Homebrewing an 8877 Linear Amplifier, junkbox style

[w8khk]

Per Tom's (K1JJ) encouragement, I will take a shot at documenting one of my recent projects.  A few months back I decided to use junkbox parts to build a new 8877 linear amplifier.  The only parts purchased for the rig were the tube, the socket, and the plate inductor. The inductor was a find at Dayton in 2009.   Maybe also a few wire ties, terminal lugs, etc.  Nothing else major was purchased for the project.

I built the rig in a 19 inch rack that formerly housed a Hewlett Packard 7976A 6250 BPI vacuum column tape drive.  Many of the other parts wer scrounged from scrapped computer hardware.  For example, the blower motor and squirrel cage fan were taken from a scrap HP 7925 disc drive, 120 MB, the size of a typical clothes washing machine.  The 5 KVA 130 pound isolation transformer and power distribution unit was from an HP 3000 series 44 minicomputer.  The 230 pound plate transformer was from a scrap Temco 5KW shortwave broadcast transformer.
The HV oil cap was MARS surplus, Davis Monthan AFB, Tucson, AZ, it is 280 UF at 5 KV.  NO PCBs!  Incidentally, the yellow circuit breakers on the bottom left corner of the power control panel were taken from a B-47 in the Davis Monthan boneyard prior to it meeting the aluminum shredder!

I will attempt to present a picture story of the construction of this amplifier.  All stainless 6-32 and other hardware was saved while scrapping obsolete HP computer hardware.  Most of the wire harness was originally cabling in an HP 3000 minicomputer.  The wire ties were carefully removed, and were long enough to re-use for most of the new harness.  I like to use terminal strips and lugs instead of plug and socket connectors.  This makes it easier to troubleshoot and modify over time.


The first photo shows the front view of the top of the rack.  Top is the RF final, next is the power control. Not yet present in this photo is the small panel is the metering, T/R switching sequencing, and overload control. It will appear in separate photos later.

 Second photo shows the bottom of the front without the dress panels.  Here you can see the power distribution box, 5KVA isolation transformer, and the main filter capacitor.

The third photo is just a front view of the RF deck.  This front panel view has no lettering, but the full rack view has Brother P-Touch labels applied.

[w8khk]

The first photo appears like a typical kit.  All the separate parts and panels punched and ready to assemble. 

The second photo shows the plate tuning and loading capacitors mounted on a bracket ready to install in the RF deck. 

The third photo shows the back of the amplifier, partially assembled.  You can see the recycled fan motor and squirrel cage blower in the amplifier plenum (used Bud MiniBox) on the left.  The large hole is clearance for the plate of the 8877 tube.

[w8khk]

The 8877 needed a socket.  While at Dayton in 2009 I found some scrap PC boards that included sockets suitable for 829B or 4-65A tubes.  I bought several for one dollar each.  

In the next two photos you can see how the socket was modified, using the original tube retainer springs on an inverted socket to make the grid grounding hardware.

In the first picture, the original socket is on the left, and the modified socket is on the right.  The filament choke is in the background on the left.

The second picture shows the 8877 tube installed in the modified socket, with the PW Viking II in the background.  That one followed me home from Huntsville, along with a Ranger and a Valiant, in 2008.

[w8khk]

The first photo shows the plenum with the top cover and blower removed.  Here you can see how the 8877 socket is mounted.  The blower pulls air in from the bottom, pressurizing the plenum.   Air exits through the plate radiator and through the socket to the underside of the plenum.  A piece of G-10 glass epoxy board with no foil is cut to cover the tube plate hole, slightly clearing the circumference of the tube radiator.

The second picture shows the plenum cover and blower installed, with the tube in the socket.

The third picture is more complete, with the glass epoxy board around the tube, and the plate choke and bypass caps installed.

RF enters through the BNC connector on the left.  Filament power enters via the terminal strip.  Two terminals on the strip are reserved for metering the filament voltage at the socket pins.  The HV safety connector is on the right side of the plenum.

