Print Friendly Version

Retired


Image of prototype board/assembly

The following are photos of serial number 1:

     


Gallery of built units

Project Status

We want to thank all of you for your patience and understanding, and we want to thank Ten Tec for the great service that they are giving us. Ten Tec proves again that they are a company that stands behind their work. Thanks again, Doug, James and Paul for NorCal.

October 30, We are making steady progress with the kitting of the NorCal 2030 project. This weekend we were able to get 2 more bags finished, and now have a total of 6 of the 10 bags of surface mount parts done. 4 more to go!! The cases are being shipped and should arrive this week. It looks like we will miss the projected Nov.1 shipping date by a couple of weeks. Thanks for your patience and understanding.

72, Doug, KI6DS, Paul, AK1P, and James KA5DVS.

Description

The NorCal QRP Club is pleased to announce the availability of the clubs newest transceiver project, the NorCal 2030. This kit was designed by Dan Tayloe, N7VE, and uses the famous Tayloe mixing scheme to produce a CW transceiver with excellent specs. The radio is CW only, and you may build it for 20 or 30 meters and parts are included to build it on either band. But you will have to choose which band you build it on. The kit comes with all parts, double sided, silkscreened, solder masked, plated through boards, connectors, hardware, and a custom silkscreened case. The kit is not a beginner’s kit, as it has approximately 400 surface mount parts, which are 1206 in size, and the IC’s are all SOIC8, SOIC14, or SOIC16. Experienced builders should not have a problem building this kit, but we do not recommend that beginning kit builders take on this project.

The cost of the kit is $160 delivered in the US, and $175 delivered DX. You may order it via paypal, or by check or money order in US Dollars. No individual confirmation of orders will be sent, but you may go here to check to see if you are on the confirmed order list. We will be doing only 200 kits, and when they are gone, they are gone. The target delivery date is Nov. 1, 2005, providing that there are no delays due to parts. To order a kit click here for ordering information.

Designer's Comments

There has been a lot of talk lately about SDR receivers. While I think the flexibility of these radios are very exciting, the wide bandwidth of these receivers tend to make them look more like superhets when it comes to large signal inference. The typical approach both SDRs and superhets have to solve this problem is to make the receiver more deaf. Attenuators are kicked in and RF preamps are kicked out. Gain is reduced. While large signal performance does go up, sensitivity goes down. The NC2030 is designed so that it can receive very weak signals while at the same time rejecting very strong adjacent signals.

The receiver local oscillator is run at a 1x frequency rather than the more typical 4x clock used for this kind of quadrature detector. I am using a 90 degree RF phasing section (basically a C/L/C section) to generate the second of the 2 phase clocks the receiver uses in the detector. This saves a lot of power, but at the cost of the opposite sideband rejection varying over the band. For the 20 and 30m rigs, this is typically > 45 db of opposite sideband rejection.

The receiver LO is a combination of a 3 MHz PTO and either a 13 MHz (30m) or 11.059 MHz (20m) VXO. One of the hardest things on a DC receiver is to get a consistent transmit-receive offset across the band. My old HW-7 (my first novice rig along with my DX40 and Drake 2-c), had an RX-TX offset that that was about 400 Hz at one end of the band, and 700 Hz at the other end. The VXO provides the RIT and the RX-TX offset, which, when mixed with the 3 MHz PTO, provides the desired uniform RX-TX offset across the band.

I really like the PTO portion of the receiver. This is an idea that I stole from an earlier NorCal kit. The PTO provides band spread in a fashion similar to a 10 turn pot. It takes about 7 to 8 turns to cover a 65 KHz section of the band. In addition, the use of a PTO for tuning allowed me to keep the Q of the VFO tank circuit high (no low Q varactor), allowing the receiver to have a very low phase noise LO necessary to complement the high signal level performance of the receiver. Many receivers tested by the ARRL have "noise limited" high signal receiver performance due to insufficient LO phase noise performance.

Analog VFOs may be "old school", but they supply a level of performance that is very difficult to achieve using other means. I would love to use a DDS. A DDS has great phase noise properties, but the spur problems generated by the use of a DDS chip, even the newer ones, make it difficult to justify it use in a very high performance radio. Plus, it would use more power than the rest of the current radio!

Two of the comments that I received over and over again was both the cleanness of the audio quality and way strong signals tended to "pop up" in without any warning. On most receivers, you can hear a strong signal coming as you tune the band. This radio has a total of 14 poles of low pass filter (9 poles of R/C active filters and 5 poles of SCAF) to keep strong signals out of the bandpass.

There are a lot of things that go into good, crisp, clean audio, but I think that the selection of the Butterworth filters in the audio help this a lot. I like signals that roll off fast as they exit the audio passband. However, the filter responses that roll off fast tend to be made of sections that have high Q.

