I realize that better references are available.

I was thinking of resurrecting a long-defunct Fluke 731 that I have in the
garage.  The reference IC is bad in it, and (1) I don't want to throw it
away without a meager attempt at repair, and (2) I don't want to spend a lot
of money on it because it's in pretty bad physical condition.

I was hoping that someone could lend a touch of advice on those old
references.  I have a few 1N827A reference zeners, and a washtubful of
transistors.  Maybe something could be cobbled together that would get the
731 back in operation.  Maybe not to original specs, but close, which is
better than nothing.

Thanks for your reply,
Dave M

Jack Mcmullen via volt-nuts wrote:

they would not be as well matched as if they were on the same chip

On Sat, Sep 12, 2015 at 3:12 PM, Dave M dgminala@mediacombb.net wrote:

--

John Phillips

On 12/09/2015 23:12, Dave M wrote:

There has been some of those references available second hand from ebay,
mostly pulled, but some may be counterfeit, so look for sellers with
pictures showing it as old.

Dave wrote:

(1) the absolute tempco (mV/C, not ppm/C) of the transistor B-E
voltage and the Zener voltage need to be very accurately matched over
the expected temperature range.  (2) the temperature of the
transistor and Zener substrates must be identical to within very
small parts of a degree C.  The only practical way to come even close
with respect to (2) is if both devices are on a monolithic
substrate.  Even separate dice thermally mounted to a very
thermally-conductive substrate in one package is a less than optimal
solution (although that is how the original reference was
constructed).  Starting with packaged components will just end in tears.

I concur with other advice you have received -- ditch the 1960's
technology and use one of the excellent low-drift, low-noise voltage
references that are plentiful these days.  The 731s just have the
reference IC mounted out in the open, which feeds a 1960s op-amp
(LM301A in the case of the 731A, LM308A in the case of the 731B)
mounted out in the open with 1% MF (731A) or precision WW (731B)
resistors, also mounted out in the open, to produce 10v.  If you do
nothing but use an LM399 reference and change the op-amp gain
resistors to suit, using the same type resistors that are in your
unit now, your 731 will be better than the day it left the factory.

If you replace the op-amp with a modern precision part (I suggest the
OPA277 -- others may try to talk you into an auto-zero op-amp, but I
think that would be a mistake in this application), and upgrade the
gain-setting resistors to ones with better tempcos, it should be very
much better than the day it left the factory.

None of these changes requires any more design or construction
expertise than making a new reference IC on the pattern of the
original, and unlike that effort, they will actually improve the 731
instead of ruining its performance.

Best regards,

Charles

On 13 Sep 2015 09:15, "M K" m1k3k1@hotmail.com wrote:

mostly pulled, but some may be counterfeit, so look for sellers with
pictures showing it as old..

Several sellers have more than 1, but with one photo, so you would have a
hard time arguing that your board is a fake just because it looks newer
than the photo. I believe it is standard practice to photograph a genuine
article but to ship a fake.

In any case, it is probably possible to make a board look older.

A used 3458A fetches several thousand USD, so I find it hard to believe
that many would be broken for parts. But even if there is a reason 3458As
are being broken up, why are other parts not showing up as often as  the
voltage references?

I just looked on ebay for "3458A board" There are 10, of which 6 are the
voltage reference and the other 4 being various other 3458A boards.
Looking for "3458A PCB" changes it to 6 references and 2 two other PCBs. It
seems a bit suspicious to me that the one board that has wider usage than
the others is available in roughly 2~3 times the quantity of all other
boards put together.

We why don't we see 3458A front panels, power supplies etc,  showing up
much - why all the voltage references?

The fact that there seems to be a market for 3458A reference boards
suggests to me that there would be a market for a voltage reference board
based on the same chip, but designed as a stand alone board.

I can think of a few improvements over a 3458A board.

1. Option to run from batteries to keep it as stable as possible when there
   are power failures.

It should also make it possible to ship a board to another volt-nut,
keeping it powered in transit.

I don't know the power consumption of the reference,  and I am aware that
there are some issues shipping batteries,  but these seem much more relaxed
if the battery is inside equipment.

2. Option to run chip at a lower temperature than in a 3458A, as some have
   claimed is better.

3. Temperature sensor that recorded min and max temperatures during
   shipping.

4. Ability to read time board has been powered on.

I for one would buy such a board, if someone produced one, either as a
blank PCB or populated with components & adjusted. Specialist components
would be more readily available is there was a number wanted.

Dave

In message CANX10hC6YXVNqMa+mo2K1TmdCDA0XBQa6G0O=ndcYiggBW_QOw@mail.gmail.com, "Dr. David Kirkby (Kirkby Microwave Ltd)" writes:

There's actually one in france for EUR1500 right now...

--
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.

We upgraded about 20 3458As  to the 4 ppm reference for a local equipment
rental co.
Others may have done the same to and are selling the old boards that may be
even more stable than 4 ppm because they have aged for years.

The difference between references is aging and screening.

On Mon, Sep 14, 2015 at 4:37 AM, Poul-Henning Kamp phk@phk.freebsd.dk
wrote:

--

John Phillips

On 14/09/2015 12:08, Dr. David Kirkby (Kirkby Microwave Ltd) wrote:

A 731 is not the same beast at all as a 3458, it uses a very different
reference indeed, I was talking about the IC itself in the 731, not
buying a completed pcb.