Wireless Set No. 52 Canadian - Project

[James D Teel II]

Our house flooded three times while I was growing up in SW Oklahoma. I can empathize with what you're dealing. I'm glad to see you back on and I pray the damage wasn't too severe.

By the way, I appreciate all the help you've given me with my radio. Check your email for the results.

[David Dunlop]

I am happy to help whenever I can, James.

I was able to get the first set of Sender Resistance Tests reviewed today with my now nicely calibrated RCA VoltOhmyst VTVM, after letting it warm up and stabilize for half an hour.

On the first go back on 10 April 2022, 18 of the 72 tests in this set came back noticeably off specifications and/or ‘odd’. Of those 18 reviewed today, five now came back on specification and the others were noticeably lower in value, or unchanged. So roughly 30% of those 18 readings are now resolved and the rest will need to be investigated.

The Socket for Valve V1J (an ARP-3) has half of its terminals hidden behind the Microphone Transformer T2A, which forms part of the V1J Grid Circuit. For the 10 April 2022 testing, I had chosen to test the V1J Socket terminals from the top, and the probe pins felt a bit too sloppy in the holes, so this time, I decided to remove the T2A Transformer from the lower rear chassis wall of the Sender. This is easily done, as it turns out, by simply removing the two sets of hex nuts and lock washers, as shown in the attached photos. I just let it hang carefully on its leads until the tests were completed from the bottom side of the V1J Socket and then reattached the T2A Transformer to the chassis.

Now onto rechecking the suspicious results for the second test.


David

[David Dunlop]

It was a stressful, but productive evening yesterday. I ended up spending it with my ears tuned to the sump pump in the furnace room, listening to hear if it would kick in, while my eyes were on my VTVM, getting the testing completed for the second Sender Resistance Tests.

I should take the time here to clarify a few points. All three resistance tests require the valves to be removed from the Sender chassis. The first test was a basic pin by pin check from each valve socket to ground, with, from time to time, the Mode of Operation Switch and the Power Switch in specific positions. This second set of resistance tests is taken from each of the terminals on the two 8-pin Connectors on the upper rear of the Sender, and the 7 terminals on T3A, the Modulation Transformer for the V5D Plate. Again, all readings are to ground and from time to time, the Mode of Operation Switch, Power Switch (and this time) Sender Meter Switch must be in specific positions. A total of 52 tests are involved in this set.

On the first work through, I found 21 results that were suspicious, so last evening, I focused solely on redoing all of these with the VTVM and randomly spot checking a few of the others to see if they were reliable. When all was said and done, two of the 21 tests reexamined now produced good, normal results and are now off the ‘Red Flag List’. For the remainder, all of the ‘Infinity’ readings disappeared and I got good solid resistance values back for my efforts. These are all higher than they should be, and a few are alarmingly so, but it is all data I can work with now.

It is still too early to know exactly what is going on in this Sender, but one interesting observation so far is that when the Mode of Operation Switch is in RT Mode, this seems to produce more than its share of very high readings, when compared to CW, M.C.W. and BREAK IN Modes.

In the meantime, I will try and get the last Sender resistance Test redone today.


David

[David Dunlop]

I finished rechecking the third and final Sender Resistance test this morning. All is now quiet on the Low Pressure Front.

This set of tests involves testing resistance from selected pins on the seven Sender valve sockets to just three terminals on either of the two 8-pin connectors on the back of the Sender. Some of these tests might require specific settings for the Band Selector Switch, Mode of Operation Switch or Power Switch.

22 tests make up this set and in the initial work through, I had 7 suspect results. Of those, three now came out spot on and the remaining four read excessively high, but not as bad as the first round. Interestingly, these four readings all come from the V1J socket for the Voltage Amplifier.

So now that I have believable data, the next step is to look at the actual circuits on the schematics to see what components relate to each test in the three sets of results, and start looking at those parts on an individual basis.


David

[David Dunlop]

I had my fingers crossed there was logic behind the layout of the three sets of resistance tests (there are no guarantees with electronics design) and I was happy to discover the testing sequences in each test indeed follow the Sender Block Diagram, starting at the front end with the Master Oscillator Tuning Circuit built around V5A, a 6V6G. They then added in V6A (the VR 150 Voltage Regulator) following into the Amplifiers V5B and V5C and arriving at V7A (the impressive 813 Power Amplifier). Testing then moved to the Audio Input Circuit V1J (ARP-3) and V5D, which ultimately end up at V7A as well.

Starting with the first test set, the first thing I noticed was I had good results through the first four blocks of the Sender (V5A, V6A, V5B and V5C) with the exception of Pin 1 on V5A, V5B and V5C.

A quick look at results on the third test set showed good results across all these valves as well. The problem showing up with the results on Pin 1 was that each test to ground should have produced a reading of 1 meg worth of Ohms. My readings were 1.50 meg taken from the top of the sockets; a little too high for my liking. From the bottom side of the chassis, access to the V5A socket is completely blocked and parts of the other three are tricky to get at as well. The real annoyance, however, was that Pin 1 on all of these valves is an open pin, never used by the valves and the designers take advantage of this feature to use the Pin 1 terminal on the socket as a connection tag to solder leads and wires from other components to, in order to complete other, often unrelated circuits. The issue with this is that when you look at the circuit schematics for any one of these valves, connections to Pin 1 do not show up. They are blank. The only way to find out what is connected to them is to actually look very carefully at the actual terminals on the sockets and hope the connections can be easily traced to identifiable components.

