Sudden Transceiver

While perusing a 2013 issue (#156) of the GQRP publication called SPRAT, I noted a clever project called the Sudden PSK receiver authored by Steve, g0fuw. It had two active devices, one being a 2N3819 arranged as an RF amplifier and the 2nd a NE602 arranged as a Product Detector. My first thoughts – why couldn't’t this be made into an SSB transceiver?

I have a history with such an undertaking, as I turned a Wes Hayward W7ZOI, high performance receiver project as described in Solid State Design for the Radio Amateur, pp 104-105, into a full fledged 20M SSB Transceiver. Thus, I knew it was possible to make such conversions. Fast Forward -- the Sudden PSK receiver is now a Sudden Transceiver! Here it is in all its glory.

[Later you will see how the Sudden Transceiver was used with the Digital Modes. Imagine that, a totally homebrew, scratch built FT8 rig? The import here is that the Sudden Transceiver is now an experimental platform. ]

We are indeed fortunate for one of the Sudden Transceiver builders has been keeping notes on the progress of his build and is sharing those notes with the greater ham community you can find those here Many Thanks to Nick Tile, G8INE who authored the notes. Nick is also building the Sudden Transceiver using Surface Mount Techniques and a CNC Milling approach. Nick has also generously agrred to share his Gerber Files of the board he has generated. You can contact Nick at the following email

There are three questions asked about component values in Nick's write up. Make it 100 nF and the 470 UFd is an electrolytic and you can upsize the 10 Ufd to 100 Ufd.

It is always a good idea to know where you are going before you start a journey. That is unless you intend to wander aimlessly in hopes that some desired end state would be achieved. I am a "noodling type" of guy and thus must have a road map before I start.

Essentially I am looking to develop a design that works (and works well) on 40 Meters; but circuit constants will be provided for 20 Meters. The 40 Meter band was selected, given the sunspot low, and very likely offers the opportunity for more contacts (and more fun) at 5 watts versus what might result trying to run with the "big guns" on 20 Meters. Frankly my experience here on the left coast of the US has shown a greater tolerance for QRP operation on 40M versus 20M. But that may just be the nature of the left coast.

There was some effort to make the 40 to 20 transition somewhat painless. In the design of the Band Pass Filters a few simple changes will move the networks to 20 Meters and all that would be required is a new Low Pass Filter for 20 Meters. The Arduino/Si5351 combo has constants that will continuously tune 160M to 10M but the default starts at 7.2 MHz. A new default start up is a simple code change. But let us focus on 40 Meters and get the rig working on that band first and then YOU can move it to any band with the exception of 17 Meters which is NOT RECOMMENDED given the 9 MHz filter being used.

Having recently decoded the "how to" have two VFO's in a rig so that is another goal. I want to avoid plagues such as the Dishal Dystopia or the AADE Build your own SSB filter and thus opted for the GQRP club 9.0 MHz Filter. Frequency control will be via an Arduino Nano and the Si5351 PLL board. For a display I consider the seasick green 16X2 LCD to be a persona non-grata and therefore, we will go with the POSH 160 X 128 Color TFT. [POSH is an Olde English term for Port Out Starboard Home which relates to the most desirable side of a ship in making a journey from England to India. Thus those who went POSH traveled in style.] That said the only POSH part is the far better display and NOT the cost. I buy the Color TFT Displays as used on the Sudden for less than $5 delivered to my QTH and the 16X2 LCD is only slightly less costly.

At this juncture I would like to take a moment and bask in a bit of elitist glory. This rig is truly scratch built, not a kit and in fact includes many aspects totally designed by me as well as being hand built by me. Today there is much buzz about rigs like the uBitx and the QCX and rightly so -- they are indeed superb rigs but essentially kits !!

There is nothing wrong with kit buidling. BUT imagine your pride in scratch building something that was not a kit and like the Imprimatur of the old Catholic church it bears YOUR stamp of approval: Nihil Obstat (meaning No Objections).

You can build this project with common parts that are readily available. So that is not an excuse. Similarly the rig is built in modules so that each module can be built and tested. True it is not Plug and Play but that is what is unique about being a ham. We have a curious nature, we push the envelope and we discover things. I would have never believed had I not done it myself that a serious radio transceiver could be built using primarily two NE602 IC's. Press on and do your own basking.

