K5DKZ/dx60c.html

A DX-60 CLONE

Here is a DX-60 built into a DX-35 case. It uses a neutralized 6146B in the final. I run 800vdc on the final plate and 100vdc on the screen. Under a full load of 180ma, the plate voltage drops to 740vdc for an input of 133 watts. Power output is over 100 watts on 80, 40 and 30 meters; and about 90 watts on 20, 15, and 10. That is pushing the final close to its maximum limit. It has been running that way for several weeks now without problems. Since this is a CW only transmitter, the final may last as long as I don't key down for an extended period. Then it does not have to be loaded to 133 watts input. It works very well at 90 watts input with over 50 watts output.

The front panel looks much better with the black accents
This project started with a very heavily modified DX-35. I rescued this transmitter from the trash about 20 years ago. It was a mess. The mods were poorly done. The cabinet had been repainted several times. After sanding, the cabinet showed signs of at least four colors. Grey, yellow, black, and dark grey.

When I got it the case was painted black and the front panel was a bad shade of puke green. From the overspray on the parts inside, it appeared that the final coat of black, on the case, was applied with the case still installed on the transmitter.
The only reason I did not give this mess a decent burial was that it still worked. At least it did work 20 years ago when I rescued it.
After it spent ten years sitting in a hot garage, it quit working. The old can style electrolyics were shot and the overloaded transformer finally died. I began taking it appart. I saved the parts, at least the ones I thought were good, and stored the mess in a box.
Recently, I retrieved the box intending to resurect this abused rig and build something useful. I do not need a medium power CW rig, but the thought of these parts being lost forever overcame my reluctance to create something useful.

My original intent was to restore this DX-35 to as original a condition as I could manage. Before doing that, I did a little research into the Heatkit DX series of low power, tube, CW transmitters. Here is what I found.
DX-20 - Very simple 6CL6 oscillator driving a 6DQ6 amp
DX-35 - 12BY7 oscilator driving a 12BY7 buffer driving a 6146
DX-40 - 6CL6 oscillator driving a 6CL6 buffer driving a 6146
DX-60 - similar to the DX-40 but with neutralized 6146, grid block keying, built-in low pass output filter, and simple pierce oscillator.
The 12AX7/12AU7(6DE7) screen modulators are about the same in all these rigs and provide marginal phone performance, at best.
Power specifications are as follows
DX-20 50 WATTS INPUT
DX-35 65 WATTS INPUT
DX-40 75 WATTS INPUT
DX-60 90 WATTS INPUT

All except the DX-20 use a 6146 final amplifier. The power supply in the DX-60 is a full wave doubler running off a 260vac high voltage secondary. The plate supply to the final in the DX-60 is 625 volts. At 144ma plate current, it comes to a 90 watt input.
The DX-60 is the better of these medium power CW transmitters. Besides a more convenient physical layout and higher power, it incorporates a neutralized final, grid block keying, low pass output filter, and simplified oscillator circuit. The pysical layout of the DX-60 is not just more convenient, it also provides a fully shielded enclosure for the final amplifier section.
There was not much I could do about implementing the superior physical layout of the DX-60, but I could certainly implement all the electrical improvements.
One of the modifications that was done to the original DX-35 was the addition of a couple of 6DE6 single diode damper tubes to allow the use of the full high voltage secondary in a bridge configuration. Evidently this did not work in the long term with the stock power transformer. Even though it worked when I got it, the transformer was shot after I tried to use it some years later. Still, the resulting higher plate voltage is not a bad idea as long as the power transformer can handle such use.
Instead of doing a complete DX-35 restoration, I decided to rebuild this rig incorporating the best of the features found in the Heathkit DX series. I liked the extra shield used in the DX-35 to isolate the final section from the rest of the above chassis components, so I kept that. I rebuilt the rig to the DX-60 schematic including 6146 neutralization, grid block keying, and the pierce oscillator.

