Repair Notes of a Ten-Tec Corsair 560

(first posted:   fall of 2011)
(last update:   22 Feb 2014)


The UMARC (University of Maine Amateur Radio Club) has a Ten-Tec Inc. Corsair 560 for HF work.

I have had to work on this radio twice, from 1997 or so up to 2011.

As near as I can figure, this particular radio was produced sometime after 1982. This seems likely since an online manual I found for the radio has a date on one of the schematic revisions. The schematic seems applicable to the hardware. The schematic is for the TIME BASE 80974 Rev A board, which is what is installed in the Digital Readout module of the radio. Down in the bottom right corner of the schematic the date "11-4-82" is written. This brackets the production of the radio between 1982, when the revision was made, and 1987, when the Corsair II was released.

Common to each repair is a flaky digital frequency readout. Unfortunately my notes for the first repair have been lost, so I'm working the 2011 repair from weak late 1990's memory. Back then I thought it was due to oxidized connections. I recall I looked high and low for a failed component but never found one. It was an intermittent problem. The problem seemed to have gone away, during other troubleshooting.

As it turned out, during this repair session, the flaky Digital Readout was an easy fix, although the cause was hard to find.

An additional failure in 2011 is that the power supply trips out early, as the Drive knob is turned up. On many bands the power supply trips around half way (well below 100 watts output). However, if the ALC is backed down from fully CW, this can also cause the power trip.

By the way. Always be sure that the power port of a radio is terminated in an appropriate manner before attempting RF power tests. In this case the radio was connected to a 50 Ω high power resistive load.

Being a radio that is exposed to operators without much experience, I figured the finals had been damaged. Folks new to antenna tuners often make mistakes while matching the antenna to the power source, and the final amplifier often bears the bunt of such abuse.

Output Power Chain

Upon closer inspection the problem may be a bit more complicated. A look at the signal driving into the final amplifier, showed some severe distortion as the power is increased. Probing around with the oscilloscope revealed no distortion up until the output transistor Q2 of the Low Level Driver 81037 board. Q2 would be mounted to the black heat sink, which is empty, shown below:

Q1 position (removed)on the Low Level Driver 81037 module

Ten-Tec Corsair 560 "LOW LEVEL AMP 81037" with Q2(MRF476)
removed, to be replaced by an NTE235 or Toshiba 2SC2166.

Q2 is an MRF476(pdf). This device was made by Motorola, when Motorola still actually made useful things, but that is another story. Fortunately, it crosses to an NTE235(pdf), which is still available from NTE. There is anecdotal evidence that the MRF476 also crosses to the Toshiba 2SC1678 or 2SC2166 but I have yet to confirm this. The 2SC2166 seems the likely candidate. It is listed as a cross, and is available from RF Parts Company, as is the 2SC1678.

Below is the schematic for the Low Level Driver 81037 board. The manual for the Corsair 560 is, or was, available on line as a pdf. Ten-Tec has archived a bunch of manuals here (as of Feb 2012).

LOW LEVEL AMP 81037 Manual Schematic

Ten-Tec Corsair 560 "LOW LEVEL AMP 81037" schematic.

An NTE235 was tried in this location but it seems to have suffered the same fate as its predecessor, the original MRF476. For some reason there is a huge loss of gain, although the transistor sort of works. It seems likely that something is going on down stream (FINAL AMP 80565) that causes this transistor to fail.

There just doesn't seem to be anything up stream of Q2 which could be responsible for its problems. But I am keeping an open mind.

I'm still hunting for the cause...

Found The Cause!

It is now late Jan 2012, and the cause of the loss of gain at Q2 has been found. It has been staring me in the face... as much as any intermittent failure stares one in the face.

The key was that the intermittent failure turned into a full blown failure across three different devices. First was the MRF476, then the NTE235, and finally the 2SC2166. The full failure didn't show up until the second of two NTE235 devices had "failed".

At that point, I figured something crazy was going on, and I swapped in a 50Ω resistive load. The resistive load was an attempt to reduce the number variables. There is no reason for the circuit to fail if it is driving its characteristic load, and the load is passive, and free of any reactances.

With the resistive load in place, if the circuit didn't work, there was one of two possibilities:

The device in Q2's position was broken.

or...

Something else in Q2's circuit or before it was broken.

Swapping in a brand new 2SC2166 for device Q2 yielded no change in behavior. Still messed up. I now had three different devices telling me that the transistor was not the problem... OK... so I'm a little hard of hearing...

