IV-11 VFD clock, some questions

[Gideon Wackers]

I want to build a four digit VFD clock (three of them actually) without using drivers like the max6921. I have two possibilities at this moment, one is this http://i.imgur.com/D4FGaV1.jpg and the other being this one http://i.imgur.com/5fMc7ty.png . Which would be best?

Another question is the resistor that I will need:

0.0035/25*3 = 0.00042
60/0.00042 = 140k = 130K or 150K ohm  for R1       and 100k for R2

For the grids the same circuit applies but with a slightly different value for R1 due to the different current. 

But this implies that I use 60 volt. A few people said that the tubes are very bright at the recommended 50-70 volt for multiplexed tubes so if I lower the voltage to 40 volts I would suddenly need only 100K for R1. Will I bump into problems if I lower the voltage to lets say 40 volts but when my resistors are calculated for 60 volt?

I want to make the filament supply with a 7806 and wires them in series, AC will be generated by four bs170 mosfets. Do I still need a filament resistor if I match my voltages this exact?

Suggestions are welcome, my PCB can be roughly 9 by 5 cm so maybe there are other (cheap) alternatives to using resistors as drivers? I first got this idea because everything had to be stuffed onto a smaller pcb so maybe now that I have more space I can now use some IC's instead?

[Spencer]

For my IV-11 clock, I use the HV5812 IC to multiplex the display. 

I only have a resistor on the filaments, 30ohm 2watt, two tubes are in series with +5V with no resistors on the segments. I run my tubes around +35V and that is plenty bright.

I tried running the filament off the 7805  and it became very hot to the touch so I use the MC34063 to stepdown from +12V to +5V DC.

I attached the schematic that I use in case that helps.

-Spencer

---

From: Gideon Wackers <gideon....@student.uhasselt.be>
To: neoni...@googlegroups.com
Sent: Tuesday, November 19, 2013 7:09 AM
Subject: [neonixie-l] IV-11 VFD clock, some questions

--
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/62997dbb-f1ac-4b11-8e24-42f3eec74ad6%40googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Gideon Wackers]

Thank you for your reply, sadly the HV5812 is SMD and I am going to do this on simple strip/hole PCBs. But you are using two tubes in series on 5 volt so that leaves about 2 volts for the resistor. My plan is to hook them all up in series (4 tubes) so I require 6 volts which can be delivered by the 7806. The result is (at least in my head) that I do not even need a filament resistor because I require no voltage drop. 

But I think that I am going to drive them at a lower than 60 volt supply because again someone mentions that 30-40 volt is bright enough. I think I will use a simple DCDC convertor off ebay to make 35 volt ( http://www.ebay.com/itm/XL6009-DC-DC-Adjustable-Step-up-boost-Power-Converter-Module-Replace-LM2577-/130987148386?pt=LH_DefaultDomain_0&hash=item1e7f714462 ) just for simplicity because I am not that much of a software guy. Maybe this allows for easier driving methods such as a transistor array like the uln2803? 

I also hear a lot of different opinions on DC or AC filaments, You are using DC (but only two tubes in series), do you notice any gradient?

[Dylan Distasio]

Hi Gideon-

I believe the HV5812 is still available in a PDIP package:

http://www.onlinecomponents.com/supertex-hv5812p.html?p=12648889

On Tue, Nov 19, 2013 at 9:42 AM, Gideon Wackers <gideon....@student.uhasselt.be> wrote:

HV5812

[Gideon Wackers]

Yeah but that would cost me 50 dollar in shipping :( (to the netherlands). What about my driver circuit on soms transistor array, driving 35 volt should not be that much of a problem?

Op dinsdag 19 november 2013 15:56:58 UTC+1 schreef William Lee:

[Gideon Wackers]

Slight update:

I just found a "local" retailler (mouser) which also sells them, but 20 euro in shipping costs... That is more than what a dozen of tubes is going to cost me. 

[Gideon Wackers]

I have looked around and maybe the ULN2803 is an option, or can someone confirm that my transistor driver is going to work?

And any opinions about the filament resistor and/or driving 4 tubes in series with DC instead of AC?