Take for example a six pole Chebychev with 1 db of ripple. Designed with a cutoff of 1 KHz, this filter has 57 db of rejection at 2 KHz. However, this filter requires three two pole sections: 350 Hz w/ Q=0.76; 750 Hz w/Q=2.2; and 1000 Hz w/Q=8. It is the high Q of this last section that causes this filter to ring. Superhet cw crystal filters tend to have this problem as well. The sound produced is kind of a hollow ring. Listen to the audio files above and the ring/no ring difference is obvious. Noise from poor band conditions tend to excite this ringing tendency to produce a sound that is very hard to listen to for very long. Fox hunters know what I am talking about.

If a Butterworth filter is used instead of a much sharper Chebychev, more sections are needed to get the same response. For example, it takes a 9 pole Butterworth filter to get a similar 56 db rejection at 2 KHz. All nine poles are at 1 KHz, and the Qs needed are 0.53, 0.65, 1, and 2.9. Since the high Q section is now only 2.9, the Butterworth filter has greatly reduced ring compared to the smaller six pole Chebychev. Again, listen to the sound files for the difference that a sharp, low Q audio filter can produce.

The entire receiver runs off of 3v. 5v is used in a spot or two primarily for the frequency counter, the keyer chip, and the driver for the class E finals. A switching supply is used to efficiently convert the 12v supply to 3 and 5v used by the rig. The switching supply is kept in a separate enclosure attached to the back of the rig. The receiver draws a bit less than 30 ma at 3v, but thanks to the switching supply, the 12v current drain is only 11.5 ma. The actual current drain will depend on the actual supply voltage used. It draws less current at higher voltages, more current at lower voltage. The max voltage input is 15v.

I have run the rig at as little as 6v, but the 5v switching supply tends to drop out and the keyer stops working. If you really wanted to work at 6v, the 5v LDO regulators could be feed from the supply voltage instead of the switcher 5v output. The current design seems very reliable down to 7v where it will put out over 1w.

The transmitter runs class E in order to be more efficient. Most of my class C amplifiers have only been about 40% efficient. In this case, taking the drain of the entire radio into consideration when transmitting, the whole rig efficiency is about 65%. Even though the finals are very cheap (three BS170 MOSFETs, a ruggedized version of a 2n7000, perhaps $0.50 to replace all three), it is no fun to blow finals in the field. Class E finals seem to be very susceptible blowing when transmitting into an accidental short circuit. The finals have been provided with both over voltage and over current protection to keep this from happening.

As mentioned above, the receiver is a phasing type DC receiver. The detector uses a high performance "Tayloe" quadrature detector to produce the I and Q (in phase and quadrature) audio channels needed to allow the receiver to reject the USB side. I have discovered that I did not actually invent this detector. The detector existed in a more complex form roughly 10 years before I came up with the idea. In essence I came up with a simplification of something that already existed, making it capable of lower noise operation.

This receiver does not use AGC. There is a lot of debate about the use of AGC for cw work, and the two camps seem about equally split. However, it is very hard on the ears to be wearing headphones (this is a headphone only rig) and get thumped with a 3v pk-pk signal. A loud, comfortable signal is only 20 to 40 mV pk-pk. This radio uses a diode limiter to limit the audio to only 0.3v pk-pk. This is a loud signal, but tolerable. Although the diode clipping will create distorted square wave audio signals, the SCAF low pass filter that follows this limiter cleans up the higher order harmonics to the point that you can often not tell that the signal has been driven into diode limiting. For such a loud signal, reduce the audio gain until you get back to a more pleasant signal. The diode clipping is a welcome relief when confronted with sudden loud impulse noises such as thunderstorm lightening flashes or high power band "swooshers" (what do you call these guys that key down and sweep the band?). Such signals when clipped are no longer such a bother to the ear. A limiter inherently has a much faster reaction time than almost any AGC.

One small point in this design was the transmit-receive switching. Most QRP rigs use a simple set of back to back diodes to keep the transmitter energy out of the receiver. In most NE602 style receivers, this is not a problem. However, in a very high performance radio such as this, these diodes would start conducting on large signals and absolute ruin the IP3 (third order intercept) performance of the receiver. This receiver switch uses instead two MOSFETS (BSS123s) as switches to the receiver front end. On transmit, the series MOSFET is turned off creating a high impedance path to the receiver, and a shunt MOSFET is turned on, shorting the receiver side of the series MOSFET to ground. In receive, these two MOSFETs change state to allow the signal to flow to the receiver.

I think this is the 6th generation of this design since I started working on this seven years ago. The big hold up this entire time has been getting the design laid out. Trevor Jacobs K6ESE took up this task and has done a fine job. The bad news is that it took so long. The good news is that the performance has been steadily going up and the current drain has been going down.