For the rebuild I used a Chicago PHC-150 transformer. This is a potted transformer, slightly larger than the stock DX-40 transformer. Its high voltage secondary is rated to deliver 380vdc at 150ma continuous using a full wave centertapped configuration. Although that is only about 60 watts, it is a continuous duty specification. For intermittant service as in a CW transmitter, this transformer is capable of at least twice the 60 watt rating.

The unloaded HV secondary provides 700 vac. A bridge circuit using two 1N4007s in each leg is used as the rectifier. The HV+ is smoothed by four 400mf/300vdc electrolytics in series. The lower voltage B+ is obtained from the centertap of this transformer and is filtered by a filter composed of two 33mf/450vdc capacitors and an 8hy/300ma choke. The HV+ measures 750 volts under a 150ma load. B+ measures 300 volts under a 50ma load. Both loads were in place when the measurements were taken.

The bias voltage is obtained from a small low voltage transformer that has dual 120vac primaries. The extra primary is rectified by a half wave rectifier and filtered. This transformer was located on a scrap of terminal board that was mounted just above the terminals of the main transformer.
The bias supply complication to obtain grid block keying is well worth the effort. When using cathode keying as in the DX-20, DX-35, and DX-40, grid voltage is obtained from the power absorbed by the preceeding stage. If the preceeding stage fails to provide power, the following stage will not have any bias and may draw excessive plate current. The most common drive failure is trying to use the rig without a crystal or with the crystal switch selecting a non existant crystal.
Grid block keying applies a cut-off bias to the grids of all three tubes in the RF string when the key is up. When the key is pressed, cut-off bias is lowered to a proper operating value. Since the tubes are always biased at either cut-off or an operating value, loss of drive to any section in the string will not result in excessive plate current.
The pierce oscillator used in the DX-60 eliminates a couple of capacitors and an RF choke that are needed in the colpitts circuit used in the other DX series transmitters. The circuit description of the DX-40 claims that the colpitts oscillator was used because it is rich in harmonic content. However, only the 80 and 40 meter signals are used because the buffer takes care of multiplying to higher frequencies. Thus, harmonic content from the oscillator is not needed. The pierce circuit is not only simpler but it will also work better when sluggish crystals are used.

The 6146 neutralizing stub is merely an 8 inch length of solid wire (#18 copper will work fine). About five inches of this wire is brought up through a feedthrough alongside the 6146. The other end of the wire is connected to the B+ side of the buffer tank coil. Neutralization is adjusted by bending the wire closer or farther away from the 6146 envelope.
DESIGN PROBLEM SOLUTIONS
METER
The DX-60 (as well as the other DX series transmitters)uses a 0-1ma meter. I was fresh out of 0-1ma meters. However, I did have a 0-15ma meter. Meter shunt design usually requires knowing the full scale deflection current of the meter as well as the internal resistance of the meter coil. After studying the DX-60 schematic, I became aware that both the grid and plate current shunts were designed to work with a 0-2 volt dc meter. The 0-1ma meter had a 2000 ohm resistor in series with one lead, turning it into a 0-2 volt dc meter. I was able to make the 0-15ma meter work as a 0-2 volt dc meter simply by changing the 2000 ohm resistance to 120 ohms. The solution works well and the orginal current shunts were installed per schematic.
40 METER OSCILLATOR COIL

I ended up using a 1/4 inch diameter, ceramic, slug tuned form. It took about 30 turns of #26 enameled wire, scramble wound on the far end of the form so that the tuning slug would have maximum effect. I had to add a 50pf fixed capacitor across this coil to get it to tune 40 meters.
MULTIBAND BUFFER COIL

