TDS3000B TDS3054B strange type rotary encoder repair/replacement

[james_s]

The number of power on cycles is stored elsewhere, the hours is the only thing that I found no other way to set. There may be a way to write it via ethernet console, I don't know. Programming the part manually is pretty trivial since there are only 5 bytes that matter, it could be done in a breadboard with a bunch of jumper wires.

[Jwalling]

by the way how come I have not lost those numbers if the chip's battery cannot keep the time/date?

Battery has enough voltage left to power SRAM, but not enough to run the oscillator.

[analogRF]

by the way how come I have not lost those numbers if the chip's battery cannot keep the time/date?

Battery has enough voltage left to power SRAM, but not enough to run the oscillator.

yes, makes perfect sense. Why didnt i think of that

[analogRF]

if there is a way to change the model number and smapling rate through LAN and GPIB...i think there must be a way to set the hours and No of power cycles, too.

Does anybody know?

[james_s]

Battery has enough voltage left to power SRAM, but not enough to run the oscillator.

That was exactly what mine was doing. When I cut the battery out it measured 0.6V, pretty remarkable really that it was enough to keep the RAM going.

I attached a new battery and it worked for a while and then failed. Later I discovered I had tugged a little too hard and cracked the solder joint at the PCB but I had already gone down the path of designing an adapter to install the DS1744 so I took that route in order to verify my design.

If you want to try replacing the battery while preserving the RAM contents, wire up a socket to a 3.3V power supply and plug the chip into that while you carve into it and cut the battery out. With care it's not too hard, I've "fixed" a number of other Dallas chips without issue.

[analogRF]

how do you take that battery out? to what points do you solder the new one? I mean, do I necessarily need a spot welder or something? because i dont have one.

[james_s]

I stuck the chip into a cheap IC socket to protect the pins, then I used a razor knife to cut the plastic potting cup off the side with the battery (confirmed with a magnet) and then started digging into the epoxy with a screwdriver. A heat gun or hot air tool is very helpful to soften the epoxy, then I pried up the corner of the nickel strip and used pliers to peel it off the battery. After that I dug around the battery a bit more until I could pry the whole battery out enough to cut the strip welded to the other side. Then I simply soldered wires to the ends of the strips that were formerly welded to the battery and soldered those to a CR2032 holder which can be glued right on top of the chip or mounted remotely.

Just be careful not to yank too hard on the strips as you're prying the battery out of there, that's how I broke the solder joint on mine. At some point I'll dig in further and repair that but my replacement adapter is working perfectly well so I plan to just leave that in the scope.

[analogRF]

My TL866 programmer can read and write DS1220 and it has exactly the same pin out as the DS1742. It seems to be the same NVRAM as ds1220 plus the clock which only reserves the top (or bottom?) 8 bytes as clock registers.

so I am thinking of buying a DS1742 and then reading mine and writing it into the new chip as a DS1220. what do you think the problem would be?

people have done this for DS1486 firmware IC on TDS700 series by reading/writing it as a DS1250 (offset by 8 bytes for the clock registers maybe)

[james_s]

My TL866 programmer can read and write DS1220 and it has exactly the same pin out as the DS1742. It seems to be the same NVRAM as ds1220 plus the clock which only reserves the top (or bottom?) 8 bytes as clock registers.

so I am thinking of buying a DS1742 and then reading mine and writing it into the new chip as a DS1220. what do you think the problem would be?

people have done this for DS1486 firmware IC on TDS700 series by reading/writing it as a DS1250 (offset by 8 bytes for the clock registers maybe)

Been there, tried that, it doesn't work.

I wasn't entirely sure why until I lashed up a rig to manually twiddle the address and data pins with switches and monitor the outputs and then studied the datasheet. You have to set a bit in a control register then write the desired data in the timekeeping and configuration areas and then clear the bit you set to enable writing. The pinout is the same as a plain NVRAM but it requires a more complex procedure to work with it. Once the battery has gone dead it defaults to frequency test mode where the LSB data line is pulsing at 512Hz.

[analogRF]

my theory was also flawed because DS1742W is a 3.3V chip and DS1220 (and DS1742) are 5V chips so in any case R/W with TL866 would not have been possible

but there is GQ-4x programmer which can be had with a  pretty good price and it has listed 1742 and 1742W specifically in the list of supported chips.
Does anybody have any experience with this?

Also Xeltek SuperPRO 610P lists 1742/1742W as supported but that is an expensive gear not worth the money...

[james_s]

I was able to read my DS1742W in a TL866, pretty sure it just hit it with 5V which eventually would likely harm it but in the short term it didn't seem to do anything. I found an interesting behavior though, the first time I read it I got garbage, but if I read it twice I was able to successfully dump it. Writing however did not work, at least not in the clock and config areas.

I tried a borrowed GQ-4X also and that didn't work for me either but I've had some general problems with that unit and suspect it may have a fault.

In the case of the DS1742 the easiest thing I found was to just twiddle it manually, I used a FPGA board but an arduino or even just a breadboard with some jumpers or dip switches would work fine. I recently picked up an Arduino mega clone with the thought of making an adapter to read and write these NVRAM/RTC chips.

[xymox]

This was a very useful thread.. I was able to easily use a razor blade and cut loose some encoders and clean everything which restored the knobs to perfect functionality. I used a fine tip on my weller to spread back out the plastic to hold them in place. Easy peasy..

But then... I was able to run it for a bit and a new problem appeared. After warm up, the whole front panel board seems to get wacky. LEDs flicker, it fails self test. This is a seperate issue from the encoder knobs being dirty. I had not noticed it as I never had it on long. It takes like 30 mins. Then it starts failing lightly then will go into a flashing EVERY LED mode and no longer accept any knob or button.

As its a thermal issue it could be a number of things. Just wondering if anyone has seen this before. I would hate for it to be the front panel IC itself.. I am thinking maybe a solder joint. I will get it back out and look it over better.

YEA those Dallas semiconductor things are just terrible. I run into them on the Tek Waveform monitors and its fatal as the unit will no longer come up unless you use special software, which, you cant find anymore. Even the Tek museum does not have the software. So all the NTSC waveform monitors will all die..