So, you have a FLUKE 5440b with worn out VFDs !.....

[lowimpedance]

 And as the title says, I do have (or did have!) an old 5440b with the customary shagged Vacuum Fluorescent Displays. Due to many many logged hours of operation no doubt.
Also some bodgy bludgers had been poking around and left their mark on several PCB's.

So what to do about the displays, clearly obsolete, (Duh!) , so of to that favorite site we all like to peruse in search of that next bargain etc etc. And what do we find, quite a good number of glass displays which look rather promising.  So after some careful searching cross referencing and guessing I decided to select these.

http://www.ebay.com.au/itm/142264193419?_trksid=p2060353.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT
and,
http://www.ebay.com.au/itm/281725949550?_trksid=p2060353.m1438.l2649&ssPageName=STRK%3AMEBIDX%3AIT

As the best compromise for fit and most suitable. Both displays filaments also measured the same as the originals.

(Side note that If you want the HP34401 VFD should also work as a numeric display replacement  ,  just a bit smaller but has the required number of digits, just ignore the annunciators and check the filament resistance first!).
 Of course they are not just simple drop ins and a fair bit of cutting of the front panel was needed to accommodate the variations of the glass dimensions. It could be possible to mount the 'star burst' numeric display closer to the PCB and still fit behind the panel, but I wanted the display to be in the correct depth from the diffusing screen as well as reusing the original foam rubber spacer.
The dot matrix display unfortunately was not pin compatible to the original footprint and also the evac. nipple was on the side, so 96 pins of wiring joy and a little panel gouging were needed to complete the fitting. The power up result makes the effort worth while though.
 Despite the effort required to fit the displays I think the 5440b deserves that time needed.

Some photos of the new and old VFDs

[lowimpedance]

Some pics of the panel mods and wiring in of the dot matrix display.

[lowimpedance]

Compare the old shagged numeric VFD to the new dot matrix display.

[lowimpedance]

Ahh the satisfaction of looking at nice bright displays again   
Suppose I should do an external cal now. 

Plus a bonus bodge pic!. 

[lowimpedance]

And on another PCB the same resistor showing its prime days being well passed .
Which is probably the reason for the bodged resistor on the pic posted in the previous post.
Owners of these calibrators might want to check this next time the covers are off.

[CalMachine]

Very nice work!  That resistor bodge is hilarious    I'm going to be trying to get a VFD display I ordered to work as a replacement for a broken display on one of our 4808s.  I hope it comes out as nice as this did! 

[jfphp]

We have developped a couple of years ago a drop in replacement front panel with modern LCDs for the 5440B and sold some in Europe through EBay. We can supply all over the world.

[lowimpedance]

 Neat replacement. When researching the best approach to replacing the dead displays I did find this 'option' and did also think of LCD re designs myself, but could not justify the time spent going in that direction for hobby purpose only.
 And once I found the VFDs that were a good fit to replace the originals it was much easier to choose this solution over the LCD. The numeric display used is pin for pin compatible and could be fitted without milling out plastic, all be it at around 5mm deeper behind the panel and given how bright it is this should not be an issue with readability. Also the dot matrix could be fitted with some "creative lead bending" to minimize extra wires needed. Only have to cut a small relief hole for the display evacuation port.

[TiN]

Neat! If use blue OLEDs, than even look will be quite similar to "real" 5440B. 
That just reminded me reused VFD in 3458A, posted by bbs.38hot.net voltnut 

[lowimpedance]

LOL  , So he can pump up some' background measuring music' on those long drift measurements ?!....... and get the balance and boom just right too. 

[Pipelie]

 , lowimpedance
here it is my modification.

[Pipelie]

yeah, it's much smaller than the original display,but it work anyway, and looks good to me.
the resistor R1 and R2 need  adjusted according the Filament voltage of VFD you choose.

[MadTux]

Not much otherwise interesting to do, so I finally finished Fluke 5440B double display transplant, was gathering dust for maybe 5 years now.

Original VFDs suffered from some kind of catastrophic filament failure, both of em had burned out filament, no idea why.
New ones are Itron DC20026H2 and FG169K1A (with first 4 segments disabled).
Btw, don't get discouraged by VFDs with blue filter coating, the DC20026H2 had it as well, but it dissolved as I was cleaning the PCBs in IPA/toluene mix. 

Front cover still needs some milling, until everything fits back togehter, since the PCBs are a bit bigger than original display.
Obviously, the PSU still has a whack, but that should be easy.
Eventually, if I have all 3 of them, 2x 5440B 1x 5442A in good working order, I need somebody with 3458A, Fluke 720A and a good voltage reference, to get them all calibrated.

[wn1fju]

Fortunately my 5440B VFD is still OK.  But to hopefully keep it that way for a long time, a small toggle switch was mounted to the front panel and the switch wired to interrupt the 29V and 5V lines coming from the control board to the display board (via a ribbon cable).

Now, for instance, when I let the 5440B warm up for several hours before using it, I simply throw the switch and turn off the display.  In fact, most of the time, I don't even enable the display when I am punching in numbers.  Or I briefly toggle it to verify the numbers.

Don't know whether or not cycling the power on/off is harder on the display than leaving it on, but at least the modification made sense to me at the time.

[MadTux]

After replacing shorted tantalum cap and a couple of blown fuses on the big transformer, all 3 of them pass the 1100V smoke test  ;-D
The 5442A one on the right is the one with new VFDs, difference in brightness is clearly visible, smaller numbers IMO look better too.

[MadTux]

After long warmup and a internal calibration, top and middle Fluke 5440B agree to a few ppm, except for 1100V range, with about 35ppm error
Used Keithley 2010/2015 as null voltmeter, since direct measurement with LM399 toys only displays error/noise of the voltmeter.
Bottom one is way off, something in the % range. Will see, wether a new calibration fixes it. Passes self test however, so hopefully nothing too serious.

Difference     Middle/top                          Middle/bottom
11V                    7µV(0.7ppm)                        67mV
22V                    14µV(0.7ppm)                      135mV
275V                  460µV(1.7ppm)                   1.7V
1100V               35mV, drifting(35ppm)        11V

[MadTux]

Just for info and to save other guys cash, Fluke 5440B works just fine with cheap AT28C16 as NVRAM.

The "MAIN CONTROL FAULT / CHECK MAIN NV MEMORY" can be fixed with a bogus calibration, doesn't mean the NVRAM has timing issues/broken/faked chip. And it doesn't need these damn expensive DQ52B13H EEPROMs as replacement.

5440B works just fine with cheap chinesium AT28C16, the 5440B just wants some vaules written in there, so run bogus calibration after replacing NVRAM, so that 5440B is happy. Then let it warm up a couple of hours and do the real calibration.