QCX with Blown Finals - Repair

Paul, KB0***[1], handed me the assembled QRP Lab’s QCX kit since being successful fixing a few other QCX and QCX+ kits. He mentioned there was a transmit problem after building it around the first of 2018. Since not working right, it had been sitting in his closet for over 3 years.


[1] The name and call of “Paul” and “KB0*** are fictional and being used to protect the actual person’s identity.

A quick visual inspection was done after cleaning off all of the flux on the PCB’s solder side. There was an attempted repair made along with missing PCB traces in the immediate area of the finals – Q1, Q2, and Q3. Maybe it was going to be all right, but something “told me” that it needed some help. Also noticed was the firmware (f/w) being version 1.00e which is a very early version, one that definitely needed to be updated.

When first turning on anything like this for the first time, I always check the current draw…

  • Receive current draw (with voltage set at 13.8 V): 122 mA

    • Right about where it should be based on past experience with the QCX and QCX+

  • Transmit current draw: 155 mA

    • Something was seriously wrong with the output. It should be right around 500 mA as seen previously.

    • That output correlated with seeing ~2.9 Vpp on the O’scope where it should have been ~40 Vpp.

The transmit current and output voltage levels were noted as being quite low. The finals, Q1-3, were highly suspected as having been blown since having seen the condition before with a QCX+ kit for another ham friend. Still needed to verify along with seeing if the TX/RX switch Q5 was working right.

Well, I needed to get some new current firmware version PICs ordered – one was needed for this QCX, and needing another one for me (since having version 1.06 in mine). Needed to get that in prior to getting the QCX rig fixed so that I could get the GPS-based frequency alignments completed when nearing done.

About 4 weeks later after receiving the new PICs:

    • Removed and replaced IC2, ATmega328P, from the original firmware version 1.00e with version 1.07a.

    • Repaired/rebuilt the “traces” for Q1, Q2, and Q3 along with installing new BS170 FETs (500 mA versions that I had previously ordered).







  • Powered-up the QCX, current draw nominal at 115 mA (volume was at minimum).

  • Performed the alignments as per the QCX manual: Peak BPF (C1), I-Q Balance (R27), 90-degree audio phase shift (I-Q Balance, 27; Phase Lo, R24, and Phase Hi, R17). All of the adjustments seemed to be “good” or very close noting it had been setup pretty good initially.

  • Frequency calibration performed with the QRP Labs QLG1 GPS Receiver for that Ref Osc, and Sys Osc.

Now, checking the transmit side…

  • Oops, something still going as only seeing ~3.2 Vpp on the output.

  • Verified the output at IC3a (IC3.3) being good - ~ 5 Vpp.

  • Figured that Q5, another BS170 is probably “smoked” too having seen that one before. So, removed and replaced Q5 with a new BS170 FET (on the 500 mA version that I have)

Also, while in that area of the board messing with Q5, the key shaping circuit’s components were updated from the original values to the new values (that started with version 4 of the QCX instructions, and incorporated into the QCX+).

  • Changed out R41 from the original 470 W to a 150 W resistor (1%, ¼ W), R42 from a 1k W to a 1.2K Wresistor (1%, ¼ W), and C31 from the original 1 µF to at 2.2 µF capacitor.

Well, there was still something amiss with the output as only seeing ~7.5 Vpp. The LPF was even checked out (for the toroids’ number of turns, the capacitors’ values, and the PCB traces). Replacing Q5 was a step in the right direction with output level coming up a bit, but there was something else wrong. In checking-out the key shaping circuit, Q5 was not behaving right when probing with the O’scope probe.

  • Removed and replaced Q4, a MPS2907, with a new MSP751 transistor. It was seen that earlier QCXs had Q5 as a MPS2907 (as this QCX was), Q5 has been a MPS751 for a while.

Wow! That was the last piece of the “low output puzzle”. In transmit, Paul's QCX was now putting out about 40.5 Vpp which is about 4.5 W out. Good stuff!

The QCX is now good in both receive and transmit. Done!

What was worked on and ….replaced (screen capture of a portion of the QCX’s schematic)…..