Upgrading and modifying an old Dentron GLA-1000B.  Adding 160m i.e. 

The Making Of GLA-1000B PRO

I have an old Dentron GLA-1000B, (user manual here, 2.6MB pdf) never used and bought used and in not working condition some 15 years ago. I decided to make something of it but I had to use what was available here preferably locally. Some ideas I got from SM0VPO and PA0FRI for the work and some I added myself.

Tubes  
The old 6LQ6 were gone but fortunately I had a few PL519 in store. They are bigger than the original tubes and won’t fit under the cover with the original tubesockets. I decided to use original sockets and rise the cover 1cm.

PL519 has thicker socketpins than the original tubes so the original sockets must carefully be enlarged with a small screwdriver pressed in the pinholes. Be careful not to widthen the holders too much!

PL519 are designed for serial heating and has a 40V 0.3 A heater. I put all four in series with 230V and a capacitor of 6uF 250V AC and a small 4W 50ohm potentiometer for fine tuning. This pot is located on the back panel. There will probably be small differences between the tubes, at most about one volt. Make sure each tube gets as close as possible 40V heating.

The original C3, C4, C5 and C6 are left as they are grounding one side of the heating pins.

In the future I will replace these PL519 with EL519 which have a 6.3V heating and add regulated DC for the heating.

PL519 has the cathode on pin 9 so the folie to pin 3 must be cut. When finished pins 1,2,3,6,7,8 are all grounded and heater current is connected to pins 4 and 5. Every cathode has a 20ohm 1W resistor parallelled with 0.1uF 400V capacitor. These resistors generates the negative feedback and can be adjusted for linearity.

Input circuit  
I removed the original pi-filter inputs, as they don’t have the warc-bands. I adapted ideas from PA0FRI and SM0VPO for the input  The 100ohm 50W non-inductive resistor is made of five 22ohm 20W resistors in TO- 220 in series. That was the only resistors I could find here locally. SM0VPO made this resistor of 100 10k 1/2W resistors in parallell which is a good solution too. The original cathodechoke is removed and the bifilar transformer replaces it. The 330n capacitor over the 10k/5W resistor can be a 0,1u but not smaller. The 33nF could be 0.1uF, but again I had this one in my drawers.

Bias  
The old 50W zener was missing and I replaced it with a few 1N5408 diodes in series to ground.  I use 5, but with a decent high-voltage they could be 6 or 7 of them. Look for about 80mA current with no input signal. This construction allows for adjusting the idle current as it is easy to add or remove the diodes My GLA-1000 has an unused contactpair, so the bias could be switched to cutoff in standby. The 10k 5W resistor to ground handles this.

Power supply  
The original transformer is too small for these tubes. It gives only 800V AC secondary and after the rectifiers and filtering 1100V DC unloaded. Loading to a few hundred milliamperes lowers the voltage to 900V.

I replaced all 3 filter capacitors as they were very old. One of them was warming up in the test.  The filtering is better with 4x470uF in series. 150k 2W resistors are connected over every capacitor.

I might in the future replace the transformer to get the maximum power from the tubes.

Control circuit  
I removed everything except the diode and capacitor for the +12V supply and built a new.  The green LED is on all the time. The red LED lights up when transmitting

Fan  
I replaced the 110V original fan with an old 220V fan. I reduced the fans rpm (to make it quiter) with a capacitor of 660n 250V AC (two 330nF/250V~ capacitors in parallell). The air stream seems to be sufficient and the fan is not very noisy.

Metering  
The original meter is a  200uA meter.

I checked the current and voltage metering. The high resistance resistors in the schematic are tuned to calibrate voltage meter to 1500V. For some reason the original shunt resistor R7 vanished in smoke, so I replaced it with resistors in parallell, something just under 0.1 ohm. The current scale is thus also original.

The diodes over the meter are for safety.

The meter lamps are dimmed just a little with the 27 ohms resistor (I love sofisticated dimmed meters, LEDs and lamps…)

Tank circuit
I worked a lot on the tank circuit. At the beginning I was keeping the original coils. I measured the variable capacitors and found out they were smaller than I though.

I wanted to have 160m so more inductance was needed.  I made a coil on a toroid T200-2 and connected it in series with the other coils. It worked fine but there was very little space for the coils, so I removed the original 80/40m coil and replaced it with more windings on the same 160m toroid. The result worked fine. Fortunately the original plate choke seems to work fine on all bands, even 160meters.

Latest mod is going back to use original coils and add a toroid for 160m. See schematic for details and the pictures of the tank-circuit. The 40-80m coil is rised on a ceramic standoff to shorten the connection wires. The toroid is placed between the variable capacitors. So far I don't have separate taps for the warc-bands. I use 14MHz position for 10MHz, 21MHz for 18MHz and 28MHz for 25MHz. Some improvement in output could be gained by carefully chosen tappositions.

The very latest (and probably final) tankcircuit is a new-wound toroid for 160-80-40 and 30m. its wound on a T-200-2 with 16t of 2mm Cu and 21 turns with black tefloncoated wire taken from the  GLA-1000. (black wire running to the old removed mains switch and back to power inlet circuitry. The inductances are 7.8 resp. 5.8 uH and filled the toroid all around. See picture on the schematics. 

Here are a few different toroid- and coildesigns I tried in the tank-circuit.

The taps for 15m and 10m must be moved. The 10m tap is now 1½ turns from the hot end and the 15m tap 3½ turns from the hot end.

I replaced the original switch SW2 with another 2x11 ceramic switch to get more positions. So far I have not taps for the warc bands

The original bandswitch is put in the DL/switch place to connect additional loading capacitors for 160m and (in my case)  for 10MHz work.

The original door-knob 200pF capacitor is switched parallell with C9 on 160m and 80m.

The white things on the switch are 300pF loading capacitors successive connected in parallell to add loading capacitance. The last capacitor in the row is a ceramic 1nF capacitor.

Misc  
I added a box around the tubes (on the pic the trial made of paper...) and turned the fan to blow into the box. This cools the tubes very efficiently.

I added a mains switch to the back plate and used the hole for the other UHF-connector as it is not in use anymore.

Results  
I am quite happy with the upgrade. Key down gives 350-400W out.  At the moment I use an IC 735 which gives more than needed driving power

Future  
There will probably be lots of small changes to this construction. Stay tuned!

Have fun

Staffan
OH2BAW

This is the version with original tank coils preserved and a toroid added for 160m (The quick and dirty method)

The blower blows in to the tube compartment made of stiff paper (sic!)  :) The warmed air comes out from the top. This cools the tubes very efficiently.

This is the latest tank circuit version with a toroid for 160m - 30m and the original airwound coil for 20m - 10m. This works very fine and will probably be the last tank circuit version.

Here are pictures of the modified power supply circuit board. I simply drilled new holes for the capacitors. Fortunately most of the pins hit useful copper, I only had to make 2 jumpers (red) as shown. There are 150k resistors over every capacitor. The 2 green resistors and the white ceramic resistors are all parallelled and replaces the burned out original meter shunt R7 connecting the -B to earth. The caps are 470uF at 400V.