GENERAL

REDUCE HEAT from your Collins S-Line Receiver

Designed to replace 75S- series 6BF5 audio amplifier tube directly. No additional parts to install. Just plug in. Total power consumption is 2.5Watts. It measures 2.1 inches high , 1.0 inch wide and 3/4 inch thick It weighs about 1 oz. (see picture). Delivery is 1-2 weeks.

The AF amplifier tube 6BF5 in the Collins S-Line gets so hot, it turns dust particles into carbon and deposits a coating on the cabinet cover. Today, a power MOS-FET can substitute for the 6BF5. The MOS-FET amplifier shown here is designed to consume 2.5W. The audio power output is acceptable for a normal receiver operation.

POWER MOS-FET for 6BF5

Most N-channel Power MOS -FET made for at least a 200V switching application can be used for this project. An IRF730 was used because I had many of them. An IRF620 is also a very good candidate and available from Mouser Electronics. The common properties are: 1) they are normally “OFF”, 2) start conducting around 3V, and 3) go into the “ON” state near 8V. The “ON”state is simply a saturation condition. Somewhere between 3 to 8V there is a linear region where this amplifier will operate.

DESIGN

The main objective of this project is “low power consumption”. The power consumption should be no more than 6BA6 or 6DC6 to make this device attractive. These amplifier tubes consume about 2W. The maximum power to this device is set at 2.5W. The voltages available from the 6BF5 socket of my 75S-3 are 145V at the plate terminal and 134V at the screen grid terminal. The negative bias of control grid is not compatible with the MOS-FET. A positive bias is needed to operate in the linear region.

The drain current is 2.5W/145V = 17.2mA. It is a very low value for a transistor, which is capable of handling many amperes. This means the MOS-FET is barely turned on. The operating voltage is very near the turn-on threshold voltage, which is around 3V. Regarding the drain voltage requirement, let us look at the load. It is 2.5Kohm output transformer and the linearity of the MOS-FET near the drain current of 17mA may not be very good. A useful range of drain current swing can be no more than +/-50%. That is 17.2mA x 100% = 17.2mA. The voltage swing at the drain becomes 17.2mA x 2.5Kohm = 43V peak to peak. 

I do not think we need 145V drain voltage. Any excess drain voltage is wasted as drain dissipation. I feel comfortable setting the drain at 60V. To get the drain voltage of 60V, the source should be at 145V-60V = 85V. This requires a resistor (R4) at the source terminal. This resistor, like a cathode bias resistor in vacuum tube circuit, will give a negative bias voltage to the gate. That is OK, because a large positive 134V is available from the screen grid terminal.

COMPONENTS VALUES

The circuit is shown in the picture. The sample picture has a red LED in series with R4. Source resistor R4 will produce 85V at the source end. The R4 value is 85V/ 17.2mA = 4.94Kohm and the power rating is 85V x 17.2mA = 1.46W. For R4, use three 15Kohm1/2W in parallel. The MOS-FET has shifted 85V of excess drain voltage to this resistor, thus reduced its own power consumption by 1.46W. The gate voltage is 88V, 3V higher than the source. When the gate is forced to assume 88V, the source has no choice but to settle at 85V by passing 17.2mA through R4. 

This is an auto bias scheme. We don’t have to know the exact gate bias voltage, as long as it is near 3V. The current though the R4 will automatically adjust to get the right voltage to the gate. The input capacitor C1 blocks DC voltage and passes an AF signal. The R1 and R2 provide gate bias voltage. R3 is there to ensure stable operation. 

Any value near 1Kohm is OK for R3. Since there is no DC loading, a simple divider will provide the gate 88V from 134V. R2/R1 = 88V/(134V-88V) = 1.91/1 = — 180Kohm/ 91Kohm R2: 180Kohm and R1: 91Kohm Higher values of R1 and R2 will give less loading to the AF voltage amplifier 6TA6. C1: 0.01uF / 200V AF coupling C2: 1 to 5uF / 200V AF bypass A 7pin miniature tube was used for the base. When you break the tube, you should be very careful not to damage the base part of the tube.

