On Sat, Aug 16, 2014 at 12:44:53AM +0100, Tony wrote:

Hello Tony.
Gordon already recommended the same kind of plugs I use, but I think
it would be better if you buy them from the manufacturer, because
I found many cheap imitation around there and they are really not the same.

They are made by Hirschmann, and if you prefer the solder/crimp version,
you should buy the BUELA 30. Red color is 930727101 and black 930727100.
They make also the female KUN 30 and the male BUELA 300 with lateral insulation
to avoid unwanted contacts when coupling them.

I normally buy them from Farnell, where you could also find the SLS425-SE/Q
that suggested Todd.

Best regards,
Andrea Baldoni

Randy,

Congratulations on your 'new' 3458A.

The 'CALRAM' back-up was just recently covered.  There are two options:

1. Remove the NVRAM, read its contents with a programmer, program a new
   NVRAM and install the new NVRAM.  I chose that option, installing sockets
   for the 'CALRAM' and the two other NVRAM's that are used for other,
   non-calibration, purposes.  Be sure to let the NVRAM cool all the way down
   before trying to read it.  Learned by experience with that one.

2. Read the 'CALRAM' via HPIB and the MREAD command (many steps involved
   but can be done as a program with your laptop/desktop), use that data to
   program a new NVRAM (I suspect there is a way to 'program' the NVRAM by HPIB
   as well but I don't recall ever reading anything about it on the list),
   remove the old NVRAM and install the new NVRAM.

When I read the contents of the 'CALRAM' with a programmer and with the
MREAD command, they matched exactly.  However, when you do an ACAL, the
CALRAM contents change.

Does your unit have the single EPROM or the 6 EPROM version of the firmware?
Also what version of firmware do you have?  The latest is REV 9.  REV 8 is
'out there' on the net.  If you purchase an exchange display, and you have
the single EPROM version of the A5 Outguard Controller Assembly, the REV 9
EPROM is $24.10 per the Keysight website.  Might be worth adding to the
order.  The 6 EPROM version is $220.

Also, what is the serial number?  You can get a good idea of age by the
serial number.  Also, the various covers inside often will have dates of
manufacture on them and the chips, etc., will have date codes, including the
NVRAM's.

Also, you can add option 001 by plugging some memory chips, specific details
I don't recall right now.

As far as the reading when shorted, I recall the Calibration Manual calls
for a piece of copper wire to short the input terminals.  It looks like a
flying 'U' when properly fashioned.  The Calibration Manual gives the
details, IIRC.

Good luck.

Joe

-----Original Message-----
From: volt-nuts-bounces@febo.com [mailto:volt-nuts-bounces@febo.com] On
Behalf Of Randy Evans
Sent: Friday, August 15, 2014 8:40 PM
To: Discussion of precise voltage measurement
Subject: Re: [volt-nuts] New HP3458A

I would like to thank everyone that replied to my query on what to look for
in my "new" HP-3458.  I did find that the AUTO ZERO button does work fine,
thanks to Bill's comments.  I had looked in the manual, just not far enough.

The display is still perfectly readable but i would like it to be "perfect".
I am particularly interested in the exchange display for $272 (better than
$700 for a new one, as Todd suggested).  Does someone have a contact number
at Keysight that I could call (my experience is that one can spend a lot of
time calling around until the right person is finally found).

I certainly need to figure out how to copy and replace  the NVRAM - I lost
the calibration on my Datron 1082 by not realizing the memory backup battery
died and now I need to calibrate it myself (the HP3458A should make that
doable, I hope).  Any suggestions would be greatly appreciated.  I believe
it has been covered before on Volt-Nuts and I need to go through the past
discussions to find it.

One question I have for the group is what should the display typically show
with the input shorted?  I see a reading of about -.0025 mV.  That seems
rather high.  I tried several different banana cables (gold plated, tin
plated) used to short the input terminals  to see if thermocouple effects
might be responsible but there was no change in the reading.

Still learning.

Thanks,

Randy

On Fri, Aug 15, 2014 at 8:06 AM, Jason Watson watson.jason@gmail.com
wrote:

careful).

volt-nuts mailing list -- volt-nuts@febo.com To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.

