Wow, I can't imagine that any engineer who was able to get hired by HP
wouldn't understand why JPG isn't appropriate for screen shots or other
line art. (or anything else that isn't a continuous-tone natural scene,
which is what JPG was designed for, and the only thing for which it
works well)

The options weren't great back then, unfortunately.  The .PNG spec was brand new and hadn't caught on yet.  Unisys was shaking everyone down for royalties on the .GIF patent.  I'm not sure .BMP existed at that time -- I believe it did, and it would have worked, but it was mostly a Windows-specific format.  Dunno if the Unix and Mac people would have been happy with that.  .TGA and TIFF would have worked but the files were large and cumbersome, and as I recall they weren't used widely outside the desktop publishing business.  .PCX and .LBM would have worked well since RLE compression is a good fit for this type of file, but again they didn't have a lot of support across different platforms and industry sectors.

So that left .JPG.  While definitely not the 'right' format, if you crank the quality level up high enough, it could yield tolerable results, and unlike .GIF it could be freely implemented.  Arguably HP should have paid the danegeld for the .GIF license and called it good, but it's hard to second-guess them at this late date.

-- john, KE5FX
Groups.io Links: You receive all messages sent to this group.

View/Reply Online (#95310): /g/HP-Agilent-Keysight-equipment/message/95310
Mute This Topic: /mt/30423918/102401
Group Owner: [email protected]
Unsubscribe: /g/HP-Agilent-Keysight-equipment/leave/defanged  [tamhan@...]

?

On Mar 21, 2019, at 3:47 PM, Reginald Beardsley via Groups.Io <pulaskite@...> wrote:

Battery life on the early Dallas parts was 5-7 years. I had to replace one in a 5-6 year old Sun workstation. Dallas stopped making them and for a while Maxim made them. At present the only NVRAM source appears to be ST Microelectronics.

I am waiting for a pair of chips to arrive from the UK as there is no stock in NA. They are backordered until June. I was able to get a pair for my Tek 11801 courtesy of a member of the TekScopes list. RS sold out 34 chips in 8-10 days while I was waiting for my 11801 to arrive. I got the last two.

The only demand for these is repairing old instruments, so ST appears to make a small run once a year. The current ST chips use a lithium cell, so shelf life should be reasonably good. But the cells are very low capacity (40-120 mAh) and can easily be drained if inserted into conductive foam. They are widely available on eBay, but the freshest I could find were US NOS parts of 2002 manufacture.

Reg

My recollection, from having developed with the Dallas
RTC versions of their NVRAMS, is that you have to coerce
the time out of the device by reading and writing a
coded sequence to the RTC bytes, to unlock the RTC bytes.
Unless you do that, the data bytes where the clock appears
are just data.

This has to be done every time you read or write the RTC.

I've done so many different of these sorts of devices that
I could easy be remembering this wrongly.

-Chuck Harris

Dave_G0WBX via Groups.Io wrote:

If you spend some time with Google, you'll find people have managed to
disconnect a failing battery, and wire in an external cell. Not trivial
to do, but it can be done.

Some types you can actually persuade the top-hat part of the package
off, and make a neat job of it too.

The problem with some "device programmers" is that they are not aware of
the RTC feature, so will fail to verify the top few bytes that change
with time, as they see it.

Plus, we've not found one yet, that can turn the clock or battery off,
if not needed (after a SRAM test, and then to be put back on a shelf again.)

There are also now, FLASH RAM types in compatible packages (minus the
RTC) that are a better match for some needs.

Regards.

Dave B

If you spend some time with Google, you'll find people have managed to
disconnect a failing battery, and wire in an external cell. Not trivial
to do, but it can be done.

Some types you can actually persuade the top-hat part of the package
off, and make a neat job of it too.

The problem with some "device programmers" is that they are not aware of
the RTC feature, so will fail to verify the top few bytes that change
with time, as they see it.

Plus, we've not found one yet, that can turn the clock or battery off,
if not needed (after a SRAM test, and then to be put back on a shelf again.)

There are also now, FLASH RAM types in compatible packages (minus the
RTC) that are a better match for some needs.

Regards.

Dave B

?

My experience with a HP 70900A LO and a similar range YIG (ie, not correct part number) from a signal
generator of the same era was that I had to play with the resistor divider (DC) network to get it into range
so the diagnostic software could change the control word.? The YIG had the "plumbing" pointing in a
different direction (straight out the side)? so the RF shield could never be installed, but it worked.
The manual said (roughly) at x volts you should have y to z Ghz out, if not change the resistors...

It took two years to find all the HP 9000 pieces to load the diagnostic software.? Doing calibration of the
HP 70000 units requires so many obscure and expensive HP pieces of test equipment I gave up collecting them.

Jim
On ?Thursday?, ?March? ?21?, ?2019? ?09?:?57?:?08? ?PM? ?EDT, Peter Gottlieb <hpnpilot@...> wrote:


Today I received a new A11 phase lock board for my misbehaving 8753ES (thank you
Pete M!).

A lot of measurements will follow as I have lab time, but I thought I'd run a
simple "plug it in and hold ears" test.

Test 48 now passes.? The 1 V/GHz waveform test on the analog bus has the same
shape as before, but the starting voltage is below 3.5 volts instead of 3.6 as
in the original board.

I still have some intermittent lock issues below 10 MHz but this is major progress.

I have a lot to do, including Root Cause Analysis of the failure of the original
phase lock board.? Perhaps the precision 6.2 volt reference zener has gone out
of tolerance, it would only need to be a few percent off.? I will see.

What this means is that when the troubleshooting guide says the waveform should
be as they show, it means literally in this case.? A waveform starting at 3.6
instead of 3.5 volts may cause failure.

Much more to follow!

Peter