Racal Dana 1992 teardown

[muvideo]

Hello,
recently I won the auction on two Rb frequency standards, no real need of them,
I just bought these because I was curious and the cost was very low.

Too many times I burn money on things that I don't really need, worse usually these
things need repair or cleaning and I burn an huge amount of time trying to find use for
them, repairing etc..., and usually these things need other things and so on... 
(edit: this time I was very happy with the results, more to follow )

Anyway I started thinking how to use these clocks, and started searching for a decent
low cost frequency counter. My purpose was to use the Rb clocks as timebase for the
counter, so my only need was a decent resolution.
And here is it, the Racal Dana 1992, it arrived in "as is" state, the seller only assured me
that the counter would power on, no mention of options, not able to test it.

When It arrived, I changed the voltage setting for european voltage, attached a power
cord and started the thing, being careful to not touch with bare hands the metal case,
you never know...

It came to life showing model number, some other things and after random numbers.
The first thing I noticed was that part, almost half, of buttons were not responsive.
I had read about these faulty switches previously, so I started to tear down this equipment.
The first thing I noticed after opening the case was the beauty of construction, all metal,
very simple and rational, a pleasure to see, so I started taking pictures.

More to follow...

[muvideo]

Ok let's open it...

[muvideo]

Now the problematic part: the switches, first the board dismantled,
after a closeup of switch dissected.
Seems that the usual conductive rubber fails, the switches measure
few kohms closed.

[muvideo]

A local seller has the usual cheap pushbuttons, I'm thinking to modify them
as follow, if someone knows a better way let me know.

[muvideo]

And now the interesting part, if I understand correctly I was lucky to find in
a OCXO, The racal dana 9462 oscillator is a 5MHz high stability ovenized,
doubled to 10MHz.
Should be equivalent to option 04E, if someone knows more please share
info.

Edit: stupid I am, it was written on a big label on the side of the instrument:
1992/02M/04E

Here is it...

[tinhead]

you will find here user/service manual/schematics and option descritions

http://www.ko4bb.com/manuals/index.php?dir=Racal

Btw, when you have a chance make a dump of the EPROM and upload to ko4bb.com incoming,
this might help other ppl to repair their old gear.

[muvideo]

After some hours of work, I swapped all the switches,
tested the basic self tests and all passed. I connected
one of the Rb clocks to input and it reads
the 10MHz frequency with less than 0.2Hz difference,
so there is a high probability that the OCXO is fine.
Now I have to build the psu for the rb clocks and
mount them in a suitable box.

More photo to follow.

you will find here user/service manual/schematics and option descritions

http://www.ko4bb.com/manuals/index.php?dir=Racal

Btw, when you have a chance make a dump of the EPROM and upload to ko4bb.com incoming,
this might help other ppl to repair their old gear.

Thank you tinhead for the links, now I don't have an eprom reader, when I'll have the
possibility I'll extract and upload the firmware.

Fabio.

[IanJ]

Hi,

Wow!....a Racal Dana 1992.... I used to use one of these back in the late 80's.....really great to see one again.

Thanks for posting.

Ian.

[muvideo]

Hi,

Wow!....a Racal Dana 1992.... I used to use one of these back in the late 80's.....really great to see one again.

Thanks for posting.

Ian.

I'm happy you liked it, it seems a fine instrument for it's age,
later i'll upload some photos of it measuring a pair of Rb oscillators.

Fabio.

[muvideo]

A little update:
for the last 10 days I left it powered, in the mean time
I found an home for the two FE5680A Rb oscillators I bought
on the bay.
When I shoot these photos the Rb oscillator were powered from
a pair of days, the measurements are with 10 seconds timebase,
the display shows the last 9 digits, the first two digits "10"
are on the left not shown by the display. The least significant
digit shows mHz of the 10MHz signal.
The first photo is the highest frequency Rb measured using the
slowest one as timebase, the difference is 15mHz.
Other two photos show the frequency of the two Rb
standards measured using the internal oscillator.
The last 3 photos show the highest Rb standard measured
from different positions of the meter, is the earth magnetic
field or the gravity?

Edit: -------
reading around, the internal oscillator that is quartz based
is acceleration-sensitive, the Rb Standards should be more
susceptible to magnetic fields, and less influenced by acceleration.
So the frequency shift should be from gravity acceleration.
--------------

Not bad from a (more than) 20 years old instrument, I was impressed
by the small frequency difference, I don't know the absolute error from true
10MHz , but I suspect that all 3 are very close to the real frequency.

Fabio.

Edit: I had previously indirectly measured the frequency difference of the
2 Rb oscillators using a dso: one on the frist channel and the other on the
second, it took about 60seconds for the two waveforms to return to the same phase.

[SeanB]

Nice unit, looks like it was made in 1989.

I used to use an older Racal Dana 6 digit multimeter, was made in the 70's, had a 7 segment filament display. Slow conversion rate ( 1 per second) and loud autoranging relays, but it just worked.

[im_a_human]

I had one of these Frequency counters up to about 3 years ago. I bought it second hand back in 1993. I had to replace all the switches as they had a habbit of packing up after a few years.

[T4P]

Nice unit, looks like it was made in 1989.

I used to use an older Racal Dana 6 digit multimeter, was made in the 70's, had a 7 segment filament display. Slow conversion rate ( 1 per second) and loud autoranging relays, but it just worked.

You mean nixie tube displays don't you ?