CONCLUSION

Finished device met my original goal without any modifications. At higher volume levels the sound distortion becomes obvious. It probably is due to poor linearity of the region where this MOS-FET is operating. No attempts were made to find better MOS-FET for this application.

May 25, 2002
Don Kang

Price: $20 ea for USA address shipping and all others S25. Above prices include shipping cost.

Send payment to: Don Kang
15520 On Orbit Drive
Saratoga, Ca 95070 USA

Postal money order, cashiers check, cash and international money order are accepted. 

Any questions, email to: donkang@ieee.org

When you put S-line radio in upside down service position, the PTO tube may hit the table top. The tube can be damaged. Build your own low profile 6AU6/7543 This is a fun project. Only 4 parts are used. The circuit, prototype, and finished unit are shown in the picture.

N-channel J-FET operates like a vacuum tube. For T1, N-channel J-FET 2N5485 or 2N5486 is selected. 2N5486 is a higher Idss version of 2N5485. Both are made for RF applications. Other J-FET, 2N5457, 5458 and 5459 made for AF/low frequency applications also work in this circuit. Again their differences are Idss. Idss is a drain current measured at a pre determined drain voltage when gate is shorted to the source terminal. 

With J-FET made for RF, the PTO frequency goes up about 2khz, on the other hand with J-FET made for AF, the frequency goes down 2-3khz. These frequency differences are coming from the junction capacitance of the J-FET. T2 works as a drain voltage extender. The J-FETs mentioned here do not have high drain voltages. High voltage NPN transistor 2N5551 or MPS-A42 is good for T2. The Zener diode Zd sets T1 drain operating voltage. It can be 8-15V. I used 12V 1/4W Zener diode.

A typical “R” value for 2N5485 is 200 ohms. However, because of wide range of Idss, other J-FET requires resistor value adjustment. A 500-ohm variable resistor will be used to adjust the “R” value. Connect an oscilloscope at the VFO output connector on the radio chassis with a 100-ohm shunt resistor. Set 500-ohm VR for 1.5V peak to peak output reading on the scope. Output of as low as 1V peak to peak seems acceptable. If the output is less than 1V even with zero resistance, check the output with original tube. If the tube gives higher output, try different J-FET. Higher Zener voltage can increase output also.

When RF J-FET is used, higher output frequency can be corrected by adding a small capacitor between pin #5 and pin #7 made by two insulated wires twisted together. In the picture two AWG #30 kynar wires were twisted. About two twist turns gave right frequency correction. For lower output frequency, the correction should be made at the PTO.

Any suggestions, improvements, and questions send e-mail to donkang@ieee.org.

How to Extend the Active Life of a Collins 75S- Receiver

By Don Kang 3-2-2005

Abstract

Heat is the major contributor for components degradation in the radio. By replacing three oscillator tubes, two audio tubes and two lamps with solid-state equivalents, the total power consumption of the radio is reduced from 70 watts to 40 watts. Those tubes handling RF/IF signal are not disturbed. Thus the vacuum tube radio performance is maintained.

1.0 Introduction

We all know that the vacuum tubes have limited life. Other components also get aged as well. Heat is the primary enemy. An obvious remedy is to install a cooling fan and many users have actually installed a fan on the S-line cabinet cover. Even today, original designers will resort to the use of a fan if no other simple solution is available.

Here is my approach to this problem, and two self-imposed restrictions.

No circuit modifications
(b) Incoming signal stays in the original vacuum tube circuit until it becomes an audio signal.

With the above restrictions, six items will be replaced with solid states devices.

