Power Input to Noritake Itron VFD GU128X64-800B seems shorted?

[qcarver]

Hello All.

The story: I found a convection oven out-back next to a grease dumpster. Eeprom had a `Dunkin sticker on it. SOOOPER greasy, took it home dissasembled. Found (see subj) cleaned it gently w/ dishsoap and a fine brush, then later rubbing alcohol. Left to dry for a week. Tried to power it with 5VDC from a liberated ATX power supply. Nope, noth'n! I start poking around and find that the 3 pin power connector 5V and DC seem to be shorted.

The questions. I'm a sw guy but not afraid of tinkering in hw. I have used two different meters (connectivity and ye olde beep mode on multimeter.) If power and ground are shorted, they are shorted right? I mean, I'm not charging caps "C33" or "C34" (see pic) right? So this is bad right? Is there some classic thing that might be a problem? FUSE1 and FUSE2 (pic) are conducting okay. Ideas. I mean, it could be some grease somewhere, but as you can see - she cleans up nice. Thanks a ton.



QC

https://www.eevblog.com/forum/repair/noritake-display-dead/?action=dlattach;attach=137475

[macboy]

Seems like your continuity tester is backwards, with positive on the gnd. Many ICs will present a short circuit when connected backwards. Also keep in mind that a basic continuity tester is a very simple device with a simple output, don't assume that if it lights up, there must be a short circuit.

I have used several Noritake VFD displays (but not this exact one), and none of them show anything on the screen when only powered up. You need to send the commands to initialize the display and some data to be displayed.

[qcarver]

Hello, thanks. So I took your advice and assumed I just wasn't getting the GPIO data right. Lol, lead to me buying a 4-channel scope and reeeeeally getting to know the spec sheet. Alas. no phosphor is excited.

So... I went back to the scrap heap where the rest of the convection oven was, and cleaned up the CPU and the connected it to the VFD and the front panel. I also probed the power supply to see what was expected. Bear in mind that the VFD connects to the CPU board with both a 26-lane ribbon cable and a two wire power line. I checked that it was 5 Volts.

My thinking was, if the CPU still works and even it can't drive the VFD then the issue is w/ the VFD hardware. video: https://drive.google.com/file/d/1A_RRZevqIPgm2I1jqORPFBvMSUdyTUnP, you can see the CPU is working b/c it responds to keypresses, alas, no gorgeous VFD blue-ness was to be seen.

I started probing and scoping the VFD, and here's what I notice. Hopefully I can get some advice on this. The 21 or so pins that go from the Noritake backing board to the phosphor matrix DO SHOW BEAUTIFUL SQUARE WAVES at 5 volts! Yay! However, the 3 pins on either side that excite the mesh show no voltage at all. Typically w/ a VFD those alternate a higher volrage... like 17? Volts DC.

So my question is, do you have any ideas how to repair the power supply? I guess I could go through and readup on all the components I see on the back. Maybe there is a quicker way?

[macboy]

The 3 pins on either side if the VFD glass are for the filament, not grid. This should be powered by approximately 3 or 4 volts AC, not DC (though rarely it is DC). You need to measure the voltage with AC mode, as it will likely show zero in DC mode. And it's likely that the DC-AC converter operates at a few kHz, so the meter needs to be have at least several kHz AC bandwidth. Many modern cheaper meters go to about 1 or 2 kHz, not high enough.

[qcarver]

On a scope, yeah, there is around 4-to-7Volts of excitation running about 7kHz. Interesting that most of the time captures look like the top picture but about 8 times a second I see the ringing that is in the bottom picture (see attachment.)

Settings are AC coupled Probe 1 goes to the left side trio, and Probe 2 goes to the right side, I also connected the scope ground to ground line - wasn't too sure what AC means in this case, but ground wire doesn't seem to change anything.

So to review, I have data on the parallel inputs to the glass and I have excitation on the filament inputs. Still no blue magic. Does that excitation look okay? and more thoughts? -thx

[macboy]

No VFD glow can mean the seal is broken and the glass has lost vacuum. If that is the case then it is done. Inside the display in one corner you should see a black patch/splotch approximately 1/4 inch or 5 mm wide. This is the "getter", a reactive metal intended to remove (react with) any small amount of oxygen left inside after creating the vacuum. If this patch has turned white-ish it is a clear sign that the seal is lost and the VFD is dead.  Post a photo if you are unsure.

Also, the waveforms on the filament don't look right. You should see a clear square(ish) wave of a few volts P-P. The 7 kHz frequency sounds ok, it's a reasonable frequency for this DC to AC converter. The short spikes you are seeing could be what you see at each edge of the square wave, if the ground lead isn't connected. Then you would only get the very high frequencies (in the fast rise edges) through capacitive coupling only.

If you power the VFD from a floating power supply then you can hook the ground lead to one end of the filament and the probe tip to the other. This is a little risky since I'd the VFD power supply isn't truly floating you could short out the filament supply, potentially damaging the circuit. An alternate and safer probing technique is to use a dual channel scope, set to display channel 1 minus channel 2 (the difference). This provides a simple differential probe, which will work just fine at the frequencies involved here (don't expect much at multi MHz though). The vertical V/div should be set the same for both channels. You can test this psuedo-differential probe using the scope Cal output. Connect one probe tip to ground and the other to the cal signal, and you should see the expected square wave. Then connect both to the cal signal. You should ideally see nothing, but you will likely see just a little ringing/glitching at the fast rise edges of the square wave, due to the imperfect common-mode rejection. You may tweak the compensation trimmer of one probe to minimise the visible waveform. Now, Connect both probes' ground leads the circuit ground, and probe both ends of the filament. You should see a square-ish wave.

[qcarver]

Thx, didn't know that about the white. There are two quarter-inch circles of black out of the visual field on the top-left and bottom-left. Definitely black though.

The scope is a HANMATEK (note the "M" it's a knock-off) DOS1104. I did per your safer suggestion and took the difference of 1 and 2 (and then 2 and 1). In both cases a sort of noisier version of the ringing pattern is present. In the first case it's about twice the size of the source wave forms, in the second it's about the same size. Definitely NOT a square wave. So, perhaps the lack-of a square wave is the problem then?