each of blue, red, and green.  The 'festoon' part on those is simply clear plastic, not the more colorful flying saucer shape vintage stuff.

--
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/229D8E26-B4E4-465D-BD7D-49FC11C2A8B4%40gmail.com.

[Mac Doktor]

Bubble light thread coming up in a bit. I'll move this topic to that...

On Nov 28, 2022, at 11:32 AM, Nick Andrews <nickja...@gmail.com> wrote:

I have a few of both types. One string of the Silvestri 70s/80s-produced smaller ones that run on a typical smaller-socket light string (box looks like old K-Mart style printing).  These are a spherical holder that the glass vial slides into, and an opening that slides over the smaller incandescent bulbs on a typical light string.

This tiny form factor is nice BUT getting them hot enough to bubble is the devil's mischief. I've found that putting some kind of insulator around the base helps. Leaving them in the styrofoam does the trick sometimes. I've never gotten them to work on a tree.

Then I have 9 of the older Noma style screw-in replacement bulbs made in the 90s for Builders Square, 3 each of blue, red, and green. 

These are candelabra base, the same as a nightlight bulb.

Sadly, like everything else these days "NOMA" is now just a registered trademark owned by Who Knows Who, inc. All bubble lights are made in an anonymous Chinese factory and put into a bespoke box. 

They did this back on the old days as well. NOMA would sell cheaper sets through other channels using a generic brand name. Collectors can tell who the real manufacturer was by the UL number on the AC cord tag. That didn't change.

The 'festoon' part on those is simply clear plastic, not the more colorful flying saucer shape vintage stuff.

In this context "festoon" means the wire and sockets. The term isn't used much anymore. The word "string" is common parlance.

  You can find them on ebay sometimes relatively inexpensively.

Unfortunately, eBay has banned all sales of bubble lights. 1946, 2016—doesn't matter. Apparently, methylene chloride (dichloromethane) is now much more hazardous. 8/

You can probably find them locally. Walmart hasn't had them in years. Lowes maybe. They turn up at random in stores that don't sell tons of Xmas stuff like the big guys. Bed, Bath & Beyond, pharmacies, the seasonal aisle at a grocery store.

Terry Bowman, KA4HJH
"The Mac Doctor"

https://www.astarcloseup.com

"I’ve seen things you people wouldn’t believe. Attack ships on fire off the shoulder of Orion. I watched c-beams glitter in the dark near Tannhäuser Gate. 

"All those moments will be lost in time like tears in rain."— Roy Batty, Blade Runner

[Moses]

Hi Everybody,

I'm happy to say the first prototype PCB has arrived and has been assembled! Some pictures here:

http://www.neonixie.com/IV-9-6D-RR/

Some details..

- I purposefully picked all through hole parts as this will be in full kit form.

- I ended up choosing the TI TPIC6C596N driver. It does everything I need and is pin compatibility with PN '595N and possibly a few others. I've been testing software that keeps the unlit segments 'warm' by PWM'ing them to just barely invisible..  not so sure if this is going to make a big lifespan difference. These are being driven by a 74AHCT125 for the logic level shift (3.3v to 5v).

- The controller is a ESP32 original dev board from Espressif, the same one I use for my previous nixie clock. This gives me a known good working part number and software carry over from my prior nixie clock. Most of the features of that clock will carry over directly to this one, including the BH1750 lux module and AHT20 temperature/humidity sensor.

- Input is 5v via a common DC barrel jack. I did order a few USB-C input jacks, even found a through-hole one. I decided against using it in a kit due to the more delicate soldering required, as well as the unknown quality of all the USB cables and power supplies I generally run into. This runs the IV-9s directly and optionally through a diode to run them at a slightly lower voltage of ~4.2v if desired.

I am ordering more parts and should have a few beta boards out before the holidays. Hopefully production kits will be ready sometime in January!

Thanks to everyone who provided advice and feedback!

Regards,

-Moses

On Thursday, November 17, 2022 at 3:06:41 AM UTC-8 hall...@hotmail.com wrote:

If Richard is in-So am I 😊

      Peter..insidiousnixies:Utube                                    

 

From: 'Ian' via neonixie-l
Sent: Thursday, November 17, 2022 5:52 PM
To: neoni...@googlegroups.com
Subject: Re: [neonixie-l] Re: Any interest in IV-9 numitron clock kit?

