Ensoniq VFX SD mainboard troubleshooting

[Rasputin]

The forum login timed out on me so I will re-write this shortly:

1. The Cable is OK, all voltages are correct at the display end - while it is not connected.

2. Vbb is shorting out at on the display board. Resistance between the Vbb and GND pins is 0.00 Ohm.

3. Very possible short with the filament too - Resistance between pins 5 & 6 of the PCB connector is 4.8 Ohm.

Cable OK, Shorts on the display board.

TIP:
No need to go brutal. using the point of a nail file, you can unlatch the pins from the connector housing
and pull them out. they click back to place when you push them back in.

I'm glad you discovered how to unlatch the pins because that's definitely handy, but what I was specifically getting at was that the filament voltage doesn't go straight from the display power header directly to the display. There should be another trace that goes somewhere else, so if something is shorted then it could be at another location than the VFD proper. So cutting the traces or lifting the pins of the actual display itself would isolate it. But who wants to cut up their PCB unless they have to?

I'm not 100% sure what would make both VBB and the filament go down together, or if it's just a coincidence both are out. Mooger could probably tell you, but maybe we can puzzle it out.

[GiliRose]

OK, I get you. You mean to actually isolate the filament from the PCB.

I can eyeball the +9VAC trace from the connector all the way to pin 74 of the display (top left corner).
Yes - it branches off via the central 10uF capacitor, then to a diode and some other stuff.

The -9VAC trace goes direct from the connector to pin 41 on the display (top right).
I couldn't see it branching off anywhere.

My best bet before cutting traces would be to remove the 10uF capacitors filtering the Vbb and +9VAC lines.
That - together with eliminating the VFD itself - should give a clue where to continue looking.

[Rasputin]

OK, I get you. You mean to actually isolate the filament from the PCB.

I can eyeball the +9VAC trace from the connector all the way to pin 74 of the display (top left corner).
Yes - it branches off via the central 10uF capacitor, then to a diode and some other stuff.

The -9VAC trace goes direct from the connector to pin 41 on the display (top right).
I couldn't see it branching off anywhere.

My best bet before cutting traces would be to remove the 10uF capacitors filtering the Vbb and +9VAC lines.
That - together with eliminating the VFD itself - should give a clue where to continue looking.

Obviously anything that is a commonality between VBB and 9VAC is the ideal place to look. Are you sure you got the 10uF capacitors polarity correct when you preemptively replaced them?

But yeah, in my opinion you're barking up the right tree.

[GiliRose]

Capacitor polarity is correct.
Unless a (-) with flowing white arrows stands for positive where you come from .

[GiliRose]

Hi Rasputin -
You wouldn't happen to have a schematic of the display/keypad board, would you?

I have a suspicion a few of the tiny Tantalum/Ceramic capacitors along the top of
the board might be shorting to ground: C30, C27, C25, C24, C22, C10, C41
(unless a resistance of 0.3 ~ 0.5 Ohm is expected).

Hopefully, someone with a working VFX-SD could verify these values to be within range?
Any idea what Farad readings these are?

Thanks

G.

[Rasputin]

VFD THEORY OF OPERATION

Ensoniq used an NEC FIP (Fluorescent Indicator Panel) series display, the type of which is known generically as a vacuum fluorescent display (VFD).

The VFD can roughly be conceptualized as a combination of grid position and character segments. Each grid corresponds to the horizontal location of a character, and since this is a two-line (80 character) display then each grid actually controls two characters -- one for the top line, one for the bottom. The shape of a character is determined by 14-segments which can be either lit or not lit. In other words, which segments are designated to be on or off at any given time will determine the shape of the character displayed.

Note: Even though it is classified as 14-segment, there are actually 16 controllable features per character, so it's a slight misnomer when examining how it works.

This particular VFD has 40 grids which means there are 40 distinct character positions from left-to-right. There are two TL5812 driver ICs and each 5812 handles half of the grids, so 20 grids per driver. One TL5812 takes the left half of the display and one 5812 takes the right, but it should be noted that information is relayed from the left 5812 to the right 5812 so if the left 5812 dies then the right will not function either.

Note: Since this is a two-line display, there are two characters per grid. Since there are 40 grids positions available then a total of 80 characters are controllable.

There are four TL 4810 driver chips which control the segments for the pair of characters currently being written to (determined by which grid is enabled) -- it's a two-line display so both the top and bottom line characters are selected simultaneously by any given grid. Each character segment has been given a letter designation of either A, B, C, D, E, F, G, H, J, K, M, N, P, R and two additional controls called CR and DP. There are two characters per grid so a total of 32 signals are needed (8 per TL4810).

