KSB2 problem found, and solved

I finally solved the problem with my KSB2 board. It turned out to be a poorly soldered joint on one of the transformers that terminate the filter. I found it by writing down everything I knew about the problem and using that to systematically exclude different elements in the system.

The final clue was that receive audio was poorer through the KSB2 filter than through the variable filter on the RF board. Poorer, but not absent. That pointed me back to the filter.

I was ready to rip the filter out and rebuild it from scratch, but decided to have one more look at all the joints using an X3 magnifier. I could just see that one of the leads from T1 primary had an annulus of solder and that I could see the copper end of the wire. Sussed: I obviously hadn’t fully stripped the insulation when I made the transformer. A continuity tester confirmed the problem: high, but not infinite, resistance.

It was a 5 second job to re-solder the joint: leaving the iron on long enough for the remaining insulation to bubble off. Just for completeness, I re-did all the other solder joints as well.

The result is that the board now works fine. I’m not completely happy with the performance, or of the filter alignment, but that’s for the next post.

More tales of woe debugging a KSB2

For a couple of months now, I’ve been trying to get to the bottom of the problem with my KSB2 board: the SSB adapter in my K2. Not continuously of course: I’ve had all sorts of other things to do. I’ve built the KIO2, KDSP2, and KAT2 modules as well. However, these were really just ways of resting my brain from the toil of trying to get the KSB2 working properly.

The problem manifests itself as an almost, but not quite, total absence of RF when I use SSB. The K2 works fine on CW.

Looking at the circuit, when the KSB 2 is installed, it intercepts the Rx IF path with a controlled bypass around the built in variable bandwidth IF filter. On Tx, the audio from the Mic (which is unused on CW) is processed and turned into a DSB (Double Side Band) signal at the Intermediate frequency before passing through the new filter backwards (which filters out the unwanted sideband) and then passing back into the main circuit.

The user can choose whether the received signal passes through the variable bandwidth filter on the RF board, or through the fixed filter on the KSB2 board. On transmit, only the fixed filter on the KSB2 board is used.

Thus, the fixed filter is used on Rx and Tx. I reasoned, wrongly I now suspect, that because I could choose the fixed filter on SSB receive, and hear signals, I could exclude the filter as being the source of the problem.

That left the audio processing section of the KSB2 board, and the ALC (Automatic Level Control) circuit. The latter is different from that used on CW and serves to adjust the power of the RF signal to match the level chosen by the user. Because we are talking about a single side band signal, all of the output is in the frequency-translated audio signal. The output power is directly proportional to how load you talk into the microphone. To get better “punch” to get through interference, fading and other signals, hams often raise the Mic gain to the point where they risk “over-driving” the Power Amplifier and sending out a distorted signal. This is “a bad thing” but is often by hams.

As the Mic gain is increased, the PA tries to maintain itself in a linear configuration (to minimise distortion) by “feeding back” a proportion of the output signal to an earlier stage and use that signal to lower the gain of the PA. This is the ALC signal.

Most manufacturers have a front panel meter that indicates the magnitude of the ALC signal and leave it to the user to adjust the Mic gain to a point where the ALC is at their desired level. Elecraft has chosen to adopt a smarter approach. The user uses a menu to choose a desired level of over-driving and the rig then operates automatically to maintain that level.

The K2 uses different ALC circuits on CW and SSB, so I suspected this section of the KSB2 because if it wasn’t operating properly, it could suppress the drive to the PA too much and result in insufficient RF power: exactly the symptom I was observing.

I spent ages trying to understand the ALC circuit and determining if it was working properly. To get anywhere, I had to assure myself that all the preceding circuits were working properly: Mic processing, DSB generation in the mixer and sideband removal in the filter. To do this, I needed a standard to work against.

From the elecraft list I determined that 100mV of AF at 1kHz should produce 100% of rated power at whatever level I chose. i.e. if I wanted 5W I should get 5W. I set up a signal generator to inject a 100Hz sinewave at 100mV pk to pk into the mic socket and then used my oscilloscope to trace the signal through. It soon became apparent that the problem lay after the first mixer. Although I was seeing 1V pp at the output of the first mixer, I was seeing virtually nothing in the filter.

I don’t have the test gear to debug the filter, so I have reached out to the email list for assistance.

More modules built


After a break, I’ve been back building more K2 modules. I’ve built the KIO2, KNB2, KDSP2 and KAT2 modules over Christmas. That’s the lot.

All but the KATU module worked straight away, but I’ve got a problem with the latter. I’ve still got to resolve the lack of Mic gain on SSB and absence of RF in 20m before I try and debug the ATU.

