Drake TR7A experiences and modifications

On this page you can find some experiences and modifications made to the Drake TR7/TR7A transceivers as they where gathered by my friends Rien PA0TRT, Wim PA3AJI and myself. Our goal is to present information not found elsewhere on the internet. From some other known changes our interpretation is shown. There is more interesting info available but not yet ready for publication, so look at this site from time to time to see what is added to the page. We are working on it. For those who entered this page via the Drake Web Ring, the rest of my homepage is dedicated to QRP.

History

Since November 2004, I am the proud owner of a Drake TR7 transceiver. Not quite a QRP rig, but long before I was infected by the "QRP virus", I bought a second-hand Drake R-4C (S/N 28408) receiver and completed the 4-Line with a T-4XB (S/N 15769C) transmitter some months later. The power supply for the transmitter is a homebrew copy of the AC-4 mounted in a MS-4 loudspeaker unit. I rewound an old transformer that seemed suitable to get the secondary currents and voltages needed. It still works fine.

The TR7 (S/N 3157) is derived from the inheritance of the late PA3BBT, a good friend who passed away too young. Because his radio equipment was stored for several years, the TR7 needed serious cleaning. Two of my other friends, Rien (PA0TRT) and Wim (PA3AJI) do have a TR7(A) too, so we could compare the rigs for differences in design and construction. My rig is the oldest one, PA0TRT has S/N 9345 and PA3AJI, a TR7A, S/N 11178.

One can find a lot about Drake equipment on several websites of the World Wide Web. That has been a great help for me to become familiar with the TR7. Meanwhile I changed and repaired quite a few things to upgrade my transceiver to the -if possible-level of later versions the R.L. Drake Company sold.
As always, Rien's knowledge of electronics and his professional measuring equipment is of great value for me. It helped us to determine the sense and nonsense of noticed modifications.

When I started cleaning and modifying, I created a spreadsheet in which all repairs and modifications are kept up in order of appearance. Click on the image below to see the complete spreadsheet.

to be continued ...

Fan replacement

One of the first changes is no mains power wires to reach the TR7. This is for pure safety reasons and to cancel hum in the audio circuit. Here in the Netherlands (Western Europe) we use 230V/50Hz-AC so the amount of hum is significantly higher (6dB) than in the USA where 115V/60Hz-AC is used. Due to that, the 115VAC cooling fan FA-7 needed to be replaced too. We used a 12VDC computer fan running at 8VDC to reduce the noise. By the way it feels much more ensuring when working on the TR7 if you know that there is no mains power around!
The 12VDC computers fans have the same dimensions for mounting on the rear panel as the Drake FA-7. That was no problem, and the idea was to use the (not common in Europe) 110VAC mains connector for powering the computer fan. The left picture below shows the new fan with the old connector.

To achieve the highest possible noice reduction, the 12VDC computer fan runs at about 7.5 Volt. A LM7808 voltage regulator is mounted to the right side of the 115VAC chassis part with some super glue. To prevent damage to the fan due to reversed polarity, a 1N4007 rectifier diode is added in series with the regulator output. With this construction, the fan replacement looks like original without having to saw or file in the rear panel of the TR7.

Upgrading product detector

Until now, January 15th 2005, the most far-reaching modification is the upgrading of the 2nd IF/Audio Board. The version #1 product detector with 1N4148 diodes, is replaced by the version #2 circuit with an MC1496. The attenuating pi network for the out going BFO signal level on the PBT/Reference Board must be increased to match the correct level for the new product detector circuit. R1058 is changed into 68 ohm (was 220 ohm) and R1060 is removed (was 68 ohm).

I removed all components from the PC-board used by the old circuit. Then, using pictures of a version #2 board as an example, I constructed the MC1496 circuit on a piece off perforated experimenter board. The sub board is mounted close to the surface of the 2nd IF/Audio Board with three small pieces of copper wire. Other connections were made with thin wire to +10V DC (11/13), BFO in (11/40) and Audio out (11/31). The input comes from the secondary tap of transformer T1102.

Unfortunately there is an error on page 2-69 of my Service Manual. Capacitor C1142 (0.05uF) is drawn from pin 10 of the MC1496 to ground. That is wrong, C1142 is a coupling capacitor for the BFO signal and should be drawn from pin 10 to BFO IN (11/40). Perhaps Drake has modified this page in later editions of the Service Manual. I don't know, but here is the correct schematic in PDF-format. Components between the dashed lines are on the sub PCB.

Two resistors were missing on my board. R1187 (10K) in series with diode CR1119 and R1188 (180 ohm) in series with diodes CR1116/CR1117. Both are in the +10V lines to switch for CW and AM modes and without these resistors a sharp click is heard in the speaker when rotating the mode switch. Due to the charging of the empty electrolytic capacitor (C1147 and C1151, both 22uF/16V).

The S-meter circuit is the last part I changed on the 2nd IF/Audio Board to upgrade to version #2. Some minor changes were made around the Sens. Adjust potentiometer R1105. Resistor R1101 (220 ohm) is replaced by two 1N4148 diodes in series with CR1104 to ground. Diode CR1101 (AA119) is replaced by a 4.7k resistor.

Huff Puff Stabilizer for the Drake TR7(A) transceiver

This circuit is based on the stabilizer by the late PA0KSB. I have made some modifications and additions to make the circuit suitable for the Drake TR7. Read for the full explanation of the stabilizer the original publication in English QEX February 1996 or in Dutch Electron December 1996 and January 1997.

Description:

I have used the 40MHz crystal oscillator that is already built in the TR7 as reference for the stabilizer circuit. A buffer amplifier BF494 amplifies the signal to a level high enough to drive the digital mixer 74HC74.

The PTO output (as the VFO is called by Drake) is amplified to drive the binary counter 74HC4060.

The Drake PTO has already a RIT input line that is connected to a varicap diode inside the PTO. This same line can also be used for the stabilizer. The RIT control line and the output of the integrator CA3140 are added together in the summing amplifier NE5534 and go the original RIT input of the PTO.

I have added a reset circuit to the original design. When the TR7 is switched on, the output of the integrator is always set to the midrange value of 4V. A manual reset input is also available. When the reset input is made high (+2 .. +10V) the integrator will also reset to mid range. So far I have not connected this manual reset line, because the total drift was never large enough to reach the limits.

Two comparators LM258 measure the output of the integrator. If the voltage comes within .5V of the upper or lower limit the LED lights up. Optional the output of the comparators can be connected to the reset circuit. However if this auto reset takes place the PTO will jump from its set frequency. This may not always be desired.

Construction:

The stabilizer was constructed on a 75x75 mm piece of perforated experimenter board. All components are at one side while at the other side all connections were made with thin bare wire and isolated wrap wire where needed. There is no printed circuit board available. My layout can be used as a guideline to other design. The board is mounted under the parent board of the TR7 (see photos). At one side the board is mounted with a 5mm spacer, using an existing screw position in the parent board. The original screw was replaced by a longer one to hold the board.
At the other side of the board an isolated 5mm stand off prevents the board from touching the parent board. At the inside of the bottom cover. I have glued a plastic isolation sheet to prevent the board touching the bottom cover. All connections between the TR7 and the stabilizer are directly to the parent board, only one wire has to be removed from the parent board and extended to the stabilizer board. A 100pF capacitor is mounted directly between the connection point of the PTO to the parent board and the coaxial cable. This to minimize the capacitive loading of the PTO by the cable.

Adjustment:

After completion the RIT center position has to be readjusted.
Hold the reset line high. Exactly center the RIT control on the TR7 front. Switch the calibrator on. Switch PBT on. Center the PBT control. Switch the RIT on.
Tune the receiver to exactly zero beat to one of calibrator points. Switch the RIT off. Adjust R24 on the parent board exactly zero to beat. (R24 can be reached through the small access hole in the parent board just above the stabilizer board on the photo).

Results:

Without the stabilizer my TR7 drifted after switch on over about 500Hz in 5 hours. With the stabilizer build in, no drift took place. The locking points appear to be about 20Hz apart.

Latest version:

In July 2005 I designed a PCB for a slightly modified circuit to be build into two other TR7's and my R7. There are some minor changes to the previous circuit build on the experimenter board.

1. The range comparator was left out. (not needed)
2. Reset can be done with the CAL switch on the TR7. When pressed, the +10Rcal voltage becomes available at point 10/5 of the PBT/Reference board. This logical change resets the stabilizer circuit. During RIT ON mode, the stabilizer keeps working.

About the stability: The PTO remains locked in a frequency grid of 20Hz at the low side of the PTO range xx000.0 KHz and 23Hz at the high side xx500.0 Khz, as long as the control voltage does not reach the limits. Reset to mid range is by pushing the CAL switch on the TR7 on/off or by switching the power off/on.

The picture below shows how the board is mounted in Carel's TR7. See how the required wiring is close to the PCB.

The board is a single sided 74x68mm type. Bitmaps of the copper layer and silk screen layer are available for personal usage. The parts are all available through the normal channels. Make sure to use good low leak polyester or MKM type 2.2uF capacitors. The design is made with Sprint Layout from Abacom. The picture below shows the "photoview" option, with disabled ground plane.

Rien, PA0TRT

Multi turn trimmers for PBT control voltages

Not unique, but very effective is the replacement of the sensitive potentiometers on the Power Supply Board by multi turn types. These variable resistors are for the adjustment of the PBT control voltages and the +10V supply. The exact adjustment of the old potentiometers is difficult and can be easily disrupted. We used Bourns 3296W and 3296Y top adjust types. Adjacent resistors are moved to the backside of the PCB due to the format of the multi turn potentiometers. Also here differences in PCB layout. On my board the potentiometers were mounted 180 degrees rotated from the boards of the other two TR7's.

Mounting the 10-turn trimmers on the same way like my friends boards, would cause a problem when placing the cabinet cover, so I placed them square to the PCB, a bit lifted for a better attainability with the trimmer tool from the front side of the TR7.

Improving audio response

When surfing the internet for Drake TR7 modifications, I found several more or less the same information about improvement of the audio frequency response. I gathered those info and changed the components on the 2nd IF - Audio Board as described. I'm not an audio purist, but I heard some undefined things in the audio sound after the changes that I didn't like. So, there were two ways to find out what could be improved in my rig:

1. Measure the audio response
2. Use a spice modeling program to calculate the response

We did both. The audio signal from a generator was injected via 10uF to pin 11/31 Audio Out. An AC voltmeter was connected across a 8,2 Ohm resistor connected to the headphone jack to eliminate the changing impedance of the loudspeaker during measurements. Using a logarithmic scale ranging from 120 - 5000 Hz, the AC voltage was measured in 21 steps and the results were placed in an Excel spreadsheet to display the curve of the audio response. By changing the values of some components and doing the measurements all over, we could see the difference in audio response. This method works, but costs a lot of time.

The second method was done by Rien, PA0TRT. He used the spice modeling program Micro-Cap V to simulate the audio circuit. Those kind of programs can perfectly be used for calculations where stepping of the component values is required. We found out that the components who really effects the audio circuit to get a flat (-3dB) response at 100Hz to 3Kc, were different from the described mods found on some ham pages.

We experimented with some values and finally changed the following components:

• C1150 = 150nF (50nF+100nF)
• C1152 = 32nF (10nF+22nF)
• C1167 = 4n7
• C1173 = 220pF
• R1178 = 600e (270e+330e)

To reduce noise in the audio, transistor Q1113 (2N3904) was replaced by a BC549B (Note: BC549B has different pinout). Q1111 was changed for the same reason too.

Below is a screenshot of the AC analyses from both circuits. In red the original TR7 audio circuit and in blue the modified circuit.

Rien, PA0TRT

Switched Capacitor Audio Lowpass Filter

The Drake TR7 has always an audible hiss that causes fatigue when listening for a longer period of time. Correcting the audio amplifier does not give sufficient result. What is needed is a steep audio lowpass filter like we find in modern transceivers with DSP filters. To build in an internal DSP filter would be a daunting task. The use of an external DSP filter could be an answer. However I preferred a simpler solution.

Maxim™ produces single chip switched capacitor audio lowpass filters that uses a minimum of external components. The beauty of these filters is the very steep response of a 8th order elliptic lowpass and the capability to tune the filter with a single capacitor. There are two versions of these chips MAX7400 with a shape factor 1.5 at >82dB of the corner frequency and the MAX7403 with a shape factor of 1.2 at >58 dB of the corner frequency. I used the MAX7400, because I had that one available. Tuning the filter is very easy by varying the timing capacitor of the internal clock frequency that is 100 times the corner frequency of the filter. The capacitor is by example 12pF for a corner frequency of 2830Hz and 33pF for 1030Hz. It is also possible to use an external clock.

Some experiments showed good improvement in the audio response of the TR7. But what I wanted was a filter that could be automatically set with each operating mode CW, SSB and AM. The idea to use variable capacitors or an external clock from something like a NE555 timer as clock did not please me at all. The solution was simple, replace the clock timing capacitor by a variable capacitance diode. Now control could be done simply by a variable DC voltage.

To maintain the full dynamic range of the filter circuit, a higher signal level is needed as that is available just before the audio amplifier IC in the TR7. A leveling and DC shifting amplifier is placed before the filter. This is done with a NE5534 low noise operational amplifier. After the filter the level is attenuated to the original level. A jumper is provided to bypass the filter for test purposes.

The filter frequency is automatically selected when the operation mode of the T7 is changed. The frequency presets are set with three potentiometers, one for each mode. Optional an additional external potentiometer and switch are provided for variable mode. When the variable mode is selected the presets are overruled.

The filter frequency can set by each mode potentiometer to once personal preference. I have set CW for 1000Hz, SSB for 2500Hz and AM for 3000Hz. In the variable mode the filter can be adjusted between 800Hz and 3000Hz. The high frequency limit of the range is adjusted with trimming capacitor.

The circuit is build on a piggy-back pc board that just fits above the 2nd IF and audio board. All connections are directly on the board. The filter board is mounted with stand off wires that are soldered on the top ground plane of the 2nd IF and audio board and the free copper surfaces at the back of the filter board. To create some extra space I did remove the 115VAC fan connector from the back panel of the TR7 (I use a 12Vdc fan).

Here are some results of the old and new noise responses made in Spectrum Lab™, a great program by DL5YHF. All measurements are direct from the load speaker output of the TR7 to the PC soundcard input. Noise spectrum response shown 0 Hz to 5kHz at 10db/div; antenna disconnected and RF gain fully clock wise.

1. 2400Hz SSB filter, audio filter off. The passband of the SSB filter is visible.
2. 2400Hz SSB filter, audio filter 2500Hz. The noise above the 2500Hz passband is dramatically reduced.
3. 300Hz CW filter and audio filter off. The passband of the CW filter is visible.
4. 2400Hz CW filter, audio filter 2500Hz. The noise above the 2500Hz passband is dramatically reduced.
5. 300Hz CW filter, audio filter 1000Hz. The noise above the 1000Hz passband is dramatically reduced.

Hover the pictures to read the tooltip info, or click for a large screen photo.

What remains is the (proto-type) diagram and PC-board design. Click on the left picture to open the diagram in PDF-format. The PC-board is designed with Abacom's™ Sprint Layout 3.0 and is available on request. A bitmap of the layout is included in the ZIP-file below. To keep the board compact and to avoid a double sided layer, there are two wire bridges. One in the middle of the board and the other under IC2 The board dimensions are: 42x52mm.

Disclaimer:

I provide as much as possible information for this design. I do not supply any parts or printed circuit boards. Modifications done to the original Drake TR7 are to the full responsibility of the owner. However I am always interested in your experiences.

Rien, PA0TRT

References:

• - Maxim Integrated Products

RF preamplifier for the Drake TR7

The Drake TR7 and TR7A are due its fully passive front end design not as sensitive as most modern transceivers. In the R7 and R7A receiver this was improved with a switchable preamplifier between the Input Bandpass filter and the Up Converter board. Both TR7 and R7 use a similar design. Where the difference is that the R7 uses individual bandpass sections for each frequency band. While the TR7 uses combinations of lowpass and highpass filter sections to achieve preselection. In the TR7 the signal path is bi-directional, because of it's receive and transmit capability. I considered that it should be possible to add a similar pre-amp to the TR7. Using the same basic circuit as the R7 preamp was the easiest to realize the project. The RF transistor is a Philips BFW16A, what is a good replacement for the original 2N5109.

Switching the amplifier in and out the TX/RX path is done by two miniature relays. When the amplifier is off, the RF transistor is cut off from it's bias source. The power supply is directly from the +13.8VRX line on the LP filter module filter unit. The "Store" switch on the TR7 front was used for the amplifier on/off function. The amplifier is positioned in the coaxial connection between the Highpass Filter module and the Up Converter board. Just remove the the coax from the connector at the Highpass Filter module.

I have made a small female connector from a 15mm piece of an old brass ballpoint filling and inserted a pin of an IC socket with some epoxy clue in one end. Make a small dent in the tube so that the coax from the Up converter board just fits in the tube. Solder about 5cm RG174 coax to the center pin and tube. Finish off with some shrink tubing. At the other end I used a piece of RG174 coax and folded the braiding back over the outer isolation. This way it will just fit in the connector on the Low pass filter module.

Results:

• The gain of the amplifier is about 10dB and is flat within 1dB up to 30MHz.
• The measured TR7 noise floor is about -130dBm. With the amplifier switched on, this improved to about -134dBm.
• These values are similar to my measurements of the original amplifier in a R7 receiver.

Do not overestimate the use of this amplifier and preamplifiers in general, because in most cases the background noise on the HF band is high and far above the noise floor of the receiver, but on a quiet band it will just lift a weak signal enough to make it readable. In other cases leave it off, because it will do no good on the receivers dynamic performance.

The preamp in my TR7 is build on a piece of perforated experimenter board. I also designed a neat 60x37 mm single sided PCB. Bitmaps of the copper layer and silk screen layer can be downloaded. The design file (Sprint Layout 3.0 from Abacom) is available on request.

Rien, PA0TRT

References:

• - Drake TR7 Operator's Manual
• - Drake TR7 / TR7A Service Manual
• - QEX February 1996
• - World Wide Web

Board extractor

With the original TR7 extender board kit came an extractor tool. My kit was missing this handy item. I made a replacement extractor out of an old steel wire coat hanger. The same one that make magic antennas for cars and television.

1. Cut 160mm wire out of the straight bottom part of the coat hanger.
2. Bend wire two times 46mm from both ends to get a U shape.
3. Bend wire 6mm from the ends under 135 degree angle.

Rien, PA0TRT

Antenna!

I heard someone once mention, "It's difficult to make an antenna that does not work at all". This item was given out as freebee at the NAB2003, the Electronic and Media show in Las Vegas, that I visited.

Rien, PA0TRT

Just for fun!

Rien, PA0TRT (left) and I in front of the old R.L. Drake Company building at 540 Richard St., Miamisburg, Ohio. We joined the Dayton Hamvention in 2005 and our hotel was only a few miles away from this location. We both were curious to see what was left of the building where a lot of hamradio equipment has been assembled!