Digit Brightness, Classic Rev6, IN-8 Tubes

More
6 months 3 weeks ago #12772 by Ian
I agree Jim, a short video could well help. Maybe there's just something that is getting lost on the way.

And Jedidiah, do you really have to wear those welding goggles indoors?

:)

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

More
6 months 3 weeks ago #12773 by Ty_Eeberfest
On one hand brightness is indeed subjective. On the other hand this isn’t his first clock. www.tubeclockdb.com/forum/12-arduino-nix...-wi-fi-not-available

A video of normal operation AND a video of healing mode are what I’m wishing for at the moment.

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

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

More
6 months 3 weeks ago - 6 months 3 weeks ago #12774 by Jedidiah
Hi all. Healing mode is very bright. I measure 19.4 volts across the 3k resistor, for a current of 6.5 ma. Not sure if that was the intended target for healing, but that's the measurement I get.

I've attached a couple of photos and yes, brightness is a bit subjective. I can tell you though in working with the clock and the IN-8's, I'm very aware of the reduced brightness, especially in a well lit environment, which is why I began this topic. If I put the clock in a dark corner, or dim the room lights, then I agree it is bright enough as is.

Back to the factual side, I'm sure we can agree that the IN-8's should be operating at about 2.5ish ma, which would require about 7.5 volts (average) across the 3k resistors. My measurements show 3.5 volts and Ty's show similar on a different clock with the same driver circuit. And we know that altering the anode resistor value does not help at all.

Since we have pretty much exhausted the hardware side, I would really appreciate some guidance on the software side to get the needed 7.5 volts across those anode resistors...

I really appreciate all of the help!!

PS: In the photos, the tube to the right is being directly driven with 175v and a 15k anode resistor.

Attachments:
Last edit: 6 months 3 weeks ago by Ian.

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

More
6 months 3 weeks ago #12775 by Ty_Eeberfest
Those pictures are fake! Nobody keeps their bench that clean.

Kidding.

But seriously, those are pictures of an UNO All In One shield but per the thread title I've been under the impression the whole time that we were dealing with a Classic Rev 6. It shouldn't make much difference, but still... what's up?

This is not a software issue. I promise. If the software was defective why wouldn't there be a whole bunch of people complaining of dim displays? And why aren't my own clocks dim? Your waveforms and voltages are pretty much the same as mine and my clock isn't (to my eye at least) dim. If your software and mine were acting different the waveforms / timing should differ.

I've been testing a few things. Based on how my clock behaves I don't agree that changing the anode resistor to 1.5K does nothing. I rigged something with a switch so I can switch one tube between 3K and 1.5K quickly and I see a very definite increase of brightness at 1.5K. The tube that's on 1.5K is is visibly brighter than its neighbors. It's not glaringly brighter but it is definitely brighter.

You won't ever get a full 2.5mA thru the tubes. 2.5 * 6 = 15mA (the pulse current needed to give a tube 2,5mA avg) which is slightly more than the optos can source. IOW if you replaced the 3K with a wire jumper either the opto would limit to ~14mA or the opto would fry - not sure which. I'm basing this statement on what Ian told us about the optos HERE

At this point I'm kinda stuck. Hopefully Ian or Jim will come up with something I haven't thought of. Or maybe Torsten will drop by and offer his opinion. I've been over and over that drive circuit and I can't see any way changing an anode resistor would not change brightness, at least a little bit. That's just plain weird,

IAN - Why am I seeing ~60VDC on the anodes that are "off"? Verified my reading using healing mode so this is not a MUX / meter artifact. Is this because the EL817s are operating above collector-emitter breakdown voltage? It doesn't seem to affect anything but I'm curious about the unexpected voltage.

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

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

More
6 months 3 weeks ago - 6 months 3 weeks ago #12776 by Ty_Eeberfest
Oh BTW, those app notes I linked yesterday show some characteristic curves for some Burroughs Nixies. Shouldn't be drastically different from Russian tubes. The curves seem to show that a Nixie is not a purely resistive device. Its apparent resistance varies depending on voltage or current so in effect we are chasing moving targets. This helps explain some of the weirdness we are seeing... the tubes apparently behave in a somewhat non-linear manner.

Or IOW - a 10% change (for example) in anode resistor doesn't necessarily produce a 10% change on current or brightness.

Look into it later when the dust is clearing off the crater.
Last edit: 6 months 3 weeks ago by Ty_Eeberfest.

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

More
6 months 3 weeks ago - 6 months 3 weeks ago #12777 by Torsten Lang
Hi everyone,
by chance I stumbled in here and read through the discussion. At least it looks to me that Jedidiah's clock behaves fine.

The measured values of Ty and Jedidiah seem to be consistent and match the values expected from the schematics and the IN-8 datasheet. BTW, it seems that the 2.5mA is the absolute maximum rated value while for 180V they recommend about 1mA. And the measured average is slightly above.

But in pulsed operation the limit is 5mA for a pulse length of 1..2ms. So at least here the clock drives the tubes beyond the limits given in the datasheet.

The multiplexing itself needs some guarding window when switching over to avoid ghosting effects, but from the scope traces it looks that this window is quite small, so there should be no significant brightness loss because the actual on time is a bit less than 1/6th of the total cycle time.

So, currently I don't have a clue why the direct driven tube on Jedidiah's foto is so much brighter than the multiplexed ones except that there may be a nonlinear relation between the current and the brightness. Perhaps you could try to change the multiplex scheme to shorter on times but then another effect will become more and more relevant: The tube needs up to 30μs to strike and up to 50μs to fully turn on.

But to make it more comparable you should at least adjust the resistor in your direct drive test setup so that the current really matches the measured average of the clock. The brightness differences look more pronounced on a digital photo than they would appear to the human eye.

Regards,
Torsten
Last edit: 6 months 3 weeks ago by Torsten Lang.
The following user(s) said Thank You: Ty_Eeberfest

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

Moderators: AccutronTy_EeberfestIan
Time to create page: 0.110 seconds
Go to top
JSN Boot template designed by JoomlaShine.com