[time-nuts] HP 5334B C-channel--was Re: Which HP Frequency Counter?

Thu Mar 20 10:31:22 EDT 2008

I was the project manager and chief EE on the HP5334B
project.  The 5334A had a C channel using an HP made
divide by 10 prescaler that had a factory cost of ~$100.
In the 5334B, I replaced this with the Fujitsu MB506
divide by 8 prescaler, which cost something like $2.
The firmware was changed slightly to account for the
different modulus.  This is the ONLY difference in
firmware between the 5334A and 5334B counters.
I considered using the NEC uPB581/2 prescalers.  This
class of prescalers is based on "dynamic" flip flops,
as opposed to static ones.  These flip flops are only
intended for prescaling a clean signal from a local
oscillator in a synthesizer.  In a frequency counter
application, they work OK on a new clean signal, but
will miscount on noisy signals.  "Noisy" here refers
to broadband noise, not close in phase noise.  Prescalers
are especially sensitive to low frequency noise.

Now it can be told that we used to test all C channels
in all models with an HP8660 synthesizer.  This was
not one of HP's best designs, and it has a lot of
broadband noise.  [The designers of this unfortunate
product redeemed themselves with the 8662, one of HP's
flagship products].  In order to properly count the
1.3 GHz signal from the 8660, it was necessary to
use a high pass filter to keep the 8660's broadband
noise from corrupting the measurement.  This was also
the case for the previous HP-made divide by 10 prescaler.

I put a LOT of effort into evaluating various prescalers
and trying to put in mitigation measures such as rolling
off the low frequencies before they could get to the
prescaler.  I eventually decided that the task was
hopeless with off the shelf prescalers.

At the same time, one of the other designers in the lab
was working on the 5386 counter, and naturally we compared
notes.  This counter used an HP-made static flip flop.
The FF used in the 5334A was made at the Santa Clara bipolar
silicon fab.  The FF used in the 5386 was made in the Santa
Rosa fab.  Whether it was the process or the circuit design,
the Santa Rosa FF was absolutely bullet proof.  It made
error free measurements of the lousiest signals.  The designer
of the 5386 delighted in finding new signals to measure and
inviting me to a "bake-off" to see who's counter did better.
Of course, I always lost these contests miserably!

I would encourage owners of 5334B's w/o the C channel option
to consider trying modern static flip flops from vendors such
as Micrel and OnSemi, rather than installing the MB506.  You
can easily glue an SMT prescaler to the board upside down
in the footprint where the MB506 goes, and then connect the
"dead bug" with little wires to the MB506 connections.

Rick Karlquist N6RK

Didier Juges wrote:
> The HP 5334A or B with the C channel option has a sensitivity spec of -30dBm
> (from memory) up to 1 GHz or so. The C channel has 15mV rms sensitivity at 1
> GHz.
> 
> The C channel option is rare, but it only requires 3 parts (a diviser and
> two dual-shottkys) and a connector (and a hole in the front panel) to add it
> to a 5334B for instance. I have two 5334Bs (one is broken) and both have the
> other parts required for the C channel except for these three. There is a
> socket for the diviser, the shottkys have to be soldered in.
> 
> The C channel input is activated on pressing the '9' key.
> 
> Unfortunately, the MB506 diviser (Fujitsu) seems a little hard to find. If
> anyone has a spare, let me know. A surface mount equivalent should not be
> too hard to put in.
> 
> The HP 5316 has similar characteristics and also has a 1 GHz option, which
> is also rare.
> 
> I also have an Advantest TR5823 counter which has the 1.3 GHz input with
> 20mV rms sensitivity at 1.3 GHz.
> 
> The HP models have reciprocal counting, so they work MUCH better for low
> frequency signals.
> 
> Didier KO4BB
> 
>> -----Original Message-----
>> From: time-nuts-bounces at febo.com 
>> [mailto:time-nuts-bounces at febo.com] On Behalf Of 
>> Jean-Christophe Deschamps
>> Sent: Saturday, March 08, 2008 10:49 AM
>> To: time-nuts at febo.com
>> Subject: [time-nuts] Which HP Frequency Counter?
>>
>> Dear group,
>>
>> I consider buying a used lab frequency counter / timer --preferably
>> HP/Agilent-- covering from few mHz to perhaps 400 MHz.  I'm 
>> worried that models handling high frequencies seem to be 
>> limited to under 100 mV input signal max. I don't want to 
>> destroy an input channel each other day when 
>> calibrating/repairing some instrument.  Also are there 
>> instruments with high impedance input? 50 Ohm is not quite 
>> right for investigation in the guts of most designs.
>>
>>
>> Is it possible to find a not-too-old model under $800?  I 
>> would like to find something in France or UK or Europe, but 
>> it seems hopeless in this budget.  I would favor a repairable 
>> model (with available service docs & schematics).
>>
>> Your expert advises are more than welcome!
>>
>> --
>> <mailto:jcd at q-e-d.org>jcd at q-e-d.org  
>>
>>
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at febo.com To unsubscribe, 
>> go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
>>
>> No virus found in this incoming message.
>> Checked by AVG. 
>> Version: 7.5.518 / Virus Database: 269.21.7/1319 - Release 
>> Date: 3/8/2008 10:14 AM
>>  
>>
> 
> No virus found in this outgoing message.
> Checked by AVG. 
> Version: 7.5.518 / Virus Database: 269.21.7/1319 - Release Date: 3/8/2008
> 10:14 AM
>  
> 
> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
> 
> 