 |
|
 |
|
HBR-2000 HF 160 to 6 meter High Performance All-Mode Transceiver Latest Updates: January 19 2009: Added photos of 100 Watt Amplifier stage. "100 Watt Amp.". January 12, 2009: Added note regarding compensationg VFO Drift:VFO and Frequency Control December 20, 2006: Revised Receiver Measuremetns. New Measurements December 18, 2006: New Receiver Front End. New Front End December 4, 2006: New VFO details. VFO and Frequency Control April 17, 2006: New section added for Front Panel Design Front Panel Layout March 2006. QST publishes article on the HBR-2000  This page provides a brief description of the receiver portion of a HF (160 to 6 meters) High Performance Transceiver that I built. I named it the HBR-2000. HBR is short for homebrew and 2000 is the year that I first heard a signal from the receiver speaker. The transceiver is under constant revision and therefore complete circuit diagrams are not provided. This section, hopefully, will be an incentive to those who read it to build their own amateur radio equipment. I am not an electrical engineer, just a true amateur radio operator who likes to learn and build radio equipment with my hands. |
Front Panel View
|
2. Design Process The design process began by deciding on features and specifications that would suite my particular operating style. With other hams in close proximity, the receiver had to have a very strong front end. As well, I often operate in various contests so good selectivity is important. Smooth QSK operation was also considered an important feature. Though I operate mostly CW I still wanted the capability of operating SSB. In addition, being able to listen to AM is a feature that I would appreciate. I began by drawing a block diagram of the receiver portion. I broke it into several modules so that I could build one section at a time and then test it before proceeding with the next stage. Each section, when completed, was enclosed in a box made of PC board copper clad material. The circuits are designed for an impedance of 50 ohms going into and out of each box. BNC connectors are used for all RF connections between the individual boxes. DC and control lines enter the individual boxes via feed thru capacitors.
Since I am not an electrical engineer I am aware of my limitations as far as designing circuits. Why re-invent the wheel when someone else has already gone though the trouble? Therefore, where suitable, I used circuits from various designs that have been published in amateur radio periodicals such as QST, QEX, The ARRL Handbook and books such as, "Solid-State Design for the Radio Amateur" by W1FB and W7ZOI and "Introduction to Radio Frequency Design" by W7ZOI. You may find further information regarding these publications at the following Web site: The latest ARRL technical publication "Experimental Methods in RF Design" by Wes Hayward, W7ZOI, Rick Campbell, KK7B, and Bob Larkin, W7PUA is is highly recommended for anyone contemplating the thought of building their own radio equipment. This work is successor to "Solid-State Design for the Radio Amateur" which was first published in 1977. EMIRFD, is 512 pages of fascinating reading and includes a CD-ROM with design software, listings for DSP firmware, and supplementary articles. © 2003, published by American Radio Relay League (ARRL). (ISBN: 0-87259-879-9) Revised first edition. © 2003-2009 #9239 -- $49.95 There is a wealth of information in this publication applicable to homebrewers! When I found a design for a section that I liked and a PC board was available I purchased it, otherwise I used what is referred to as "ugly construction". Ugly construction is the process of using copper clad PC board material for a base to build electronic circuits on. Once I had chosen the circuits for each section I asked a friend to look it over and make suggestions and point out things that I hadn't considered. It is a very good idea to be friends with an Elmer, i.e. someone that knows more than you and is experienced in not only designing but actually building electronics circuits. 3. BFO Product Detector and Audio Module Working from the back to front seemed to be the way to go, so I started with the audio and detector module. I had previously used the R-1 PC board from Rick Campbell, KK7B's article, "High-Performance Direct-Conversion Receivers", in August 1992 QST to construct a DSB transceiver for 6 meters. (see above menu bar for the 6 meter design, DCP-10) Please note that the LM387 OP AMP is no longer in production and I have not found another OP AMP with similar pin-outs. I was very pleased with the audio quality of this design. The R-1 circuit includes a double balanced mixer which is used for the CW/SSB product detector. Since I intended to include the capability to receiver AM, I added transistor switches to select either the product detector or the AM detector. For the BFO I used diode switched crystals to cover, USB, LSB, CW lower and upper sideband and another one for Digital. The output of the BFO is amplified to provide +7dBm to the product detector, 150 mV to the transmitter balanced modulator and a low level output for the frequency counter. The first time I connected the BFO to the Product Detector and IF stage, I noticed that there was considerable BFO RF power being picked up by the AGC circuit on the IF board. It was only after I had enclosed the BFO, Prouduct Detector and Audio stage in a RF sealed box made of PC board that I was able to elliminate AGC pickup of the BFO signal.