So before being able to dig any deeper into the resistance test results, I needed to get a better look at what was going on underneath the V5A socket. The attached photo shows what this section of the chassis looks like from the bottom, covered with a heavy metal plate with the Modification Card Holder mounted to it.

David

[David Dunlop]

I tend to be very careful when taking on something new, that I know nothing about and gaining access to the metal box hiding this circuit is a prime example of that action. I had no idea at all what was inside, or where and how it was mounted. I could see an access hole in the lower rear chassis wall for purposes of tuning the coil, the V5A Holder (socket) on top of the lower chassis and that was about it There also had to be a switch lurking inside somewhere for purposes of selecting Crystals.

The first discovery was this piece of metal did not exist in the Illustrated Parts List at all, first or second Issue. It was not a cover, a plate or a shield. Not an encouraging start. I then thought about the existence of the Modification Card Holder fitted to the bottom and wondered if the installation instruction for the holder might shed some light on things. It did. It was called a ‘Plate’, four screws held it in place (2 top and bottom in the photo in the previous Post), and the Plate was to be removed carefully.

After looking closely at how the Plate was installed in the Sender chassis, I could see the outside end was slipped in behind the lower left rear side of the Sender chassis rail, in behind the heavy reinforced rail the Sender slides in and out of the Carriers No. 4 on. The inside end of the Plate just slipped in beside an adjoining chassis compartment wall. So, with the Sender resting on its right side, putting the heavy rail at the top, I removed the four retaining screws and lock washers. The plate was a snug fit. I had to very carefully insert a slot head screw driver blade just into the open slot either side of the Plate at each lower side and ease the Plate out enough to be able to grab the lower edge with my fingers and pull it free, but not out. While holding it in place, I then slid the screwdriver blade carefully up the slot to the top of the Plate to keep it flush with the chassis rail. If this end of the Plate slipped inwards, it hung up on something unknown. Once you can clear the inside end of the Plate from the partition wall it attaches to, you can move the lip of the Plate inwards enough to allow the upper end tucked behind the chassis rail to drop free, and remove the Plate.

The four photos here illustrate the location of the four screws the Plate is mounted to the Sender Chassis with, what the bottom of the Plate looks like once removed and what the inner (top) of the Plate looks like, full of dirt and dust bunnies.

The real surprise in this exercise was the discovery of the hole, roughly centered in the Plate, which is effectively blocked by the Modification Card Holder. My initial though was it being another access hole for tuning something. NOPE! Turns out it is a ‘finger hole’ thoughtfully designed into this Plate by Canadian Marconi to make removing and installing this Plate an easy task to accomplish. My guess is that any surviving Senders in the UK will still have an easy time working with this Plate. Canadian owners, not so much.

David

[David Dunlop]

Finally, my first look inside this section of the Sender chassis. What struck me was how clean everything was, with little or no dirt and most important, no sign if heat discolouration or burnt smells lingering about.

The most notable thing was the pristine looking L31A COILS, RF sitting front and centre, complete with its CMC Part Number stamped on the side of it – CMC 119-209. I am not sure what prompted me to do it, but I decided to run that number against the Parts List at the back of the Operators Manual. Did that ever turn up the excitement!

According to said Parts List, L31A had Part Number CMC 119-107. The number stamped on the assembly in my Sender was nowhere to be found. So next stop was the 1945 Illustrated Parts List. It showed L31A as being, COILS, RF, 1.75-4-mc, MO, No. C1. It had a 1-1/32 inch diameter with two riveted 6-32 mounting screws at one end and no sign of a central, adjustable core. The ZA-Number reference given was ZA/CAN 4355.

A pail of tea later, I decided to check the 1948 Issue of the Illustrated Parts List for a possible explanation. Bingo! The same reference for ZA/CAN 4355 was found, but this time, directly below it was a second item, ZA/CAN 4221. It was a stink smaller diameter and was fitted to its mounting Bracket, that also held the central, adjustable iron dust core. So this complete assembly bears the CMC 119-209 Part Number.

My thought is the original design, in theory, allowed for replacement of either the coil, or the core, but the amount of work required to do either, was more than making the Coil and Bracket a single item replacement. Interestingly, the Bracket itself does not show up at all in either the 1945 or 1948 Illustrated Parts Lists. Having confirmed at this point that all the components in the Master Oscillator Tuning Circuit were replaced during the Senders last overhaul when in service, I then traced the connections to Pin 1 on the V5A Holder (socket) and found nothing worth a red flag. To be on the safe side, I retested the resistance of this Pin circuit to ground after getting a good clean contact point on the pin and now found the reading to be a steady 1.25 meg. I liked that reading and took the time to recheck Pin 1 on V5B and V5C in the same way. They both now showed 1.25 meg as well. So I was glad in the end to have opened up the shielding plate for this circuit and had a closer look.

The surprises were not quite over, however.

David