I used a 1 inch diameter, 2-1/2 inch long bakelite coil form. Seven terminals were installed at intervals along the length of the form and #26 enameled wire was used to wind the coils. The schematic shows this coil shunted by two capacitors. A fixed 120pf is connected from the B+ end of the coil to ground. A 21pf variable is connected from the plate end of the coil to ground. These two capacitors are effectively connected in series across the coil. Capacitors in series behave like resistors in parallel. The effective capacitance of these two capacitors is about 25pf with 21pf being adjustable. This relatively low capacitance requires relatively high inductance to resonate in each band. The end result is a somewhat low Q circuit and guarantees that the peaking capacitor can only be adjusted for a peak in the band of interest. A larger capacitance with an appropriately smaller inductance might cover more than one band at the high end of the frequency spectrum.
My coil ended up with the following turns per band.
10meters - 4 turns
15meters - 10meters + 3 turns
20meters - 10meters + 15meters + 4 turns
30meters - 10meters + 15meters + 20meters + 7 turns
40meters - 10meters + 15meters + 20meters + 30meters + 7 turns
80meters - 10meters + 15meters + 20meters + 30meters + 40meters + 30 turns
These turns were determined using a grid dip meter and installing the 21pf variable in series with the 120pf capacitor across the coil. The final turns above were determined after the coil was installed and the the transmitter operated in the tune position. Details on this are described below.
COIL VERIFICATION UNDER POWER
Final frequency determination was done after the transmitter was completed. In the tune position, the screen of the final is grounded. For this testing, the HV supply to the final was also disconnected.
All tests were done using one 3.59mhz and one 7.066mhz crystal. A better method would have been to use a VFO.
The 7.066mhz crystal was used to peak the 40 meter oscillator coil by adjusting for a peak of grid current. Make sure you find a real peak. The grid current should rise to a maximum and then fall off while the oscillator coil slug is still about midway in its adjustment range.
The buffer plate coil was adjusted so that a peak in grid current was obtained with the peaking capacitor at about half mesh. Coil adjustment of one turn (plus or minus) is accomodated by the use of terminals for each winding. The 80 meter crystal was used to calibrate 80 meters. The 40 meter crystal was used to calibrate all other bands except 30 meters. Calibration on 30 meters requires a 30 meter crystal.
A frequency counter was used to verify frequency as the peaking control was calibrated for a grid current peak at mid mesh of the peaking capacitor.
During this testing, the buffer plate voltage was kept at 250 vdc with the buffer screen fixed at 150 vdc. Under these conditions there was plenty of grid drive in all bands (grid current peaked at from 2ma to 4ma). It was decided not to implement the potentiometer to vary screen voltage to the buffer. This pot is a better grid drive control but there was no place to bring the adjustment out to the front panel without drilling an extra hole. Grid drive is adjusted by detuning as in the other DX series transmitters.
On 40 meters the rig worked fine even before the neutralization was adjusted. However, it was flakey on 10 meters until the amp was properly neutralized.
NEUTRALIZATION
There are several methods to adjust the neutralization. I used the method described in the DX-60 manual. Here is a brief description of what was done.
Plate voltage was remove by disconnecting the HV connection below the plate RF choke. Merely pulling the plate cap will not work because the circuit needs to be in place for the adjustment.
A length of hookup wire was formed into a dual loop pickup coil. One of the loops was placed over the 10 meter section of the tank coil. The other loop was placed over the coil of a grid dip meter.
The transmitter was turned on, set to 10 meters, and peaked for maximum grid current. This was done in the 'tune' position which grounds the final screen grid lead but allows the oscillator and buffer stages to function normally.
The grid dip meter was then tuned to frequency and the plate tuning was adjusted for what would normally have been a dip in plate current. In this case it gave a peak on the grid dip meter.\
Using the grid dip meter as an indicator, the neutralizing wire was moved, first toward the final, then away from it. Doing so varied the reading on the grid dip meter. The wire was then re-adjusted for a minimum indication on the grid dip meter.
The plate tuning was readjusted for a peak, the neutralizing wire was re-adjusted for as much of a null as possible, then left alone.
That is it. After this adjustment the transmitter settled down and operated smoothly on ten meters.

The schematic can be found here. There were no changes needed from the original DX-60 schematic.
Note, when you click the blue word here, the schematic will be presented in a new browser window and may be very difficult to read. Click on the magnifying glass to make the schematic larger. Now it will be so large that it will not fit the screen. Scroll around to the parts of interest. To get back to the project page just close the schematic window (click on X ). You can get a decent copy of the schematic by saving it as jpg and printing it.