Red Hot NorCal NC20 "Manhattan Style"

Impressed by the 2N2/x projects of Jim Kortge, K8IQY using the "Manhattan" construction style I started rebuilding another famous QRP transceiver, the NorCal NC20. This RTx is designed by Dave Fifield, AD6A and sold as a kit named The Red Hot Radio RH20. At work I found some old enclosures containing two PC-boards, one in each U-shaped part of the cabinet. Besides that, the inside of the cabinet has a gray, conducting layer that seems to be excellent for shielding. After having read the article from K7QO about "Manhattan Construction" and encouraged by Jim, K8IQY I started building in January 2002. My friend Rien, PA0TRT bought me a punch at Harbour Freight Tools during his stay in the USA last year. From 1/32" copper clad board I punched a few hundred 3/32" pads. At this moment (Easter 2002) the VFO, AGC, audio circuits and complete Tx are finished. The output is a clean 5 Watt sinus and within a few weeks the whole transceiver should be ready. I have made an impression of my work with a simple webcam. All pictures speak for themselves. Besides a discrepancy in pin connection of two different PN2222A product specifications (reverse emittor/collector), through which I mounted the transistor in the wrong way, everything works great.

May 2002

After a few weeks break, waiting for some parts, I decided to continue my design work starting with the Rx/Tx/Mute circuit and than the rest of the receiver. The advantage of building backwards is the possibility to test each circuit for proper functioning. I spend the Whit weekend designing the parts layout for the Rx and building the product detector and IF amplifier. Due to the given dimensions of the PC-board, it was hard to place all Rx components in the available room. Whatever I tried, I couldn't fit the 2N4427 with heatsink in my design. There just was not enough space even with the reserved room for the keyer circuit. After melting solder for several weeks I now had the problem I was afraid of. Mostly one builds a circuit and thén you make (or buy) a proper cabinet. I did it the opposite way because of the nice shape and dimensions of this enclosure !

So I have chosen now for an alternative solution by using a BFR96S transistor that should be suitable for this post mixer amplifier function too. The transistor will be placed, with some conducting paste, directly against the copper-clad board. The whole 'ground plane' will act as a heatsink.

June 2002

Some response of leading QRP hams after my message to the QRP-L email reflector list.

"WOW! What a splendid job you did building your NC-20. Bravo my friend. As nice of a Manhattan-style building job as I've ever seen, including my work, or Chuck Adams. Wish I could help you with the filter, but don't have any crystals on the frequency that you need. 72, and hope to work your wonderful rig one of these days. Thanks for sharing your effort."
Jim, K8IQY

"Gang, I recommend you look at Carel's photo's of his homebrew NC20 if you are into the Manhattan Style thing (like I am). Absolutely top-notch construction, and the photos are virtually a step-by-step on how to do it.
Great job, Carel. Having held Jim Kortge's original 2N2/40 in my hands, I can say you built yours with the same level of care and beauty. Only a few have reached that level of "art."
72, Paul NA5N

"Very, very, nice work Carel. Thank you for sharing your work with us. I've thought of doing a Norcal 20 and your work has made me seriously consider it."
73, John N1QO

Dit dah dit, dit dah.... the first sounds from RA1CM are reaching my ears. Yés, the receiver is working too! This time I had some more serious troubles. I have been building up the complete Rx without testing any circuit because of the absence of the wiring between both PC-boards. When powering up the Rx only a slight noisy sound was heard. After peaking the front-end filter for maximum noise, things were beginning to move. The DC-voltages of all transistors were OK, but the receiver seemed to be almost dead. Time to pick up the one and only original example, the Red Hot NC20 build by my friend Wim, PA3AJI.

I compared every stage using an RF generator and oscilloscope and the problem was quickly found. The roofing filter crystal was defect. There was some mechanical movement in one of the leads, probably caused by myself when bending them. After replacing the crystal by another one, the shack was suddenly filled with the oh so whished sound! Compared to the RH20 my homemade Rx is doing an (almost) equal job except for the IF filter response of course. I didn't use optimal matched crystals. Out of a batch of 18 crystals (8.867.238 Hz) I picked the most equal ones. The next step is to match them properly and measure the response.

For all my effort in rebuilding this NorCal NC20 using the "Manhattan Style", Mister Red Hot Radio himself, Dave Fifield, AD6A offered me a set of matched 9MHz crystals along with an Audible Frequency Annunciator (AFA) chip and a TiCK-1 keyer to complete my work. Ain't that great...! I'm anxious to know when the postal parcel arrives here.

On Tuesday June 11th 2002, 10:15 pm the first QSO was made with PA3AJI. Due too his help by putting his RH20 at my disposal, he earned to be the first one to hear my homemade NC20. The rig sounds FB. A beautiful 599 report. No chirp at all, even without the dual regulator mod for the VFO.

August 2002

I never received the components that Dave promissed me, but the rig works fine with the parts I've used. Lately I found a link to another Manhattan style NC20. Jeff Hecht, K8GD has build one on a 5x7 inch PCB and also made a diode ring mixer using 1N5711 hot carrier diodes and FT37-43 cores. See his work here.

April 2003

After several months of just using the NC20, I jumped into the world of PIC programming. Until now my K9 (K1EL) keyer chip was put into the socket placed on the upper board op the rig. Because I like the K9 features and the way it works, I decided to experiment with the sourcecode of the K8 model which is available at Steve, K1EL's website.

Being a newbee in micro controllers, I had to learn a lot. The original code was written for the 12C509A which is a One-Time-Programmable chip. Not a type to learn how things work. Luckily Ian M. Wilson, K3IMW modified the code for a PIC16F84A. His job and some nice changes worked out by Jerry B. Hall, W0PWE were a great help to me in migrating the code to a PIC12F629. Several changes are made now like 11 characters for the callsign (need that outside the Netherlands like: DL/PA0CMU/m), a quick speed mod (W0PWE), better shaped sidetone (less thumb, by changing GP4 from output to input between dits and dahs), usage of the pushbutton to stop the input of Call & Message and some other personal defaults. The keyer works great and now I'm studying how to save defaults to EEPROM.

September 2003

During the monthly meeting of the local VERON Radio Club in September we did some measurements with a Advantest R3465 Spectrum Analyser. Unfortunately one of the pictures is lost so we have to measure again some time. Special thanks to Henk, PE0SSB who did the measurements.

Photo Gallery of the building process[hover over thumbnails to view]

The start of this project was somewhere in December 2001 gathering components, a suitable cabinet and PC-board. During Christmas Holydays I colored all components on a copy off the RH-20 silk screen to know where every component was situated that belonged to a certain part of the tranceiver. The first soldering steps were on Januari 6th 2002 and I've made nearly 100 photo's of the building process. At that time I didn't have a digital camera, only a very simple webcam. Despite that low-tech gadget, the quality of the pictures is acceptable.

In May 2005 Rien, PA0TRT and I visited the Hamvention and FDIM in Dayton. As promised to K8IQY and K7QO, I took my NC20 to America to show it in Dayton. Not knowing there was a Building Contest, I left mine judged by the jury and won the second price, an Elecraft XG1 test oscillator kit.

...