Vacuum fluorescent display rejuvenation...SUCCESS

[willb]

So I just bought a Fluke 6060A synthesized RF generator from eBay, broken. It was sold with as "fan spins, nothing on display/bad display". All it needed was a filter cap for the +37VDC rail, and a pin re-seated on a connector for the +6.2VDC bias for the VFD. Now that I have it up and running, I have another issue...

Both VFD's are very dim. This unit seems to have a lot of hours on it. The displays are legible, but dim. I have pretty much given up on finding replacement displays for my unit. The voltages seem mostly to be within spec (6vac for the filament, DC bias is a little low at 6.05v). Would a slightly low bias dim the display considerably?

Would it be possible to rejuvenate a VFD in the same way as you would with a CRT by slightly increasing the filament voltage for a few seconds? Anyone have any experience with bringing back old VFD's?

Thanks!

[xrunner]

Both VFD's are very dim. This unit seems to have a lot of hours on it. The displays are legible, but dim.

I was given an older stereo by a neighbor, because they were going to throw it out. I was going to use it to power a set of outdoor speakers ... anyway it hadn't been used for many years and the display, which is the same type as you are having trouble with, was very, very dim. I left it on for several days and it just got better on it's own. 

[SeanB]

Boosting filament voltage will not help, it is a case of the phosphor on the anodes having degraded with constant power. You could increase the anode voltage a little, to the upper end of the anode supply range, but this may kill the driver chip. Try cleaning the inside of the display filter in the case, often there is a fine dust film there, which dims the display.

[envisionelec]

Yes, and I've done it many times with good results. The filament oxidizes which reduces the number of electrons that can be emitted at the same drive current. Driving it with higher current (constant current power supply is necessary) to white hot condition for 5-10 seconds will breathe new life into the display. It's unlikely that the phosphors are damaged if the whole display is dim. You might find a few "shapes" such as a leading zero or the decimal will be permanently dimmed, but an overall rejuvenation will minimize brightness differences.

It's not a permanent solution, but it does give you a few years. I did it to both of my bedside clocks and they've managed to look good for almost four years of 24/7 operation.

[willb]

I've already cleaned the front displays and filters with no improvement.

I'd be happy with a few extra years of life out of the display. I do have several HP power supplies that will do CC. I've never done any sort of rejuvenation on a CRT/VFD, only read about it, so the procedure is totally new to me. What would be the best way to do this? Directly hooking up the filaments to my DC supply? How much current should I use?

[willb]

Well, I was bored this afternoon so I decided to play around with it...SUCCESS!

I hooked up my DC power supply to the filaments with the 3.0VAC transformer taps disconnected. I did 3x short ramps with higher current till they started to slightly glow, each ramp for a few seconds. Hooked everything back up, and fired it up. The display was much much brighter, 100% better.

My RF generator is now perfectly usable! Thanks!

[Fraser]

Thanks for this. I did not know it was possible with VFD's 

[envisionelec]

My RF generator is now perfectly usable! Thanks!

Looks great!

[envisionelec]

Thanks for this. I did not know it was possible with VFD's 

I found it by accident when I used to take stuff apart as a kid (maybe 10-11* years old) in the mid 1980s. I didn't know how to make a VFD work, but I knew *something* needed the 30V from the power supply and the old MM series clock driver datasheet (from the library, pre-internet) said it was good to some 60V. So I accidentally applied high voltage to the filament. Once I learned that there needs to be about 30V between the filament and the segments did I realize that I had improved the brightness by a huge amount. I put that one in my brain's "little black book".

*I always encourage parents whose kids are into electronics that they really can comprehend some pretty serious concepts. If I could do it on my own without any external support (actually quite the contrary...but that's another story), imagine the kind of success they could experience with a bit of help.

[willb]

Well, maybe I spoke too fast. When I went to use it this morning, nothing came up. It blew the main line fuse. Upon further inspection, the +18VDC bridge is shorted, and one of the 7815 fed from another rail/tap is bad (output is 11VDC with load, 23 without). When I supply +18VDC and +15VDC from my bench supply, the unit powers up and does output RF (hooked up to my spectrum analyzer), I can vary the frequency/modulation/amplitude to spec throughout the range, but nothing is coming up on the front displays. Either something got taken out when one of the rails went bad on the display circuity, or the display itself is bad (caused by me???). Only thing is, the display latches run on +5VDC, and the data drivers at +37VDC, none of which were affected by the bad rails...

It went from being perfect to broken again, over the span of 12 hours, while sitting on my bench. Ugh, I think this thing is cursed.

