Fluke Scopemeter 123 display dim

[echase]

My oscilloscope display has a large dark patch covering most of the LCD like a large screen burn. It is just legible but difficult to use. Edges are OK. The backlight CCFL is OK. Also has squiggley yellow marks around the edge outside the active LCD area. Suspect the glass panes are delaminating.

Not found anyone else with this problem using Google.

Suspect that a new display is the only solution but not worth paying official price for it so has anyone found an OEM solution? Or have to buy a dead 123 on eBay and hope the screen is good. Or get a Smartscope which is half the price of a new display.

Not impressed by this unit's reliability as have already had to replace the battery, the keyboard foil, the battery current sense resistors and a 32kHz crystal early on in its life. Have not been able to get it running on batteries for years.

 

[kellys_eye]

Disappointing situation for a Fluke..... do you know what the size/resolution is? Any identifying details that might help locate a replacement?

Found an earlier thread that reveals some details:

https://www.electronicspoint.com/threads/fluke-123-replacement-lcd.187922/

<edit> hmmmm.... there is 'Flukeview' software for displaying on a laptop (possible last gasp option) but this may have a signal that could be interpreted by a small μP and converted to display on a modern LCD (colour even?) fitted as a replacement.... just a thought.

The signal (video I presume) is a serial stream via an IR interface as shown on page 166 of the service manual.

Service manual:

http://assets.fluke.com/manuals/123_124_smeng0000.pdf

 

[echase]

Disappointing situation for a Fluke..... do you know what the size/resolution is? Any identifying details that might help locate a replacement?

Found an earlier thread that reveals some details:

https://www.electronicspoint.com/threads/fluke-123-replacement-lcd.187922/

<edit> hmmmm.... there is 'Flukeview' software for displaying on a laptop (possible last gasp option) but this may have a signal that could be interpreted by a small μP and converted to display on a modern LCD (colour even?) fitted as a replacement.... just a thought.

The signal (video I presume) is a serial stream via an IR interface as shown on page 166 of the service manual.

Service manual:

http://assets.fluke.com/manuals/123_124_smeng0000.pdf

Many thanks. I will post a pic of back of screen but has few markings.

I do have the RS232 connection lead that I have never used so could try to use a laptop or tablet with that. Not sure if the windows software duplicates the inbuilt screen and buttons. But rather defeats the point of this scope which is as a battery powered free standing machine. If I can be tied to mains and laptop I already have a Picoscope.

My 123 also seems to have lost ability to trigger off channel A so is yet another reliability problem.

Advantage of the Smartscope is that, whilst it does need a second device like tablet as a display it can all be battery powered so portable. Then I guess the Picoscope can be too, although bulkier as does not support android tablets.

 

[kellys_eye]

Fault tracing the Fluke should be relatively easy given the full schematic available via the document I linked to.

The display is definitely compromised - spares seem to be impossible to source unless you want to pay exorbitant prices so it should be considered 'beyond economic repair' unless you have some 'love' for it.

If you have the serial lead then you could 'sniff' the packages using a simple port monitoring program and, if you're lucky, identify the protocol(s) and/or the parts needed to redirect to a modern TFT screen.

There are heaps of 320x240 colour LCDs out there that could fit in the space inside a Fluke - if you managed to do it I reckon there'd be a healthy market for kits as that range of scope looks to have lots of issues relating to the screen.

 

[echase]

Looks to me that the LCD has has six Hitachi HD66204TFL 80x80 column and row drivers to give a 240 x 240 set of dots. These six chips are driven from a 21 way ribbon connector using under 10 of its ways, with the rest being earths or providing, together with a couple of simple op amp chips, voltages for the LCD supplies. So the Fluke’s motherboard must be just transmitting a set of points to activate as x,y coordinates. If I put those 10 ways into a PIC microcontroller which transmits the data and timing into something to drive a different LCD dot matrix display would it work? Just need a cheap 320 x 240 display that just takes row and column data rather than “draw me a letter C of size N and font M at position x,y”


Am not very competent at graphic display protocols but do know how to programme PICs and get them to drive character displays. The data/enable/clock lines on those character displays are similar in concept to this parallel driven HD66204TFL ones.


http://pdf1.alldatasheet.com/datasheet-pdf/view/113565/HITACHI/HD66204.html has the datasheet for the row drivers and an example of a multi-chip larger display which looks broadly similar to the Fluke circuit. They have all the 4 bit data lines to all chips shared on the same 4 ribbon connector lines, as does the Fluke. The chip is not programmable so the protocol of which bit of data turns on which pixel is fixed. “All” one need to do is the find out how the motherboard is commanding a particular pixel on and get the PIC to transmit that to the same pixel on the new LCD. May be even easier if the new LCD has a serial interface. Some 3.2" ones available with SPI interface for Raspberry Pi or Arduino for say £20. As not square like the original probably can only get a 3.2" in whereas original is 4".

