Dallas DS1486 and DS1250 Replacement designs

[Tantratron]

The 784D does not boot and the numeric display oscillates between "6" and "8".   The scope behaves the same way if the DS1486 socket is empty.

Wether on C serie or D serie, during the boot sequence check (15 steps from 1 to E), there is specific step number 11 (B) which makes a local verification of DS1486 (will write sequence of 1 2 3 4 5 6 7 8 then read if same values). If there is no DS1486 installed or the DS1486 has an internal failure, it will generate .B error by choosing DIP switch check mode value 11 (decimal). Otherwise it will go alternate 6 and 8 as you seem to report but I recommend that you specifically DIP switch choose step number 11 (decimal) or B (hexa) which would raise the failure .B where DOT means failure. If the test pass then you will have only B flashing.

On a side note, if there is no DS1250Y installed then no problem because these TDSxxx-C/D can actually run only with DS1486.

[Tantratron]

That sounds like the DS1486 is in test mode. This causes one of the pins to oscillate at a frequency, 512hz IIRC to allow you to measure the frequency of the internal crystal. This will of course result in unpredictable behavior of whatever device it is installed in. I ran into this some time back when I was trying to program a similar Dallas RTC in a regular EPROM/flash programmer. Read the datasheet and it explains how to enable and disable test mode, I ended up using an arduino mega and wrote a little program to do it.

From the DS1384 datasheet: "SQW - Square Wave (output): This pin can be programmed to output a 1024 Hz square wave signal.
When the signal is turned off, the pin is high Z."

With my own replica, I do have 1KHz square wave on pin-32 but do you confirm if this is the case the device cannot work ?

Albert

[Chalcogenide]

If anyone is interested, I came up with a design that replaces a DS1245 or DS1250 with a non volatile MRAM. It's more expensive than an SRAM + battery backup solution, but it should be the ultimate "set and forget" replacement since no battery for backup is needed. I only tested it with a TL866 and on a Keithley 2001, but if anyone is interested let me know and I will share the BOM and gerbers. I also have a couple of partially populated boards if someone wants to give a try.

[Tantratron]

If anyone is interested, I came up with a design that replaces a DS1245 or DS1250 with a non volatile MRAM. It's more expensive than an SRAM + battery backup solution, but it should be the ultimate "set and forget" replacement since no battery for backup is needed. I only tested it with a TL866 and on a Keithley 2001, but if anyone is interested let me know and I will share the BOM and gerbers. I also have a couple of partially populated boards if someone wants to give a try.

Hello,

A few questions as an owner and repairing few TDSxxx/C/D oscilloscope.

What are the memory speed access of your prototypes -70 -100 or -120 (ns)

Any hope to have a design based on MRAM technology with internal RTC to replace the DS1486

What pins form factor does you prototype uses, flat like legacy DALLAS or rounded

Technically it weems we can replace a failed DS1486 by either a DS1245Y or DS1250Y where of course we loose the RTC feature. I have tried using one legacy DS1250Y-100 and it worked.

Note that the DS1250Y in the TDSxxx/C/D main board does not need to be installed, the oscilloscope only needs the DS1486-120 installed to properly boot and run (check this other post and information https://www.eevblog.com/forum/testgear/i-noticed-nvram-replacements-for-the-tds784d-on-ebay/msg3610951/#msg3610951). Of course you'd loose the save memory option but we can live without it or download waveforms via GPIB-USB.

Since you're based in Italy, I'm based in France then I could try on TDSxxx/C/D if you do not have one.

Grazie, Albert

[Chalcogenide]

The MRAM I chose is the Everspin MR0A08BYS35 (DS1245) or MR2A08AYS35 (DS1250) which are 35 ns parts, plus a voltage translator TI TXS0108E which I guess shouldn't add more than 20 ns at the most. The spare boards I have were assembled at JLCPCB and the only things missing are the MRAM (which I only bought one to fit on the Keithley 2001) and the pin headers (I have used round pins on mine but I guess you could solder in whatever kind).
I have no plan on tackling the DS1486 because at that point using a battery is unavoidable so more conventional SRAM based approaches are probably the way to go.
PM me for more info.

[Tantratron]

plus a voltage translator TI TXS0108E which I guess shouldn't add more than 20 ns at the most.

