Marconi 2955 Test Set Repair Story
The Marconi 2955 Receiver & Transmitter Test Set
This is my story repairing one of these endlessly useful pieces of test equipment from the 1980's.
Like the one in the photo, mine is an early example, with no letter suffix in the model number. And not as clean and tidy looking either - in fact it was pretty grubby outside and very grubby inside. Some of the buttons have been used so often, that the lettering has rubbed off. So a well used example for sure.
At switch on, the cooling fan would run, but nothing else - cold and dark as they say. These units have a 10MHz reference oscillator, oven controlled, and, as well as providing an RF reference, is used to derive the clock signal for the microprocessor. For some weird reason it is fitted to the EHT Supply board for the CRT - but anyhow, checking the output signal on this board showed no activity.
The Reference Oscillator
Removing the EHT Board is not too difficult and then the oscillator module, housed in a small greay plastic case, can be removed. This is it out of its housing:
As you can see, it is made of discrete components and consists of a crystal oscillator and an oven controller, set to maintain 60C. Powering it up on the bench quickly showed the output transistor was dead. Richard on groups.io has managed to work out the schematic:
With the oscillator repaired, the unit powered up - happy days. Basic functionality seemed ok, so time to run the self test and see what it reports:
Hmm ... ok, so a few problems to sort then...
The items from 2.0 to 2.3 are all about AM modulation, so let's check that out...the AM modulation output look like it was working ok, but the levels may be a bit off. I have a copy of the service manual, so quickly ran through the calibration routines for the AM section and...,
tada...AM issues sorted. These remaining ones are all about FM modulation: I'll come back to that.
With the unit up and running, it was time to calibrate the frequency. of that reference oscillator. I have a home made GPS disciplined oscillator in my workshop, so used that to trim the 10MHz reference oscillator to be better than 10Hz on frequency. Nice.
Front Panel Issues
Quite a few of the panel buttons were difficult to use and, once again thanks to groups.io, I found that Mouser have buttons exactly the same as the originals available as a new part: 611-D6L90F1LFS and are around €1 each. So ordered 25 of them.
To get the front panel pcb out, you just need to drop down the front panel (and disconnect one coax connection)
Whilst doing this, I found that the two of the five yellow LED's had stopped working, so decided to replace all 5 - with blue ones :) (looks much nicer imho)
To be honest, it's a bit tedious replacing the buttons, even though they are fairly easy to remove, so I just replaced the worst ones - around 10 in all.
I'm not using FM that much, so decided to ignore the errors on that part and check out the RF side of things.
The RF Power Input
The RF output from the two front panel connectors all checked out fine.
It is important to know at this point, the difference between the BNC connector and the N-Type connector; the N-Type has a 20dB high power pad directly connected to it, so can be used to measure transmitter outputs. Mine is rated at 30watts, others are rated at 70W or even 100W (although rumour has it that the same parts are used in all cases...).
The BNC input bypasses this pad, so is only able to handle low power inputs.
For Receiver testing, this arrangement means the max output from the BNC connector is 0dBm, but from the N-Type is -20dBm (because of the fixed 20dB pad).
I had an HF Ham Radio on the bench around this time and used the 2955 to calibrate its frequency settings and also to check the power output. Not sure exactly how it happened, but I manage to destroy the 20dB pad on the N-Type input. Ah. I had already seen a video of someone replacing this pad with a modern part, so set about doing the same. (Hot air gun required for this.)
This photo shows the 20dB pad with a very clear burn mark on it, and in fact, the ceramic substrate has a crack right across it.
Here is a photo of the new part fitted (has the blue writing on it):
This one is rated to 150W - ok, the cooling provided in here is not up to that really, but short bursts of 100W should be fine. And clumsy fingers won't burn it out again.
EEPROM Issues
I still had those annoying test failures relating to the FM side. My first thought was that it could be the same as with the AM side, that I needed to calibrate. However, FM calibration is a bit of a tedious job. There is a look-up tabled stored in memory that has 300 entries and, to calibrate, you need to access the maintenance menu (which involves hitting .......), select the memory location, make an adjustment, and then come out of maintenance mode to see if the adjustment was right. Takes FOREVER.
(And by the way, check out this video on an easy way to measure FM deviation: XXXX)
But.... I discovered during this process that two of these locations were not returning any value. The schematic helpfully tells you the memory maps for the EEPROMs, so I removed the relevant one and checked it on a programmer. Sure enough, there were a few locations that were blank. So copied the contents, added some rough values to the missing locations, and burned a new chip.
With this fitted, all self tests passed. Happy days again!
Output Signal Quality
...more to follow....
Comments
Post a Comment