AWA OT CW Contest Operating

Last week was the AWA OT CW Contest. I had hoped to be operating using my 815 transmitter but it continues to run off on it's own as a tuned grid-tuned plate oscillator. I switched to my 6J5/6L6 transmitter ( http://www.io.com/~nielw/2tube_xmtr/2tube_xmtr.htm ) and a NC-101X ( http://www.io.com/~nielw/nat_list/nc100.htm ). The pair worked well together and I had four enjoyable QSOs including WA9QNN, W0LGU, WU2D and AA4RM.

More on my 6L6/815 Transmitter

Next Wednesday, January 23, is the AWA Linc Cundall Memorial OT CW Contest. It is about time I get my 6L6/815 Transmitter on the air. Two big stumbling blocks had to be worked on: 1) three different high voltage requirements, 815 B+ around 500 VDC, 815 screen around 250 VDC and oscillator B+ around 150-200 VDC and 2) a tendency of the 815 to run off on it's own as a tuned grid-tuned plate oscillator.

The original transmitter design called for voltage dropping networks to get the right screen and oscillator voltages. In the end I added a VR tube to even out the voltage to the 6L6. In the 1941 design the 6L6 plate voltage would drop from 500 to 250VDC at key down. The oscillator would over drive the 815 as the voltage dropped, creating harmonics and spurious RF output. It was bad enough that the SWR would bounce up at the beginning of each dit or dah indicating RF temporarily not on the design frequency. The VR 150 at the left rear of the chassis along with a 10K resistor network underneath now keeps the oscillator B+ at 150 VDC, key up and key down. I found the 815 screen voltage to be more forgiving than the 6L6 plate voltage. There I added the recommended power resistor network to drop the screen voltage to an acceptable level. Neither of these solutions are elegant. Between them they waste 25 watts of power but they do work.

The instability of the 815 amplifier is still a problem. A tin can (actually Wolf brand chili can) shield around the 815 grid coil helps a lot, especially if the final is lightly loaded. I still have to tune up carefully, though, or the output looks pretty trashy.

815 Transmitter Postmortem

After weeks of vacation, etc I finally got back to my 815 transmitter. To recap ... the major mods since the original design shown in the ARRL handbook and my blog entry dated March 19, 2007 have been to 1) add a loading capacitor between one side of the output link and ground 2) Replace R4 with an RFC to a 200-300 VDC supply and 3) replace the 815 screen voltage dropping network R5/R6 with an RFC to a 200-300 VDC supply. The final layout is shown in my May 28 blog entry.

Tests on 40 meters have been a little disappointing. The 815 runs at only 35%-40% effeciency into a 50 ohm load. 50 watts input gives only about 20 watts out. (whoops! I checked my Drake watt meter agains a Bird. Turns out that the Drake watt meter is reading way low. The Bird says I'm putting out about 30 watts. This is about the power output I expected.) Also the transmitter is not real forgiving. If misadjusted, it goes into self oscillation or generates a trashy sounding signal.

This one will go to the back of the shelf. (but after discovering the watt meter error, only while it waits for a power supply)

I've powered my new transmitter up on 40 meters.

For use during this test phase I built an adapter to allow the use of B&W 5 pin JVL transmitter coils instead of winding coils on my National coil forms. I'm also temporarily connecting the bias and voltage dropping resistor networks to the back 12 terminal barrier strip. This way I can easily play with the resistor values to get the right voltages.

The transmitter loads up to 40 watts input and 30 watts output with no sign of self oscillation

I've still problems, though. The 815 screen current won't come down. It appears to draw around 60 mA rather than 15. High screen current causes the screen dissipation to be exceeded, maybe damaging the 815. Right now I'm looking for another 815 to compare readings with.

815 Xmtr parts

Here's the collection of parts needed to build my transmitter (the power supply is separate). I'll mount the male 8 pin plug on the back of the chassis for power. The barrier strip is for metering points and for the key connection. I'll be able to connect in a plate modulator here also. In the final transmitter the two 25 watt pots will be replaced by wire wound resistors.

Additional 815 xmtr Mod

Another feature missing from most pre-WWII transmitters is some sort of spotting capability. Two reasons drive the need for spotting capability today. First, with transceivers being the norm, operators rarely tune the band looking for a response to a CQ. If answering a CQ, you need to be close to on frequency. Operating in an open spot on the band is the second reason for a spotting switch. Choosing the right operating frequency/crystal requires spotting capability.

