Help figuring out oscillator circuit for frequency meter.

[TERRA Operative]

I'm currently repairing an old Japanese 'Digipet-60' Nixie tube frequency meter made by Advantest back when they were called Takeda Riken in collaboration with Toyomura.
(I'm currently in the process of obtaining any info on the unit, Advantest/Toyomura are searching their archives).

I'm poking the little stand-alone oscillator board and it seems I might be getting a weird output from it.

Shown below is the schematic of the oscillator board and the output waveform. Can any of you wizards let me know if that waveform is correct (I was maybe expecting a 10Mhz sine wave?) and give me any pointers?

So far, I replaced the crystal with new incase it was broken, and tested all the components and they all seem to test ok..

[borghese]

What you see is the 100 Hz ripple. First check the supply voltages and replace the electrolytic capacitors

[TERRA Operative]

Yeah, the electrolytic is new, same brand and latest equivalent type.

I'm also running it from my lab power supply too, at exactly 5VDC. The scope probe hooked up to the output and 0V.

[TERRA Operative]

Ok, a development!

The awesome guy at Advantest here in the Japan headquarters was able to scan their last remaining copy of the manual for me, including a schematic! Domo arigatou Advantest-sama!
You can see the oscillator just left of center on the schematic.

Looks like some of my measured values were a little off from the schematic values, so I might try replacing the parts with new to make sure they are within spec. 
(Looks like I put 510pF instead of reading it properly as 50pF too, derp...)

[basinstreetdesign]

It looks like your scope horizontal sweep speed is several orders of magnitude too slow.  Crank up the time/div to at least 100 nS/div to see the individual cycles of the waveform.  Don't use delayed sweep.

When the sweep speeds of those older Tek DSO scopes is turned slower, the sampling rate slows down too.  Then the sampling of the scope cannot capture the high frequency content of the signal and the display aliases.  This can look like a very low frequency version of the actual signal wave form that just WILL NOT lock to the trigger point or may look like random crap all over the screen.  Turning the sweep speed up also turns up the sampling rate until the waveform is captured with fidelity.

[orbanp]

Hi Terra,

The funny signal on your scope could be from ground loops, the scope's input is referenced (tied) to the ground.

A simulation shows that the oscillator stage (the first transistor) would run.
I used a "present day" crystal model, not sure how it would differ from your original crystal.
For the transistor I used a modified 2N3904 model. The 2SC269 is a 100HMz Si transistor, with a minimum beta of 40. In my 2N3904 model I changed the beta to 100.

What made a difference in the output was the biasing of the second stage, the output transistor.
From DC point of view it is a "badly designed" circuit, the biasing is not too good, it depends solely on the beta of the transistor.
Even with a beta of 100 the transistor is saturated, the collector is nearly at ground level.
When I changed R52 (in the scanned schematics), the base biasing transistor, to 220k I got about 5Vpp "squarish" output signal.

Hope this helps.

Regards, Peter

[Andy Watson]

The funny signal on your scope could be from ground loops, the scope's input is referenced (tied) to the ground.

Yes, it could be. But

If we are to believe the measurements on the scope it is a ground loop  of 19.6Vp-p - that's a huge induced voltage.

I think basinstreetdesign has hit the nail on the head. Pay particular attention to the word "aliasing".

[cncjerry]

I agree with the prior 2, either aliasing or aliasing combined with a 1X probe in 10X mode.  Either way, with a +5Vcc and no output transformer or inductor and feeding TTL logic, how are you going to get 19V P2P?  Unless the +5Vcc regulator is really shorted out some how.  But the OP said he was using a power supply...


Jerry

[TERRA Operative]

Hmmm yeah, I got the circuit hooked up to the scope again and had a play around, it's looks like it was a PEBKAC error on the old TDS210....  basinstreetdesign was right.

That'll teach me to mess with stuff at 2am when I should be asleep...... 


Next step is to figure out why the rest of the device isn't working, but I think I have that traced to a bad Q32 (2SC288A) as the input signal stops dead there.
Just waiting on a replacement part to arrive to swap it out.