Finding replacement for defective oscillator in ADI-2 Rev 1.4

Hi again,

a followup: Since the last post, I have replaced the oscillator and my ADI-2 works again.

First, I desoldered the oscillator to check if it works correctly at a full 5 V via a lab PSU. It did not and still oscillated at about 41.66 MHz with no load other than a 10X oscilloscope probe. I have a theory regarding why the oscillator oscillates with this specific frequency: I think that the crystal inside the oscillator is intended to run at its third harmonic with a base frequency of 25 MHz / 3 = 8.33 MHz. However, due to damage or old age, the circuit inside the oscillator instead resonates at the fifth harmonic, which is 8.33 x 5 = 41.66 MHz.

Regarding the replacement: The oscillator in the ADI-2 comes in a standard DIL-8 package, so I took my chances and ordered an Abracon ACH-25.000MHZ-EK. This is a high-precision (30 ppm) part that accepts supply voltages from 4.5 V to 5.5 V, so it should operate correctly even at the low 4.57 V supply I measured. I am unsure about its drive capability, which seems rather low at 15 pF typ., but comparable parts had the same spec. Also, it seems the oscillator only drives one input pin of the FPGA, so the load capacitance is most likely well below that value.

After the replacement arrived, I soldered in the new oscillator, applied power to the ADI-2 and lo and behold, it just worked. So this concludes this issue.

For those interested, some more information: To remove the oscillator, I first desoldered all pins but the GND pin using a desoldering gun, making sure they were loose inside their respective through-holes. Then, heating the GND pin, I slid the oscillator out. The reason for me to do it this way is that there are large GND planes inside the PCB which suck away heat from the desoldering gun to quickly for it to be able to make all solder inside the through-hole fluid enough for it be able to be sucked away. With the oscillator removed and using a large chisel tip and some additional flux, I was able to clean the through-hole easily with desoldering wick. Both chisel tip and desoldering wick are hard to impossible to use while the oscillator's GND pin is still inside its through-hole, so removing the oscillator first is almost obligatory.

Also, SMD oscillators should be usable for a replacement, since there are pads below the oscillator package! A standard 7 x 5 mm SMD oscillator should fit, with some pad space to spare for handsoldering. The pad dimensions are

      FPGA
            ~1 mm
 ___         ___
| 4 | ~4 mm | 3 | ~2 mm
|___|       |___|

~3 mm
 ___         ___
| 1 |       | 2 |
|___|       |___|

This is the pinout:
1 Enable/NC (is directly connected to Vdd pad)
2 GND
3 Output
4 Vdd

Best regards,

deadsnare

Thanks for the kind words! I hope that this helps should somebody else encounter the same or a similar problem.

As a nice boon, the ADI-2 I've repaired exceeds its ADC SNR rating by a whopping 3 dB, noise is about -113 dBFS RMS unweighted without any connection on the inputs. Really impressive for such an old device