A while ago I saw a review of the Penntek TR-35 and I thought to myself that this could be one of the best options for a super-portable POTA CW transceiver. It is incredibly small, has very low power consumption, has great audio, pleasantly narrow CW filtering, has four bands that are heavily used in POTA (40m, 30m, 20m, and 17m), and is a kit to boot. I ordered one with the optical encoder option (now standard) and then life got in the way and it sat on my shelf for longer than I'd like to admit. Then one weekend I had plans fall through that suddenly gave me unexpected free time, so with a desire to spend some time behind my magnifying glass and soldering iron, I pulled down the TR-35 from my shelf and got to building.

The first thing I noticed was that the kit was organized better than any kit I had ever done. Each step on the manuals had a sealed bag connected on a strip that kept them all together. This eliminated one of the more tedious first steps for building a kit: part inventory. It meant that instead of having a big pile of parts, you only had to deal with what you needed for that step. Absolutely fantastic. The kit consisted of an upper board with the user interface elements (display, knobs, LEDs, and switches), a lower board with the RF components, and a specially made case. Construction started with the upper board and installing the 16 MHz crystal which needed to be installed with a gap from the board. Next came a pin header and the OLED display which was cleverly suspended from the board with standoffs and connected via leads from sacrificial resistors. Then LEDs, capacitors, jacks, switches, potentiometers, and the tuning encoder were added. The assembly was straightforward, although the encoder did require some fairly fine solder work with only a few millimeters between the pins. Finally the switches and knobs were soldered using the front of the case as a jig to ensure alignment. It sounds complicated, but it was fairly straightforward in practice.

Next came the lower board assembly, starting with the relays and a polyfuse. Then some jacks and potentiometers were added along with some FETs and a series of capacitors. Then it was time for the final amplifier, an IRF510 FET complete with a heat sink, mica insulator, and a plastic spacer to provide electrical isolation between the FET's metal tab and the mounting hardware. This proved to be a tricky step as the initial placement of the FET requires it to be bent over. When I first did this there was enough tension from the bent over leads that even with the plastic spacer the FET was shorting out to ground. I had to disassemble things a bit and bend the leads so that it was no longer pressing against the mounting screw. After a few attempts the continuity function on my multimeter was no longer beeping at me and we had success. The next step was the installation of the many toroids required for the radio. There are many people who do not like winding toroids. I don't mind winding them and find the process somewhat meditative, but I understand how they could be frustrating. For the toroid averse, this kit provides the answer: all the toroids needed are pre-wound at the factory and the enamel pre-stripped as well. If you can install a capacitor on a board, you can install the toroids here. Simple as can be. The toroids were then followed by a bank of 4 filters that needed to have their cases grounded as well. Then we were near the finish line and needed to do a bit of assembly to get the pin headers lined up so that the lower assembly would be electrically connected to the upper assembly. After a lot of solder connections were made, the two boards were mated and the microcontroller was installed and we had an assembled radio.

Now it was time to check my work. There were resistance checks for the 3 power supply rails to ensure there were no shorts, and then I hooked it up to an antenna, power supply, and an external speaker and powered it up for the first time. Right away the display came to life and showed 7030 kHz. I could hear sound too and no "magic smoke" had been released - a welcome result. I went through several tests of functionality ensuring switches did their thing, knobs changed the desired functionality, and then made my way to the encoder. While it could be clicked to change the tuning steps, it wasn't changing the frequency on rotation. Something was wrong. Remembering the tiny pins that needed to be soldered earlier and the fact that it had partial functionality, I powered everything down and disassembled the boards to inspect the pins. I saw some pins that may have been a bit light on solder. I had been conservative on the solder to avoid bridging the fine pins. I gave each pin a bit more solder, gave a continuity test to each pin to make sure I hadn't bridged them, and then re-assembled the board. After getting everything connected the encoder changed the frequency as it should have done from the start. Success.

Next it was time to bias the final RF amplifier, a process that involved using a digital multimeter in current measuring mode to measure the receive current drain (in my case a very small 79 ma) and then to transmit at the very lowest power output which caused the current draw to go up about 11 ma. From there you adjusted to have the draw go up another 5ma and the PA was successfully biased. Further adjustments were made to set the CW sidetone level to a pleasant volume with a potentiometer. One thing that is unique about this kit is the sidetone you hear is actually your transmitted signal, giving you silky smooth QSK. If you're using an unprotected lithium battery, you can even set the voltage for a low power warning. I will be using a regulated supply or a battery with a BMS, so I skipped that step. Finally, you adjust the CW narrow bandpass filter to be properly centered by listening for peak amplitude and then you're done. All that was left to do was assemble the case, install the knobs, LED lenses, vinyl switch covers, and the rotary encoder.