[Adam Jacobs]

Hi Gideon,

I think that you have some confusion regarding how VFDs are driven. VFDs are not nixie tubes or anything even similar. Nixies are not vacuum tubes, they are cold-cathode (neon) tubes. With nixies, we place ~180vdc across the anode & cathode via a current-limiting resistor. The reason for current-limiting in a nixie is because as current increases, nixie impedance decreases, causing the neon tube to rapidly begin dissipating a catastrophic amount of heat. Nixies, being neons, regulate voltage to their maintenance voltage. Any reasonable voltage above the striking voltage will work fine.

VFDs are triodes: Hot-cathode vacuum tubes. Ideally, the filament is driven with an AC supply of the designed voltage (via the two filament pins, usually a couple of volts) and the anode segments are driven with a voltage regulated DC supply at the designed voltage (usually 20-30v for direct and ~60v for multiplex). The grid is driven exactly the same as an anode segment and is used to turn the display 'on'. There are no current-limiting resistors used for driving the anode segments or the grid!

In my VFD clocks, I used the trick (which I learned from here) of driving the filaments with DC. This works fine if you are using individual numeral VFD tubes, I wouldn't try it if you are using the big multi-numeral VFD display tubes. The (known) current draw of the filament is used for calculating a resistor-divider. One leg of the filament is tied to ground, the other leg to +5vdc via the resistor. If you go with this approach be sure to do your math on the front end. These filaments draw a lot of current, which means a lot of heat dissipation in the voltage-divider resistor. Driving this setup via a linear regulator supply would need a very large heat-sink. I would recommend using a switching supply like the LM2575. Obviously, if you have an AC supply that is the correct voltage for the filament then no resistor is needed.
For a schematic: http://elbastl.sweb.cz/6-digit-VFD.zip

I liked Maxim IC's 6921 driver IC. I know you specifically called it out as not an option, but it worked well for me. Specifically, I used the PLCC package because I work exclusively in protoboard and there are easy DIP PLCC sockets. There are numerous advantages in my opinion to the dedicated VFD driver chip. Instead of 8 dedicated GPIO pins (7-segments + grid), the 6921 uses a standard SPI interface. This would be even more advantageous on 16-segment VFDs.

One thing you mentioned: Yes, 60v applied to the anodes will be VERY bright, even if multiplexed. I had to move my VFD clocks to bright places like my desk at work. a LOT brighter than equivalent nixie designs. I spent some time dialing back the supply voltages before finally giving up and accepting that VFDs are just designed to be bright.

-Adam

[Spencer W]

Only on longer tubes will you truly notice gradient so you might see it if you do 4 or 6 in a series. 

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/msgid/neonixie-l/9fd4e8e1-8c62-47ad-8807-d2bca4078b90%40googlegroups.com.

[Gideon Wackers]

I know the difference between a nixie and a VFD tube but I have to say that your explanation is very nice to summarize all the loose bits of information in my head. 

So If I have a 6 volt supply for my four filaments (in series) I would not need a resistor. I think the 7806 won't get that warm with roughly 400ma going through it but I can always take a simple switching psu for it or attach it to my (metal) enclosure. I will simply use a DC filament supply then to keep things simple (this project is going to be complicated enough). 

My only remaining concern is the driving method, I would like to order all my parts off ebay and/or tayda electronics. Places like mouser etc have (in my opinion) ridiculous shipping rates so I would like to avoid them. Another point is that the max6921 costs around 6-7 dollar a piece +1 euro for a plcc socket, I am trying to keep the costs down a bit. So a method to avoid using these drivers is preferred. 

HV5812: cheap but shipping costs are 40-50 dollar
MAX6921 expensive and/or high shipping costs

Op dinsdag 19 november 2013 20:06:50 UTC+1 schreef Adam Jacobs:

[Gideon Wackers]

Okay else I will just take one of those step down converters off ebay http://www.ebay.com/itm/5-pcs-DC-DC-LM2596-Power-Supply-Step-Down-Adjustable-Converter-Module1-3V-35V-/130984321712?pt=LH_DefaultDomain_0&hash=item1e7f4622b0

For that price I can't go messing around with DC-DC converters myself and I can just match the voltage to my tubes. 

Do any of you use some sort of protection circuit for the filaments?

