IV-12 power supply for filaments (considerations)

More
3 months 3 weeks ago #13023 by 64bittz
Hi,
I was designing my own VFD clock out of those russian IV-xx tubes and needed to design the supply for the filaments.
First, I read in many many forums that AC is better to drive filaments, mainly because it reduces lighting gradients across the display. Single digit displays should be somewhat immune to this. Does the DC driver hurt the filament in some way or is it just brightness related?


Second, I found on this forum this schematic



and I read that the anodes need to be slightly negative with respect to the cathode (0.5v). Can someone confirm? Also is 0.5v really enough or should i go higher?


Thanks
Attachments:

Please Log in or Create an account to join the conversation.

More
3 months 3 weeks ago - 3 months 3 weeks ago #13024 by Ty_Eeberfest
Have you looked here:
www.nixieclock.biz/Manuals.html

If you scroll down the page you'll see a section "IV-3A/IV-6 VFD Clock" with a link to a schematic. That design is somewhat "beta" but a few people have built it and it does work.

Purists insist on AC filaments. My own VFD experience consists of one IV-18 clock. The filament (and it's a looong one) is fed pure DC and it works fine. No discernible brightness gradient at all.

EDIT: Apparently DC does not harm the filament - can't think of why it would. But my IV-18 clock with DC filaments has been running 24/7/365 for 13 years (about 113,000 hours continuous operation) with no problems and no sign of display tube aging.

You lost me with the anode-cathode stuff. The filaments are the cathodes, running at a couple volts. The segments are the anodes, running at about +50 volts. Having the anodes at a lower voltage than the cathodes is a contradiction anyway. By definition the anode is the more positive node.

Are you going to do direct drive or multiplexing? For multiplexing you'll need tubes with grids (IV-6 maybe). The schematic I linked above is multiplexed and it appears to drive the grids at the same voltage as the segments (anodes).

There is a guy on this forum named Torsten who did some pretty extensive work with VFD clock designs a couple years ago. Hopefully he will see this and respond.

Look into it later when the dust is clearing off the crater.
Last edit: 3 months 3 weeks ago by Ty_Eeberfest.

Please Log in or Create an account to join the conversation.

More
3 months 3 weeks ago #13025 by 64bittz
Hi, thanks.

I’m driving all in static mode and I’m using IV-12 and IV-6 tubes.

I know that the filament doubles as the cathode, that’s why I wanted to give it a positive offset so that the segments have a slightly negative offset with respect to it.
Having a lower voltage on the anode is not a contradiction, it just ensures that the segments are fully off when not used, it was just a detail.

I looked at the schematic you pointed me to and I was about to say that putting filaments in series is not a great idea, but now that I think of it it endures that the current on all of them is equal.

My main problem right now is to know if DC on filaments wears them out faster than AC.

Please Log in or Create an account to join the conversation.

More
3 months 3 weeks ago #13026 by Ty_Eeberfest
For your AC vs DC wear-out, look at the edit I made to my post above. I'm guessing you read the version that got emailed to you which didn't include the edit.

So you meant you want the anodes to go below the cathode when the anodes are off? Do I understand you correctly now? Couldn't you just pull cathodes down to ground with, say, 15K or so pull-down? Drive them to +50 to turn on, float and let the pull-downs do their job when off?

Look into it later when the dust is clearing off the crater.

Please Log in or Create an account to join the conversation.

More
3 months 3 weeks ago - 3 months 3 weeks ago #13027 by 64bittz
Everything is clearer now. Thanks a lot for your feedback from you own clock that really calms me. I will drive filaments with dc.

I am no expert in this and I only read about this online, but from what I know the negative anode voltage is needed because even with 0V on the anodes the segment might glow just a bit or maybe even do some harm to some component of the tube in the long term. Anyway here is the source: www.noritake-elec.com/technology/general...mation/vfd-operation

Just half a volt over GND should be enough from what I read on the other post on this forum.
(I actually use 28v because static mode needs <30v, otherwise it would damage the segments according ti the datasheet)

On a different note, I translated a couple of IV-xx tube datasheet from russian. They are available on the tube tester archive and on sparktube. I translated IV-1, IV-3, IV-6 and IV-12 datasheets.

Thanks a lot for the replies


EDIT:

I forgot to say I’m driving the segments with a HV5812 which is push-pull. This reduces the on->off transition time because the anodes themselves are little capacitors.

I have most of the clock figured out, I was simply missing the theory of operation of the filament. Now that I’m more confident with DC, it’s just a matter to have 0.”something” volts offset on the filament. I think I might simply use some voltage reference or so. Just need to figure out how to handle 400mA
Last edit: 3 months 3 weeks ago by 64bittz.

Please Log in or Create an account to join the conversation.

More
3 months 3 weeks ago - 3 months 3 weeks ago #13028 by Ty_Eeberfest
Okay, after reading the info you linked I think I understand what you are getting at with the negative bias. Here is how I interpreted the graphs and calcs they present

In order to.ensure that an anode (segment) turns off completely and constantly, the anode must be at a lower potential (voltage) then the filament. All of the discussion of the "cutoff voltage Ek" is talking about tubes running with true AC (not DC or pulses) filaments (take another look at the graphs and you will see it). AC has a zero-crossing every half-cycle (obviously) and with the center tapped scheme they seem to like for their AC filament, every negative half-cycle the filaments would go below ground, thus setting up a situation where the filament is negative relative to the turned off / grounded segment. I can totally see how this would produce unwanted glow on "off" segments. Their example with the zener and capacitor puts a positive DC offset "underneath" the AC sine wave - their AC is actually riding on top of a DC bias. Thus zero crossings are not crossing thru ground, they are crossing thru the zener voltage, and the filament never goes negative relative to an "off" segment.

That was a lot of words, I probably should have made a sketch instead. But the big take-away I got from that info is that this bias/offset/Ek business does NOT apply to DC filaments. That would explain how my IV-18 clock and Ian's IV-6/IV-3a clocks - both having DC filaments - can work properly as designed... with no bias stuff, just turning the segments, and in Ian's case the grids too, by applying B+ and turning them off by removing B+.

Look into it later when the dust is clearing off the crater.
Last edit: 3 months 3 weeks ago by Ty_Eeberfest. Reason: minor typos

Please Log in or Create an account to join the conversation.

Moderators: AccutronTy_EeberfestIan
Time to create page: 0.101 seconds
Go to top
JSN Boot template designed by JoomlaShine.com