Drake Home Page       Drake Technical Tips Menu

Restoring A Drake TR-7 To Full Output Power

Written By: Floyd Sense / K8AC

A few months ago, I acquired a TR-7 from someone on eBay and, of course, it was said to be in excellent condition.  When I tested the transmitter section, I found low output on the higher bands (only 40W out on 10 meters) and sending CW for a minute or so resulted in the power out dropping to zero.  A complete alignment didn't improve the situation and I began searching for a solution.  I read everything published on the Web on the TR-7, and asked questions on the Drake reflector about the power output level.  The general consensus was that one should be happy with 50-75 watts output on 10 meters and that "they were all that way".  That wasn't a very satisfying answer, and a few knowledgeable TR-7 owners suggested that things could be better. To make a long story short, after many hours of work I was able to bring the power output up to what I considered to be a proper level, following tips from a few of the reflector members as well as using information from the Drake documentation by VE3EFJ.  Since I was unable to locate a single source for all the information necessary, I've prepared this page for other TR-7 owners who might have a similar problem.   

The Test Setup

Power measurements for this project were made with an old Bird 43 wattmeter with a 250 watt HF slug and a Heathkit oil-filled dummy load of around 49 ohms.  When working on the PA "brick" outside the TR-7, the drive signal was provided by my trusty Boonton 103F signal generator which has a nice stepped attenuator and output level meter allowing you to supply an RF signal of known level into a 50 ohm load (up to 3 volts).

One nice feature of the TR-7 is that the entire PA unit can be removed from the TR-7, including the heat sink, so that you can work on it independently of the TR-7.  Perhaps that should have been obvious to me, but it wasn't and I thank Garey Barrell for suggesting that approach.  He told me how he had worked on the PA using a signal generator to supply the required drive.  To do that, all you need to do is:

1. Remove the PA brick from the TR-7, unsolder the input and output coax lines, unbolt the two DC supply leads and unsolder the red lead that connects to the terminal post near the +DC lead.  

2. Connect a wattmeter/dummy load to the output terminal posts

3. Connect the signal generator input to the input terminal posts

4. Connect (+) and (-) leads from a DC supply capable of supplying 26 amps or more

The red lead removed in step 1 supplies +13.6 VDC to run the pre-driver board and the driver transistors and feeds the regulator that supplies the bias for the finals.  So, all you need to do to transmit with the PA brick is to connect the post where the red lead was connected to the nearby +DC screw terminal and then apply power.  Removing the connection puts the PA in its "receive" mode state.


The PA unit sitting on the bench, in operation.  The white and yellow leads near the top of the photo go to the ammeter to measure the "bias current".  The red lead connecting the + power lead to the small post just below it sets the unit to "transmit" mode.  The signal from the generator is applied via the coax seen at the bottom of the photo.

Solving The Problem

The first thing I did was to replace the electrolytic and tantalum capacitors on the ALC and PA boards to eliminate those as a possible problem.  There was no benefit from that effort.

Upon close inspection, someone had been into the PA before me and had replaced the final transistors, the driver pair, and the pre-driver (version 2 board).  The final transistors were found to be MRF-421s, the drivers MRF-475s, and the pre-driver an MRF-476.  After connecting everything as described above, I applied drive from the signal generator and attempted to achieve 100 watts output on 7 bands while monitoring the drive required from the generator.  Here's what I observed:

The large amount of drive required on the higher bands told me that the problem was indeed in the PA brick somewhere.  Furthermore, there are back-to-back silicon diodes across the pre-driver input and the signal clipping of those diodes undoubtedly would introduce some undesirable distortion of the waveform.  A scope trace of the signal coming out of the pre-driver board indeed showed a pretty ugly waveform on 10 meters.  At this point, I decided to tackle the problem of the power dropping off sharply if I held the key down more than a few seconds.

In VE3EFJ's well-known document covering Drake Mods (section 12.2, TR7 Mods and Tech, "Late Model Driver Boards"), Wayne described a thermal runaway problem that would produce the same symptom I was seeing.  I replaced the MRF-475 transistor (no longer available from RF Parts) with one from Communications Concepts, Inc. (http://www.communication-concepts.com/) and inserted the resistors in the emitter lead as described by Wayne.  That provides a degenerative bias that Wayne claimed would solve the runaway problem.  