 

I’m in

On 17 Nov 2022, at 04:28, Richard Scales <ric...@scalesweb.co.uk> wrote:

I have built several numitron kits from nixieclock.biz and am a big fan of numitrons, I have only had them running far a couple of years and then for only half a day at most. I have heard that their tube life is 'long' but have not seen any actual numbers. I have also re-worked the design for other numitrons like IV-19 and DTF104B.

<RCS_5954.jpg>

- Richard

On Wednesday, 16 November 2022 at 22:26:36 UTC mo...@neonixie.com wrote:

Hi Everybody,

I seem to have enough IV-9 tubes to where I can produce a batch of 100 or so 6 digit numitron clocks. Thinking ESP32 powered like my last one (Thank you to all those who gave suggestions, the kit came out great!). I'm thinking maybe try for a full through-hole solder kit with some shift registers.

I know nixies are more popular.. but anyone interested in a numitron clock kit?

Oh, one question I do have, anyone have long term experience (ie, years?) running this particular tube? How do they fare?

Thank you!

Regards,

-Moses

--

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/421b1cfd-f165-49ad-a63b-1afae649b1ebn%40googlegroups.com.

<RCS_5954.jpg>

--
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/6975019D-E714-4B91-94DC-878BE396B531%40yahoo.co.uk.

[Ian Sparkes]

Nice work!

I'm not sure if it is just your picture, but they look over-driven to me. It's perhaps just the exposure - I guess D1 is a Zener to provide some stability on the drive to the filaments, maybe you need to drop the voltage on that. The Zener is a pretty wasteful way of stabilising the voltage - you're going to need to provide for about an amp at full blast (42 segments at around 20mA = .84A), and to dump the excess over the life of the clock amounts to a lot of energy.

That's also a fairly hard set up you have there to find a nice case for, with the big dev board standing way up. Also for kits, you're going to want to put those drivers in sockets and they will also stand proud and too close behind the tubes to make an attractive casing easy.

Here's my take on it - feel free to nab a few ideas there. (These are not for sale at the moment - stock trouble).

https://www.nixieclock.biz/StoreNumitron-6Digit.html

https://www.nixieclock.biz/StoreNumitron-4Digit.html

BTW, when I was still delivering these, I would usually deliver them with the USB connector and the NeoPixels already soldered and tested.

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/KK9P2jsFon8/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/dbf54f53-6b8c-4a09-8766-a7c4c9b2b887n%40googlegroups.com.

[Moses]

Ian,

Yes, the picture is overexposed. I've found it's pretty difficult to take good pictures of numitrons!

The diode is just a plain silicon version, I put a jumper pad on the bottom to bypass it if desired. When in line it drops about 0.8v, so my input of 5v becomes 4.2v and the segments are using about 20ma each. Bypassing it and providing 5v to the segments gives 24ma. There is PWM dimming, so a maximum current option is easy to add to the firmware (on top the the default brightness settings). During normal clock operation it's running about 600ma, so I'm burning about 0.5w in the diode. It's not very hot, but I will add some more copper on the PCB if the diode remains. I did think about adding a switching converter, but that adds more parts to solder. Assuming a switching converter would be 90% efficient.. I go from burning 0.5w to 0.2w. Is saving 1/3 of a watt worth the additional parts cost? I'm not sure. I'm trying to go with a low parts count without sacrificing much.

Yes, it's a bog standard 'clock in a box' type kit, haha. But I think I'm ok with that right now. Could move everything to the bottom.. that's always an option.

I used to socket everything.. but now my thinking is with modern semiconductors, once it works it's likely good for a few decades I would hope! It's feels like I'm adding another point of potential failure.

Cool clocks on your site!

I did look at the Wemos D1 too.. but decided for this iteration I'll go with the controller I already know. For something smaller I was also looking at the Adafruit QT Py as well as the Seeed Xiao boards.

I couldn't figure out how you power your IV-9s.

