Tek 1502B TDR restoration project and more...

[sorenkir]

Hi,

I have acquired this obviously non properly functioning TDR with the intention of fixing the display.


The LCD screen is a 256x128 pixels panel from the beginning of LCDs with greenish high voltage backlight and heating system for cold temperatures (as the instrument is supposed to be used in the field).

According to the service manual (easily found on the Internet; the one for the 1502C includes the full schematics, while the 1502B's does not), the 'image bitmap' to be displayed is simply written regularly by the main CPU into a static 4096x8 bit RAM located on the display board. The display processor on its side reads this memory to drive the LCD panel.

My plan was to sample the write operations (Address + Data) to this RAM with a micro-controller and then to drive a simple small color LCD display I have (used in black & white). Then, I could find a bigger display to fit into the instrument.

The first part of the project was OK, and I managed to have a functioning display with some very simple glue logic and a few program lines. The TDR was functioning OK, with a clean trace and the ability to measure cable length was OK (picture below is after the incident):


Then, while probing some connector to search for power supply to power the micro-controller, I think (but I'm not sure) that I accidentally shorted two power pins together.
As a result the TDR then powered-up with an error message:


Also the displayed trace was flat with a lot of noise.


The error message is undocumented, but by using the troubleshooting charts and a USB logic analyser, I identified a defective 74ALS161 counter that I replaced (with a 74LS161, not ALS).
This brought back some life to the instrument: the pulse can be seen, but the trace is very noisy.


When the trace is moved vertically, at some positions, the line is less noisy.


The trace is digitized internally by sampling a 'video signal' with an ADC (top right):


The video signal at the input of the ADC seems OK, so I think the ADC is damaged 
The part seems to be available, but quite expensive (around $20), so I want to be sure that my diagnosis is OK...

Now the restoration project has switched to a repair project!
Michel.

[texaspyro]

Tek eventually replaced that first gen LCD with a newer one that did not have the issues of the original.  The newer LCD backlight is whiter than the green one. 

The backlights are an electro-luminescent panel.  I have replaced MANY of them.  I found some EL panels that were longer than the ones Tek used.  You can trim them to fit with scissors (seal the cut edge with kapton tape.

Also,  Tek has (had?) a surplus components store on Ebay that sold off old inventory.   I bought around 100 of the newer LCD assemblies... all they had...  but have used them all.

The 1502B/C sell for stupid prices.  Apparently there are some aviation repair applications that spec their use and are required for compliance reasons.   Hence the demand and people willing to pay big bucks for them.

The 1503B/C uses the same LCD as the 1502...

[sorenkir]

Hi,

I have replace the ADC (ordered from eBay seller hifiic, $15.50 shipped). And now, the trace is clean:


Step 2 (unexpected fix) completed, now, back to Step 1 (screen replacement).

I have orderer and receive from HAOYU Electronics a 480*272 LCD display with SPI interface and FT800 Embedded Video Engine which might physically fit in the instrument.
The screen is smaller, the frame is a bit too large and the resolution will require some software adjustments, but that's all I could find:


It seems that the frame can be cut/resized:


Now, back to software development...

Michel.

[BFX]

It seems to be great project 
I'm waiting for next part

[tecman]

First get the 1502C service manual.  It is essentially the same, except for the battery setup.  Next, the display in the unit may be good.  There is a  pot for LCD Contrast (I believe) that often will cure the "blown out" screen look.  I have a 1502B and 1503B and both had "bad" screens.  A simple adjust of the pot brought them back to perfect condition.

paul

[sorenkir]

Hi Paul,
Thanks for the suggestion. I have the service manual and I can tell you that the contrast pot. does nothing on this permanent discoloration (visible even when the instrument is off, see first picture) !

Next part is to make sure that I can write an efficient enough software to drive this new display, which is quite different from the small one I have used for prototyping.

Michel.

[sorenkir]

Some progress:
By chance, the 1502B video memory is organized as a L1 bitmap, and the FT800 graphic controller of the LCD screen can display directly such a bitmap.
I just had to double each pixel in each direction to map the 256x128 bitmap to a 480x256 area in the FT800 graphics memory and voilà:


Of course, the image is truncated on the right: 480 is not 256x2, so 16 pixels of the original image are lost. But, except on these help screens it is not very annoying (in general, the most right of the trace is not very useful):


Now, I have to work on mechanical integration of the LCD into the instrument.
And a few software optimizations: there are some visual glitches sometimes..
Michel.

[capt bullshot]

Thanks for that idea!

I've replaced the fubar display of a 1502C with an STM 32F469 Discovery board:
http://www.st.com/content/st_com/en/products/evaluation-tools/product-evaluation-tools/mcu-eval-tools/stm32-mcu-eval-tools/stm32-mcu-discovery-kits/32f469idiscovery.html

It has a nice TFT display and was an easy fit into the 1502C front panel (mechanically). The 1502C display / front panel apparently is somewhat different from your 1502B, the LCD controller chip (an OKI MSM6255) is located on the main board and a film cable is used to connect from the main board to the display board (having a lot of column / row drivers on board).

So I've connected the MSM6255 LCD driver outputs to the extension connector of the F469 discovery (only a few signals required, in particular: FRP, CLP, LD0, LD1, LD2, LD3, UD0, UD1, UD2, UD3 and /CS) and wrote some kind of frame grabber SW to capture the display output, upscale and display it.
The CLP frequency is 625kHz, for a modern MCU like the STM32F496 at 180 (or whatever) MHz clock frequency not a big deal, even using pure bit banging techniques.

