Nixies in EETimes

[fixitsan]

http://www.eetimes.com/electronics-news/4396235/Microchip-s-Nixie-Tubes-light-up-Design-East?cid=NL_MCU&Ecosystem=microcontroller-mcu#180032

[Nick]

Hugely over-engineered or what?

On Tuesday, 18 September 2012 09:35:23 UTC+1, fixitsan wrote:

http://www.eetimes.com/electronics-news/4396235/Microchip-s-Nixie-Tubes-light-up-Design-East?cid=NL_MCU&Ecosystem=microcontroller-mcu#180032

[Dekatron42]

Definately over-engineered, it says "created by John Day and Sean Cappy of Microchip" so it is no wonder they use a lot of Microchip IC's......

 

/Martin

[jb-electronics]

I don't know if it is over-engineered. They have PWMable RGB LEDs on every Nixie tube. So either use a small, cheap PIC controller per Nixie tube (for $2 each or so) or buy an expensive RGB controller.

Jens

--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To post to this group, send an email to neoni...@googlegroups.com.
To unsubscribe from this group, send email to neonixie-l+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msg/neonixie-l/-/AJIZpzkI8ZwJ.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[fixitsan]

I don't know if it is over engineered, and I can hardly argue it is when I used one microcontroller per each B7971 back in 2007 (project still available here http://groups.yahoo.com/group/smartsockets/files/ )

i think they could have done away with the main controller pic and assign it's functions to one of the tube micros and same too for the RTC, but coming from Microchip, it is designed to use as many of their products as possible I guess !

[Michel]

The PICs being so dirt cheap, I designed my first 6 controller board
just a few weeks ago. That is a board not bigger than 10x6cm. I even
find it more convenient to take a PIC rather than a simple NE555 for
example. That would have sound crazy a couple of years ago, I know....

The only thing that is a bit inconvenient is programming 6 controllers
on 1 board. You either need to program them before soldering them, use
tricks that allow you to multiplex them to a single programming port
or use 1 programming port for each controller. For my circuit I choose
the latter one, it wasn't my first choice but it ended up to be the
best choice.

Michel

[threeneurons]

Chris,

You beat me to it, but brought to our attention, by the guy who brought us the "Smartsocket". Another bit of irony, Microchip are the ones who brought us "Charlieplexing", so you could do this with the minimum of I/O. 8 bits can give 56 outputs [(8*8)-8]. 44 for the cathodes, 6 for LEDs, plus another 6 for other stuff.

[fixitsan]

True Mike.
But then they don't sell so many chips if they promote that idea :)
Maybe the marketing department is just waking up to the idea of selling by example !

[fixitsan]

On Tuesday, September 18, 2012 11:57:16 AM UTC+1, Michel wrote:

The only thing that is a bit inconvenient is programming 6 controllers
on 1 board. You either need to program them before soldering them, use
tricks that allow you to multiplex them to a single programming port
or use 1 programming port for each controller. For my circuit I choose
the latter one, it wasn't my first choice but it ended up to be the
best choice.

Michel

Couldn't you daisychain them via the serial port and use a form of bootloading ?
It isn't fast doing it like that though.

Microchip have a service where they program the code onto your chips before they send them out...I've looked at it but it wasn't cost effective at the time, I don't know what it costs these days...here's the link to the UK upload page (you can change country using the 'change' link) http://www.microchipdirect.com/programming/default.aspx

[NeonJohn]

On 09/18/2012 05:28 AM, Dekatron42 wrote:
> Definately over-engineered, it says "created by John Day and Sean Cappy of
> Microchip" so it is no wonder they use a lot of Microchip IC's......

That's kinda old-think, wouldn't you say? While the multicolor fading
LEDs are a bit tacky, the throwing of handfuls of processors is very
appropriate.

I'm taking the same approach to my latest induction heater design.
Whereas my last one contains one processor, lots of code that took
months to test for all possible boundary conditions and is a bear for an
incoming technician to come up to speed on, in my current design I have
6 Atmels and an FPGA. Each $2 processor handles its own little area but
does it well, very fast and with simple code. The FPGA handles the
really fast stuff like desat protection of the power IGBTs and then
informs the master processor what it did via SPI.

This project is coming together approximately 3 times as fast as the
previous one, has many more features (such as ethernet and USB remote
control) and will cost about the same. The savings on code testing
alone justify this approach.

One little example. I needed AGC (automatic gain control) to level out
the signal coming back from the resonant tank of the heater. This
signal ranges in voltage from about 20 volts to over 2000. I whipped
out a conventional analog AGC circuit using an op-amp and a FET plus
some resistors and it worked fine. But at about a buck fifty for the
very high speed op amp and 85 cents for the FET, I realized that I
could do it cheaper in a processor.

Now an AT90PWM sits there using its 12 bit DAC to measuring the incoming
signal. It PWM DACs a control voltage to a simple bipolar transistor
that shunts one of the resistors in the voltage divider. The few lines
of code use the built-in hardware PWM processor to generate whatever
base current is required to keep the processor incoming voltage the same.

I'm a slow programmer but it only took a couple of hours to knock out
the code to make it work.

I get a couple of vital functions out of the deal too. One is an "AGC
lock" signal (via SPI) that tells the main processor not to try to
phase-lock if there is insufficient signal. Another is with a little
integral thrown in, the output is always the same voltage. The analog
AGC did OK in this regard but the digital version is perfect.

If I were designing a Nixie clock right now (wish I had the time!), I'd
be taking the same approach. One processor per tube, a processor for
feature control such as remote control and motion sensing and a master
processor to glue it all together.

Hmmm, this got kinda long but I felt the need to counter the inevitable
"overengineered" comments with a little NewThink.

John

--
John DeArmond
Tellico Plains, Occupied TN
http://www.fluxeon.com <-- THE source for induction heaters
http://www.neon-john.com <-- email from here
http://www.johndearmond.com <-- Best damned Blog on the net
PGP key: wwwkeys.pgp.net: BCB68D77

[fixitsan]

On Tuesday, September 18, 2012 5:05:29 PM UTC+1, NeonJohn wrote:

If I were designing a Nixie clock right now (wish I had the time!), I'd
be taking the same approach.  One processor per tube, a processor for
feature control such as remote control and motion sensing and a master
processor to glue it all together.

John

Is it wrong of me to think that you would take the same approach for a 4 digit 7 or 14-seg LED clock too, John ?  If not, what is the difference ?

[Dave]

Are these guys going to share or sell this  board and the uP code?

On Tuesday, September 18, 2012 4:35:23 AM UTC-4, fixitsan wrote:

http://www.eetimes.com/electronics-news/4396235/Microchip-s-Nixie-Tubes-light-up-Design-East?cid=NL_MCU&Ecosystem=microcontroller-mcu#180032