Op dinsdag 19 november 2013 20:34:13 UTC+1 schreef Spencer:

[John Rehwinkel]

> Do any of you use some sort of protection circuit for the filaments?

I just use a higher voltage supply and a series resistor. That way, the difference in current between a cold filament and a hot one is less, so the starting surge is minimized. It's less efficient than just driving the filaments at their rated voltage directly, but safer. Another approach is to use a current regulated supply instead of a voltage regulated supply - then you can skip the resistor.

Note that if you're using a 7806 anyway, the efficiency is the same - it's a linear regulator, so the heat shows up in the regulator instead of the resistor. In this case, I'd either use a 7808 and a resistor (reducing the heat on the regulator and gaining some protection), or wiring it up as a current source:

http://www.circuitstoday.com/fixed-positive-voltage-regulators

As it happens, the 7806 does have the ability to limit current, so it won't give the cold filament more than 1.5A or so anyway, which helps a little (it's still more than 3x the rated current, however).

- John

[Gideon Wackers]

Yeah I think I'll go for a 

Op dinsdag 19 november 2013 21:07:42 UTC+1 schreef jrehwin:

[Gideon Wackers]

Yeah I was just thinking of wiring two filaments in series and feeding them with a 7805, that means a 10 ohm 1 watt resistor. That way I'll have the best of both worlds; minimal gradient effect, easy DC and resistor protection

So now I only have to figure out wether my two transistor driver is a good way to drive my IV-11 tubes :)

A big thank you for all the help until now, I've only got my driver question remaining :)

Op dinsdag 19 november 2013 21:15:05 UTC+1 schreef Gideon Wackers:

[John Rehwinkel]

> So now I only have to figure out wether my two transistor driver is a good way to drive my IV-11 tubes :)

I was hoping someone more versed in the intricacies of transistors would address this, but they haven't, so I'll take a stab at it.

> I want to build a four digit VFD clock (three of them actually) without using drivers like the max6921. I have two possibilities at this moment, one is this http://i.imgur.com/D4FGaV1.jpg

This one has Q1 in a "common base" configuration, which does not provide current gain. It seems to me like it would put Q1 into its linear region (due to the drop across R1), so you might have to fiddle with the R values to get Q2 to turn on and off reliably. Since you're dealing with low currents, it should work fine.

> and the other being this one http://i.imgur.com/5fMc7ty.png .

This is a more traditional "common emitter" configuration, where both transistors can give current gain. Instead of the usual series base resistor to limit current, there's an emitter resistor which does double duty of limiting both the base and collector currents.

> Which would be best?

I'd slightly prefer the second over the first (although there are reasons you might want to use the first one, involving current limiting and the like).

They both work essentially the same way: the PNP transistor is kept turned off by its base bias resistor to V+, until the NPN transistor turns on, and pulls the base down toward ground, thereby turning on the PNP transistor, which then provides V+ out of its collector to drive the VFD anodes/grids. The NPN transistor's emitter resistor limits the current flowing through both transistors (it does not control the amount of current available to the VFD, except in a sort of indirect fashion - normally the transistors are fully off or fully on).

> Another question is the resistor that I will need:
> 0.0035/25*3 = 0.00042
> 60/0.00042 = 140k = 130K or 150K ohm for R1 and 100k for R2
>
> For the grids the same circuit applies but with a slightly different value for R1 due to the different current.

As I said, those resistors don't really control the current into the VFD. However, the VFD doesn't need to be current limited, so all the resistors do is protect the transistors and microcontroller.

> But this implies that I use 60 volt. A few people said that the tubes are very bright at the recommended 50-70 volt for multiplexed tubes so if I lower the voltage to 40 volts I would suddenly need only 100K for R1. Will I bump into problems if I lower the voltage to lets say 40 volts but when my resistors are calculated for 60 volt?

Again, the resistors are only there for current limiting, and they're not terribly critical. If the transistors turn on and off reliably, and there isn't too much current through them, you're fine.

- John

[Gideon Wackers]

Thank you very much for this reply, I'll just order a few values and see what works. 

But I am missing one thing to complete the story; why are current values given in the datasheet if there is no need for current regulation? or should those be seen as normal values that you will see during operation just as with the filament current. 