[Fraser]

Likely just bad luck.

Check the PSU for bad caps.

[willb]

Well, it's up and running. It's my fault that it died, I always seem to have bad luck.

I mixed up the screws that hold the top cover, I accidentally put a longer one right at the power supply, shoring out the +18VDC rail directly to chassis ground, which is what initially took out the main line fuse (the screw touched the positive solder pad for the filter cap for the rail). The bridge rectifier for that rail also took a hit and cracked in half. I put in a new cap, new rectifier, new 7815 and fired it up. I was getting RF output again, as I was before, but still no display. I checked my filament voltages on the display PCB directly, and I was missing the 6.2VDC bias. The diode which provides that voltage was shorted to ground as well. It was originally a 1N753A, so I threw in a 1N4735A, fired it up, and the display was alive again.

I'm hoping it works for more than 12 hours this time, hahahaha.

[n6rob]

Well, I was bored this afternoon so I decided to play around with it...SUCCESS!

I hooked up my DC power supply to the filaments with the 3.0VAC transformer taps disconnected. I did 3x short ramps with higher current till they started to slightly glow, each ramp for a few seconds. Hooked everything back up, and fired it up. The display was much much brighter, 100% better.

My RF generator is now perfectly usable! Thanks!

willb;

Thanks for this information.  I am interested in doing something similar to a display (VFD) in an older stereo.  Can you please describe the process you used in more detail?  I am especially curious about what voltage and current settings you used.

If anyone else has suggestions I would appreciate it.  Also, if you can recommend sources on the internet; I have searched but found little.

Thanks.

[Art]

Hi:

I'm trying to rejuvenate a VFD display on an 80's VCR. When it's running, the display is dim, and on one terminal of the filament there's -20vdc on the other filament terminal there's -20vdc. I have a couple of these VCRs and they all read the same voltage.

So what I did was unplug the vcr, hook up a plug-in power supply (21vdc @ 150ma) and connected it to the two filament connects on the VFD. The filaments (I guess that's what they are) look like about ten horizontal lines, they glow in orange.  I left them like that for about 60 seconds. Then tried the VCR. The display is markedly brighter, but not up to where I'd like it to be. It seems that after about half an hour it starts to fade a bit.

I also found out I could jumper -30vdc from the VCR circuit board to one of the filament legs and it would run a bit brighter (still not as bright as I'd like) but things in the background of the display are starting to show up.

So I'm wondering, if I "cook" the filaments at a higher voltage will the fix be brighter and last longer?

       

[dadler]

Any updates on the longevity of this solution with the Fluke VFD in particular?

I have a couple of early Agilent devices with dim VFDs, but I don't want to try this if it will yield an bright, early supernova-like demise.

[willb]

Still working strong!

[dadler]

Ah excellent! Thanks for the photo.

[TheSteve]

Gave this a shot on my Agilent 33120A - wasn't able to see any change in brightness, but then maybe I didn't cook the filament's enough. I did certainly have them glowing nicely. Wasn't sure how much is too much and a slightly dim display beats a dead display anyday.

Anyone else experimented with this?

[Raff]

I have successfully rejuvenated the VFD displays in pinball machines following this, http://www.pinrepair.com/sys1/#recharge

So far all are still working well.

Raff

[Brumby]

Well, it's up and running. It's my fault that it died, I always seem to have bad luck.

I mixed up the screws that hold the top cover, I accidentally put a longer one right at the power supply, shoring out the +18VDC rail directly to chassis ground, which is what initially took out the main line fuse (the screw touched the positive solder pad for the filter cap for the rail). The bridge rectifier for that rail also took a hit and cracked in half.

So ... you literally screwed the power supply!   


(Sorry ... I just had to laugh.  It's not like I've ever done something like that ....  )

[Kjo]

I'd be cautious attempting rejuvenateing by raising the cathode heaters much above the nominal voltage. In my experience success or failure will depend on the underlying reason for a dim display. Assuming no power supply issues, VFDs fail by several mechanisms. Microwave oven displays often fail from phosphor fatigue on elements that are always on. Boosting voltages can help these in the short term.
But another failure is anode poisoning. This causes dimming of all anode elements regardless of activity. The Fluke VFD units are susceptible to this mechanism in my belief. They age based on total powered time, not individual anode segment activity. I think it is a product of the VFD manufacturing process. These VFDs are not sealed like a vacuum tube. Rather they are glued with a glass like seal. I think the getter process is not perfect, and contaminants eventually poison the emission from the anode segments.
I had a dim Fluke 8840a VFD that I carefully boosted the filiment voltage. It only made the display dimmer. Eventually it just went dark. No amount of filament or anode voltage would light it. Not being a chemist, I think there are contaminants that develop within the display that coat or chemically modify the anode phosphors.
I'd like to hear from an expert on this.
(I eventually scored a OEM NOS module on eBay!)