They are: 

(1) 100khz Xtal marker oscillator, 6DC6

(2) Main tuning PTO, 6AU6/7543

(3) CW BFO, 6DC6 (NA for 75S-1)

(4) Detector and audio voltage amplifier, 6AT6

(5) Audio power amplifier, 6BF5

(6) Two lamps for S-meter and main dial, GE47

Heater current: 6DC6 0.3A x 2 — 0.6A

6AU6/7543 —- 0.3A

6AT6 ———– 0.3A

6BF5 ———– 1.2A

Lamps (0.15 x 2) — 0.3A

The total current saving is 2.7A and the total power reduction is 2.7A x 6.3V = 17.01 watts.

2.0 Finding Solid-State Devices for The Vacuum tubes

Many thousands of discrete solid-state devices have been produced in the years following the vacuum tube era. Today, most of those discrete devices have become a thing of the past. Some of them are more difficult to find than vacuum tubes. And unlike the vacuum tube, the appearance or the part number does not reveal what kind of device it is.

2.1 An insight of solid states (semiconductor) devices

There are two types of charge carriers, electrons and holes in the semiconductor. A hole is simply a missing electron. If the electrons are dominating, they become the majority carriers, and the semiconductor is called N-type. When holes are dominating it is a P-type material. Many solid state devices have an amplifying function. They are called transistors. The transistors are classified into bipolar and unipolar transistors based on the operating principle. For the bipolar transistor, the minority carriers are injected into the base region by a forward biased PN junction. The minority carrier in the base controls the operation of the device and both carriers are participating, thus bipolar operation. For the unipolar transistor, the input is a reverse biased PN junction and the electrical field at this junction controls conductivity of the channel which connects input and output. No minority carriers are involved. It is a majority carrier device and it is commonly called field effect transistor, FET.

Because the input of the bipolar device is a forward biased PN junction, its impedance is low. On the other hand, the input of the junction FET is a reverse biased PN junction, and its input impedance is very high. The only contribution for lowering the impedance is the junction leakage current and the gate capacitance. There is also the non-junction type FET called the Insulated Gate FET or IGFET. Actually the input of IGFET is a capacitor. The gate is sitting on an insulating material. If the gate is a metal and the insulating material is an oxide layer over a silicon, it is an MOS capacitor. This kind of unipolar device is known as MOS FET.

Unlike the vacuum tube, transistors have two genders, N and P type, based on the material as discussed above. For the bipolar transistor, they are the NPN and PNP. For the FET, they are the N channel and P channel. The FET is further classified by its operating condition (mode) as either an enhancement mode or a depletion mode. If the transistor is normally off with zero gate voltage, it is an enhancement (off) mode and if it is normally on with zero gate voltage, it is a depletion (on) mode.

The FET requires a conducting channel to pass current from the input (Source) to the output (Drain). For the enhancement mode, there are very few charges in the channel. To make it conduct, the gate voltage is increased to induce charges into the channel. The gate voltage at which the channel is about to start conducting or pass current, is the turn-on threshold voltage.

For the depletion mode, there are already lots of charges in the channel. To turn off the current in the depletion mode transistor, the charges should be depleted. The gate voltage which will deplete all the charges is called the turn-off voltage. However unlike enhancement mode, for the case of depletion mode, Idss ( the drain current when the gate voltage is zero) is used to characterize the FET. The initial charges in the channel are built in during the fabrication process, and it is not easy to control. Thus the Idss spec is made very loose. Very often different part numbers are assigned to cover the wide range of Idss.

2.2 Solid States Circuits

A depletion mode N-channel FET works in principle like a vacuum tube. It can be a MOS FET or a junction FET. Many FETs available for an amplifier application do not have high drain voltage. By teaming up with a high voltage bipolar NPN transistor, the drain voltage can be extended. Other important specs to consider are gain and Idss. They should match to the tube circuits. Because of loose Idss spec, some form of bias tweaking is needed to duplicate tube function.

A pentode tube can be considered as a triode and a buffer combination. The circuits shown below are starting point. A combination of one of these input circuits, and one of these output circuits will replace the tube function.

