Send DX Hints or Kinks to any of the Blog Authors shown below.

Thursday, February 23, 2017

ALPHA 87A Repair for W8OM

It did not take long for a second WVDXA member’s Alpha 87A to show up on my bench needing some TLC. This one is 1991 vintage. It is in nearly mint condition. The main difference with this second Alpha from the first is it is equipped with the auxiliary fan for supplemental cooling of the power transformer. Per owner W8OM, this particular Alpha 87A belonged to W8AH (SK) who used it on RTTY. Al may not have been the original owner. I don’t know how many amplifiers W8AH ultimately had, but there are several still operating within WVDXA. Even I had one, an Alpha 76CA which went from W8AH while he was alive to K8RK (SK) and then to me in 1985. Currently, this Alpha 76CA is making power like new for WVDXA member Frank, KA8SYV, the new owner. These Alpha amplifiers just keep on kicking.

W8OM’s amp presented with an intermittent soft FAULT 20, high filament current. The amp faulted on my bench during the initial warm up timer interval. During the initial 60 minute power up session, the amp generated FAULT 20 about 4 or 5 times, so it happened frequently.

There is a related Alpha 87A filament soft FAULT 18, which is low filament current. FAULT 20 seems to be the more common of the two. This is a well known problem and is usually related to oxidation of the pins on the Molex connectors.

In the last few years, W8OM’s amplifier has a history of generating FAULT 20. Bob had extensive correspondence with the factory and followed their recommendation to clean the Molex wiring connectors at least twice. The fault generation calms down for a while, but always returns. Now it is to the point where the amp cannot be depended on to work when needed.

This amplifier is a work of art and it is designed to be bullet proof. Of the monitored parameters, this filament current monitoring feature seems among the least useful. There is no mistake by the operator or antenna malfunction that can cause the filament to see higher or lower than normal current. There are tube or wiring failures that can occur, but a majority of FAULT 20 problems are related to wiring harness issues and not tube failures; this is what Alpha has said.

Although FAULT 20 is a common Alpha 87A problem, before troubleshooting the current issue, the wiring diagram/schematics were reviewed. The 3CX800A7 filaments operate in parallel from the 13.5 VAC winding of the plug- in transformer. One side of the filament winding goes direct to a common side of the tube filaments, via the wiring harness. The other common side of the tube filaments goes directly to chassis ground in the Tube Deck compartment. The second side of the filament transformer winding (FILAMENT RETURN) goes to chassis but through a series 0.3 ohm 5 watt resistor R7 (located in the Tube Deck compartment). Thus, the common FILAMENT RETURN is lifted above chassis potential by the small voltage drop developed by the current flowing through resistor R7.

The small AC voltage differential between the FILAMENT RETURN common and chassis ground is picked up by a FILAMENT SENSE line running from the Tube Deck chassis ground over to the LVPS board through the wiring harness. All of the filament related lines (FILAMENT, FILAMENT RETURN, and FILAMENT SENSE) are contained in a wiring harness and connect via J3 in the Tube Deck enclosure and J7 on the LVPS board.

The small AC Filament Sense voltage is rectified in the LVPS board and the resulting DC voltage feeds an OP Amp driver circuit, also located in the LVPS. The OP Amp driver outputs a proportional separately derived DC voltage to the microprocessor board. It is the output voltage from the OP Amp driver which triggers the FAULT 20, or the twin filament FAULT 18. Note the FAULT 18 protection renders it impossible to troubleshoot the amplifier with only one tube in the socket, a technique that is occasionally useful in amplifier troubleshooting.

FAULT 20 is triggered by high filament current, greater than 3.5A, which translates to a higher than normal voltage on the FILAMENT SENSE line. The voltage developed across R7 in the Tube Deck board is the primary source, but not the only one. Anything that adds resistance in the FILAMENT SENSE control line between the LVPS and Tube Deck chassis ground adds directly to the resistance of R7 and causes a higher FILAMENT SENSE voltage to be developed for a given current level. This is why the Molex connectors are such a problem; the contacts tend to oxidize, thus increasing the connector resistance.

