Hi All
Over the last year Nick Stock and I have been working on a clock using six E1T decade counting tubes. I have been working on the electronics and firmware and Nick has been working on the overall presentation and case for the clock.
Here's a preliminary view of the clock:

This show a development test case using laser cut acrylic. Our
final "built and tested" case will be black anodised machined
aluminium. Over the top will go a clear acrylic cover to protect
the tubes. The PCBs are green solder mask prototypes but the
final boards will be black.
The project has reach the point where we are about to order the
PCBs and we will then work toward having the clock available. We
should have the clock available all the way from just bare PCB
sets to fully built and tested.
The main points about the electronics design are - SAM3X8C 32
bit microcontroller, firmware written in C using Atmel Studio
(GCC compiler), PIR to sense room occupancy and to shut
down/wake up the clock, chimes based on playing stereo WAV
files, bling-bling 3 colour led modules, IR handset and rotary
encoder user controls, automatic time setting using WIFI or GPS,
RTC using TCXO, power supplies etc
All will be open design and open source firmware.
The purpose of this email is to try to gauge interest in the clock to help us get the right number of PCBs made. So I would appreciate an email from any one who has some E1T tubes and might want to build a clock with them. Such an email will be on a no-obligation basis. If you choose not to buy any PCBs or a kit then that choice is still yours to make.
It is unlikely that Nick and I will be selling any E1T tubes as we only have enough for our own clocks and for what we hope to sell with built and tested.
If you want to see more about the clock, then this link will take you to a work-in-progress case 3D render:
And this link will take you to a dropbox where you will find some rough documentation, Eagle files, work-in-progress 3D printed parts, DXF files for acrylic cutting and so on:
https://www.dropbox.com/sh/6eoeioya5ljh116/AAClQipdvQUYd7n6a6jbGigua?dl=0
If nothing else, you will see from the dropbox what we feel being "open" is all about.
So, if you might be interested then drop us a line, there is no obligation on your part, the only obligation is on us to keep you informed of progress and, eventually, the kit options and prices.
Enjoy! Questions?
Nick and Grahame
The purpose of this email is to try to gauge interest in the clock to help us get the right number of PCBs made. So I would appreciate an email from any one who has some E1T tubes and might want to build a clock with them. Such an email will be on a no-obligation basis. If you choose not to buy any PCBs or a kit then that choice is still yours to make.It is unlikely that Nick and I will be selling any E1T tubes as we only have enough for our own clocks and for what we hope to sell with built and tested.
Yep, hence my caution on how many PCBs sets to buy. Easy to buy
too many. So I thought I would ask in case anyone is sitting on
some with no clock to build.
You have to fight on ebay for them. I understand Jan
https://www.die-wuestens.de/eindex.htm might have some. Whatever
the source, they are not a cheap tube - and the clock uses six of
them. I've never got my hands on an East German clone but I have
every expectation that they will work.
Nice work guys. If only tubes were more available.
Terry
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/28209447-d817-4367-a9c3-4c2612a028f0%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Hi Terry,
I think I can answer all of your points. On the run into this project I built a lite placer pick and place machine https://www.liteplacer.com/ and built and tested a large toaster oven for reflow soldering. I'm converted to the dark side and mostly use SMD parts now. For example, this is the SAM3X8C 4 layer microcontroller plug in board:

The lite placer builds it in 10 minutes (I don't run it at full
speed) as against a few hours and serious back ache... So
pre-fitted SMD parts is going to be my preferred route.
Your other points:
1. The E1T needs two pulse inputs. One to reset the tube to zero
and one to step the tube forwards. The reset pulse is a a fairly
brute force effort and not hard to generate. The step pulse is
more tricky requiring a sawtooth waveform for a certain peak
voltage, a defined rise time and a defined decay time.

In investigating E1T designs I found clocks by Dieter Wächter http://www.tube-tester.com/sites/nixie/different/e1t-clock/e1t.htm and and Ron Dekker http://www.dos4ever.com/E1T_clock/E1T.html . The reset pulse generators are more or less the same, the step pulse designs are quiet different. So I built and tested both. (BTW I have corresponded with both Dieter and Ron and acknowledge and reference their work in the the clock documentation.) I found Dieter's design, shall I say, more fussy. Using Ron's design I can step all of the E1Ts I own without firmware tuning except two tubes. Those two tubes are most seriously worn and are effectively dead. To keep choices open, the E1T driver board will take either Dieter's or Ron's design but I'm planning on fitting Ron's. Lastly, I'm not asking the tubes to run at there design speed of 30kHz, effectively I only use them to count at 1kHz at the most.