The transistors selected for this project are:

2N5484, N-channel depletion mode junction FET (Mouser Electronics p/n: 512-2N5484)
LND150N3, high voltage N-channel depletion mode MOS FET (Mouser: 689-LND150N3)
2N5551, high voltage NPN transistor (Mouser: 610-2N5551)
FQPF1N50, High voltage N-channel enhancement mode power MOS FET (Mouser: 512-FQPF1N50)

These are my selections from Mouser Electronics. There are many other choices you can make. When an external DC source pin with AC grounding is available, the base voltage can be stabilized by a Zener diode as shown in the output circuit (B) and the emitter terminal can supply a stable drain voltage to the preceding FET stage.

3.0 CONSTRUCTION

3.1 Tube Base – I

One obvious base source is the tube itself. At first I was very hesitant to destroy perfectly good tubes. The picture shows how I destroyed a tube to get the bottom part. The tube envelope near the bottom was scratched by a small grinding tool (1). This helped somewhat to avoid total destruction. Wrap the tube with vinyl film food wrap and gently hit the upper part of the tube with any small metal tool. Hold the pointed end of the tool when you use the kind of tool shown at (2). Make sure the inside pins are available for secure connection (not easy). Also external connection to the pins can be made.

3.2 Tube Base – II

Another way to make the base is using a tube socket. All the metal parts were removed from the socket. The right size pins are inserted from the top. The pins shown in the picture are from a D-sub connector (1). Only 1/4 inch thick socket (4) worked in this case because of the body length of the pins. When circuits are built on the base, make sure that it is working and the pins are properly mating to the socket on the radio. Then the pins are glued with an epoxy adhesive (5).

A third possibility is to try to find a source of 7-pin tube base sockets. I was told that they use to be plentiful, but are hard to find now.

3.3 100 khz Xtal Marker Oscillator, 6DC6

The LND150N3 and 2N5551 combination were used for 6DC6. No bias adjustment was needed for the input FET.

The output signal is not as strong as the vacuum tube counterpart, but I do not see why a strong signal is needed for a marker. It may be the case that the original designer could not find any capacitor less than 1 pico farad.

3.4 Main Tuning PTO, 6AU6/7543

Pin 6 of the PTO tube is AC grounded. This terminal is clamped at 14V with a 14V Zener diode. Any voltage from 8 to 20V will work. Use of the Zener diode helps to stabilize the PTO frequency especially for those earlier 75S- radios which do not have B+ voltage stabilization. Again no bias adjustment was needed for the 2N5484. For the right output voltage and frequency compensation of the PTO, please see June 2002 Album section, at the www.collinsra.com web site

3.5 CW BFO, 6DC6

The 2N5484 is a basic oscillator. The 2N5551 NPN transistor is an amplifying buffer and at the same time it extends drain voltage. Both transistors required bias adjustment for a proper operation. I used 20 K ohm variable resistors for the bias adjustments. I found that 10 K ohm at the 2N5484 source and 4.7K ohm at the 2N5551 emitter gave right output value. 

You may find slightly different resistor values for your circuit due to the transistor gain difference. The bias level should be set to produce 1.6v to 2.2v at product detector cathode. The 20 pico farad capacitor between pin6 and pin7 is to compensate frequency difference. If this capacitor is not used, the BFO frequency can still be adjusted as instructed in the factory manual.

3.6 Detectors and Audio Voltage Amplifier

For the 6AT6, the replacement is one to one. The gain of the FET amplifier seems a little bit low compared to the 6AT6. It can be easily compensated by the AF gain control. If you are using very old speaker, make sure that the speaker has adequate efficiency. Today many high efficiency speakers are available. You do not need a HIFI speaker.

The two diodes are Schottky diodes. Much smaller surface mount two-diode chip is also a good choice. I used 1N5819’s but any silicon Schottky signal diode will work.

3.7 Audio Power Amplifier

The original design of the AF power amplifier is shown in the February 2002 Album section of the www.collinsra.com web site. A few revisions are made here.