The prime suspect for causing FAULT 20 is excessive voltage on the FILAMENT SENSE line. There are other potential causes. The tube filament could be shorting - an unlikely event. There could be a cracked or cold solder connection on a tube, especially on the grounded pin side of the socket. Resistor R7 could be increasing in value, thus raising the FILAMENT SENSE voltage. There could even be issues in the OP Amp driver circuit, raising the monitored voltage to the microprocessor. History has shown the Molex connectors are usually the cause of FAULT 20.

The amplifier was carefully inspected. This is a filament fault so focus was placed only on the filament circuits, components and connectors. These components and connections are:

J7 on the LVPS board.
J3 in the Tube Deck compartment.
R7 in the Tube Deck compartment.
The wiring harness between LVPS J7 and Tube Deck J3.
The 3CX800A7 tube sockets.
If no problem were found in any of the above areas, attention would have been paid to the OP Amp driver circuit for the Filament Monitor and the plug in connections from LVPS to the Microprocessor board. Fortunately, the above component focus resulted in a problem resolution.

Resistor R7 was the correct value and showed no signs of overheating. The filament pins on the tube sockets looked good. There was no sign of excess heat or discoloration in the Tube Deck compartment. Molex connector J3 in the Tube Deck compartment was removed. It was immediately noticed this connector did not seem to have a tight fit on the mating pins. The connector was re-installed and an ohmmeter check was made between chassis and the three conducting clips on the back of the plug, FILAMENT, FILAMENT RETURN, and FILAMENT SENSE. The FILAMENT SENSE connection would not read when touched with the ohmmeter lead, unless force was applied to the plug. A slight touch of the ohmmeter lead would cause a good reading on the other two pins. 

Suspicion was high a problem existed within this J3 plug pin connection. It was felt the plug perhaps did not have sufficient clamping force around the pin on the FILAMENT SENSE line. If this J3 contact were marginal between the plug and pin connection, any oxidation of the pin would exacerbate the issue. Any higher than necessary resistance in this connection would raise the FILAMENT SENSE voltage and could trigger FAULT 20. It was decided to thoroughly clean J3 and see what effect this would have on FAULT 20 generation.

A Q Tip was soaked in DeoxIT® cleaner and used to vigorously clean the mating male pins on J3 in the Tube Deck and J7 on the LVPS board. DeoxIT® was also dropped into the female plug connections. Last, each plug was inserted and removed from the mating connector about 20 times each. A power up test of the amplifier resulted in completely stable operation, with no faults, over a 4 hour operating period broken up into three sessions. These tests included RF application, plus repeated keying of the amp via an external footswitch. The amplifier is currently in a stable state and will be returned to W8OM.

If, as suspected, the contact force between J3 plug and pins is less than optimum, this FAULT 20 generation will likely occur again. It might be months or it might be years, but it should happen again. The amp will be returned to service now but if the problem recurs, W8OM knows which particular plugs/pins to clean. If it really becomes problematic, the J3 plug could be replaced. Another fix would be to bypass the FILAMENT SENSE J3 plug connection altogether, by directly grounding the line to the chassis in the Tube Deck enclosure. If the amp would have been unstable after cleaning, this is what I would have done. Another mitigation could be to insert a thin shim between the J3 plug and receptacle halves, forcing lateral pressure on the pin connections. This could improve the connection enough to mitigate the problem. This solution was not tried. Both bypassing the FILAMENT SENSE connection at the plug and shimming the plug/receptacle remain future options, along with replacing the plug, if frequent problems with FAULT 20 continue to occur.

It might be asked, why was the plug cleaning limited to just the filament plugs, when all of the Alpha 87A Molex connectors are known to oxidize with age. The plug connections are fragile, as is the wiring harness. K8YYY’s amp recently had two broken wires in the harness. W8OM has cleaned all the plugs within the last year or two. With no other issues being manifest, it was decided not to disturb the other connections now.

Post written by: Charlie, N8RR

No comments: