Topics

TDS784C


Steve Hendrix
 

I didn't research this as well as I might have, because I really didn't expect to get it. I made a fairly low offer on eBay for a TDS784C, fully expecting it to be rejected. Pleasant surprise when it was accepted. The unit arrived last Wednesday and I've been so swamped with work, family, and volunteer work that I just finally got it out of the box yesterday. Very nice unit, exactly as advertised, and appears to be fully working. It's physically quite a bit bigger than I anticipated, so I need to do some long-overdue housecleaning in the lab to make room for it.

It came as stated with no cables or accessories. Power cords I have in abundance. I have some 100 MHz scope probes available, but of course those won't let me use the full 1 GHz bandwidth. I plan to use another of my KISS-488 interfaces for screenshots and PC control, so that's covered. I have the crippleware version of WaveStar that came with my TDS220 so will try that with this unit. Can anyone here point me to a good reliable source of 1 GHz rated probes (that won't cost more than I paid for the scope!).

Any other pointers, pitfalls, etc. from current or past users of this scope would be very welcome. I've seen it occasionally mentioned here, but maybe the relative lack of chatter suggests it's one of those instruments that just plain works.

Thanks for reading!

Steve Hendrix


JJ
 

Hi Steve,

You need to be concerned with the NVRAM chips - DS1486 clock chip and the DS1650/DS1250. They have internal batteries that hold the calibration constants and options info. The batteries expiration date ended many years ago. So, if those chips haven't been changed, you need to attend to it. Many articles on backing those chips up using the GPIB interface, the internal disk drive, and other ways on this site and forum.tek.com. But, hopefully, someone swapped out the chips already.
Great scope - good luck, with it! I have a TDS754C.
Best,
John Justin


 

Any other pointers, pitfalls, etc. from current or past users of this scope would be very welcome. I've seen it occasionally mentioned here, but maybe the relative lack of chatter suggests it's one of those instruments that just plain works.
Hi Steve,

Make sure that it passes SPC from the utility menu - at least 10 times with no failures.
Also run the full self test from the same utility menu.

Jay


Steve Hendrix
 

At 2020-01-28 09:00 AM, Jay Walling via Groups.Io wrote:


Make sure that it passes SPC from the utility menu - at least 10 times
with no failures.
Also run the full self test from the same utility menu.
I'll probably think of it right after I hit "send", but what does SPC stand for in this context?

Steve Hendrix


Siggi
 

On Tue, Jan 28, 2020 at 8:39 AM JJ <jajustin@...> wrote:

You need to be concerned with the NVRAM chips - DS1486 clock chip and the
DS1650/DS1250. They have internal batteries that hold the calibration
constants and options info.
It's my understanding that the calibration constants are stored in EEPROM
on the acquisition board. There's no reason to worry about the options, as
it's easy enough to restore those with GPIB or other means.


Stephen Hanselman
 

Speaking only from experience with TDS 540x, yes the calibration is kept on the acquisition PCA. I think, but have not taken the time to prove, that the primary use of the NVRAM is to store trace data. Everyone of the NVRAMs I have read seem to have a pile of "empty" space in them along with a bit of stuff relating to options, date, time, and the like.

Just as easily though I could be FoS.

Steve

-----Original Message-----
From: TekScopes@groups.io <TekScopes@groups.io> On Behalf Of Siggi
Sent: Tuesday, January 28, 2020 7:05 AM
To: TekScopes@groups.io
Subject: Re: [TekScopes] TDS784C

On Tue, Jan 28, 2020 at 8:39 AM JJ <jajustin@...> wrote:

You need to be concerned with the NVRAM chips - DS1486 clock chip and
the DS1650/DS1250. They have internal batteries that hold the
calibration constants and options info.
It's my understanding that the calibration constants are stored in EEPROM on the acquisition board. There's no reason to worry about the options, as it's easy enough to restore those with GPIB or other means.


victor.silva
 

Steve,

I don't know what you consider a low ball offer but I usually make brutally low offers on liquidator's equipment because usually it's so overpriced.
I usually start at a max of 10%~20% of the asking price and go from there, unless of course I see a stellar deal that I'm sure will be snapped up very quickly.

For example, I've seen liquidators list 2465Bs for $1000 and I'll offer $125 and they have accepted.
Many liquidators have no idea of the value and they simply go on ebay and list their stuff for the highest price they can see.

Good luck with your TDS784C. You may eventually need to convert to LCD because those shutter type LCD/CRTs will eventually fail.