[w8khk]

The first photo shows an incomplete view of the power control section.   The large variac controls the plate voltage.   It is a 120 VAC variac.  The plate transformer is set up for 240 on the primary.  Via a panel switch, the transformer may be fed with 120 via the variac, referenced to neutral, or one side of the transformer may be connected to one side of the 240 line while the other side is fed via the variac.  Thus it is possible to provide 0 to 240 volts to the transformer with a 120 volt variac.  The smaller variac is used to control the fan speed.  A series variable resistor is used to adjust the filament voltage on the 8877.  

A high quality high-current rotary switch (upper Left corner of panel) was salvaged from a Tektronix 175 High-Current Adaptor for a Tek 575 Transistor Curve Tracer.  This rotary switch enables a series of diodes to be switched into the cathode circuit, providing for continuously adjustable cathode bias.

The other pictures show two views of the power control panel, including the relays.

The original large variac was designed to turn the entire calibration dial.  There was no knob on the dial.  I modified the dial drive, so that the calibration dial is mounted stationary on the panel.  I then made a flange to mount a knob.  Having found no suitable knob, I was looking at a jar of machine screws on the workbench.  A Skippy peanut butter jar.  The lid was just the right size for the knob.  It was blue, that would not do.  I spray painted the knob black, and mounted it to the brass flange I made for the variac shaft.  In the full front view pictures in the earlier posts you can see the finished knob.

[w8khk]

This photo shows the completed metering, sequencing, and overload control section.  Again, scrap aluminum for the chassis and panel came from obsolete HP computers.  The power transformer was from an HP 2645A terminal, circa 1975.  It used an 8080A processor.  

The front panel screwdriver adjusted pots are used to set the meter calibration and overload threshold.   This is all done with low voltage on the deck, using a test power supply, and a calibrated digital meter in series with the meter circuit.  The filament voltage meter is calibrated while using a true RMS meter attached to the sense terminals on the back of the 8877 plenum chassis.

T/R switching is handled on this chassis as well.  I always leave room for growth and modifications.

[w8khk]

In this first picture, you can see the back of the rack, with the plate transformer and oil cap on the bottom shelf.  On the side of the rack is mounted one board with the HV rectifier string (full wave, center tapped, with surge resistors) and meter multiplier.  Another small chassis has the  string of bleeder resistors.

The second photo is a close-up of the rectifier and meter multiplier assembly.

The third photo is a close-up of the bleeder string, before wiring.

[w8khk]

This photo shows the rectifier string and bleeder stack wired, with the protective plexiglas shields.

[w8khk]

Here are three views of the back of the rig, wiring complete.  The HV lead to the RF deck is not shown, but all other harnesses are present.

[KL7OF]

Wow Rick...Very nice...Gotta like Skippy peanut butter parts...

[flintstone mop]

How much power to the people are you gonna have available? Your P.E.P.
Drive needed for whatever power the tube and P.S. is capable of?
Has this been fired up? The construction is pretty solid and looks like it's ready for some juice.
I do not understand the use of the isolation transformer. It is actually isolating the AC mains? Or is it re-purposed in the electronic circuits as something else?
You do some nice work there. The amp should last forever!
Fred

[W2VW]

Very nice. Thanks for posting.

[w8khk]

How much power to the people are you gonna have available? Your P.E.P.
Drive needed for whatever power the tube and P.S. is capable of?
Has this been fired up? The construction is pretty solid and looks like it's ready for some juice.

I am attaching a more recent picture showing the completed rig, front view.  In the shelf is a Kenwood TS-830 slop bucket transceiver.

The rig is intended to do legal limit.   Power transformer and other components are much larger and heavier than the newer hypersil products, but the cost was not an issue, other than hauling them from NJ to FL then to GA.  No worry about stressing components if running extended key-down tests into the dummy load.  The variac allows me to use the power supply for this rig, or the 4-1000A, or other rigs as needed.  The rig has been operated and has not presented any problems.    The tornado that struck my neighborhood several weeks ago took down my antennas.  The trees nearby need some major trimming before I will be able to get the OWL dipole back up where it belongs.

