HP ESG1000A Repair

[wavebits]

I am trying to fix my ESG1000A signal generator and need some advice. The unit had two problems. Missing RF output between 250-500 MHz and error 605. I was able to repair the missing frequency range by replacing a bad divide by 2 NEC scaler chip on the synthesizer module. It now has RF from 250K to 1GHz.

The error "605 DSP FW Download Failed" has proven to be more illusive. The error occurs at boot-up and anytime a change is made to the modulation settings. If FM modulation is turned off and back on, or the modulation frequency or depth is changed, the error clears then re-appears. The micro must be trying to send data to the DSP chip to set it up,  but is failing to do so. No modulation is therefore possible.  External modulation works fine with an audio signal applied to the EXT-1 input. The DSP provides a very nice function generator for modulation with multiple waveforms including dual sine with independent frequency and amplitude levels.

The documentation is sparse on this unit. The error message is explained in the manual as If this happens "contact HP" No schematics are available to mere mortals for the ESG series.

Has anyone ran into this problem and offer any troubleshooting tips?  All supply voltages are present and WNL. The DSP appears to have some external RAM connected to it. I don't know if the main FW is corrupted and sending bad data to the DSP or there is a hardware problem between the main processor and the DSP or possibly the DSP RAM is bad. Other than the DSP related functions, the RF portion  is working normally with proper output levels over the attenuator ranges. No other errors appear in the queue.

Thanks, Dave K7DMK

[wavebits]

I did a little more troubleshooting tonight. Looking at the ADSP-2101 DSP chip with a scope showed no activity on the data or address lines. VDD was OK at 5.1V Checking the clock output pin showed no clock. Probing the crystal revealed the xtal was not oscillating! I temporarily bridged a junk box 14 MHz crystal across the old one while looking at the clock pin and bang, a nice clock pulse. Digging deeper in the junk box I fould a old Ziatech CPU board with a proper 16 MHz crystal.  All is well.

So a bad DSP clock crystal was the problem. This was a free standing crystal, not an oscilator module.  I have personally not seen many failed crystals in the past.

Dave, K7DMK