 

[echase]

kellys_eye, how would you like to work with me on this? I am in Hampshire. Don't know if this forum takes private messages.

Correction to above post, it’s not 6 Hitachi HD66204TFL as that is just a column driver and there are only 3. The row drivers are 3 Hitachi HD66205TFL. Each group of 3 trigger the next one in the chain so that it cycles through the 240 rows and columns without needing a signal from the motherboard to move to the next chip. The main control signals that the motherboard delivers is a start signal for the frame, the 4 bit data for the columns and a couple of clocks to synch it all.

In one frame it must cycle down all 240 rows in order and the column drivers turn on the pixels that are needed in sequence for the first 4 columns because it’s 4 bit column data. One 4 bit word per row. That frame is then repeated 59 times more for the remaining columns to fill the whole screen. Not clear how the pixel stays on till the next refresh, maybe it does not but is simply pulsing briefly once in every 60 x 240 cycles and the eye smooths the pulses out.

If a PIC is synched to the row sequence and reads all the column data into an array it will have the complete map of pixels. Somehow can that be then output on SPI to the new display? As most are colour and the Fluke is monochrome just need to tell every pixel to be black or white. Would help to know the refresh rate of the screen so can tell if a PIC can collect all that data and output it in a new format fast enough. Suspect it’s in the 25-60Hz range. PICs can go up to 32bit and at least 64MHz clock so should be OK. It needs to read 60 x 240 4 bit words per refresh = 720kHz data rate at 50Hz. I've probably made a mistake in the maths here somewhere but seems doable with a PIC.

I can use my Picoscope to measure the real figs.

 

[kellys_eye]

kellys_eye, how would you like to work with me on this?

As much as the idea intrigues me I neither have a Fluke Scopemeter to work on nor the space in my schedule to give it the attention it deserves.

My idea to do this as a profitable venture may* still be valid - if you (or someone else) wanted to progress it as the alternative is a VERY expensive replacement display or a junked piece of equipment that hasn't (yet) reached it's end of useful life...... or has it?

*With desktop DSO's now in the sub £200 range - I think there are examples of handheld devices out there that are priced less than a replacement screen (Fluke) and/or the whole scope itself:

https://www.tester.co.uk/unit-itd1050cl-handheld-digital-oscilloscope

Building a screen-replacement for the Fluke may only cost £50(???) but that's still 25% of the cost of a whole NEW 'scope - and that's without adding on the profit margin you'd want to take yourself.

 

[echase]

Thanks for reply.

Have done more research on this and the biggest problem to achieving £50 may be that the cheap displays have a controller chip that can take fancy commands like "draw a circe of certain size." They aren't so good on "turn on these specified 4 or 8 pixels at the same time". So the fluke one is turning on 4 pixels at a time whereas the cheap ones only do one at a time unless you specify it as a circle, line etc.

Thus I think I will hit a speed bottleneck where I can't input the pixels fast enough to keep up. Ways around this may be to only refresh on every 4 or 8 new screens and dump the data for the others. Any idea how that would affect the display? Presumably not at all if it's a stable trace but may be problematic for a unstable trace or a transient. Perhaps causing gaps in the trace. Or I find a display that can take multiple pixel commands, but suspect either not available or expensive. Or a super fast controller like at least 4 times the fluke solution.

 

[kellys_eye]

Have a look at the LCDs that they use in the cheapo DIY-build 'scopes from China - these kits cost around £12 'complete' - the screen is a separate plug-in module with, I think, direct access rather than via I2C interfacing.

https://www.ebay.co.uk/itm/Android-...206999&hash=item48a531fffc:g:4XwAAOSw3WxaFijN

another option is the STM32 kits with screen too. I have one of their development boards programmed for SDR use and the direct drive screen can easily cope with scrolling 'live' waterfall displays of received signals.

https://www.ebay.co.uk/itm/STMicroe...129615?hash=item2a89a838cf:g:G0UAAOSwySFZgX--
not suggesting you use the complete board - but the display itself is useful. Download the development pack for this board to get the full schematic/interface and part numbers.

 

[echase]

Is that STM one monochrome? If so that screen may work. Biggest speed bottleneck may be colour as have to transmit a 16 to 24 bit colour word for every pixel. Unless there is one which enables a block command of one colour but not found one. On monochrome only one bit needed for on or off so transfer speed is at least 15 times faster.

 

[kellys_eye]

No, the STM is full colour. Nothing stopping you tying the RGB data lines together to get white though.

 

[echase]

Looked at VGA RGB and composite video screens but all are taking an analogue level in so not easy to implement without an extra A/D chip. Even TTL RGB is as far as I can tell really a multi stepped voltage level to define colour.

But the common raspberry pi etc. cheap screens with RGB are easier as genuinely a digital i/f. Although needing 18 or more bits per pixel to define colour it should be easy with hardware to set all to 1 or 0 depending if scope wants each pixel on or off. So so don't need a processor to generate all 18 bits.

Have ordered a monochrome lcd with i/f very close to what fluke generates which should solve mine. But there is no guarenteed supply of these unless buy 100s at a time so may not provide a fix for others. A colour one, although harder to drive has more secure supply.

Found out the reason mine is bad is that the rear glass sheet is cracked. As it's protected against damage by multiple layers of glass, plastic or pcb on on 8r other side it's not external forces that cracked it, more likely poor quality in production.

 

[kellys_eye]

Not sure how relevant it may be but Dave over at EEVBlog has an article on getting your own LCD manufactured. It's not as expensive as many would think and if you bought a 'minimum' quantity you may find the surplus getting you a profit from others with similar issues.

 

[tester]

My oscilloscope display has a large dark patch covering most of the LCD like a large screen burn. It is just legible but difficult to use. Edges are OK. The backlight CCFL is OK. Also has squiggley yellow marks around the edge outside the active LCD area. Suspect the glass panes are delaminating.

Not found anyone else with this problem using Google.

Suspect that a new display is the only solution but not worth paying official price for it so has anyone found an OEM solution? Or have to buy a dead 123 on eBay and hope the screen is good. Or get a Smartscope which is half the price of a new display.

Not impressed by this unit's reliability as have already had to replace the battery, the keyboard foil, the battery current sense resistors and a 32kHz crystal early on in its life. Have not been able to get it running on batteries for years.

I have had 3 Fluke scopemeter 105B with the same dark patch as seen in the previous post photo. The problem was that the polarizing filters that sandwich the glass LCD panel were getting old and the adhesive backing was also deteriorating, hence the cracked look.

To fix this I removed the screen and with an indelible marker I marked which side is front and which side is back and a arrow pointing up. This is because the filters are placed in different angle of polarization and it helped me figure out what those angles were.

Now using a single sided razor blade I peeled off the plastic polarizing filter from both sides.( one at a time). I used metholated spirits on the dry adhesive backing and found it came off much easier( made it gummy again so I had to do a lot of scraping with a razor blade). It took about an hour to peel off the filter and clean the screen. Being careful not to crack the screen.

Once I peeled off the old filter I determined that the front filter was 45 degree diagonal polarization and the rear filter was a vertical politicization. Having the marking of "front" and "back" and which way was "up" helped in this. The angles may be different with other model scopemeter.

I purchased filter material from ebay 10 sheets for $20 US . Search for

((10) Nitto Denko NPF Polarizer Film TEG Sheets 44% for LCD Screens 11.4" x 8.6")

On the Scopemeter 105 I cut 103x103 mm filters with correct polarization and then stuck it to the clean glass LCD screen. The filter I purchased had an adhesive backing.

You must also make sure that the zebra strips are perfectly clean as with the edge strip on the LCD glass as any impurities will mean lines across your screen when turned on.

I think this should get you by and good luck. I fixed 3 meters like this so it works.

Wouldn't be the first time Aussies have come to the rescue.