So I guess you generate via another chip the 3.3V from the 5V required to properly voltage match interface ?
The real problem with these legacy DALLAS then Maxim DSxxxx NVRAMs and RTC's, they were all designed for 5V bus and power supplies. Now to make a replication is complex because most components are 3.3V only so you need to add other chips.

The memory access time could be an issue as well where the legacy DSxxxx has battery controller with Vdd controler voting who supplies energy to the memory. The power up of old main boards and PSUs (i.e. TDSxxx oscilloscope) can be tricky and not meet the transients of new replicas then boot fail.

[Chalcogenide]

Yes, there's an LDO to supply the 3.3V to the various ICs on the board. I am hesitant to provide the design files before verifying that it works in at least some other instrument... in the 2001 the Dallas chip is only used as a buffer to store measurements and it's neither accessed particularly fast nor strictly required for boot, so I can only attest that my design behaves like a normal Dallas chip when used in these conditions. The MRAM has an internal "UVLO" of sorts to deal with power-up and power-down, but I have no idea whether it's enough; the MRAM has a startup time of 2 ms after VDD reaches 3.0 V (so given the dropout of the LDO, when the 5V rail reaches something in the order of 3.5 V). It is probably fast enough but I am not sure.

[Tantratron]

Hello Chalcogenide,

Well technically and as mentioned before, if we focus on TDSxxx/C/D oscilloscope and decide to drop the RTC feature, it works to install inside the DS1286 slot either DS1250Y or DS1650Y. The way the DS1486 is wired does not actually use the 512K, only the 128K. So it seems replacing a DS1486-120 by either a DS1250Y DS1650Y DS1245Y or DS1250Y works.

Note that i've only tried DS1250Y and DS1650Y so cannot confirm about the 2 others but some members seem to have confirmed DS1245 or DS1250 do work as a replacement for DS1486.

However there is a strange thing which I still do not fully understand hence my caution. Installing the DS1250Y clone designed by cuebus inside the DS1486 does not work. It should since after all this a memory with battery backup but maybe the SRAM used is too fast, too slow or too whatever.

With the TDSxxx/C/D the processor board is 68040, the data and address bus is clocked at 2.5MHz

Albert

[madao]

I confirm: DS1245 works in my TDS784D and  TDS754C
Clock  doesn't running..  It is in most case not important.

regards
mattt

[Tantratron]

I confirm: DS1245 works in my TDS784D and  TDS754C
Clock  doesn't running..  It is in most case not important.

Hello Matt, what specific DS1245 model did you try (DS1245Y or DS1245AB voltage rating.... -70 -85 -100 or -120 memory speed) ?
Thanks, Albert

[whatboy]

At last I manage to make 3 sets for my 3 oscilloscopes TDS 684 A, B and C, but only could program one set, already installed and working, the other two sets gives me errors with the DS1250, when programming in my LAB TOOL-48UXP programmer, at first I didn't knew why, but after testing the chip CUEBUS made and kindly shared, I noticed that the DS1312 is not switching the voltaje from the battery, the chip is dead, if you put iton the programmer it will supply 5V to the memory, but once the 5V are gone, no BAT voltaje is switched to the memory, so all info written is gone. And yes, if you read it after, it will give you all 0's, I do not know if I kill the DS1312 or if it came that way, did not put it wrong, not the battery, have to buy more and see. I still am happy to manage to set up at least one osci!

[Teichhermelin]

Could the board be used for other DS types?

[whatboy]

You can check it out with the datasheet of each chip!

[cuebus]

Check the orientation of the DS1312S-2. Note that it faces pointing down with pin 1 facing the middle of the pcb. The SRAM and controller should be facing each other. In hindsight this might have been bad design but it's been about 2 years since I designed this...
Did you get your DS1312S-2 from a reputable supplier? It's still a current part, so there should be no problems getting genuine working parts from Mouser, digikey, etc.

[whatboy]

There is a line on pin one, so I did put it right!, and no, I bought them from ebay, the seller was from USA thought, next time I will buy from mouser!