There are some blank or partially blank frames output by the display controller, causing nasty flickering. I've used the /CS signal to detect the partially blank ones, and a simple SW check to detect completely blank frames, resulting in a clean and stable image.

Didn't take any photos of the mod, and the SW is a quick and dirty hack (re-using a STM provided example).

[coromonadalix]

congrats nice job   

[texaspyro]

If you have the thermal printer for the 1502/1503A/B/C, the drive roller surface like to turn to black goo and run all over the insides and the print head.    I have repaired the rollers by cleaning up all the goo (lots of fun),  putting a couple of pieces of thin heat shrink over the platen rod to build up the diameter a bit, and then using some 3M cold shrink to replace the drive roller surface.   Sorry, I don't remember what part number of cold shrink that I used...

[her50]

texaspyro

I have a print that the platen has melted also.  Do you know the diameter of the platen before it melted?  Why did you use the cold tubing instead of heat stink?

Harold

[texaspyro]

Unfortunately I don't have the measurement anymore.

The cold shrink is rather thick (maybe 1/8 inch) and gives a nice, smooth, level, grippy surface.   Heat shrink tends to be a bit slick.  Also the shrunk surface can be a bit lumpy.

You could try heat shrink.  Add or subtract layers until the paper feeds at the correct rate.  Once that happens you could leave it as-is or measure the "roller" thickness and then re-cover the roller with heat shrink (if needed) and then a cold shrink outer layer.

[her50]

Does anyone have a reliable source for the roll paper for the printer?

Harold

[texaspyro]

Any office supply store should have calculator rolls that can be made to fit.   Also check Ebay for rolls for medical EEG/ECG machines.   Make sure they have the "sensitive" side in the correct orientation (in or out).

[Patrick]

Hello,

Could you share the schematic of the pins on the 1502 you connect to for reading the RAM R/W events, please?

[pvfb]

I found a company in the UK that manufactures and re-manufactures rubber rollers, they were able to identify and strip the rubber off a few YT-1 rollers for us and then replace them... my notes say 14.6mm dia and 40mm long...

I've used the heatshrink technique on the tape drives in HP4951A/Bs, had to try a few different type to find some with a suitable surface, not a "permanent" solution but good enough to read a few tapes and IIRC print or transfer the data we needed to the later diskette based version.

[Roman oh]

This being an “old” thread,  I wasn’t sure whether to breathe more life into it or start another on the same topic.
I think the former is more useful for others that follow....

I have a 1502B which appears to work in all except the display, which is bright, good contrast, but all the display data appears to be compressed into a few vertical lines (divided into a top and bottom half of screen) that skip from place to place across the screen. As if the “horizontal scan” was a bunch of slowly changing random positions instead of a steady ramp, with all the vertical data squashed into a single line.

Now, here’s the interesting thing. When I apply pressure, twist, etc to the two boards that constitute the display itself, a good image will occasionally flash up. This suggests to me that the underlying electronics is operating ok, and it’s a matter of dirty connections. I, like many of us, I’m sure, have had instances of flaky elastomeric connectors which respond well to a good alcohol clean. So that’s what I propose to try.

There is a small conventional elastomeric connector at one end, sandwiched between the two circuit cards that make up the display. This should be fairly straightforward to access and clean. The wide printed flexible at the other end is bolted down with a clamp arrangement on the top pc, and appears to disappear at the other end into a potted assembly on the display panel itself. This may be a bit more challenging.

I note in an old post that at least one person (texaspyro)  has at least replaced the backlight, which necessarily involves disassembly of the display, despite the caution in the Tek manual cautioning against this.  Others no doubt have also been down that path. So I’m hoping that someone out there with experience in this can give me any pointers as to what to do, and what to NOT do?

I would be very grateful for any advice.

EDIT: I did the easy part and cleaned the small elastomeric connector on one end (deOxit Gold on the fingers and isopropyl on the rubber) and up it came. Now have a functioning unit with a crisp display!

Roman

[Larry818]

Hi, I know this is an old topic, but it helped me out, so I thought I'd add this here where others might find it.

I have a 1502c and 1503c.  The 1503c was dim, but you could use it with a flashlight, the 1502c had that "burned" look pictured at the top and was unusable.  I replaced both luminescent panels with cheap ebay parts, there's a guy selling "A6" panels with 12vdc power supplies.  I just took out the original EL panel and heater (I won't be using these in the snow), and ignoring the original power source from the undocumented display module, I just connected the power supply the ebay panel came with to the TDR's 15vdc bus after the power switch.  The panel is quite bright at 15 volts.  You'll need to retain or replace the 20k thermistor or the screen will flicker badly.  One of mine was corroded and just fell apart.  The LCDs on both were fine.

The 1503c's heater and el panel just came out with some conductor snipping.  The 1502c's el panel was "bonded" to the back of the LCD, but peeled off with minor difficulty and didn't leave any residue.  The display module was the same part number for both, so I don't think it was a 1502/1503 thing.

The A6 el panel can be cut and I cut them with the existing cord oriented towards the gap at the end of the assembly.  It's just big enough to cover the whole display.  A bit of kapton tape and it's good.  I did use the original heater with a notch cut out to clear the cable attachment as a spacer to hold the new el panel up against the lcd.  I just used VHB tape to hold the new power supply in the case.

I can take pictures if anyone is interested.

[_Wim_]

I can take pictures if anyone is interested.

Interested!

[rodd]

I am interested also.
Would you please post the pictures?
Thank you