PTF 4211 Time Standard/Temex SRO-100 tear down and "hopefully" repair.

[CatalinaWOW]

The which one is right question was getting me, with several signal generators and counters all giving different answers it is time to get another piece of gear into operation.

The gear is a PTF 4211 frequency source.  It is a basic model discipline rubidium clock unit providing a 10 MHz output, and capable of being disciplined by a 1 PPS input.  Operation is dead simple.  Plug it in and the power light comes on.  If something wrong is detected a fault light goes on.  A third LED lights to indicate that the Rubidium clock is locked in, providing a correct output. 

The symptoms are that the locked light only intermittently lights, and the output 10 MHz wanders around by several hundred Hz (approximately 1 part in 10E4 in a unit that is supposed to have on the order of 1 part in 10E11 stability.)

Rather than show my so so photography skills will just attach the brochure page which has a photo and specs.  This particular unit has no options installed.

Next post will go into the innards.

[CatalinaWOW]

So first step is to open the box.  A simple beast it is.  A standard power brick, a Temex SRO-100 rubidium standard, a circuit card to distribute the signals and to divide them out to other frequencies when those options are installed.  The circuit card also drives the indicator LEDs.

A quick look at the signal from the rubidium clock to the divider board showed that the source clock was the source of the problem.  I spent quite a bit of time reviewing TiN's thread found here on the similar Temex LPFRS/LPRO found here https://www.eevblog.com/forum/repair/repair-temexspectratime-lpfrslpro-10mhz-rb-generator-(drifting-like-mad)/ in preparation for diving deeper.

On a side note, while the connectors for all of the options are installed on the main board so I looked to see if the signals are present and it would just be necessary to install cables to get these options.  Unfortunately the board has a number of components not installed so none of the options are easily activated.

While the unit should be relatively easy to assemble, the reverse is not true.  The major components are mounted on an aluminum sheet that drops into the case and is bolted down at the four corners.  After this the IEC power connector is pressed in and the indicator board installed.  These steps eliminate access to two of the hold down nuts.  The IEC connector is so close to the corner of the case that it is very difficult to get to the locking tabs so they can be depressed and the connector pushed out.  The second picture shows this area after successfully getting the connector pushed back.  Access to the indicator board screws is poor but not a terrible problem.

After pulling this assembly out I found that the rubidium clock was only held by two of the four screws intended, and these two were not even tightened to the finger tight stage.  This interface is the primary thermal interface for the rubidium clock so I am sure it was cooking at well above intended temperatures.  Whether this is an indication of a prior repair attempt or a manufacturing failure is unknown.

Next post will dive into the Temex.

[CatalinaWOW]

So the first step was pulling the case off of the Temex.  As you can see the lamp lights.  Basic operation is there.

Logging the output showed something new.  The Temex was consistently indicating locked.  Even though locked it showed much higher than specified drift.  It stayed locked for a few hours and then lost lock and started wandering around. 

The locked frequency showed far from 10 MHz on my counter, but I don't really know how accurate the counter is.  That is why I want to get this thing working.

The configuration of this unit is similar to, but with many key differences from the LPRO unit in the other thread.  As shown in the photos instead of a flex harness assembly it is two distinct circuit cards with a connector joining them.  The lamp and readout cell are spaced substantially farther apart.  Most importantly, all of the adjustment pots are in different locations.

I hooked up the serial port and examined the status reports as discussed in the LPRO thread.  All values seemed appropriate.  I attempted jiggling all of the pots as this had been a successful fix for some on the other thread.  Nothing changed, but since it was now locking consistently (for a while) I guessed that different thermal conditions out of the case were related to the problem and buttoned it back up.  No locking whatsoever.

Pulling it back out of the case brought back the locked but drifty condition.  After a fortunate occurrence while monitoring the serial port I discovered that ambient room light was bringing the detector output within its locked range in amplitude.

Now chasing an aged crystal moving the frequency out of lock range I did an "OK" calibration on my counter using other methods and found that it was indeed reading high.  After recalibration logging the rubidium showed that it was never in the capture range.

Now aggressively moving the various pots I find nothing changes the frequency, and am considering options like replacing the crystal. 

Any suggestions would be welcome.