[Tantratron]

At last I manage to make 3 sets for my 3 oscilloscopes TDS 684 A, B and C, but only could program one set, already installed and working, the other two sets gives me errors with the DS1250, when programming in my LAB TOOL-48UXP programmer, at first I didn't knew why, but after testing the chip CUEBUS made and kindly shared, I noticed that the DS1312 is not switching the voltaje from the battery, the chip is dead, if you put iton the programmer it will supply 5V to the memory, but once the 5V are gone, no BAT voltaje is switched to the memory, so all info written is gone. And yes, if you read it after, it will give you all 0's, I do not know if I kill the DS1312 or if it came that way, did not put it wrong, not the battery, have to buy more and see. I still am happy to manage to set up at least one osci!

Thank you where it is interesting that you bring the issue on DS1250 replica or emulation.

On my side as reported in this other thread https://www.eevblog.com/forum/testgear/i-noticed-nvram-replacements-for-the-tds784d-on-ebay/msg3613318/#msg3613318 among 3 complete kits (DS1486 and DS1250), only one set is still working now since last june.

I really have no idea on what root cause or multiple root causes but wether DS1486 or DS1250 now we have enough members which has good soldering skills to not confirm 100% replica effect. In my case, all the programming was done using the tekfwtool application through the GPIB-USB interface from National Instruments. Others have used an external programmer, I do not have one so far but unless I'm wrong, many times I've repaired different TDSxxx/C/D the reprogrammed their orginal Dallas DS1486 and DS1250Y, it worked always via GPIB.

Albert

[m3vuv]

has anyone tried doing this for a DS1230Y-100 ?

[whatboy]

UPDATE!!!

Hello, I have 3 1GHz oscilloscopes, TDS684 A/B/C the A has a broke screen (the flyback is bad) and 2 hybrid modules bad, use'em to rapair the other two ones, I got the gerbers from CUEBUS and got the parts, although I didn't thought it would work, made 3 sets, but 1 I fsk it up and burn the RTC chip, but managed to make 2 working sets and put it on the 2 working oscis... even thought I bought the wrong parts, wrong size of memory bad quality crystal and batteries!!! I was about to quit, and they work, all 3 of them, but like I say burnt one set, but since I do not have the money to fix the A one which is the one with the screen and bad modules, I only use one (C) and wanted to sell the B one but can find a buyer, it was sitting all alone and just test it today and it is working, was concern a little because it gave me NULL readings on all 4 channels, but deleted the memory and it work! boots up and everything!

So, I guess mine are working 100%

[whatboy]

you can see I had to bend the pins inwards to fit on the board!!! I really suck!!! 

[max-bit]

Does anyone have a diagram of this DS1486 adapter?
Including edit files (not gerber) and/or schematic
I would like to correct it to the DS1384 version in a 44 pin housing and use even better SRAM memories (newer and other types)

The undoubted problem here is the DS1384 chip, which is difficult to access (price!)
Couldn't some other IC DS be used?

[max-bit]

DS1558 ?

https://datasheets.stg-maximintegrated.com/en/ds/DS1558-DS1558Y.pdf

[Tantratron]

DS1558 ?

https://datasheets.stg-maximintegrated.com/en/ds/DS1558-DS1558Y.pdf

This chip will not work as many others, check the first post https://www.eevblog.com/forum/repair/dallas-ds1486-and-ds1250-replacement-designs/msg2954864/#msg2954864 mentioning DS1558. One reason being the address space of registers (i.e. data, time...), the other is that DS1384 has 64 bytes protected which are used during the boot.

Albert

[TERRA Operative]

My turn to be on the hunt for a number of replacement DS1486 chips....

I have successfully made a drop-in replacement for the DS1225Y using FM16W08 FRAM chips on a 1-to-1 pinout SOIC28 to DIP28 adapter PCB, and my next trick will be trying out something to replace DS1250Y chips, maybe flash RAM. Gotta look into is a bit deeper.

The tricky bit with the DS1486 is of course that DS1384 chip. Has anyone found a suitable substitution?

[james_s]

I made an adapter to use a FRAM, I forget which one yet but it may be the same part you used. Unfortunately it hasn't quite worked in anything I've tried it in yet due to the way those parts require CE (I think?) to go high between writes or something like that. They are hardware compatible with NVRAM but not always a drop in replacement for existing equipment unless you can modify the firmware.

[max-bit]

DS1384G such as from the adapter are still available on ebay.
I just ordered 5 pieces so I will make 5 adapters
3-4 pieces will be for sale.
Probably in about a month.
The price will be "at cost"

I still have a question if anyone has a schematic for the DS1486 replacement.
Yes I can redraw from PCB but...