--Victor


Siggi
 

On Tue, Jan 28, 2020 at 9:14 AM Steve Hendrix <SteveHx@...>
wrote:

I'll probably think of it right after I hit "send", but what does SPC
stand for in this context?
SPC stands for "Signal Path Compensation". This injects a calibration
signal into the pre-amplifiers for each channel, and calibrates away any DC
offset in the signal path (https://www.tek.com/support/faqs/what-spc).


JJ
 

Steve,
Here is the link to the discussion on using Java SW to copy your NVRAM data using the TDS784C internal floppy: https://www.eevblog.com/forum/testgear/tektronix-tds500600700-nvram-floppy-dump-tool/
There is seller on eBay that sells those two chips (or the equivalent thereof) mounted with external coin cell batteries on top of the chip. He has gotten high ratings from other TDS scope owners on their quality. He states that he will program the chips with your backup data if you want. You still need to remove the chips yourself though - so you will need a good quality solder sucker like the Hakko. There are also YouTube videos showing how some pros removed the chips - you should look through them. Some TDS owners have said that just heating the chip pins caused the chips' batteries to drain - some claimed that apparently the chips' batteries were right on the edge.of failing.

Best,
John Justin


Ragnar S
 

Hi all,

There is some calibration data in the EEPROMs on the acquisition board, but also in the NVRAMs on the CPU board.

I have been told that you should never swap acquisition boards between scopes without also moving the NVRAM contents.
Also, there are indications that there is calibration data that not even the field service calibration data will update, that was written at the factory.

You are right that the NVRAMs are also used for storing waveforms, settings and other stuff.

If you care about your scope I would _absolutely_ recommend backing up your NVRAMs and your EEPROMs.

There are some DOS programs build for National Instruments GPIB drivers, mainly tekfwtool and getcaldata, as well as the floppy based tools that John pointed out, tdsNvramFloppyTool, that also has a Java program that can verify the checksums of your dumps.

Both ways of getting your EEPROMs and the NVRAMs are a bit fragile, so I would recommend using two separate methods, or at least do it twice, and check that there is no difference between the results.
Sadly there always is a difference in the first bytes of the NVRAM dumps, because that is where there real time clock is mapped (in the DS1486).

Since I had no National Instruments GPIB interface around (but an Agilent USB one), and no DOS machine (but linux and unix machines), I did the little work it took to make them compile clean on Linux, on a Raspberry Pi in my case, fixed some simple makefiles and made a collection of the tools. Another guy wanted to run them on a Mac, which was trivial to fix as well, so I added that too.

You can find my collection here:
https://github.com/ragges/tektools

Best regards,

Ragnar

On 28 Jan 2020, at 18:23, Stephen Hanselman <kc4sw.io@...> wrote:

Speaking only from experience with TDS 540x, yes the calibration is kept on the acquisition PCA. I think, but have not taken the time to prove, that the primary use of the NVRAM is to store trace data. Everyone of the NVRAMs I have read seem to have a pile of "empty" space in them along with a bit of stuff relating to options, date, time, and the like.

Just as easily though I could be FoS.

Steve

-----Original Message-----
From: TekScopes@groups.io <TekScopes@groups.io> On Behalf Of Siggi
Sent: Tuesday, January 28, 2020 7:05 AM
To: TekScopes@groups.io
Subject: Re: [TekScopes] TDS784C

On Tue, Jan 28, 2020 at 8:39 AM JJ <jajustin@...> wrote:

You need to be concerned with the NVRAM chips - DS1486 clock chip and
the DS1650/DS1250. They have internal batteries that hold the
calibration constants and options info.
It's my understanding that the calibration constants are stored in EEPROM on the acquisition board. There's no reason to worry about the options, as it's easy enough to restore those with GPIB or other means.






Steve Hendrix
 

At 2020-01-28 10:01 PM, Ragnar S wrote:


There are some DOS programs build for National Instruments GPIB drivers,
mainly tekfwtool and getcaldata, as well as the floppy based tools that
John pointed out, tdsNvramFloppyTool, that also has a Java program that
can verify the checksums of your dumps.
Thank you very much for the uploads of the tools. I've downloaded those but will be a few days before I'll get a chance to look at them. I have at least half a day of cleaning out in order to make space to set up the TDS784C properly, and before that I'm up to my ears in alligators with work and other things. When I do get to it, I'll need to do some adapting. I like the idea of Linux, but have never had time to take that plunge as my work keeps me mostly under Windoze, despite my roots being much closer to Linux (HexDOS, PCDos, CPM, and Unix). And I'll need to adapt the I/O to go via Telnet to my KISS-488, which is the only IEEE-488 interface I kept.

