Tektronix SG5010 Audio Signal Generator Repair
Background
The Tektronix SG5010 is an audio signal generator plug-in for the TM5000 modular test equipment system. It can generate five waveforms: sine wave, square wave, SMPTE/DIN intermodulation test signal, CCIF intermodulation test signal, and sine wave burst. All these signals may be swept in frequency and amplitude. It can generate very low distortion (THD+N < 0.0008%; my own measurements) sine waves from 20Hz-20kHz. It is CPU controlled with selections made via GPIB or front panel keyboard. These characteristics made it very desirable as part of an audio amplifier test lab. These units command a high price: Several working units have recently (Dec 2023) been sold for $USD 1500.
I recently acquired a non-working SG5010 for a good price. This post documents my efforts (ultimately successful) in getting it working.
First Steps
This SG5010 as received was very dirty, did not power up, and apparently had several blown fuses. I decided that a thorough cleaning and visual inspection was needed. During the cleaning and inspecting process I discovered open fuses in the power supply. Not knowing the circumstances which lead to the fuses blowing, I decided to replace them and power up the unit.
Problem 1: Error PF310
The fuses did not open upon power up, but I was greeted with an immediate Error PF310. The manual indicated that this error indicated an analog power supply problem. “Thou Shalt Check Voltages”, and after some investigation, I found that the issue was due to badly leaking electrolytic capacitors in the floating power supply on the A13 Output Board. Specifically, C1052 (+17v line) and C1060 (-17v line) were defective. This also affected the +15 and -15 volt supplies as they are derived from the same source. As usual in cases like this, seeing similar capacitors all over the power supply boards prompted me to replace all of them. This cleared the PF310 Error. I surmise that leaking capacitors could cause the power supplies to draw excessive current, probably accounting for the failed fuses I noted earlier.
Problem 2: Error 320
This error indicates a problem with the main sine wave oscillator board: The main sine wave oscillator is not working. Further checking showed that the main oscillator was not working on any frequency and that the frequency could not be changed via the front panel. I spent some time troubleshooting this before I remembered reading that the PLL board and the main oscillator board receive the same frequency control information from the CPU (manual page 4-5 and CPU isolator schematic 5). The problem in this case turned out to be a defective optocoupler, specifically U2042. The defective optocoupler was not passing the FL FREQ CLK signal to the shift registers on the main oscillator board which prevented the oscillator frequency from being set by the CPU. Replacing the optocoupler cleared Error 320.
Problem 3: PLL Unlocked Indicator Lit
This problem was solved by clearing the Error 320 fault and getting the main oscillator running.
Problem 4: Output COM connection grounded
During testing, I found that the Output COM connection could not be floated from ground. In fact, it was always grounded, regardless of the front panel settings, measuring 0.15 ohms to ground. The problem was a shorted TVS protection diode (VR4051) which is connected across the COM terminal at the front panel. The diode probably did its job and failed “short”, thus protecting the instrument. Replacing the shorted diode solved this problem.
Problem 5: Low value carbon composition resistor drift
Prior experience suggested that I check low value carbon composition resistors for any change in value. The output board (A13) had twelve 4.7 ohm and two 2.7 ohm 1/4w, 5% carbon composition resistors. They had all drifted up and out of tolerance by as much as 30%. These were all replaced by higher wattage metal film resistors.
Conclusion
I checked all the outputs and functions of the SG5010 and found it to be operating correctly. A quick measurement gave:
- THD of 0.00017% at 1 kHz with 48kHz bandwidth and 192kHz sample rate (at least 10 harmonics)
- THD+N of 0.00052% at 1 kHz with 48kHz bandwidth and 192kHz sample rate (at least 10 harmonics)
- Both measurements done at 1.454 vrms and 600-ohm load using a calibrated Quantasylum QA403
- BH Windowing was used with a 512k FFT and 8 averages
The calibration and performance check of the SG5010 will be covered in a future post.
Thanks for reading.
Shaun Merrigan