Hi STC throughout the years around 1960 onwards I had the opportunity to use both selenium stack rectifiers for voltage and current rectification and then using the new silicon rectifiers in their place.
The most obvious difference with silicon was the increased efficiency in rectification giving higher voltages, less heating and less ripple.
The silicon rectifiers exhibited lower forward voltage drop and less reverse current leakage.

In general the stack rectifiers had a medium life in hours say 10's of years and could either punch through or just lose efficiency with rising forward and reverse leakage. 50 year old stack rectifiers could be just like a leaky capacitor and have almost no directional effect at all.
When I replaced stack with silicon I always got more voltage on load, that is the "forward resistance" was lower.
In some cases I had to include a series resistor to reduce the load voltage.
That was about the only practical thing required.

Silicon rectifiers from GEC when selected properly by voltage and current rating were bullet proof.
I used many SCR and avalanche diodes in inductive motor control circuits and they never failed.
When other suppliers took over the market I started to get failures, proving that the original designers could make cast iron products and once you started shaving corners for profit the reliability reduced.

Coming back to replacing valve or stack rectifiers with silicon, you have more voltage and less ripple on load and may look at inserting some series resistance to arrive at the rated load voltage on load.
Fred.

Thanks for all of that Fred.

I found a note that the bridge rectifier was 5D133 or GEX34 on the parts list and has 4D133 marked on it. The other was 5D3. Couldn't find anything on them on the net. Replaced the seleniums with 1 amp 400 or 600volt silicons.

Need to revisit this tester with your information. There are various resistors in series with and parallel to the rectifiers and will take some nutting out as to what extra to put in there. One thing that occurs is that given the range of valves that can be tested and variability currents involved. might be difficult.

Just revisiting this: You changed what I am assuming to be a metal rectifier, to Silicon.

Is there now an issue due to over voltage? Silicon rectifies are much more efficient?

Marc

Re over voltage with silicon rectifiers, I found information on the net at the time several years ago, probably on the British vintage-radio forum but maybe a German source, that said that this modification did not need any extra resistors. I have wondered about this, but given that there are already current limiting resistors maybe it is right. However I suspect that the voltage drop through the selenium stacks is a lot more than 0.7 volt. If I knew the drop, could just put in a series of silicon diodes to give that drop.

Some of the functions such as element shorts and filament continuity meter measurements aren't made - the fact of a deflection is enough. I need to go through the others, especially mutual conductance, and look at the resistors in series as current limiters and parallel for setting voltage, to see what might happen. There was a time 50 years ago when I could have done node and loop current calculations and figured it out. I shudder at revisiting Kirchoff's laws again, as the only electronics I did for work was to design and make a 200Tonne load cell.

It is possible that the way the tester works for mutual conductance ie wind on a potentiometer from zero to a given meter position, then wind back to zero with another potentiometer, then press the test button for the mut. cond. reading, could mean that voltage errors arising from the rectifier changes, mains fluctuations etc would balance out. It would make sense to design it that way to cancel out component drift with time.

Jimb also made this mod. to his 45C. Have you any comment Jim?

I also own a 45c and it is working fine after a good clean and overall check.
At this stage have left the original rectifiers in. But have considered that to make it more reliable and safe for the future users to change the rectifiers to silicon.
There his no regulation of the supply, no electro's.
All parameters are set up for the valve to be tested, so I can't see that a bit of extra voltage will make any difference.
The mains voltage alone could vary as much as the difference between silicon and metal rectifiers.
I'm just waiting for someone else to modify and test
JJ

Please note that I should not have said selenium rectifiers - I was forgetting about Cu - CuO types. Could be either.

Document uploaded to Post 28.

I changed a TV style EHT stick rectifier in a Heathkit OS-1 oscilloscope to Silicon (figures recorded). I had to loose several hundred volts.as a consequence. The change was due to a failure.