Time & Frequency Test Equipment

© Brooke Clarke 2007

These are some of the test equipment I use for Time & Frequency.  Sort of chronological order oldest at the top.
The Cesium standards have their own page FTS4060

Got this for practically nothing since it was being sold as a simple crystal oscillator because the oven was not working.  The problem was that the gel cell lead acid batteries had vented and the fumes etched away some of the PCB traces.  It's surprising that the FTS4060 and Datum 4065B both have provision for internal gel cell batteries!  They would suffer that same problem and be destroyed if the battery vented.  Much better to locate the batteries external to the main chassis to prevent any acid fumes from causing problems.

A very versatile design that can be factory configured to have a control voltage polarity and range to match about any existing ovenized VCO.  Also the crystal drive can be adjusted to trade stability and aging.  A replacement for the Gibbs.

Just sits on a shelf running 24/7.  Mainly used to feed 10 MHz into the reference inputs of test equipment like the SR620 counter, 8648 sig gen, HP 4395A Network, Spectrum & Impedance Analyzer, HP 33120A Function Gen, etc. 
Has it's own PRS10 web page.

I had one of these counters which can count something like 12 digits of frequency in 1 second.  A straight cycle counter working with a 10 MHz input would display 10,000,000 after one second, i.e. only 8 digits.  But there are a couple of things not so good.  The manual user interface might be called user hostile.  The 12 digits does not apply to measuring time intervals.  Measuring time intervals is what you do when working with precision frequency or time signals, not measure frequency directly.  So the 53132 got replaced by the SR620.  Note the name of the SR620 is not frequency counter, but Time Interval Counter, i.e. it was made to do TI counting.  The annual conference is called Precision Time and Time Interval (PTTI).

But (I think, but have not checked it) the TAC32 software is steup to work with either this counter or those closly related for making time interval measurements aginst GPS and applies the sawtooth correction, so they have value for that use.

This is an analog audio generator who's frequency can be programmed using a parallel digital interface.  It was designed with active components added so that reverse engineering would be more difficult.  So far no working, although new in the box.

Currently measuring the difference between a GPS 1 PPS and the output from an FTS4060/S24 Cesium Standard.  The 1 MHz signal from the Cesium is the external clock for the counter and it's 1 kHz Reference out is synchronous with the Cesium input.  By using the counter this way the time for rollover is improved from 1 micro second to 1 milli second.

The black triangles are so make it easier to keep straight where different units are located.  The great thing about this counter is that there are 15 digits.  It does not use exponent notation, but rather just puts the digits where they belong.  The current display is825us 248ns 976ps.  This is the result of averaging 100,000 readings which takes a little more than 1 day.   It's connected to an old computer that's data logging the readings.  This display turned on at 18:04 on 1 July 2007.  On 23 June at 15:37 the display was 825us 494 ns.  So the 4060 frequency is off by (825494-825248) / 700,000 = +3.5E-13.

I need to connect this to the UPC that protects this computer.  Every now and then the UPC records a glitch and every now and then the 620 has a glitch in it's data.  It may also need a bigger, or new, power supply filter cap.

The fan might be replaced by EBMPabst 624.
Factory fan is a Delta Electronics DFB0624 H, Dimensions are 60mm x 60mm x 25mm, DC24V 0.11A

Although the displayed one shot resolution is 1 pico second in actuality it's more like 20 ps.  This number is critical since it sets the stability of the counter.  When measuring high stability devices, like cesium standards the counter must have a stability Allan plot that's below the device your are trying to measure.

Use the following adjustment sequence to get the frequency spot on: 1. Set the Cal jumper to Cal Enable 2. Connect the reference to Ext Ref (rear) and Input A 3. Switch the counter to Ext Ref 4. Set the CalByte 50 for the best display (this is a very fine adjustement) 5. Switch the counter to Int Ref 6. Set the CalByte 4 for the best display (this adjustement is coarse, optimize!) 7. Set the Cal jumper to Cal Disable

The International Laser Ranging Service On using Time Interval Counters for Satellite Laser Ranging (SLR) - SR620 - HP5370B - Time of Flight Best Practices -

I've heard that the circuit in this generator came from Stanford Research.  But that's not here or there.  The nice thing about it is that you can connect an external frequency standard then whatever frequency you select is very precise.  I have a number of audio oscillators that have either analog tuning , push button or switch setting, but if you want to be spot on frequency none of them can do it.

Origionally purchased for working with military radios, it gets the most use as the external clock input for testing PIC based clocks.  Can easily generate 1, 2.5, 5, 10 or 20 MHz and with the external frequency standard connected to the PRS10 the frequency has the needed accuracy to check for software problems.  This 100 kHz to 1 GHz sig gen also has the optional pager package that adds programmable audio modulations.  The stability of the 8648 is considerably better than prior HP VHF/UHF signal geneerators and was needed for pager testing.  It's able to generate an AM modulated (1 kHz or 100 Hz) signal at the WWV H.F. frequencies that's very handy for functional testing the Heathkit GC1000 Most Accurate Clock.

This is a precision delay generator that has 4 channels where the delay can be set between the trigger event and the A, B, C and D outputs.  At the end of the delay the single letter outputs change state.  In order to get a pulse the combined outputs AB, /AB, CD and /CD are available.  Each of the delays is settable to a pico second.  Has rear panel input for a house 10 MHz frequency standard.  The two connectors at the left are the Trigger Input and the T0 output.

A very capable and powerful instrument.  Just the thing for testing PIC based Time Interval Counters.

DOA from eBay

Got this on eBay with the logical to me front panel showing External Clock Error and the rear panel switch in the External position.
But it also has the same error message when the switch is in the Internal Oscillator position.

Covers off, upside down to get access to Q502 & Q503 near the rear oscillator selection switch.





Step one check the waveform at the common point of Q502 & Q503.
100 ns period (10 MHz) signal sort of a square wave, +0.5 volts and -0.5 volts.



The problem turned out to be C512 was out of adjustment.  This cap reasonates an inductor that forms the heart of an 80 Mhz oscillator that drives the DG535.  The cap was set so that the PLL could not lock, hence the error for either INT or EXT settings of the 10 MHz switch.  You adjust the voltage at pin 6 of the U507 op amp to be 2.6 Volts by tuning C512.

Sticky Stuff

The 10X probe shown above is connected to the HP 54501 scope.  Using this scope to troubleshoot the DG535 pointed out some of it's problems:

• The digital memory survives power off and so the scope remembers all kinds of settings that can cause wrong answers.
  So you must press <RECALL> . .  then . .  <BLUE>  to reset most things to their defaults.
• Next to the buttons for the Delta V and Detla T readouts the soft label on the CRT has the channel number to the left of the cursor value.
  When only one channel has any input, say Chan 4 is being used, and if the soft label is set for another chanel, say Chan 1, then you will be using the volts per division and offset setup in Channel 1 to measure the on screen voltages, which can give very wrong answers.

          A scope like one of the Teksmall USB types is looking pretty good.  The other option is a Made in China knockoff scope.  The TDS2024B looks good.
Ended up getting the Rigol DS1052E.