More info here:
http://www.dos4ever.com/trochotron/TROCH.html
http://www.dos4ever.com/E1T/E1T.html
Electronic Counting Circuits JB Dance 1967
2. The sleep options are to run the heaters at full power 24/7 or shut them down completely when a set period of inactivity has been seen by the PIR (the heater power choice and time periods are all user adjustable without burning the firmware). I know there are various schools of thought on extending heater life including keeping them simmering on low power. The clock doesn't do the latter. My understanding is that the greatest risk to the heater is when its resistance is lowest when fully cold and so the peak current will be highest on switch on. So the 6.3V heater PSU slow starts the output, taking 7 seconds to rise from 0V to 6.3V. If you want a slower rise then it is not hard to change this period by changing single capacitor on the main board.
3) The selection of WIFI or GPS is made using a plug in board,
both cannot be fitted simultaneously. The the plug in board
communicates with the SAM using 3.3V TTL (not 5V!). So if you have
an external NMEA source it is trivial to feed that into the
connector. I use the default 4800 baud speed, another speed would
require a trivial firmware change. I have designed (but not built)
a RS232 plug in board so that a wider range of external time
sources can be accommodated.
Grahame
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/856e344f-694e-476e-baa9-7509401010ce%40googlegroups.com.
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/19b07c42-9afe-731d-f958-3444f70427ff%40googlemail.com.
The lite placer builds it in 10 minutes (I don't run it at full speed) as against a few hours and serious back ache... So pre-fitted SMD parts is going to be my preferred route.
Your other points:
1. The E1T needs two pulse inputs. One to reset the tube to zero and one to step the tube forwards. The reset pulse is a a fairly brute force effort and not hard to generate. The step pulse is more tricky requiring a sawtooth waveform for a certain peak voltage, a defined rise time and a defined decay time.
More info here:
2. The sleep options are to run the heaters at full power 24/7 or shut them down completely when a set period of inactivity has been seen by the PIR (the heater power choice and time periods are all user adjustable without burning the firmware). I know there are various schools of thought on extending heater life including keeping them simmering on low power. The clock doesn't do the latter. My understanding is that the greatest risk to the heater is when its resistance is lowest when fully cold and so the peak current will be highest on switch on. So the 6.3V heater PSU slow starts the output, taking 7 seconds to rise from 0V to 6.3V. If you want a slower rise then it is not hard to change this period by changing single capacitor on the main board.
3) The selection of WIFI or GPS is made using a plug in board, both cannot be fitted simultaneously. The the plug in board communicates with the SAM using 3.3V TTL (not 5V!). So if you have an external NMEA source it is trivial to feed that into the connector. I use the default 4800 baud speed, another speed would require a trivial firmware change. I have designed (but not built) a RS232 plug in board so that a wider range of external time sources can be accommodated.
Hi Martin
Hopefully the case will be available as machined anodised
aluminium (expensive but very nice), laser cut acrylic
(cheaper)(clear, black, etc you choose) and as DXF and SVG cutting
files for you to make your own. The case is held together with 3D
printed parts so these will be available (the E1T bases are 3D
printed) to buy and again as STL files. I will have available the
Fusion 360 files if anyone wants to modify them.
The dust cover will be welded butt joints (expensive) or held
together with 3D printed corners (Låda parts
https://wyolum.com/lada-a-custom-project-box-system/). The welded
butt jointed case is likely to be bought in but an overall
dimensioned drawing will be available as a DXF file. The Låda
version will be available to buy (flat packed!) and as DXF/SVG/STL
etc for anyone to roll their own.
Open Design!
Grahame
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/97627dfa-6524-4819-b6ae-078a6f317b2b%40googlegroups.com.
Hi Martin
I shall have a play when I get a chance. Increasing the anode
voltage is not a problem and a pot in cathode is easy to try. Do
you know if any other voltages or component values were changed?
There is too much magic involved with these tubes so I stuck
closely to the datasheet and the arrangements described by JB
Dance.