Regards,

-Moses

[Moses]

An update. I've gotten a hold of a few Adafruit QT Py ESP32 boards and have been testing them out lately. The last picture is a working mockup before I spin up another PCB.

http://www.neonixie.com/IV-9-6D-RR/

Looks like I can keep both the original ESP32 DEVKITC footprint and add the QT Py easily, giving me a few options. The Seeed Studio Xiao boards have the same pinout as the QT Py, so potentially another controller option.

Height of components was mentioned previously, and if I omit the socket for the QT Py or the larger ESP32, I can get the board height down to about 9mm, which is pretty good I would say. As luck would have it.. I believe the pinout of the IV-9 is such that you can add the missing pin of the circular pin arrangement, connect them together and mount the tubes on the 'bottom' of the board. With this small change one can solder the tubes to the "bottom" with just a simple software option!

The 74125 logic converter IC is below the ESP32, I may just move this away to the right side of the board. The BH1750 lux sensor and AHT20 temp/humidity in the corner of the board are working well.

One thing I did notice is the QT Py boards can't be too far away from the wifi router.. likely from the tiny chip antenna they use. They have a variant with external antenna, might try that too. The ESP32-C3 (RISC-V) variant needed some USB drivers on my dev machine, which was a task, but eventually worked.

I have been providing the tubes with 5v and running PWM at about 75%, this gives me right around 20ma per segment, and avoids having to provide a DC-DC converter. The actual voltage to the segments is less then 5v.. since there are losses in the driver ICs. I do have a diode on the PCB as well, which you can put in line with a solder jumper. The diode drops the 5v down to about 4.2v, giving a native 20ma per segment without PWM. I'd like to ask if anyone has any knowledge on theoretical lifespan impact of either option. I like the 5v option with PWM, can limit the maximum in software but provide the option to overdrive them if anyone so chooses. This is my first production attempt at a numitron clock, so any detailed experience from group members is greatly appreciated.

Regards,

-Moses

[Moses]

I'm taking pre-orders! Waiting on a few production parts to arrive, testing and adding a few final features to the firmware. There is an order link on the main page at http://www.neonixie.com/IV-9-6D-RR/

The firmware has several new features. I've added barometric readings from the optional BMP280, mostly data displayed on the devices web server, but I've included a drop down option to use the display as an altimeter! :)

I'm working on adding a pressure display option, likely alternating between current pressure and HI / LO pressures for the last 24 hours. There's also an option to display wifi RSSI and LUX readings, mainly for debugging. If anyone has any other ideas for an 'alternate display' I'm all ears!

I'm delaying the QT PY board sale until I can do a little bit more testing and secure sufficient stock. The original Espressif devkitc module and Xiao C3 versions are available now.

Thank you everyone for your support and be sure to read the dedication on the bottom of the page!

Regards,

-Moses

[Allen]

Moses,

I was going to place an order but have a couple of questions.

What is the difference between an ESP32 and a Xaiu ESP32C3?

Also there is no way to pick “optional items”

I’m looking forward to this clock. 

Allen

Allen Wisbey, W1SBY

On Jan 20, 2023, at 5:31 PM, Moses <mo...@neonixie.com> wrote:

I'm taking pre-orders! Waiting on a few production parts to arrive, testing and adding a few final features to the firmware. There is an order link on the main page at http://www.neonixie.com/IV-9-6D-RR/

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/f8f38848-0b8c-407d-a291-fcd527033664n%40googlegroups.com.

[Moses]

Allen,

I almost missed this post completely.. been a busy week. The clock is in production and I'm catching up on my shipments!

There are currently two offerings on the microcontroller used with the clock, one is an original ESP32-DEVKITC-VE board and the other the XIAO ESP32C3..

You can look them up, as they are standard parts.. but in the clocks case the main differences would be that the XIAO has an external antenna vs onboard with the ESP32-DEVKITC-VE. The XIAO is also a much smaller board, buying you a bit more prototype space around the board. There are more nuanced differences, but a good question and I will update the docs page to give more info about this.

The optional items are links to the actual items, I'll try and make this more obvious.

Regards,

-Moses