Op dinsdag 19 november 2013 21:36:56 UTC+1 schreef jrehwin:

[David Forbes]

On 11/19/2013 1:36 PM, John Rehwinkel wrote:
>> I want to build a four digit VFD clock (three of them actually) without using drivers like the max6921. I have two possibilities at this moment, one is this http://i.imgur.com/D4FGaV1.jpg

>> and the other being this one http://i.imgur.com/5fMc7ty.png .

>> Which would be best?
>

I'd use the more traditional design with a resistor in series with the base of
the NPN transistor driven by the CPU, rather than the emitter resistor as you
showed.

More or less like this:
http://www.dos4ever.com/nixie1/multiplexing.gif

That resistor R5 from base to emitter of the PNP transistor can be a lot lower
value, as low as R6/Vanode. It will make the anode turn off faster, the lower
resistance it is.

The transistors can be any type rated for more than the VFD anode supply.

Also, if your VFD is too bright, you can reduce the brightness by reducing the
duty cycle, which is done by turning off the anodes on for some time in each cycle.

--
David Forbes, Tucson, AZ

[John Rehwinkel]

> I'd use the more traditional design with a resistor in series with the base of the NPN transistor driven by the CPU, rather than the emitter resistor as you showed.

David has a point - the disadvantage is it uses one more resistor, the (large) advantage is the functions are now separated - R7 limits the current out of the microcontroller pin into the base of T2, and R6 limits the current through T2 into the base of T1. Since they're separate, they can be different values (as, in fact, they are). You might have to tweak R6 a bit lower, since the supply voltage will be less for VFD use (R7 is fine).

- John

[Gideon Wackers]

http://i.imgur.com/CFTb8gI.png

is this what you mean?

Eagle was being weird today so I had to draw it in lochmaster

Why is that better than the original two ideas? I would like to understand everything :)

Op dinsdag 19 november 2013 21:48:18 UTC+1 schreef nixiebunny:

[Gideon Wackers]

I was typing the previous reply while you posted it, that clears some things up. 

You mention tweaking the value, is there a way to actually calculate it? (I'm not that much into transistor calculations as you may have noticed)

Op dinsdag 19 november 2013 22:06:16 UTC+1 schreef jrehwin:

[Adam Jacobs]

FYI David:
On the clock that I built with IV-11 tubes, I initially started with
an anode voltage of 60v; which is the median multiplexed voltage as per
this datasheet: http://www.tromop.eu/cms/media/IV-11%20datasheet.pdf and
a 1/6 duty cycle (16.67%). That clock was bright. Couldn't sleep in the
room with it running, way too bright. I tried lowering the anode voltage
supply to 50v, which is the minimum voltage listed in the datasheet for
multiplexed mode. This helped, but not enough; not by a long shot.
So, then I started lowering duty cycle. I can't remember how low I
eventually took it, but I soon realized that I wouldn't be able to get
the brightness as low as I needed without introducing some very
noticeable flicker. So, I moved the voltage back to 60v and the duty
cycle back to 1/6 and moved the clock to a much brighter location. Works
perfect. :) FYI, this was with the filament @ 1.5vdc. I don't know
enough about VFDs to know if lowering the filament voltage might have
helped.
I think that one of the features of Vacuum Fluorescent Display is
brightness. To paraphrase someone wiser than me: If you find yourself
thinking up more and more convoluted mechanisms for making it work, then
that's often an indicator that you're barking up the wrong tree.

-Adam

[Adam Jacobs]

Hi Gideon,
  How many of these clocks are you planning to build? If it is just a few, then I wouldn't sweat the costs too much. Both of those companies (used to be - I assume still) are very generous with samples. I agree that the Supertex and Maxim parts can be difficult to obtain for a low price in individual quantities from normal vendors. I've got a drawer full of Supertex and Maxim samples, I'd be happy to mail you the parts. FYI, requests for samples get a much more positive response if you use a company or university email and describe a plausible project in the request. Any time that I can't find some obscure part that is listed for a project, I move to the obscure part manufacturer's website and request a sample. No problem.