While working on that board, I also checked resistor values and found that one had drifted badly (R2213) in value and another (R2202) had been changed to another value by a previous owner.  In addition, resistor R2209 in parallel with the R2210 gain adjustment pot had been cut out.  I restored the resistors to the values shown in the schematic and replaced R2209.  Subsequent testing showed that Wayne had been right on and the power runaway problem had been resolved.  Unfortunately, there was little or no effect on the power output level on the higher bands.

Garey Barrell suggested that I change the final transistors to 2SC2879s as they have higher gain at 30 MHz and so I ordered a matched pair of them from RF Parts.  I had previously installed a new pair of drivers, a matched pair of MRF-475s from RF Parts.  With the new finals, the output on the higher bands was up quite a bit - 75 watts or so on 10 meters - but still not what it ought to be.  I then checked the bias current at the jumper position on the PA board.  While the Service Manual doesn't give a suggested bias current, Garey said Drake suggested a value of 800 ma, but that he preferred something closer to 200 ma.  I measured 2.2 amps and something was clearly wrong.  Note that the "bias current" at that point is really the collector current of the driver transistors - there is no bias setting capability for the final transistors.  

Garey had me check the base voltage on the drivers in "receive" mode and while it should have been zero it read .64 volts.  He then suggested that I lift the driver base pins from the board and check the voltage again.  The voltage measured at the floating base leads was now .76 volts (slightly different on the two transistors).  So, the drivers were pulling 2.2 A of collector current with the base leads open - clearly a problem with those devices. Garey had also suggested I use 2SC1969 transistors for the drivers instead of the MRF-475s, so another order to RF Parts for a matched pair of 2SC1969s.  That device is used as the output stage in many current CB rigs, so is readily available, but it appears that only RF Parts supplies matched pairs.  I also ordered new transistor mounting insulators, as the old ones were not in very good shape.  After installing the new drivers, the bias current was observed to be 330 ma - within reason.  

Again, Garey provided the knowledge on adjusting the bias current.  By inserting an additional resistor in series with R2303, the bias on the drivers is affected, with a higher resistance lowering the "bias current".  I arbitrarily added a 47 ohm resistor in series with the existing 47 ohm resistor and since the current was now within the acceptable range I decided to leave well enough alone.

Testing the output and drive levels as before showed that the output problem was solved.  The drive required to produce 100 watts output was lower on the lower bands and a full 100 watts output could now be achieved on even 10 meters.   The following table shows the drive required to produce 100 watts out on all bands.  Compare this to the drive required prior to the transistor changes and note that the drive required on 10 meters is now far below the level that will cause the protective diodes to clip the input.



I set the ALC pot to produce 125 watts output on 20 meters with the Carrier control full CW, then switched to 10 meters and adjusted the pre-driver gain so that the ALC light just came on full.  The following table shows the output power then seen on each band when adjusting the Carrier control to a point where the ALC light came on full.  Current draw as seen on the power supply ammeter was 22 amps on 20 meters and 24 amps on 10 meters

I'm not sure why the output on 160 meters is low and I didn't notice that while the PA was running outside of the TR-7.  I assume that the power is being lost in the low pass filter section for 160.

I made no attempt to see just how far the output could be increased, but when testing with the signal generator, I had no trouble driving the output up to 180-200 watts on the lower bands.  I ended up with the pre-driver gain pot set to a few degrees CW past 1/2 rotation. There has been no trace of instability on any of the bands, and power output is steady in RTTY and all other modes.

Based on my experience, I'd suggest the following if you are attempting to restore your TR-7 output level to where it should be:
  1. Pull the PA unit and make sure all resistors are the values they should be.
  2. Replace the final transistors with 2SC2879s.  I don't know if a matched pair is really necessary, but RF Parts has them for a nominal price increase over an unmatched pair.
  3. Replace the driver transistors with 2SC1969s.  Same comment regarding the matched pair.
  4. On a Version 2 pre-driver board, replace Q2202 with a fresh MRF-476 from Communications Concepts. There may be a better substitute for this application, but the MRF-476 was available and I couldn't find any advice on a better transistor for the job.
  5. Follow VE3EFJ's advice and insert the appropriate resistors between the emitter of Q2202 and ground.
  6. When doing the transistor replacements, be very careful not to overheat the traces where the transistor leads go through the board.  If the drivers have been worked on before, chances are that the traces on top of the board are already separated from the substrate.   Luckily, those traces aren't needed as they dead-end at the transistor cut-out.  So if they're already lifting, you can remove them. 

Thanks to all who made suggestions via the Drake reflector and private emails, and to VE3EFJ, whose Drake documentation is an important part of the hobby.

73, Floyd - K8AC
Angier, North Carolina