I confess that I'm prejudiced, when it comes to active devices, and circuit failures. Active devices are almost always responsible. The LAST thing I point my finger at is passive devices. I usually say to myself "... no.. it can't be a passive device... they just don't break that way..."

Famous last words...

Now... Passive devices do fail. But most of the time there are strong indications that they have failed. Such as a burnt resistor, a cracked or exploded capacitor... some other type of obvious physical evidence.

In this case not only was there no physical evidence, there was little reason to believe that the component would fail.

Why?

Because the power dissipated by the device was one fifth the capacity of the component.

Don't overlook the simple stuff...

There was another reason why I kept chasing my tail on this problem.

I repeatedly overlooked the bias point. The bias point is listed in the manual. I also checked it, and upon occasion it looked OK. But I did not check it often, and I became so obsessed with the Q2 device, that when I did check it, and saw that it was out of specification, I did not recognize it as a problem.

At long last, I decided that the "impossible" was possible. The emitter resistor of Q2, R9, may be bad. Even though it looked pristine, and undamaged.

I lifted its Emitter connected leg (just to be sure), and sure enough it measured more or less as an open circuit. That isn't good when it is supposed to be about 3.3 ohms.

No wonder there was little gain. There was hardly any current passing through the transformer in the Collector circuit of Q2. It is hard to make much of a magnetic field when little to no current is passing through a coil. Transformer core or no transformer core.

Not a mark on the resistor, and it's a 1/2 watt device. Way over rated for the application. No reason to suspect that it was bad. And resistors just don't fail that way...

And yet it was the problem all along... go figure.


Alignment and Tuning Notes

The manual implies, by sequence of appearance, that the OSC/MXR 8975 board should be aligned first.

In practice this didn't work so well.

The first problem was that older manuals list the wrong alignment procedure for the OSC/MXR 8975 board. At least for serial number: 560 00692. This unit is a better match for the pdf based manual, available from Ten-Tec. The pdf manual does have the correct alignment procedure, as listed on page 34.

Unfortunately this procedure doesn't always yield results. The variable resistor R1 on the OSC/MXR 8975 board can be swept back and forth with no effect what so ever on the tone.

It is also true that the alignment controls (R1, C1, C2) of the RF MIXER 80987 board also seem to yield little in the way of results.

However...

You may find that if you center R1 of the OSC/MXR 8975 board, as well as the three controls of the RF MIXER 80987 board, you can align the receiver.

With all these controls centered. Turn to page 36 of the pdf manual (older manuals have the same procedure), this should be the RF MIXER 80987 section. The strategy that yielded some results was similar to this.

Start with the 21.320 MHz alignment in the RF MIXER 80987 section. Don't bother with the OSC/MXR 8975 board's 1.950 MHz attempt at a null.

Be sure to let the radio warm up a bit first. Wait at least 10 minutes, multiples of that would be better, if you can stand it.

Tune to 21.320 MHz, in CW mode.

Turn up the AF-RF AMP.

If no tone is heard, the OSC/MXR 8975, and RF MIXER 80987 boards are probably aligned OK, and there is little point in doing anything further. If a tone is heard, some tweaking will be needed.

Contrary to what the manual suggests, in the RF MIXER 80987 alignment process, adjusting R1 FIRST may yield no effect.

Better results were found by adjusting C1 of the RF MIXER 80987 board first. This will require careful, and slow turning. The null will be quite sharp. If the capacitor's rotor is turned too fast, the change in audio level will be missed.

In the 560 00692 radio the null appeared about one fifth from fully un-meshed. Un-meshed is the lowest capacitance the capacitor is capable of.

The null was difficult to optimize. Small deviations from the null brought the tone boom back.

Center on the null as near as possible.

Now go back to the OSC/MXR 8975 board's R1, and move that back and forth from its position. The null should get deeper one way or another. In any case position this pot so that the null is as deep as possible.

Go back to the RF MIXER 80987 board, and move R1 around in a similar fashion. Again the null should get deeper in one position or another. Find the deepest null.

It should now be possible to use C2 of the RF MIXER 80987 board to trim out the last of the tone. Turn the AF-RF AMP up a little more to see if the tone is still there. Keep fiddling with C2, and the two resistors to deepen the null.

Be cautious of changing C1, when the volume is turned up. Minute changes in C1 will bring back the tone at ear splitting volume.

Working these three controls is not unlike finding an antenna tune using an antenna tuner. There will be some combination of all four controls that produces the deepest null.



Power Amplifier




Image thanks to ARRL