Steve Hendrix


Ragnar S
 

On 29 Jan 2020, at 12:15, Steve Hendrix <SteveHx@...> wrote:

At 2020-01-28 10:01 PM, Ragnar S wrote:


There are some DOS programs build for National Instruments GPIB drivers,
mainly tekfwtool and getcaldata, as well as the floppy based tools that
John pointed out, tdsNvramFloppyTool, that also has a Java program that
can verify the checksums of your dumps.
Thank you very much for the uploads of the tools. I've downloaded those but will be a few days before I'll get a chance to look at them. I have at least half a day of cleaning out in order to make space to set up the TDS784C properly, and before that I'm up to my ears in alligators with work and other things. When I do get to it, I'll need to do some adapting. I like the idea of Linux, but have never had time to take that plunge as my work keeps me mostly under Windoze, despite my roots being much closer to Linux (HexDOS, PCDos, CPM, and Unix). And I'll need to adapt the I/O to go via Telnet to my KISS-488, which is the only IEEE-488 interface I kept.
I am only happy if others have use for it too!

There are only a few calls that needs to be implemented, like ibwrt(), ibrd(), ibcntl(), ibdev(), ibclr(), ibonl(), and the global status variables like “ibsta”, “iberr", and the constants to test them against, linux-gpib has the constants.

If you are a bit more lazy, or don’t have the time, you could pick up e.g. an Agilent 82357B USB-GPIB interface on eBay for around USD 70, that is what I did. They are marketed as “new”, but they obviously have some chips soldered in afterwards, there is flux left on those, and they are not always that good positioned. Also, they are supposed to have shielding paint on the inside of the plastics, but these don't. But many seems to be happy with these interfaces, and it has worked for me. I would recommend to inspect it first though - mine had a short between two pins on the Xilinx chip, but both pins were "not connected", so it wasn’t a problem and even a self test (if there is one) should most likely not have noticed.

Best regards,

Ragnar


amirb
 

Ragnar,

So, you mean the tools you posted on github here do work with Agilent 82357B? like the tekfwtool, etc...
Because everything I have seen about these tools whether here or on eevblog point to NI GPIB-USB
and people say they dont work with 82357B.
NI GPIB is too expensive but my cheap 82357B clone works perfectly for me.

On Tue, Jan 28, 2020 at 10:01 PM, Ragnar S wrote:



Hi all,

There is some calibration data in the EEPROMs on the acquisition board, but
also in the NVRAMs on the CPU board.

I have been told that you should never swap acquisition boards between scopes
without also moving the NVRAM contents.
Also, there are indications that there is calibration data that not even the
field service calibration data will update, that was written at the factory.

You are right that the NVRAMs are also used for storing waveforms, settings
and other stuff.

If you care about your scope I would _absolutely_ recommend backing up your
NVRAMs and your EEPROMs.

There are some DOS programs build for National Instruments GPIB drivers,
mainly tekfwtool and getcaldata, as well as the floppy based tools that John
pointed out, tdsNvramFloppyTool, that also has a Java program that can verify
the checksums of your dumps.

Both ways of getting your EEPROMs and the NVRAMs are a bit fragile, so I would
recommend using two separate methods, or at least do it twice, and check that
there is no difference between the results.
Sadly there always is a difference in the first bytes of the NVRAM dumps,
because that is where there real time clock is mapped (in the DS1486).

Since I had no National Instruments GPIB interface around (but an Agilent USB
one), and no DOS machine (but linux and unix machines), I did the little work
it took to make them compile clean on Linux, on a Raspberry Pi in my case,
fixed some simple makefiles and made a collection of the tools. Another guy
wanted to run them on a Mac, which was trivial to fix as well, so I added that
too.

You can find my collection here:
https://github.com/ragges/tektools

Best regards,

Ragnar

On 28 Jan 2020, at 18:23, Stephen Hanselman <kc4sw.io@...> wrote:

Speaking only from experience with TDS 540x, yes the calibration is kept on
the acquisition PCA. I think, but have not taken the time to prove, that the
primary use of the NVRAM is to store trace data. Everyone of the NVRAMs I
have read seem to have a pile of "empty" space in them along with a bit of
stuff relating to options, date, time, and the like.

Just as easily though I could be FoS.