Much higher resistor values are used for the voltage divider from pin 6 to pin 2. This is to reduce the loading of the voltage amplifier drain. An un-bypassed 390 ohm resistor is added at the drain terminal for a small amount of negative feedback to improve linearity of the amplifier. A one ampere rated power MOSFET, FQPF1N50, is used to operate at a higher current density which improves linearity and gain. Also two 1W 10Kohm resistors are used instead of three 15Kohm resistors in the source bias circuit. The maximum output power is less than that of the original 6BF5. However, it is quite adequate for normal operation.

The picture below shows some of the finished Solid State tubes. (1) is the original 6BF5, and (2) and (3) are the SS AF power amp. They are about the size of 12AX7 tube. The bases (2) and (3) are Japanese origin and they are not available in the US. (4) is the audio voltage amplifier build on base II. (5) is the PTO oscillator built on base I. (6) is the 100Khz marker oscillator built also on base I.

The broken edge of the glass base is wrapped with a fish paper and glued with 5 minute epoxy adhesive.

4.8 Illumination Lamp

Two illumination lamps are replaced with high intensity light emitting diodes. I used green and red LEDs. The colors are not the best choices but I happened to have them in my junk box. The parts numbers or their manufactures are unknown. I adjusted the current to 15ma with R in the diagram. This is a 90% power reduction from the original lamps.

I used green LEDs for the dial and red for the S-meter. The green dial LEDs are aiming about 30 degrees off to cover wider area. The S-meter LEDs are looking in opposite directions because of the edge lighting. The pictures shown are my prototypes. Do not operate LEDs in the AC circuit without a high voltage rectifier in series. Many LEDs have very poor reverse characteristics. Also use in pair in opposite current flow to make it balanced full wave operation as shown in the diagram below.

The light projection pattern can be modified by changing the LED’s plastic dome shape. By roughing the shiny surface, the light can be scattered. These LED lamps are installed in the 75S-3C as shown in the picture below.

LED lamps are installed in 75S-3C

5.0 HEATER VOLTAGE ADJUSTMENT

The reduction in heater current as a result of using Solid State tubes has caused the heater voltage to go up to 6.9V. This condition will shorten the life of the vacuum tubes in the radio. The reduction of the power transformer load also increases the high voltage output. Later model 75S- has Zener diode regulated 140V, B+ supply. The series resistor (R86, 1 Kohm) and the Zener diode (CR6, 1N3010A) will dissipate more power to maintain constant output voltage.

The graph below was plotted to see how the heater voltage changes with the AC line voltage. Ideally I want to keep heater voltage slightly below 6.3V. At 110VAC input, the heater voltage is 6.2V with SS tubes installed. The regulated B+ supply voltage is stable between 120VAC and 110VAC input and it starts to drop around 110VAC input. At 100VAC input, it is 128V on my Round Emblem 75S-3C. This means that below 110VAC, the Zener diode is not regulating.

I used a 60 watt variac for the voltage reduction. If your radio has a 140V regulated B+ supply, reduce the AC input to the radio with the variac until the heater voltage is 6.2VAC or the B+ starts to drop out of regulation- whichever comes first. For a 75S- with no B+ regulation adjust the AC input for a heater voltage of 6.2VAC. The optimum point of my radio was at 110VAC with the Solid State tube installed, and the total power consumed in the radio was 40 watts. With all the original tube installed and at nominal 120VAC input, the power consumption was 70 watts. This represents a 43% reduction in power consumed by the radio.

An alternative to the variac is to use an external power resistor (30 ohm/5w in my case) in series with the AC input line. When a small 12V filament transformer secondary is connected to the primary in series in phase, it becomes an auto transformer to boost or reduce the AC voltage by 10% depending on input/output wiring. If your AC line fluctuates, the solution becomes more complicated.

Reducing AC input voltage is not a new topic. If I find a simple way to reduce and stabilize the input AC voltage, it will be posted here in this Album section.

Please e-mail me at donkang@ieee.org for any suggestions or comments.