To better understand this block diagram it would be helpful to see it in two pieces with those being the Receive configuration and the Transmit configuration

There are three relays that basically make possible the transformation from a receiver module to a transmitter module. Two of those relays are SPDT, where in the NC position the Receiver Band Pass Filter is connected on the front end and the Audio Amplifier is connected on the back end. The LO signal goes to NE602 #1 and the BFO goes to NE602 #2. On transmit the Microphone Amplifier is connected on the front end and the follow on Transmitter stages are connected on the back end. Only now, the BFO is supplied to NE602 #1 and the LO is connected to NE602 #2. The switching of the LO and BFO signals is done with a DPDT relay that is cross connected. The relay position is shown in the NC "Receive" condition.

[Note: the relays are from OMRON and are the G5V1 series with 12 VDC coils. Be sure to use a "Snubber Diode" across the field coil. The back EMF can be quite large and you could wipe out your circuits if the diode is omitted. In case you are wondering -- it is the field collapse generated EMF that is a fundamental element in switching power supplies.]

CLK2 on the Si5351 is the BFO Generator. For the Sudden = 9 MHz (But to be technically correct 8998500 for LSB and 9001500 for USB. This seems reversed; but because of Sideband Inversion with the LO operating above the incoming signal the sidebands are reversed.)

The incoming signal after passing through the Receive Band Pass Filter is directed to Pin 1 on the NE602. To Pin 6 is fed the LO signal, at 16 MHz, which results in a mixing action producing two distinct products. One product is at 23 MHz (sum) and the other at 9 MHz (difference). Only the 9 MHz component is passed to the 9 MHz filter connected to the output on Pin 5. Note we are using the unbalanced configuration of the NE602 and thus only Pins 1 and 5 are used for In/Out.

After passing through the 9 MHz filter, the signal is directed to Pin 1 on the second NE602. The BFO signal at either 8998500 or 9001500 Hz is fed to Pin 6 where again a mixing action takes place with a sum frequency in the 18 MHz range and the difference is in the audio range regardless of what BFO is used. For Receive it is the audio range signals that are passed to the audio amplifier stages.

Recap: In receive, LO (16 MHz) to Pin 6 on NE602 #1 with a resultant frequency of 9 MHz and the BFO (9MHz) to Pin 6 on NE602 with a resultant frequency in the audio range.

Our relay switching scheme does two things: 1) Switches modules on the input and output sides of the NE602’s and 2) Redirects the LO and BFO signals.

In transmit Pin 1 of NE602 #1 has now been switched from the Receive BPF to the Microphone Amplifier and now Pin 6 is fed the BFO signal at either 8998500 (LSB) or 9001500 (USB). The signal coming out of Pin 5 (NE602 #1) is really an interesting case study of DSB.

Assume we are using the 8998500 BFO signal and are feeding an audio signal from the Microphone. The two products would be the 8998500 + the audio and the 8998500 – the audio. But the filter with a Center Frequency of 9 MHz would only pass the signals that are the 8998500 + the audio since the other component is OUTSIDE the 2.1 kHz pass band of the filter.

Now let’s look at the case where the 9001500 Hz BFO is used with again our audio range signal. The two products would be 9001500 + the audio and 9001500 – the audio. Only the 9001500 – the audio would pass through the filter with a Center Frequency of 9 MHz as the sum product would be outside the filter pass band.

After going through the Crystal Filter, the 9 MHz signal is directed to Pin 1 on NE602 #2 and now the LO (16MHz) is fed to Pin 6. Here again two products: with one at 25 MHz and the other at 7 MHz, the desired signal. The two outputs are fed to the Transmit Band Pass Filter where only the 7 MHz signal passws through BPF on to the linear amplifier chain.

The redirection of the LO and BFO signals was accomplished using a small cross connected DPDT relay so that the cable feeding Pin 6 on NE602 #1 has the LO signal on it during receive. But when in transmit mode that cable is now carrying the BFO signal. Likewise, for NE602 #2 during receive the cable feeding it's Pin 6 has the BFO signal and when in transmit the cable is now fed the LO.

Recap: In Transmit NE602 #1 has the Microphone Amp feeding Pin 1 and the BFO feeding Pin 6. Likewise, NE602 #2 has the LO being fed to its Pin 6 and the output is to the Transmit Band Pass Filter.