Note: The 4810 chip signals are not clearly split between display or character lines so any given dead 4810 will affect both lines evenly and not totally wipe out an entire line.

The segments are alternately referred to as either anodes or phospors depending upon the technical context.

The VFD also needs a filament (cathode in a technical context), in this case supplied with 9VAC. Uneven illumination can usually be traced back to a problem with the filament voltage. This voltage comes in from the PSU on display/keypad power connector pins 5 and 6.

The grid and segments are run at high voltage (+50VDC) the source of which is called VBB on the schematics. Even if the digital logic on display board are working, if the VFD driver chips do not have proper VBB then the VFD cannot work. This is specifically mentioned because VBB is carried in on its own line from the PSU over to the display via display/keypad power connector pin 3.

These display schematics should help: http://www.buchty.net/ensoniq/files/sch ... isplay.jpg

[GiliRose]

OK - This forum is closed for posting images and attachments,
so explaining things gets kind of awkward.

Rasputin - I am deducing from the link to the ESQ-1 display schematic that the suspicious elements
might be the 1N34A diodes and/or the pair caps filtering each TL4810A.

On testing, the 1N34A pair gives a forward bias voltage drop of 0.473 and a reverse bias reading of 1.29.
So I guess I will be ordering a new pair...

The ESQ-1 schematic is a bit unclear as to the designation - it might read 0.01nF or 0.01uF...
The visible markings on the board (that I can read) state "104" - which would be 0.1uF or 100nF

Thanks for the schematic, it gives certain clues as to whats going on - it is however for a different board.

Again - my issue is that something on the display board is shorting to ground.

[cutoffres]

As GILIRose as addressed a new fault without opening a new tread, it became very hard and confusing for me to read this tread. I finally abandoned to repair my VFX sd. I'm going to buy a working one and if i have time i will try swap board to found where was the fault, and will let know here.
Thanks a lot for the help Rasputin.

[Rasputin]

As GILIRose as addressed a new fault without opening a new tread, it became very hard and confusing for me to read this tread. I finally abandoned to repair my VFX sd. I'm going to buy a working one and if i have time i will try swap board to found where was the fault, and will let know here.
Thanks a lot for the help Rasputin.

This thread has definitely transcended its original intent, but just go back to your last post on April 29th and pick up there. But then again, you were planning on getting a working machine for comparison purposes regardless, right?

At any rate, you're welcome. Good luck.

[cutoffres]

I have spend hours triying to find the fault on this VFX SD. I have abandoned
I've contacted http://www.vintagesynthshop.com (buy a few parts from them the last years) and i'v been lucky, because they have accepted to swap my faulty mainboard with a new old stock one they had. So it cost to me 60$ including shipping and now the the VFX is woring again

Again, thanks a lot Rasputin for all the help. Because if i hadn't repaired the Power supply with your help, maybe the new mainboard Would not have been useful.

[Rasputin]

So it cost to me 60$ including shipping and now the the VFX is woring again

Terrific. Makes you wonder what the issue was though, doesn't it? This was the mainboard and not the display board that you exchanged?

[RNA]

Wow what an amazing thread!!

My VFX has a problem when I load the 8 Voice Chorus Effect. It throws a System Error 32 at me and crashes.
Could this be a faulty ESP chip or more to do with the software?

[PardalFerrari]

Dear Rasputin,

Sorry 4 years later asking a question about the Ensoniq VFX SD mainboard. My problem is in the DAC TDA 1541A. Checking the voltages and the data sheet, the negative voltage for the data sheet must be -15V. in the power supply there is no -15V, but next to the TDA there is an LM7915. Question: Where does he get the voltage from to send this LM 7915?
Thanks.

[Rasputin]

Sorry 4 years later asking a question about the Ensoniq VFX SD mainboard. My problem is in the DAC TDA 1541A.

You have a VFX SD with a TDA 1541? Are you sure it's not really an SD-1?

As far as I remember, the SD-1 does not supply unregulated voltage directly to the 7915 to derive the -15V rail, but uses the +/-12 analog and 5V digital as input to the 7915.

Although a bit convoluted, I am assuming they did this because they didn't want to redesign the PSU between the VFX SD and SD-1 models.

[natere2ster]

Hello all. Not to hijack this threat but I recently acquired an Ensoniq VFX (not the vfx sd) and was poking about to get audio from it. Surprisingly the keyboard calibration error went away after a "hard" reset but I am not getting any audio. Upon visual inspection I noticed R22 is burnt out. Looking at the schematic the value seems to be 220 Ohm. Can someone confirm this please. The resistor is so burnt it isn't possible to ID from the color markings on it.
Thanks!

166552285_474166960450581_2309147207258558789_n.jpg (3.45KiB)Viewed 46631 times