[AndyC_772]

[Yes, old thread, I know, but it's relevant]

Is the general consensus that a VFD which has some segments dim, but others bright, is NOT a candidate for being rejuvenated?

I've been restoring a Sony PCM-7040, which is a professional DAT machine, likely to have been left switched on 24x7 in a recording studio for much of its life. Its display is quite tired (see photo), though the segments relating to features that I presume were't used much are still bright and clear. For example, the "EXT SYNC" and "44.1" indicators are fine, but the "TIME CODE" and "LOCATE POINT" symbols are very dim, and not even uniformly so.

Any suggestions, other than trying to find a similar machine for spares?

[floobydust]

If particular segments are dimmer, then I'd guess it's the phosphor that has aged. It maybe the cathode is locally worn out? That is, low emission over that segment area, but this is a straw I am grasping at.

All we can do so far, is overheat the filament for a short while, say 60sec to "burn off" undesirables from the cathode.


For ideas, it may be worthwhile to look at CRT rejuvenators, such as B&K Model 490. In that era, many patents etc. as it was big money compared to replacing the CRT.
A VFD has a directly-heated cathode though.

Technique seems to be apply high voltage on the filaments, and also high positive G1-K voltage, with high risk.
B&K's patent here is to heat up the cathode, then simultaneously switch off the filament while then applying high +ve V to G1 for I think max. cathode current.

They say "dust" inside the tube deposits on the cathode and must be burned off.  If the cathode is worn out, depleted of barium oxide, nothing can be done.

[james_s]

Yes what you're seeing is the equivalent of phosphor burn on a CRT, there's nothing you can do about it unless you want to wear out the brighter segments to match. Years of being hammered by electrons takes its toll on the phosphors and they work less effectively.

My microwave oven displays the clock on the middle 4 segments of a larger VFD and now those segments are substantially dimmer than the others that light up when I use the timer and other functions.

[calexanian]

Very good. Now let me tell you what is happening.

In VFD displays there is a nickel alloy wire filament that is coated with barium and strontium oxide powder that serves as the emission source. At the boundary of the nickel and the oxide coating certain dopants (Aluminum, titanium, carbon, magnesium, etc) in the nickel will reduce the oxide to metallic barium and strontium and the low work functions of these metals are what give you your emission. This happens extremely slowly at normal cathode temperature of say 850 C to 950 C and during tube manufacturing the temp is raised to neat 100 C for nearly a minute depending on cathode size to speed this process. In VFD tubes the emission requirement is so low that they are run at minimum filament voltage. This also reduces the sight of the glowing filament which would interfere with he lit digit glow. At these reduced voltages that reducing effect can be weak and insufficient barium can remain to emit due to barium sublimation. Temporary raising the filament temperature or "Flashing" can restore this process and provide new barium and increase emission. This is a loosing battle however. you can typically only do it once or twice before you run out of viable material or poison the filament. As far as over voltage I cannot comment as I do not know the composition of what is in your particular VFD. In general nickel filaments are run to near yellow heat in the initial breakdown process and that occurs at about double filament voltage. I have seen literature that suggests %150 filament voltage is typical for a minute or so in rejuvenation processes. you don't want it to go much past light orange hot though for risk of it melting. You could also just bump up the filament voltage %10 and just let it fly. you may see the glow more from the filaments but it will defiantly increase emission. You are living on borrowed time at this point anyways.

Just my two cents.

[Signal Simon]

I have a home cinema amplifier, Sony STR-DB985D with a very tired display.
It's a unit I acquired for repair and to use as it seems to be the only model I could find with integrated DAB tuner.

This doesn't seem to be just a case of worn phosphors, as the rarely used segments are also very dim, just a little less so.
The heaters do seem very bright compared to other VFD displays I have seen. I have not measured the voltage as yet.

I suspected a low negative rail, but it is at -18.7V as per manual (although labelled -28V).
I am going to try and source a replacement display, the same part number seems to be used in a number of models.

Could the display have been damaged by the filaments being over-run from new (production issue)?

Simon.

[james_s]

It's possible, or the filaments may be contaminated, you could try heating them up to a yellow glow for a few seconds and then return to the rated voltage and see if it makes any improvement.