The picture shows the KDSP2, KNB2 and KAT2 modules. The KIO2 is already attached to the top cover of the radio.

Building the KSB2 SSB Adapter board

After a brief break due to some International travel, I’m back in builder mode again. This time it’s the SSB adapter for the K2. My CW is still very rusty and not to be trusted on the air waves, so I really need SSB to make good use of the K2.

This picture shows to partially completed board on the bench. The KSB2 is a lot tighter than the main K2 boards and requires a steady hand when soldering. There aren’t any SMD devices luckily, but some components are mounted very close together.

It’s done, the K2 is complete

At least as far as the basic unit anyway.


After getting over the VCO problem, I’ve been able to make steady, if slow, progress towards completion. The filters are only roughly peaked because I don’t have an RF signal generator. I will rectify that later as my next project is to build an RF signal generator controlled by an Arduino.

I still have a number of option modules to build and install: KSB2, KAT, KNB2, KIO2, KBT2 and KDSP2. None are too difficult and I can pick those off at my leisure.


Making progress again

OK, we got past the problem with the VCO.

After tracing through the signals, I checked the levels on DIN and CLK against the indicative levels in the manual. They were the wrong way up!

At this point I got on to the Elecraft mailing list and started asking questions. Luckily, Don Wilhelm W3FPR was on hand to offer some advice. Essentially, the main reason for the VCO not locking was likely to be the values of L3 and T5 being such that the voltage swing on the PLL couldn’t get the VCO to change frequency enough to lock. The advice was to either adjust L3 or change the turns on T5.

I didn’t want to disturb T5 unless it was really necessary, so I did a bit of thinking. The PLL is calibrated when the K2 is switched on for the first time. The MCU swings the VCXO through its full range and measures the relationship between PLL data and VCO frequency. It uses the results to program the EEPROM U3. What if, I reasoned, if the EEPROM calibration had not completed properly? I decided to give the rig a full reset (hold keys 4-5-6 in when switching on).

Bingo! Immediately, the VCO locked on.

Next step was to adjust L3 to set the voltage range so that the VCO works on all bands.

More problems – my fault this time

I had to tweak T5 a bit, but I got the VCO to lock OK on some bands, but not on others.

Turned out that I had put a 47nF capacitor in C74 instead of 47pF. Not much difference!

Anyway, I changed it and all bands work OK.

Lastly, for this evening anyway, I got as far as checking the BFO frequencies.

Debugging the VCO problem on my K2

I’ve been away on business a bit over the past few weeks and have only managed a couple of snatched hours on radio construction. However, I did manage a couple of hours last night to try and find out why the VCO is off frequency.
I started off by seeing how the VCO frequency changed as I adjusted the controls. The main control seems to have no effect at all. I connected my Rigol ‘scope to test point P1 and there was no change as I turned the tuning knob. However, changing bands did affect the frequency. That would seem to rule out the VCO itself as a source of error, for now anyway.
Looking at the circuitry, it seemed that the obvious next place to look was the MC14145170 PLL synthesiser, U4. This takes the VCXO output and compares it with the VCO output to generate a varying control voltage that changes the VCO frequency to keep the VCO locked to the VCXO. U4 changes the frequency in 5kHz steps and interediate frequencies are generated by varying the frequency of the VCXO using a pair of varicaps and a LTC1451 DAC.
Probing around, it immediately became obvious that something is amiss. There should be clock and data signals coming from the microprocessor on the Control Board and going to the DAC and the PLL chip. There isn’t.
The microprocessor is running OK, because the front panel is working and driving AUXBUS to change the VCO is bands.
Clock (SCK) and data (SDO) are created on the Control Board by the PIC U6 and they appear to be dead as well.
Next stage is to work out why the PIC is not generating SCK and SDO, or why the outputs are being held high; which is an alternative explanation.

Ahhh, we have a problem

Testing the K2

  • Resistance checks
    • Pass
  • Assemble the Case and plug in the Front Panel and Control Board.
    • It’s starting to look like a radio now. Isn’t it small!
  • 4MHz Reference Oscillator Calibration Check
    • The target is 12090 kHz +/- 30 kHz. Mine is running at 12099.45. Pass.
  • PLL Reference Oscillator Range Test.
    • Target range is between 9.5 and 15kHz. Mine is 13.38kHz. Pass.
  • VCO test
    • Target is to be stable in the range 8 to 10MHz. Mine is 12.099.45 MHz. Fail 🙁
  • Time to get digging.