[muvideo]

Nice unit, looks like it was made in 1989.

I used to use an older Racal Dana 6 digit multimeter, was made in the 70's, had a 7 segment filament display. Slow conversion rate ( 1 per second) and loud autoranging relays, but it just worked.

You mean nixie tube displays don't you ?

I don't know the racal multimeters but I have somewere in the "strange things bin" some 7 segments incandescent displays.

Fabio.

[lowimpedance]

Racal made some nice test gear particularly when they acquired Dana (thus Racal Dana).
Generally not sought after on the second hand market, so can be picked up cheap.
Besides the switches you should check C52 and 59 both 680uF electro's they have a tendency to dry out.
A give away is a display of the rectified mains frequency with no input (100hz or 120hz for those on 60hz).
Also check the 50 ohm (100ohm//100ohm) on each input to make sure someone has not connected a transmitter
and toasted them!!.
I have a couple of the 1991 counters, and the oven oscillator was rare, most likely a 10Mhz TCO.
Worked for them early eighties.
cheers
John

[muvideo]

Racal made some nice test gear particularly when they acquired Dana (thus Racal Dana).
Generally not sought after on the second hand market, so can be picked up cheap.
Besides the switches you should check C52 and 59 both 680uF electro's they have a tendency to dry out.
A give away is a display of the rectified mains frequency with no input (100hz or 120hz for those on 60hz).
Also check the 50 ohm (100ohm//100ohm) on each input to make sure someone has not connected a transmitter
and toasted them!!.
I have a couple of the 1991 counters, and the oven oscillator was rare, most likely a 10Mhz TCO.
Worked for them early eighties.
cheers
John

Thank you very much John, I'll check these components.
So you worked for Racal?
Reading around seems that these 1992/02M are from (large?) military contract,
and are all with this 04E oven oscillator.

Fabio.

Edit, for the record if anybody need them, I uploaded the ROMS on http://www.ko4bb.com/manuals/
as of 16/may/2012 they are in recents directory, in few days they will be checked and put
probably under ROM directory.

[slburris]

What switches did you use to replace the failing ones?  I have
A Racal Dana 1998 counter and many of its switches have
become mushy and don't spring back anymore.  Looking at
the service manual it's the same switch as one the 1992.

I can't seem to find any 2pin switches of the proper type.

Scott

[muvideo]

What switches did you use to replace the failing ones?  I have
A Racal Dana 1998 counter and many of its switches have
become mushy and don't spring back anymore.  Looking at
the service manual it's the same switch as one the 1992.

I can't seem to find any 2pin switches of the proper type.

Scott

Hello Scott, I used normal 4 pin 6mm square tactile switches,
size like these ones ebay #180909689661.
Just cut away two pins, they will fit in the holes.
They are working ok, the only small defect is that are a little
"bouncy".

Here you can see what I've done:


Fabio.

[grumpydoc]

Or something like these switches from Farnell/Element 14

I needed to do just one on a 1998 which is working perfectly now. I think muvideo cut the actuator into a cross to fit the switch - I drilled out the switch to fit the actuator.

[muvideo]

I think muvideo cut the actuator into a cross to fit the switch - I drilled out the switch to fit the actuator.

Yes I used a brass plate with a cross cut, insterted it into a soldering iron and used it as "hot die",
after some regulation found the right temp to do the work.
Tried drilling the switch caps but didnt like the result.

Fabio.

[grumpydoc]

Undoubtably the best approach - after all if you destroy one of the new switches it's no great problem. the switch caps themselves are rather less replaceable.

I tried this but found the actuator too small and fiddly to cut and wouldn't be able to reproduce a hot die at all so didn't have much choice.

[free_electron]

holy moly, that's a bucket of unobtainium parts ....
I see an asic from AMCC ... a hybrid module next to it and a custom botch from Plessey nonetheless...
7218 for display scaning an a 74922 for the keypad scanning...

[slburris]

I ran across a reference to someone using this Panasonic switch, which is
P12227SCT-ND at Digikey.  Looks promising, although you still have to cut
the shaft to fit the keycaps.

I'm not sure how best to fabricate the cross cut.  Maybe it's possible to
do a passable job with a dremel and some silicone adhesive in the keycap?

It just makes me really mad that such useful counters are rendered useless
by such a stupid problem.  I had hoped that the 1998 counter would not have
the same problems as the 1992.

Scott

[slburris]

I picked up this little beauty on Ebay a few months ago.
It's a 5Mhz OCXO with a doubler board to make 10Mhz.

I think it was a pull from a 1992 counter, but it drops right
into my 1998 counter, replacing the standard TCXO.

Normally, I don't care too much about the internal oscillator, since
I use a Trimble Thunderbolt to generate a house 10Mhz standard
and run that to all my test equipment which can take an
external reference.  But since this is my go-to counter, I
wanted to be as accurate as possible when I take it with me
somewhere.

I've got an HP5328A, but it's such a noisy beast, I almost never
use it (until now, since my 1998 developed its switch problem).

Scott

[grumpydoc]

Hmmm, I think I spotted that one - wish I'd had my Crystal Ball working though because I didn't buy it and I've now just acquired a 1992 that locked up after a few minutes use - turned out to be a faulty TCXO.

Odd that I should have two pieces of equipment have oscillator modules which failed in the same way (stopped when "warm") within a couple of weeks of each other. C'est la vie, I suppose!