[Signal Simon]

Cheers, I will give it a go.
I did get another Sony AV amp, but the display was different.
In that case, there was no display because the negative supply was missing.
The negative voltage was -33V (again as per manual) once restored.
The display on that unit is good as new and there is no visible glow from the cathode.
Interestingly, the driver IC is the same, albeit one is NEC uPD16315 and the other is Princeton PT6315

I've attached a photo of the display on the DB895D, you can see how bright the cathode wires are glowing.
The display looks readable on the photo, but when you consider the relative visibility of the unlit segments, it isn't bright at all.

It will be another week or so before I have another go.

Simon.
 

[Jamieson]

Just a quick thanks for all who contributed to this useful thread!

I used this technique to bring back the VFD on a Sony Mini Disc deck (JE500, circa 1996). In my case the filaments started to glow dim orange around 200mA. So i ramped my bench supply from 0 to 250mA, held it there for a few seconds. Repeated a couple of times. Now the VFD is uniformly bright, looks like new, whereas before it had splotchy dim areas and dim horizontal streaks.

Edit: I should add that I first checked the negative rail voltage going to the VFD driver chip and it was bang on at -32VDC per the service manual.

[elecdonia]

I have a CD player with a vacuum fluorescent display where 2 digits intermittently get dimmer. This issue is with one digit or with the other digit. They don't both get dimmer at the same time.

All external connections have been checked and resoldered. The MCU which drives the display works perfectly. All of the digits and icons in the display are nice and bright except for the 2 which occasionally get dimmer. Note: This display has 12 "anodes" and 6 "grids."

Lightly tapping on the VFD glass itself causes the problem to appear or disappear. When I am tapping it I can sometimes see something vibrating or moving slightly inside the display itself. I suspect one of the internal filament wires might not have sufficient tension. Perhaps it occasionally touches one of the grids or anodes, thus causing this fault? Or perhaps there is a tiny bit of foreign material loose inside the display which is electrically conductive?

If I tap and shake the display for a while I can usually get it into a state where all of it works.

I'm wondering if there is any chance of being able to apply just enough volts and mA to the affected filament and grid in a manner which might "burn out" the intermittent short circuit?

-E

[james_s]

I suspect it isn't the filament at all but a connection to one of the grids, or perhaps the grid is shorting to something else. If you look closely with a magnifier while tapping on the display you might be able to see what's loose, but it's unlikely that you'll be able to fix it. If it's a short to an adjacent grid you *might* be able to blow out the short, I fixed a nixie tube one time that had been dropped by doing that, but there is a risk of blowing out the wrong part and making things worse.

[floobydust]

Always first check for fractured solder joints at the VFD's pins. Being made of Kovar, they don't solder well. It's the number one problem with intermittents and VFD's. Look at the driver IC's pins as well.

Tapping will naturally make everything vibrate. If a filament wire breaks or is loose it will short to a grid and light up like a light bulb. If a grid shorts, it will smoke the driver IC. It's not like things are moving around in there without consequences.

Number two for causing (entire) dim displays is bad electrolytic capacitors.
Number three for displays cutting out is bad filament soldering connections, these are the pins at each end of the tube.
Number four for dim digits and segments is phosphor burns from aging, this you cannot fix at all.
Number five is contaminated filament wires due to aging, the entire display will seem dim. This thread was about "burning off the oxides" by slight overvoltaging of the filament for a short time.

[stojke]

Hi, please excuse me, but in the following example, where would you apply higher voltage that is on which pins?:

[EDIT]
And what makes the display brighter or dimmer? -VKK (-30V DC) ? Or is it the AC power (3.5V)?

[floobydust]

The filament pins are at each of the far ends of the tube, where AC power is fed in.
What I find works is:
1. Disconnect one end of the filament (to stop damage of the transformer or surrounding circuitry etc.) either unplug a connector or lift the tube's filament pins out of the PC board.
2. Gently over-voltage the tube filament using a DC power supply at least 300mA. The DC power supply should be floating/isolated from everything else and with a series resistor around 5-10ohms 5W. Start at low voltage and slowly ramp up over many seconds, to at most around 2X the normal voltage or so at the tube, and hold for 30 seconds with a cozy orange glow.
Don't make it light up white like a light bulb! Don't bump/bang it around when glowing orange hot!
Don't nail the tube with filament over-voltage when it's cold, the in-rush current will blow the filaments. You can't just connect a DC power supply to it at 2x voltage, ramp it up and use the resistor.
3. Disconnect power and Let it cool down for a minute afterwards.

I do another rejuvenation stage involving the grids, but it's best if the tube is out of the board/circuit for that.