Steve

-----Original Message-----
From: TekScopes@groups.io <TekScopes@groups.io> On Behalf Of Siggi
Sent: Tuesday, January 28, 2020 7:05 AM
To: TekScopes@groups.io
Subject: Re: [TekScopes] TDS784C

On Tue, Jan 28, 2020 at 8:39 AM JJ <jajustin@...> wrote:

You need to be concerned with the NVRAM chips - DS1486 clock chip and
the DS1650/DS1250. They have internal batteries that hold the
calibration constants and options info.
It's my understanding that the calibration constants are stored in EEPROM on
the acquisition board. There's no reason to worry about the options, as it's
easy enough to restore those with GPIB or other means.






Ragnar S
 

Hi Amir,

You can build them for any library with the old NI 488 API, like the NI drivers, linux-gpib, and probably several others.
It is the library that must have support for the interface you are using, the programs are just using a very simple API to the library.

I happened to use those programs with:
- A Raspberry Pi
- The linux distribution Raspbian
- A 82357B GPIB-USB interface
- The linux GPIB library “linux-glib, which has support for many interfaces including 82357B.

Another guy use them with:
- A Mac with MacOS
- National Instrument’s libraries/drivers
- A National Instrument GPIB-USB interface

Many other combinations are possible of course.

The compiled binaries for DOS/Windows that you can find on eevblog are almost always built for the National Instruments library, and as far as I know the NI Library only support NI interfaces, I guess that is why the told you that.

Regards,

Ragnar

On 30 Jan 2020, at 14:39, amirb <@amirb> wrote:

Ragnar,

So, you mean the tools you posted on github here do work with Agilent 82357B? like the tekfwtool, etc...
Because everything I have seen about these tools whether here or on eevblog point to NI GPIB-USB
and people say they dont work with 82357B.
NI GPIB is too expensive but my cheap 82357B clone works perfectly for me.

On Tue, Jan 28, 2020 at 10:01 PM, Ragnar S wrote:



Hi all,

There is some calibration data in the EEPROMs on the acquisition board, but
also in the NVRAMs on the CPU board.

I have been told that you should never swap acquisition boards between scopes
without also moving the NVRAM contents.
Also, there are indications that there is calibration data that not even the
field service calibration data will update, that was written at the factory.

You are right that the NVRAMs are also used for storing waveforms, settings
and other stuff.

If you care about your scope I would _absolutely_ recommend backing up your
NVRAMs and your EEPROMs.

There are some DOS programs build for National Instruments GPIB drivers,
mainly tekfwtool and getcaldata, as well as the floppy based tools that John
pointed out, tdsNvramFloppyTool, that also has a Java program that can verify
the checksums of your dumps.

Both ways of getting your EEPROMs and the NVRAMs are a bit fragile, so I would
recommend using two separate methods, or at least do it twice, and check that
there is no difference between the results.
Sadly there always is a difference in the first bytes of the NVRAM dumps,
because that is where there real time clock is mapped (in the DS1486).

Since I had no National Instruments GPIB interface around (but an Agilent USB
one), and no DOS machine (but linux and unix machines), I did the little work
it took to make them compile clean on Linux, on a Raspberry Pi in my case,
fixed some simple makefiles and made a collection of the tools. Another guy
wanted to run them on a Mac, which was trivial to fix as well, so I added that
too.

You can find my collection here:
https://github.com/ragges/tektools

Best regards,

Ragnar

On 28 Jan 2020, at 18:23, Stephen Hanselman <kc4sw.io@...> wrote:

Speaking only from experience with TDS 540x, yes the calibration is kept on
the acquisition PCA. I think, but have not taken the time to prove, that the
primary use of the NVRAM is to store trace data. Everyone of the NVRAMs I
have read seem to have a pile of "empty" space in them along with a bit of
stuff relating to options, date, time, and the like.

Just as easily though I could be FoS.

Steve

-----Original Message-----
From: TekScopes@groups.io <TekScopes@groups.io> On Behalf Of Siggi
Sent: Tuesday, January 28, 2020 7:05 AM
To: TekScopes@groups.io
Subject: Re: [TekScopes] TDS784C

On Tue, Jan 28, 2020 at 8:39 AM JJ <jajustin@...> wrote:

You need to be concerned with the NVRAM chips - DS1486 clock chip and
the DS1650/DS1250. They have internal batteries that hold the
calibration constants and options info.
It's my understanding that the calibration constants are stored in EEPROM on
the acquisition board. There's no reason to worry about the options, as it's
easy enough to restore those with GPIB or other means.