To add a spotting switch to the 815 transmitter I'll try a DPDT switch connected in the cathode circuits of both the oscillator and the final. In the "spot" position it will key the oscillator just as the key and also open the cathode-to-ground connection on the 815. The crystal oscillator should operate normally letting me locate myself in the band but I'll not have 50 watts of RF overloading my receiver.

Another layout

Here's another thought on layout ... move C2, the final plate tuning cap, from on top of to under the chassis. This frees up space top side and there is still enough space below for the parts around the 815. C2, a major shock hazard, is now safely under the chassis. I also found ceramic plate cap connectors for the 815. L4 is still exposed with voltage on it but it is set back from the edge fairly well and it is opposite the crystal socket.

815 Xmtr - Safety mods

Someone asked me about modifying a vintage design in the name of safety. Sounds like a good idea to me. I already modify vintage receiver projects to keep B+ out of the headphones ( see http://www.io.com/~nielw/onestep/onestep.htm ). I also try to keep transmitter B+ away from accidental touching. This design, unfortunately, has 500 VDC exposed on the final tank coil and C2. I plan a couple of things to to help. First a front panel. I'll probably use a piece of masonite painted wrinkle black. As a bonus this gives me room for a plate current meter should I decide to add one. The second modification is to move the crystal socket to the front panel. No need to be reaching around behind the front panel to change frequency. The rest needs to be handled by being careful. Obviously power down when changing bands and short the tank coil to ground before touching it. Not so obvious is the key. Cathode keying allows one side of the key to float to B+. Usually one side of a key is more protected than the other. Connect this side to R1 and don't diddle with the key contacts with power applied.

815 Xmtr Parts Subbing

As I search for parts, the topic of subbing parts has come up, particularly for C1 and C2. Miniture 140pf dual section variables are called out for both of these. In my junque box I fine slightly larger (physically) parts and, for C2, only a 100pf dual section variable. I can probably work around the 100pf problem by adding two fixed 50pf caps to L4, a 80JVL plug-in coil. The physical size increase appears to be OK for C2 but C1 is going to cramp the wiring around the oscillator stage. Can I shift C1 to tune L2 rather than L3? That reduces me to needing only a single section variable. Unfortunately is also means that I need to insolate "C1" from ground. The mounting starts to take up any space that I saved by changing to a single section cap. Looks like I'll stick with the parts I've found and just carefully lay out the oscillator stage in the space I have.

815 xmtr

I'm to the point of starting to layout my 815 transmitter. Here are a couple of figures from the 1943 ARRL handbook describing this transmitter along with a picture of my preliminary layout. My chassis is 3"x 8"x12" so I have a little more room. Good thing, since I didn't have the small size variable caps called out in the original design. I'm also allowing room for a VR150 (seen on the back left corner) in case the oscillator stage needs regulated B+ for stability. The original design used only 3.5MHz crytals to cover 80, 40 and 20. I'm going to use 3.5 MHz crystals on 80, 7MHz crystals on 40 and 7MHz crystals on 20. This leads to one change. I'm adding a switch across L1/C3. This will be used to short out the cathode coil/cap when I'm running the tri-tet oscillator straight through.

815 transmitter

According to the 1942 ARRL Handbook the 815 requires less then 2/10 of a watt of drive. A 6L6 is rated for much more than that. The 6V6 has the same base pinout but lower rating. I'll switch over to a 6V6 for the oscillator stage. My power supply will deliver about 500 VDC for the 815 but I don't want to run the 6V6 at 500 volts. I'll need to somehow drop 500V down to about 150 VDC for the oscillator stage. Depending on current requirements a VR150 and a dropping resistor might work. I don't want to use just a dropping resistor since then the 6V6 plate voltage could go as high as 500VDC with the key up.

6L6 - 815 three band xmtr

I've always been interested in the 815 based 50 watt transmetter that showed up in QST and the ARRL handbooks just before WWII. The 815 is a dual "beam" tube designed for UHF (150 MHz in 1940). I'm starting to plan a tramsmitter using an 815 as a push-pull final (as shown in Feb '41 QST) driven by a 6L6 tri-tet oscillator such as used by Millen in his 6L6-807 rack moount transmitter.