The first thing I saw that needed changing was that the 2N3819 would be replaced with an additional NE602 whose purpose now on receive would be the receiver mixer stage
The two-pole crystal filter was changed to a commercial 9 MHz crystal filter sold through the GQRP club
What resulted is what I call a single pass IF Module comprised of two NE602’s and the crystal filter. Signals are routed to the input and output of the module and pass in one direction as opposed to some currently popular topologies where in a similar module the signals are passed in both directions. Get over it, this is not a BITX!
In the receive mode an off-board, receiver RF amplifier passes the amplified received signal through a Band Pass Filter (located on the IF Module board) to a SPDT relay, also on the board connected to the IF Module. The Normally closed contacts of this relay connect the BPF to the Input side of the NE602 (Pin 1) via a 10nF capacitor connected to this relay. However, the BPF termination is 50 Ohms and the Input Z of the NE602 is 1500 Ohms. Thus we need a 30:1 impedance transformation. This is easily done with a 2 Turn to 11 turn broad band transformer connected right after the band pass filter to the NC contact of the relay. Use #26 wire on a FT-37-43 Ferrite Core. For the EMRFD and uBitx illuminati who watch such things 2^2 = 4 and 11^2 = 121. 121/4, which is very close to 30.
On the output side of our IF Module we have yet another SPDT relay where on receive the output (NC) is connected to a healthy audio amp comprised of an NE5534 driving an LM380 (14 pin) audio amp. A point about gain distribution across the rig. Too many times high power RF amplifiers are introduced on the front ends of receivers. So while you are boosting the signal you also are boosting the noise. Slightly less gain on the front end and more gain on the back end (strong audio amp) is more desirable to maintaining reasonable signal to noise ratios! Later you will see that my Receiver RF amp stage has a board mounted gain pot just so you can tweak the RF gain while holding back the noise.
On receive the LO signal (above the signal frequency by the IF) is fed to the 1 st NE602, Pin 6 via 10nF capacitor. In this mode the 1 st NE602 is a receiver mixer and the single output is taken from Pin 5 via a 10nF capacitor. Matching to the 500 Input Impedance of the crystal filter is another broad band matching transformer. The Z out of the NE602 is 1500 Ohms and thus we need a 3:1 match. This is done with a 12-turn solenoid (single winding) transformer tapped at 7 turns. The 12-turn winding connects to the 10nF capacitor on one end and to ground on the other. The tap connection at 7 turns connects to the crystal filter. For the illuminati who watch things closely --- 12^2 = 144 and 7^2 = 49. 144/49 is close to 3 which the 1500:500 Ohm transform. For those uncomfortable with the solenoid you can use a 12 turn and 7 turn winding: Observe the phasing. For those who can’t phase –use the solenoid!
A similar transformer is built on the other side of the filter and this connects (12 turn side) to Pin 1 of the second NE602 via another 10nF cap. The wiring is almost the same as the 1 st with the BFO connecting to Pin 6 via a 10nF and output via Pin 5 via another 10nF.
Pin 3 on both devices is ground and Pin 2 on both connects to a 100nF to ground. There are no connections to pins 4 or 7. Power is supplied to Pin 8 NO MORE THAN 6 VDC via a filter comprised of a 100nF at Pin 6 to ground then a 1000 uH rf choke and finally a 10 Ufd to ground.
On Transmit our two SPST relays would switch over to other circuit modules. On the front end in transmit we now have the 1st NE602 connected to the microphone amplifier and what was the receiver mixer is now the Balanced Modulator. Following to the 2nd NE602 which was Product Detector on receive is now the transmit mixer. Following the 2nd is another matching transformer (11:2 turns) and a duplicate Band Pass Filter. Again signals are passing in a single direction.
Why we use the single pass is that the LO and BFO signals are swapped in the two NE602’s. The 1st gets the LO on receive and the BFO on transmit. The 2nd gets the BFO on receive and the LO on transmit. My initial thoughts was to do this all in software where I would have the outputs for the LO and BFO swap clocks. A transmit command to the Arduino would change the CLK0 from an LO to BFO signal frequency and CLK2 on transmit would go from the BFO to LO signal. I could get it to swap but somehow in the process the Si5351 would get confused and in going from R to T and then back to R –it was not on the same frequency. I decided I would not make that a science project and so resorted to a cross connected DPDT relay that would do that routing for me. At times a hardware solution overcomes lack of software skills!

Below is a Block diagram of the prior discussion and a drawing of the cross connected relay. The BFO is either at 9001500 for USB or 8998500 for LSB as there is a sideband inversion. The LO shown in the block diagram as 16 MHz that would be for the 40 Meter version which is the primary build. For 20 Meters that LO would be 23.2 MHz for start up.

Let us review what needs to be built to have the receiver working and this is best done by looking at a block diagram of just the Receiver sections and how the pieces go together.