-0.0025mV does seem a little excessive, I think that may be out of mfr.
specifications if I'm remembering right.  Charles is right you don't want
to use banana plugs, at least for low dc volts stuff.  The service manual
has a good picture of how you should use a section of clean copper wire to
short the terminals.  Also make sure you've done an ACAL recently and that
the ambient temperature hasn't changed by more than 1 deg. C since doing
it.  The air filter is also a good call, I've seen the internal
temperatures change by 2-3 degrees just by having the air filter dirty.
Use the TEMP? command to get the internal temperature.  Use the CAL? 58,
CAL? 59 and CAL? 60 commands to get the internal temperatures when the last
calibration was performed.  There shouldn't be more than a 5 deg. C
difference or else there are adders you have to start throwing on the
accuracy spec's.

Finally after all of that the input offset voltage is still high, running a
CAL 0 command in both the front terminal and rear terminal positions with
the input properly shorted will re-'zero' the instrument.  Make sure the
cal security is turned off to be able to run the CAL 0 command.  All that
is well detailed in the service manual too.

On Sat, Aug 16, 2014 at 8:48 AM, J. L. Trantham jltran@att.net wrote:

In message 006501cfb950$68b183a0$3a148ae0$@att.net, "J. L. Trantham" writes:

There is no relevant difference between REV 9 and REV 8 worth opening
the box for.

--
Poul-Henning Kamp      | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG        | TCP/IP since RFC 956
FreeBSD committer      | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.

Since it was brought up about the possible leakage between Guard to Low, I
decided to check three 3458A. All three measured "open" resistance between
the two connectors in the open position as measured with an 8846A.

On two, I measured ~0.14-0.17 ohms with the switch closed. The first two
look fine and they are steady regardless of how I play with the switches.

The third meter measured around 100ohms and then drifted down below 90ohms
over a period of several seconds. Wiggling the button on the front of the
meter will make it jump around. This is the same indication that I had with
the bad front / rear switches on two of these meters.

The CLIP shows nothing else in between ground and guard, so it appears the
switch will need to be replaced. The switch should be a lot easier to
change compared to the bigger schadow switches. Fortunately, these are
available from Keysight for < $7 plus shipping.

Todd

On Fri, Aug 15, 2014 at 11:06 AM, Jason Watson watson.jason@gmail.com
wrote:

I meant to say Guard and Lo.

On Sat, Aug 16, 2014 at 5:47 PM, Todd Micallef tmicallef@gmail.com wrote:

On 16/08/2014 11:24, Andrea Baldoni wrote:

Thanks everybody for all the suggestions - I have to say I didn't expect
it to be quite so difficult.

Gordon: That ebay suggestion looks good but they're out of stock. The
description doesn't say if they are gold plated (my preference) or not -
they look as though they are in two photos but not in the other.

Andrea: I found the Hirschmann BUELA 30s here:
http://www.conrad-electronic.co.uk/ce/en/Search.html?search=BUELA - the
price is reasonable at  £0.79 each ($1.3), but they are nickel plated
and I would prefer gold plated. I have to agree with your suggestion to
buy branded products given the poor quality of some Chinese ebay goods,
but some of it is perfectly ok - its pot luck.

Todd: The SLS425-SE/Q is available from Farnell but they are shielded
which isn't what I'm after (sorry, I forgot to mention that) and aren't
that cheap at £3 each (+ 20% VAT) and you have to buy them in packs of 5
of one colour. The Multi-Contact SLS205s are gold plated and aren't
shielded but aren't stackable.

The Multi-Contact SLS425-A looks like a possibility; its still a bit
short at 18mm but its gold plated, solder connection, stackable and cost
£1.59 including VAT from conrad-electronics. (I don't know anything
about them but they do seem to have an large range of connectors).

I might have to buy low thermal cables as I haven't yet found any
suitable banana plugs - ie. gold plated, crimp and solder connections
and made from copper with a copper based spring. (Some have
copper-beryllium springs but brass bodies). How are the ready made low
thermal leads connected - are they soldered, crimped or both? My concern
with crimped connection is that despite a good tight connection, the
copper wires in the joint may still be prone to oxidation leading to
high thermal emfs and even rectifying properties?

Can anyone tell me how long the plug is on HP test leads? The first 8mm
or so is lost to the contacts being recessed in the test socket on the
34401A.

On gold plated contacts:

I found this interesting:
http://www.te.com/documentation/whitepapers/pdf/aurulrep.pdf
They claim gold to be the best connector coating with 2 exceptions:

1. They shouldn't be mated with tin contacts
2. They aren't recommended for making/breaking live circuits due to the
   susceptibility of the gold contact to arc damage.

I wasn't aware of these and will be more careful in future -
particularly when undertaking current measurements with my HP 34401A to
avoid damaging the terminal sockets.

Tony H