Grahame
Telefunken and other manufacturers of counters with the E1T used two simple changes to the design to be able to adjust the circuit to be able to use tubes with wider tolerances, the first was to run them in a design with up to 400V anode voltage, the design was compensated for this higher voltage, but they also had a much simpler design which just incorporated a 10k potentiometer in series with the 15k cathode resistor. This potentiometer is then used to adjust for different tube characteristics. I haven't seen Philips present similar circuits but they were probably using them too. I've tested both designs but find the one with the 10k potentiometer a lot easier to use. It let me run quite a few more of the tubes in my scrap bin, but more importantly I think it is a good way of sorting tubes that behave like new from tubes that will soon end in the scrap bin, apart from letting them run in an unmodified original circuit. I haven't found any documents describing this change but it appears in some of the Telefunken counters.
/Martin
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6047b0a3-1c0c-4129-83f5-35afbe0c4550%40googlegroups.com.
Hi
Yes, I understand that the voltages scale, I suppose that is expected. In early experiments I found some tubes would count with, say, 320V anode voltage but not 300V. This was when I was not forming the sawtooth step pulse correctly.
In the section on running an E1T counter at 100kHz, Dance says
"It is important that the first E1T tube should be selected
carefully." He presents a test circuit to select high frequency
tubes (100kHz) and the remainder that don't pass the test are used
for other counts (< 30kHz).
I don't use the A1 output anode to detect a count of '10' but maintain a position list in the firmware and send a reset pulse to move to '0' when you try step beyond '9'.
Thanks
Grahame
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/252715da-5fd1-4883-8f02-441caa3305f3%40googlegroups.com.
Good project - I'm in!
What reflow oven are you using - I have a controleo3 kit but have not chosen the oven yet...
Cheers
Nick
Hi
Yes, I've seen that 400V circuit but it is not one I've followed
through on it.
The reset will always be to zero and then I pulse the tube, e.g.
'8' to '2' is 'reset' 'step' 'step'. Each E1T uses three I/O lines
from the microcontroller: 'HT switch', 'reset' and 'step pulse'.
So each E1T can be individually selected to glow. Then two further
I/O lines can switch the 6.3V heater PSU and the +300V PSU for the
whole display. In the clock as a whole, I started with 62 I/O
lines from the SAM and I'm using 60 of them. I step pulse the
tubes at 1mS intervals so the eye sees very little of the steps,
in fact, I want to do the opposite which is to allow the step
pulses to be longer apart so the eye sees the tubes stepping.
I've run all of the E1Ts I own and not found any tubes that get unbearably hot (except when I accidentally plugged one in to 12.6V, Ooooops, it did survive). At 6.3V all are warm to the hand but not burning. Given each tube has 2W input power I wonder why some are so hot? A quick measurement suggests the back of the tube to the acrylic cover distance will be about 40mm. Heat generation and final operating temperature is something we will keep an eye on. The heaters are the bulk of the energy input to the box so 6 * 6.3V * 0.3A = about 12W. Overall the clock uses < 20W of power and Nick (as the case designer) knows that passive ventilation for this heat is required.
All good fun
Grahame
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/cc948685-169b-4322-9949-1629a854d1be%40googlegroups.com.
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/2d4cf79e-381c-416b-9b1d-d4c927ad0bf3%40googlegroups.com.
I built a E1T clock a few years ago that can be seen here: https://www.youtube.com/watch?v=J5I_L4KY4Qo
Hi
I agree with Morris. Once the clock is awake you aggressively try to shut it down again. I have a few badly warn E1Ts and one has a burn in the zero position. It must have sat glowing in the position for a very long time. Unlike an oscilloscope tube you have no brightness control as the beam current is set for you. You also cannot jiggle the image as you can with a CRT clock. With leading zero suppression, shut down and the odd slot machine routine I don't think phosphor burn will be the cause of tube failure.
Grahame
PS nice clock Morris.
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/66fdb559-1f9f-4a9e-9d2c-ce4893bf7ac4%40googlegroups.com.
As some day it may happen that a victim must be found,
I've got a little list — I've got a little list
Of society offenders who might well be underground,
And who never would be missed — who never would be missed!
There's the pestilential nuisances who write for autographs —
All people who have flabby hands and irritating laughs —
All children who are up in dates, and floor you with 'em flat —
All persons who in shaking hands, shake hands with you like that
—
And all third persons who on spoiling tête-á-têtes insist —
They'd none of 'em be missed — they'd none of 'em be missed!
And now You're on the list as well.
--
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 email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/00049398-e9e4-4bf4-b27a-fd2e2cca5d2c%40googlegroups.com.