[w8khk]

I do not understand the use of the isolation transformer. It is actually isolating the AC mains? Or is it re-purposed in the electronic circuits as something else?
You do some nice work there. The amp should last forever!
Fred

Fred, the isolation transformer is used for several reasons.  The room where I operate used to contain a complete HP 3000 minicomputer system.  This required either 208 or 240 VAC, single phase, and a 30 amp receptacle.  It used a ground and the two phase wires, but no neutral.  The isolation transformer was part of the computer system, and it also provided 120 volt receptacles for the peripherals.   

I will attach a couple photos of the transformer cover and internals here.  You can see that it provides two 120 VAC output windings, which may be placed either in series or parallel.  The input windings can be connected in various ways to comply with several input voltage ranges.

Some of the rigs I run use 120 volts, others use 240.  For example, the Hallicrafters HT-33A requires 120 volts, but does not like a long run of 15 amp wiring from the main panel.  I have a separate power distribution box that plugs into the single 240 volt 30 amp receptacle, that contains a line filter, surge suppressors, and an identical isolation transformer, breakers, and various receptacles.  This unit, and the self-contained transformer in the 8877 rig, allows me to have a balanced load on both sides of the 240 volt line, no matter what loads I place on the iso tran.  A single ground to neutral bond at the load side of the transformer ensures minimal voltage between ground and neutral for all attached devices.   While code allows a ground to neutral bond only at the entrance panel after the meter, and does not allow a ground to neutral bond on sub-panels, in this case it is considered a separately-derived system due to the transformer isolation. 

The major advantage to the transformer in my specific case is it eliminated the need to replace the three-wire circuit with a 4-wire circuit in the shack quite a distance from the panel.  The transformer and distribution box provide a single point ground, a balanced load on the two phase legs, and minimal voltage drop under varying load.

[K1JJ]

Rick,

I like the finished rack pic -  and the  8877 chassis parts assembly shot.  Orderly layout and work.


So you have an 8877 linear amp and a 4X1 plated modulated rig by 4-400A's IIRC.  Any pics and details on the 4X1 rig?

T

[w8khk]

So you have an 8877 linear amp and a 4X1 plated modulated rig by 4-400A's IIRC.  Any pics and details on the 4X1 rig?

Tom, I am glad the roller inductor is working out well for you.  Can't wait to hear it on the air! 

The 4x1 rig is not yet completed.  The RF deck is done, but the power, modulation, and control are not yet started.  I hope to get back to that soon.  I recently enclosed my carport, now I have a large garage, and have been doing restoration on three Nissan Maximas.  This weekend I am rebuilding the entire front end of my daughter's Nissan Altima.  I hope to get back to ham projects soon.  My antenna went down during a tornado a few weeks ago, and it will be a project to get it up again.  

The only rigs airworthy right now are the Viking II / 51J4,  32V3 / 75A3, Kenwood TS-830 with the 8877, and the modified Valiant / RME-69.

Long-term projects include restoration and conversion of an RCA BTA-1MX, dad's PP 304-TLs modded by four TZ-40s, and my grandfather's PP 250-THs modded by a pair of 810s.  Also have a couple SB-220s to restore, but they are at the bottom of the list.   Not bored looking for things to do at the moment!

[W7TFO]

Good job.

I love it when the meters match---but I'm picky that way

73DG

[WZ5Q]

Excellent Job Rick!
It's always good to see acute attention to detail not to mention all the great engineering practices implemented in a project. You've given me some ideas to add to my own build.
Thanks for taking the time to post your creation and the pics!

[flintstone mop]

Excellent Job Rick!
It's always good to see acute attention to detail not to mention all the great engineering practices implemented in a project. You've given me some ideas to add to my own build.
Thanks for taking the time to post your creation and the pics!

I'll have to toot my horn also.
The explanation of the isolation transformer was excellent. The typical Ham may not have access to one of these transformers. It allows mixing 120 and 240 without separate runs from the panel.
Maybe a picture of the watt meter at full monkey would be nice to see. The 8877 is a great tube!
Fred