Contact me off-list with your address & parts request and I'll see what I can do. Of course, if you're designing a clock for mass production, then all bets are off. :)

-Adam

To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/679846e1-8f29-4f22-b190-322f6477af14%40googlegroups.com.

[Gideon Wackers]

I'm going to make three clocks (Christmas presents for girlfriend, brother and sister), but I'll see if I can get samples approved with my university email. Thank you for your generous offer to send me some samples but I'll first try to get my own samples at a reasonable price. 

Op dinsdag 19 november 2013 23:09:11 UTC+1 schreef Adam Jacobs:

[Gideon Wackers]

Does anyone know whether it is a bad idea to go below the recommended multiplexing voltage as mentioned in Adams comment??

Op dinsdag 19 november 2013 23:01:33 UTC+1 schreef Adam Jacobs:

[Spencer W]

Mine run from 18v low brightness to 35v for full brightness for the last 3 years and have had no issues. 

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/msgid/neonixie-l/b8deb5d8-2cb6-47ad-87b2-999149b4b870%40googlegroups.com.

[John Rehwinkel]

> Does anyone know whether it is a bad idea to go below the recommended multiplexing voltage as mentioned in Adams comment??

It won't damage the VFD or anything, but it slows down the electrons, so you could get ghosting while multiplexing. I'm guessing the numbers in the datasheet are to produce the rated brightness. Since you don't want the rated brightness, I'm guessing you're fine with multiplexing it with reduced voltage. Reducing the duty cycle should work too, naturally.

- John

[Gideon Wackers]

If there is one thing I like it is an answer like that :)

I tried to request three samples of the supertex drivers, we'll see if they grant me some. 

Thank you all for the comments, suggestions and solutions up to this point. I got more (practical) answers here in less than 24 hours than two weeks of forums got me. 

Op dinsdag 19 november 2013 23:38:09 UTC+1 schreef Spencer:

[Gideon Wackers]

Can someone tell me how to tweak and/or calculate the value of R6?

I would like to order my parts this evening, so it would be nice to know what extra parts to order in case something needs adjustment. 

I will show the results including CNC milled enclosures (a friend can make them very cheap :D )

[Gideon Wackers]

Sadly my sample request for some HV5812 was denied, but I will go for the transistor drivers (2n3904 + MPSA92) like mentioned a few posts ago. 

[GastonP]

Actually, brightness is a needed feature when this displays are used as intended. They were designed and are still used with heavy optical filters in front, to enhance contrast and visually eliminate the tube mechanical internal structure from the users view.

We of course like to see the tube naked, but we have to live with the cost of the features :)

Gastón

[Gideon Wackers]

I could give them sunglasses with the clock :)

Might be interesting for others: there are two versions of the XL6009 step up converter on ebay, one that goes to 35 volt, and another that reaches 45 volt. Example of a 45 volt type: http://www.ebay.com/itm/2pcs-XL6009-DC-DC-Adjustable-Boost-Step-up-Power-Supply-Module-6A-Output-6V-45V-/171127883167?pt=LH_DefaultDomain_0&hash=item27d804459f

It looks like Maxim does support starving students, I'll order the transistors anyway but I will probably work with the MAX6921 now. 

[Adam Jacobs]

Congratulations!

I've always been a big fan of Maxim's sampling policy. I've even abused
it (obviously, since I have a drawer-full of their stuff). The only
times that I've had them object to any part of a sample request, I was
requesting DIP samples. Anything PLCC or surface mount, they're ultra
crazy generous. My theory (and the theory of co-workers here) is that
Maxim samples like they do because they know that these engineers who
have worked with their parts on side-projects and hobbies will select
them for big projects at work as well.

FYI, the MAX6921 is also the driver IC that Limor Fried is using in her
VFD clock kit.

-Adam

[Gideon Wackers]

Yes it is very nice but also very smart. People will like you a lot more because of it and maybe that student that gets 10 dollar in samples today later orders shiploads of IC's from your company because he now designs equipment and remembers how nice those samples were. 

But now I am looking for some arduino code to drive the MAX6921, code seems sarce and the icetubeclock uses an older software version I believe. 

Op woensdag 20 november 2013 20:58:37 UTC+1 schreef Adam Jacobs: