GW4ALG's Drake TR7 Pages

[ GW4ALG went QRT in February 2007 ]

Home ]

Up ] Articles ] Introduction ] Links ] [ Modifications ] Publications ]

View/Hide Contents Window

Notes on the Drake TR7

Contents

(1) RF Decoupling on receive audio line
(2) RF decoupling of key jack
(3) RF decoupling of AUX DC jack
(4) Varying panel lamp brightness, and foldback on receive audio line
(5) Improved AM reception
(6) PTT contact bounce
(7) TX operation at 'Power On'
8) LF coverage
(9) Improved keying
(10) PTO Braid Modification
(11) Fuse rating for 220 V supply
(12) TX Exciter Board modifications
(13) A review of the TR7
(14) New predriver board
(15) Wideband TX enable
(16) PS7 Adjustments
(17) PS7 Troubleshooting Guide
(18) ALC Improvements for the TR7
(19) Further keying improvements
(20) Homemade WARC band modules
(21) IF Filters
(22) More RTM7 Module Information
(23) The ultimate solution to key clicks?


 

(1) RF decoupling of receive audio line
Under transmitter test into a dummy load a small amount of RF pickup was found on the receive audio output wiring.  As a result, the following decoupling was provided:
(a) 0.1uF 100v capacitor across speaker terminals;
(b) 0.1uF 100v capacitor across external speaker jack J12; and,
(c) 0.1uF 100v capacitor across headphone jack J11.
(2) RF decoupling of key jack
A 0.001uF 200v capacitor was wired across the key jack.
(3) RF decoupling of AUX DC jack
A 0.1uF 100v capacitor was wired across the 'AUX DC' jack in the power supply, PS7.
(4) Varying panel lamp brightness, and foldback on receive audio line
When transmitting full power single sideband (SSB), supply ripple was occurring on the 13.6v DC to the TR7 due to a voltage drop across the two core supply leads on speech peaks.  This showed up as a fluctuation in panel light brightness, and foldback on the receive audio line.  At G4ALG the strap between the two +ve connectors in the DC plug of the PS7 was removed and a separate +ve wire run to the PS7 for the low current 13.6v rail.  The existing +ve conductor thus being used to supply the PA only.  Another -ve wire was provided between the plug and an earth tag in the PS7.  Both extra conductors were fed through a length of plastic sleeving before being laced to the existing DC supply lead.
(5) Improved AM reception
The AM position was found to be useless on receive with only the standard 2.3kHz filter.  To avoid the expense of the optional AM filter, a 'dummy' filter was provided in filter position "B".  See Figure 1.  Good results are obtained on the BC bands, and NBFM is resolved well using slope detection.

Figure 1

fig01.bmp (4850 bytes)

(6) PTT contact bounce
The PTT line was found to be very intolerant of contact bounce at the PTT switch, and poor TX/RX switching resulted.  Extra components were provided around the microphone jack, J7, to effect the modification shown in Figure 2.  The 1k resistor increases the PTT switch current to 10mA, and the 15uF capacitor sets the transmit/receive changeover timing to about 100 ms.

Figure 2

fig02.bmp (6326 bytes)

(7) TX operation at 'Power On'
Under certain conditions associated with the VOX delay circuitry, the TR7 may momentarily switch to 'transmit' at the instant of power ON.  This can be eliminated by generating a 'TX inhibit' signal at power 'ON.'  The delay circuit of Figure 3 produces a three-second inhibit signal at the 'out of lock' terminal (2/12) of the Digital Control board each time the TR7 is switched on.

Figure 3

fig03.bmp (6814 bytes)

8) LF coverage
It is not essential to purchase the AUX 7 board in order to provide 0 - 1.5 MHz receive coverage on the TR7.  Without the AUX 7 board, only eleven signal diodes are required. The diodes can be mounted on a small piece of 'veroboard', and the whole assembly mounted on the underside of the Parent Board.  See Figure 4. 

Connect an LF antenna to the ACC socket at the rear of the TR7.

When using the LF provision, rotate the bandswitch to the 1.5 - 2.0 position to extinguish the 'set band' lamp. 

Figure 4

(The pin numbers relate to the
AUX-7 connector on the Parent Board.)

fig04.bmp (18106 bytes)

For a picture of the relevant area of the Parent Board, see:
http://www.dproducts.be/drake_Museum/rx_0-1_5_mhz.htm

For  0-0.5 MHz, select AUX Program 1.
For 0.5-1.0 MHz, select AUX Program 2.
For 1.0-1.5 MHz, select AUX Program 3.

 

(9) Improved keying
The keying waveform on CW transmit was improved at G4ALG after permanently wiring a series-connected capacitor and resistor across the key jack, J15.  See Figure 5.  Also see Notes (19) and (23).

Figure 5

fig05.bmp (4990 bytes)

(10) PTO Braid Modification
'PTO Braid Modification' to cure frequency jumping problems, as directed by R. L. Drake, Ohio:

(a) Remove all connecting cables from the TR7.
(b) Remove the cabinet wraparound by removing the eight screws on the bottom, and sliding the wraparound toward the rear. 
NOTE: Perform steps (c) - (f) if the DR7 is installed.
(c) Unplug the 5 cable connectors connecting DR7 to the TR7. Carefully position these cables to the side.
(d) Unplug antenna coax and blue/shite bandswitch stepping wire from the high pass filter module and remove the grommet.
(e) Remove the DR7 hold-down screw and lockwasher.
(f) Carefully remove the DR7 by hooking the board puller under the read edge and lifting upward. Once unplugged, the DR7 can be removed toward the rear of the TR7.
(g) Clean chrome rider bar with wood alcohol or similar solvent. See Figure 6.  Leave entire rider bar absolutely clean and dry. DO NOT OIL.
(h) Solder braid to VFO fork and frame as per Figure 6. Be careful not to allow solder to flow into the braid as the braid must be flexible - hold the braid 1/8 inch from the end with needle nose pliers.
(i) Reinstall the DR7 by locating LED readout block in the proper slot in front panel and lowering the pins on the bottom of the DR7 into their respective sockets. Be sure all of the pins are aligned with the proper sockets, and the antenna coax is routed through the correct hole on the DR7.
(j) Reinstall the DR7 hold-down screw and lock washer.
(k) Reinstall the rubber grommet on the antenna coax and bandswitch stepping wire. Connect these wires to the appropriate connectors, and dress the wires and grommet into the slot provided in the chassis.
(l) Reconnect the 5 cable connectors to the appropriate pins on the DR7.  Be sure to install the connectors so that the black stripe is uppermost (facing you).
(m) Check for broken or pinched wires, board misalignment, etc., and correct any problems. Dress all leads down into the chassis.
(n) Reinstall the cabinet wraparound.

Figure 6

fig06.bmp (22142 bytes)

(11) Fuse rating for 220 V supply
Check the primary fuse in the PS7.  Many chaps with 220/240v mains voltage are happier with a 5 amp fuse in place of the 8 amp fuse fitted as standard.
(12) TX Exciter Board modifications
Those with TR7's with serial numbers below 1400 may care to make the following mods to update their TX Exciter Boards. See Fig. 2-6 in the Service Manual.
R327 from 33k ohm to 150k ohm
R333 from 330 ohm to 180 ohm
C308 from 1000pF to 68pF
C311 from 220pF to 470pF
Q301 from 2N3904 to 2N6521

The reason for the above changes is as yet unknown. Additionally, install 470k ohm across pins 1 and 4 of the microphone jack, J7.  This is to provide an extra mic. input to the TR7 for high output microphones.
(13) A review of the TR7
For a review of the TR7 by K1ZZ see 'QST', May 1979
(14) New predriver board
The Predriver Board has been redesigned, and GW3GHC has kindly forwarded the circuit diagram of the new board.  See Figure 2-37A.

 

Figure 2-37A

Click here to view

 

(15) Wideband TX enable
To enable the TR7 to transmit on all frequencies 1.5-30MHz (excluding 2.5 and 5.0 MHz bands) simply unsolder the collector of Q9001 on the DR7 board.  See Figures 4-1 and 4-2 in the Service Manual.
(16) PS7 Adjustments
The current limit control should be set to provide 30A and shut down at 31 to 32 amperes. If the power supply board is wired-in, adjust the white PC control clockwise to increase current.  If the power supply board is plug-in, adjust the white PC control CCW to increase current.  Set the output voltage to 13.6 v by rotating the yellow PC control.
(17) PS7 Troubleshooting Guide
Unit does not turn on (0v at output):
  (a) Fuse blown
  (b) Fuse holder not soldered to primary switch board
  (c) Primary wiring incorrect
  (d) Defective transformer
Apparent short circuit at output:
  (a) Large SCR (C228F3) is ON (shorted or triggered)
  (b) Check that wiring on board connector is correct
  (c) Primary wiring incorrect
  (d) Primary switch board has short to ground
  (e) Large electrolytics reversed (15000 uF)
  (f) Defective high current bridge (MDA 3500)
  (g) High current DC line shorted to ground
Unit will not handle peak loads (trips out):
  (a) Defective pass transistor (2N5301)
  (b) Open 1.8 ohm resistor in emitter of 2N5301
(c) Current pot.(white) not adjusted properly
  (d) TIP31 transistor defective
  (e) MC1723 regulator IC defective
Unit will not shut down at any current level:
  (a) Defective MC1723 (7v reference gone)
  (b) Defective 2N4402 at IC regulator
  (c) Defective 2N5060
  (d) Shorted 2N3566
Unit will not regulate (13.6v wanders with load):
  (a) Defective regulator IC MC1723
  (b) Defective pre-regulator transistor TIP31
  (c) Defective 2N3566 on PCB
  (d) Defective 1N4005 in rectifier bridge on PCB
  (e) Defective 470uF filter capacitor on PCB
  (f) Defective transformer
  (g) Large electrolytics reversed (15000uF)
  (h) Defective high current bridge (MDA 3500)
(18) ALC Improvements for the TR7
By George Cutsogeorge, W2VJN

On the air tests indicated that my new TR7 did not cut through the QRM as well as my 20 year old Collins S Line.  This was true for barefoot operations as well as when driving a linear.  In all cases the key down power output was adjusted to be the same and the VSWR was low.  Both rigs were operating with ALC and the same microphone was used without any speech processing.  The difference was traced to the time constants used in the ALC circuits of the two units.  The 32S-3 has a fast attack and a dual delay time constant. The attack is less than 1 ms and the delay is 68 ms and 1.6 seconds, proportioned 17% and 83% respectively.  The fast attack is necessary to prevent flat topping on peaks.  The longer delay time constant adjusts the average gain for the particular microphone and operator.  The shorter delay time follows the syllabic rate and brings up the gain between peaks which increases the average power.

Examination of the TR7 ALC indicated an attack time of 10ms and the delay time constant was very long.   Only leakage currents and the bias current of the op amp U1601 would discharge the ALC capacitor C1611.  This not only makes the delay time constants long, but also quite unstable with temperature, and life: and highly variable from unit to unit.

An 82k ohm resistor was added from the cathode of CR1603 to ground.  This resistor may be added by removing the bottom cover of the TR7, and soldering it between the CR1603 end of R1617, and the ground end of R1615 on the top of the ALC board.  It is not necessary to remove the board to make this addition, which leaves the attack time at 10ms but gives a delay time constant of 92 ms.

On the air tests now show approximately equal average power outputs from the TR7 and the 32S-3.  DX stations now report that there is little or no difference between the two units. Some small additional improvements may be obtained by utilizing a dual time constant similar to the 32S-3.

(19) Further keying improvements
Concerning Note (9) above, W2VJN writes: 
I installed the mod using the 330 ohm resistor and 1uF capacitor across the key jack.   I noticed that on high speed CW, the duty cycle of my keying was too heavy.   The trailing edge of the wave was extended by 15 ms.  Also the leading edge (not affected by your mod) was too soft.  The net result is keying that is not easy to read.  I found that the trailing edge could be softened without extending it by increasing the size of the by-pass capacitor on the CW oscillator.  In my unit, a 1.2uF capacitor across C1143 on the 2nd IF/Audio Board did the trick.  This left the leading edge too soft.  The problem was traced to the crystal being very slow to start. It was determined that a resistor across C1136 would cut the crystal Q down somewhat, and speed up the oscillator start-up.  I believe that the required resistor value will vary from unit to unit, but in my case a 3000 ohm resistor was chosen.

The rise and fall times are now 2.5 ms when the ALC is adjusted to threshold.  This mod reduces the oscillator output slightly, and the carrier control may be advanced about 1/2 division to compensate.
(At G4ALG, the modification in Note (9) has been replaced by the modification in Note (19).)  Update 23/11/2006: Also see Note (23).
(20) Homemade WARC band modules
For those wishing to make up their own RTM7 modules, GM3WIL has forwarded details of the required diode arrays for the proposed new amateur bands.  See Figure 7.  For 'receive only' modules, omit the diodes to pin 10.

Figure 7

fig07.bmp (13502 bytes)

(21) IF Filters
Sherwood Engineering of Colorado is marketing a range of high quality IF filters to suit the TR7, and has now added an RF speech processor, model 7-SP to their price list.  Address follows:
Sherwood Engineering Inc.
1268 Sough Ogden Street
Denver
Colorado 80210
U.S.A.
(22) More RTM7 Module Information
Tom Evans, AG9X was kind enough to write:
"
I found your RTM7 modules drawing helpful. I took the .GIF at Modification #20 and added the 0 - 0.5 MHz, 0.5 - 1.0 MHz, and 1.0 - 1.5 MHz modules to it by cutting and pasting.  May be you'll want to use the attached .GIF on your page."

A great idea, Tom!  So here it is:

 


(23) The ultimate solution to key clicks?
In November 2006, Floyd Sense, K8AC sent me a fascinating email about his own investigations into TR7 key clicks.  Floyd writes:

"I was recently checking out my TR7 (serial 8913) and noticed that it had very bad key clicks even with the ALC properly adjusted. This occurred at all power levels. In researching the Web for a solution, I ran across your TR7 info at: www.zerobeat.net/drakelist/tr7moremods.html

I first tried out the suggestion you had in notes 9 and 19, with no success.

I next posed a question about the clicks to drakelist@www.zerobeat.net which led me to a partial solution as K4OAH offered some helpful guidance. Here's that info:

The leading edge of the keyed CW waveform make take on a sharp leading edge with a period of overshoot (above the average power) and ringing if T1101 on the 2nd IF/Audio board is mis-adjusted. After properly peaking T1101, the problem was gone.

That left me with slightly better key clicks, and the CW envelope as observed on an oscilloscope showed a very rapid risetime of 0.5 to 0.6 milliseconds.  At that point, Keith Evans (G3VKW), recalled that John Basley (formerly G3HCT and now VK4OQ) had made some modifications many years ago to an earlier TR7 that fixed this problem. John went out of his way to contact an old friend in England from years ago who was now in possession of John's notes on the subject.  I installed John's modification (with a slight adjustment) and it completely cured my key click problem.  Following is the info I received from VK4OQ and the change I made to it.

"On the 2nd/IF Audio Board the xtal oscillator Q1110 and its associated buffer Q1109 are keyed by Q1103. The modification is done to Q1103.  Add on the back of the board a 15k 1/8 watt resistor between the emitter and base. Add on the back of the board a 0.1fd Disc capacitor BETWEEN COLLECTOR AND BASE. That is all there is to it but the shape will then be " Text Book" !!

73 JOHN VK4OQ"

I found that I had to use a 0.22 mfd capacitor between collector and base to achieve the near-perfect waveform.  I used a tantalum capacitor and found plenty of room for it and the resistor on the back of the board.  Here's a link to a webpage with photos of before and after CW waveforms showing just how effective this modification was for me:
http://www.k8ac.net/KeyClicks.html

All credit goes to VK4OQ for this outstanding technical work, and to G3VKW, who doggedly tracked John down on the other side of the world for information that was decades old!"

 

 
   

Steve Rawlings, GW4ALG
23 November 2006

 


Figure 2-37A ]