The only down side to this design has been that on two occasions I have heard weak short wave broadcast detection, the bain of DC receivers. In this design, this is a function of the distortion rating of the first audio preamplifier. There are devices that are better, but they require a lot of voltage and power, and are fairly expensive.

The transceiver could be set up on other bands. Some experimentation would be required. 40m is a bit more problematic in that the percentage bandwidth covered on 40m (say 7.0 to 7.050 MHz) is larger than seen on 30 and 20m, thus the LCL RF phasing strip may cause the opposite sideband rejection to degrade to only 40 db over the entire band range. However a 10 MHz crystal could be used with the 3 MHz PTO to get to 7 MHz. This would create a birdie at 7.0 MHz. You would not want to use a 4 MHz crystal as it would not tend to VXO over a very wide range.

I encourage you to read the presentations on line to get a little better flavor of the performance of the rig. To simply quote numbers at various points masks what is really going on. Some of the blocking and dynamic range plots really bring this out. This Austin presentation is a bit better in this respect than last years NorCal presentation.

- Dan, N7VE
Download:

 Austin Presentation
 Pacificon Presentation
 Pacificon Sound Files

Supporting Documents

 NorCal 2030 Assembly and Operation Manual

 NorCal 2030 Errata

 NorCal 2030 Schematic

 Receiver Schematic

 Transmitter Schematic

 Power Supply Schematic

 NC2030 parts list table (Steve Weber KD1JV)

General specifications

Direct conversion receiver with single sided reception using a high performance quadrature detector (Tayloe detector) and audio phase shifting techniques.
Opposite sideband rejection: Opposite sideband suppression greater than 45 db across the band Input supply voltage: 7 to 15v Receiver current drain: Less than 12 ma at 13.8v.
RIT tuning range ~ 2 KHz total
Spot switch
Audio filtering: 9 pole Butterworth low pass filter 800 Hz fixed plus 5 pole variable from 300 Hz to 900 Hz Elliptic low pass filter (SCAF) plus 3 pole high pass filter, 350 Hz fixed. Butterworth filtering provides clean, no ring audio passband. 14 poles of low pass filtering provides a very sharp frequency roll off. SCAF variable LPF can be used to reduce high side QRM and to narrow up the bandwidth for weak signals.
Built in CW keyer with speed pot and one memory.
Built in audio frequency counter
SWR protected power amplifier
Audio limiting at 0.3v pk-pk, headphone operation only. ~104 to 108 db sensitivity headphones optimum.
Built in 3v and 5v switching supply

Measurements taken the 30m prototype

Sensitivity: -133 to -134 dbm (~0.1 uV pk) 3 db S+N/N MDS Receiver Bandwidth: 500 Hz, -6 db down at 350 and 850 Hz
Blocking: -14 dbm @ 2 KHz, -4.5 dbm @ 5 KHz, +6 dbm @ 10 KHz, +11 dbm @ 20 KHz Blocking dynamic range: 120 db @ 2 KHz, 130 db @ 5 KHz, 140 db @ 10 KHz, 145 db @ 20 KHz (noise limited) 3rd order intercept dynamic range (IP3DR), LSB: 94 db @ 2 KHz, 102 db @ 5 KHz, 108 db @ 10 KHz 3rd order intercept dynamic range (IP3DR), USB: 95 db @ 3 KHz, 102 db @ 5 KHz, 108 db @ 10 KHz
IP3: +28 dbm at 10 KHz
Transmit power: Roughly 3w (12v) to 4w (13.8v) Tuning range: 10.1 to 10.15 MHz

Measurements taken the 20m prototype

Sensitivity: -134 to -135 dbm (~0.1 uV pk) 3 db S+N/N MDS Receiver Bandwidth: 500 Hz, -6 db down at 350 and 850 Hz
Blocking: -16 dbm @ 2 KHz, -6.5 dbm @ 5 KHz, +4 dbm @ 10 KHz, +7 dbm @ 20 KHz Blocking dynamic range: 119 db @ 2 KHz, 128.5 db @ 5 KHz, 139 db @ 10 KHz, 142 db @ 20 KHz (noise limited) 3rd order intercept dynamic range (IP3DR), LSB: 93 db @ 2 KHz, 105 db @ 5 KHz, 109 db @ 10 KHz 3rd order intercept dynamic range (IP3DR), USB: 98.5 db @ 3 KHz, 102 db @ 5 KHz, 109 db @ 10 KHz
IP3: +28 dbm at 10 KHz
Transmit power: Roughly 3.5w (12v) to 5w (13.8v) Tuning range: 14.0 to 14.065 MHz

Ordering the NorCal 2030

SOLD OUT

This item is currently sold out. We would like to thank everyone who orderded this kit.

You may also check our Order status page to check on your order and see if we have received your order and shipped it to you.