After getting it all together I started spinning the dial around the bands and I was amazed with how low the noise floor seemed to be. The CW signals seemed to pop-out from the background noise and sounded quite pleasant. There are two CW filter modes, one very wide filter around 1.6 kHz and a narrower filter at 350 Hz, albeit with somewhat soft skirts. At the narrower filter setting noise and adjacent signals drop dramatically. I started calling CQ and looked onto the reverse beacon network and saw I was being received well into Europe on the East and West nearly across the US on 40m. Sadly, there were no takers to my CQ, and I decided to give the 30m band a try.

On 30m I had similar RBN results, getting into Europe and across the United States. I heard a signal about 1 kHz down just faintly as I was calling CQ so I paused for a moment to see who it was. It turned out to be Wolf, DF2PY in Germany calling CQ. I had the pleasure of working him in the past, so I decided to return his call. He heard me gave me a 549 and I heard him a solid 559 and we had a nice chat about home built rigs (he also likes to build rigs) until we were taken out by some QSB. Not bad for a first contact on the rig. In the next day I was able to contact Curacao on 17m as well with 1 Watt. I also made contacts with Oklahoma, Mississippi, Missouri, Pennsylvania, Michigan, Florida, New York, New Jersey, and Central Virginia on the 40, 30, 20, and 17m bands. The rig is a lot of fun, and most of the stations I called I was able to work. I even broke a few POTA CW pileups by going off frequency and using the RIT control to listen to the activator on their frequency, thus making my signal stand out from the zero beat crowd. I am very pleased with the results.

So now that we know it works, what can it do? Well, the power output is variable and somewhat based on input voltage. On the lower bands I can get as much as 9 W out with the output power getting lower as the frequency goes higher. There is an RF output knob that allows you to change your power output, if desired. One thing you notice about the TR-35 is there are no deep menus to learn, almost every function has a button or knob dedicated to it. There are knobs for keyer speed, RF power, RF gain, and volume as well as an optical encoder for the VFO that is smooth and stepless. Pushing in on the VFO encoder changes your step, allowing you to go by 10 Hz and 100 Hz increments with a short press or by 1 kHz increments with a long press. When it comes to switches, you have a power switch, as well as a couple momentary switches that are bi-directional. The first momentary switch allows you to change the receiver mode from wide to narrow bandwidth if pressed up quickly, or to a full SSB width if long pressed. If pressed down it allows you to access the CW memories (there are 2) that are selectable by hitting either the dit or dah side of your paddle. A long press down allows you to program those memories. The final switch allows you to switch bands with a single fast press cycling 40, 30, 20, and 17 and then back again. If you long press up on this switch it will store the current frequency into a band memory. To recall that memory you would press up quickly twice. Finally pressing the switch down engages RIT allowing you to modify your receive frequency while maintaining your transmit frequency. Very helpful for those callers who may be calling you off frequency and not in the center of your filters. This can also be used to allow for split operation, albeit in reverse. Long pressing down will engage a dial lock, useful if you're running a frequency and don't want to accidently go off frequency. There are also LED indicators for signal reception (which shows as blue), red for a low battery warning that is configurable to your preferred voltage via an internal potentiometer, and RIT activation via an amber LED. Another useful feature is that this radio has 2 key input jacks: one for a paddle and another for a straight key. You can have both hooked up simultaneously, in fact in my shack I keep both connected to make for easier tuning with the straight key, or to switch things up if I feel like going manual.

So everything sounds pretty good, right? Well there a few things to note that are missing. There is no internal way of knowing what your output power is, and the output varies by band and input voltage. If you have a wattmeter you could make a table of dial positions and output powers, but you won't see it on the screen. Also, the TR-35 does not have an SWR meter, so you'll want to pack one, or have an antenna tuner available if you plan on using a non-resonant antenna. I have these in my arsenal, so I can deal, but it would be much nicer if the SWR were displayed either on the OLED screen or via some LED metering. Finally, I do get a significant reduction of output on the 17m band, and I've reached out to Penntek to see if they have any idea why this is. I have made plenty of contacts on 17m with the 1-1.5 W I've measured from the radio, but when 40m and 30m are 7-9 W and 20m is about 3.5 W, 1-1.5W is a big drop-off.

All in all, I have been very pleased with my new TR-35. At time of this review they are currently going for $319 as a kit or $419 fully assembled from the Penntek web site. While not the cheapest radio out there, I have been impressed with the quality of the kit and the wonderful sound that it is capable of producing. I look forward to getting it out in the field for years to come.