I had already for some time a HP8656b synthesizer running from 0.1 until 990 MHz

After owning this for several years I wanted to do some measurements with this unit but after some trials I became suspicious that it had no RF output, the instrument was no longer working

A bit confused what was going on I measured with a power meter for RF but nothing, I listened with a receiver if I could hear a carrier but nothing also, I had to repair this unit.
For me it was strange that the instrument did not show any failure indicator, from other instruments I was used to see an alarm indicating something was wrong.

On internet I searched for manuals of the unit and even better I found the operating and service manual.

I removed the top and bottom plates and took a look at the instrument internal.

I started with the troubleshooting and checked the power supply voltages, these were not OK the led for +15 volt was not lighted,
 indicating failure of the power supply, not bad at all because these are more easy to be solved, what I expected.

+15 Volt indicator led not burning

I looked in the service manual but the quality of it was not very readable, these were scanned pages and most details were not readable. I tried to look further in the documentation but leading me to nothing at all.

Also I found some Youtube movies for repair of this synthesizer but off course these had different problems.

After searching again the internet I found a service manual in better quality and also schematics which were combined to one document
making it much easier to search for the different unites and components.

I kept on searching for the defect in my instrument without result but I suspected the +15 Volt regulator to be the problem.

Measuring at a measurement point I found the +25 Volt going to the +15V regulator alive and I removed the regulator and checked it on a separate power supply.
The regulator seemed to work fine but when I put it under load it reacted strange and for that I decided to suspect the regulator.

Before ordering a new regulator I decided to test the instrument with an external power supply of 15 Volt and then all was working fine, so after that I was convinced that with a new regulator the instrument would be fixed.

A few days later the new regulator arrived an after mounting it I started the instrument with again +15 Volt failure ???

Looking at the documentation I had no clue what could be the problem and I concluded it had to be a damaged trace on the PCB. I started doing measurements with an ohmmeter to check PCB traces. These measurements leading me to an undocumented component on the PCB. It was a relay used for switching on the power supplies. It was not documented as my instrument had an optional high stability timebase and the changes were not in my documentation. I was already glad I found some documentation which was readable.

I removed the mounting screws of the PCB to be able to look at the backside of the PCB and flip it open, as written in the documentation. At first I did not succeed in flipping it open and became doubtful if that would be a good way to proceed. Somehow the board was not free to be opened but I did not see where it stuck.
The next day I tried again to flip the PCB and found why my first attempt failed in some cabling to the front panel and to the high stability timebase.
I had to bend these cabling and then I succeeded in flipping the PCB open and take a look on the backside.

I did not find any damaged traces on the backside in the neighbourhood of the power supply or the relay I found. Only way to explain the failure was to suspect a failure in the relay.
To solve it I could only proceed in dismounting the relay and see if the relay could be opened to try and clean the contact causing the failure.
I was a bit reluctant as the backside of the PCB was not easily reachable and for that I had to de-solder the relay. If the relay was the reason for the failure I realised I needed to get me a replacement which I also expected to be not so easy.

Looking on the side of the relay I saw that it had 4 switching contacts. From the documentation I knew the instrument using only 3 different voltages.
That gave me the idea that 1 switching contact of the relay perhaps remains unused.

I had to do some measuring and searching to find what contact of the relay was not in use. When I found it I realized an easy solution could be realized in adding two wire bridges between the unused contact and the failing contact for the +15 Volt power supply.
After mounting these bridges I tested the instrument, first power supply only (I loosened the connector J2) and the 3 leds for the 3 voltages all were indicating fine, so the +15V was present again. Then I mounted the connector J2 feeding the rest of the instrument to check if it was working again and it did. 

After connecting my power meter I saw indication of RF from the output and by switching on a receiver and setting the frequency I could also hear a stable and strong carrier, so I concluded that the instrument is working again.