(It is not necessary to enclose the Audio Stage in a sealed enclosure, it just turned out to be convenient in my case) When the Product Detector Audio board and BFO circuit were completed I injected a signal into the Product Detector Mixer and checked to make sure all the crystals were oscillating. I then trimmed the xtal oscillaters to the correct frequency and set the bias control for the audio circuit for linearity. Since the audio gain of the R-1 audio board is higher than needed I lowered the gain of the op amps preceding the final output audio output transistors. This module is enclosed in a box made of PC material with all leads going in and out fed through feed-thru insulators except for the input from the IF stage where I used a BNC connector. 4. IF Module In the May 1996 issue of QST, Bill Carver, K6OLG, presented "A High-Performance AGC/IF Subsystem". The specifications for this system are impressive so I ordered a PC board from Far Circuits. The board is still listed as being available. See the following site for ordering information: After building the IF system and powering it up to confirm that it was working, I built a box made of PC board to house the IF system. Note that the 43 pf cap. in parallel with the primary of T-2, an 8 turn trifilar winding, will not resonate at 9 MHz as indicated in the QST article. I used a 180 pf cap. across the primary of T-2 to resonate at my IF frequency of 8.83 MHz. When I enclosed IF board inside my homemade enclosure, the IF module oscillated vigorously. I spent hours trying to solve the problem and was only successfull when I soldered a brass shield completely surrounding the input JFET pre-amp stage and input BNC connector. In addition I soldered the top of the IF PC board to the side walls all around the enclosure and then inserted a shield across the centre of the board. Feed-thru capacitors were used for all leads going in and out of the box except for the RF lines which use BNC connectors. The IF is very sensitive. I measured the MDS at -138 dBm looking into the JFET pre-amp with a 2.5 kHz BW filter in the centre of the IF module. This calculates to be a NF of 2 dB. The AGC system is superb, especially on CW as there is no thumping, clicks or distortion even when listening to really strong signals. 
5. IF Filters I had several Kenwood 8.83 MHz crystal filters on hand and decided to use them in the receiver. The picture below shows the construction and mounting of the filters. A shield was inserted between the input and output terminals of the filters inside the PC box providing measured stop band attenuation in excess of 100 dB. Again all control leads go through feed-thru capacitors and the input and output RF points employ BNC connectors.  I choose relays instead of diodes to switch between the different filters as I did not what to chance introducing IMD via diodes into the receiver. The BW of the filters I had on hand were, 6kHz, 2.5kHz, 1.8kHz, 400Hz and 250Hz. When the 250Hz filter is selected, the 450Hz filter is automatically inserted in series with the 250Hz filter along with an amplifier which compensates for the extra loss of the two filters in series. Also, resistive pads are added to the output of all the filters except the 6 KHz filter to compensate for the different losses of each filter. I can switch between filters when listening to a signal and the output level does not change. 6. Frequency Control Since low oscillator phase noise is one of the pre-requisites to obtaining a high overall receiver dynamic range, I used a low phase noise analog VFO and mix the output of the VFO to the required injection frequency with separate crystal oscillators which are have a very low phase noise, for each of the 10 lowest amateur radio bands from 160 meters to 6 meters. The VFO tuning range is 1 MHz, (later changes to two 500Khz conseceuive ranges) I can thus tune from the low end of 20 meters up to WWV at 15MHz. As well, a 1 MHz tuning range allows me to cover the lowest 1 MHz of ten meters without having to add another crystal oscillator. An analog VFO is considered by many as old technology, however it is much easier to make a clean, low phase noise analog VFO than a digital VFO. Technology is improving daily and no doubt it won't be long before you will be able to purchase a single IC that will do the job of an analog VFO with similar spec's but until then, the analog VFO is still king! The VFO main capacitor came from a WWII aircraft transmitter called the ARC-5. It is beautifully made. The capacitor is silver plated with ball bearings on both ends of the main rotor. The reduction gear comprises two gears, one fixed and the other floating with a spring pulling them together as they mess with a worm gear preventing what is called "back-lash" or a lag when changing direction. Before putting the capacitor into service I removed the reduction gears. I then soaked all the parts in solvent for several days to loosen all the accumulated dirt and grime. Remember these capacitors are over 70 years old. I then blew out the bearings with a high pressure hose, re-assembled it and then oiled the bearings and the two reduction gears before re-assembling. The VFO is silky smooth now. The attached picture shows the main VFO capacitor with the new (November 2006) dual range VFO circuit board attached to the back.  Originally I built the VFO to cover a one MHz range (5 to 6 MHz) however I found that 22 KHz per one turn of the VFO knob was too fast when using the narrow 170 Hz audio filter. The revised VFO has two consecutive 500kHz ranges selected from the front panel with a toggle switch that operates a latching relay located next to the VFO inductor. It is important to use a latching relay as a regular relay coil when energized will heat up and since it is located next to the VFO main inductor, the VFO will drift with the increase in relay coil temperature. The dual range VFO now tunes approximately 12 KHz per turn of the VFO knob, perfect! One of the factors that effects phase noise of an oscillator is the unloaded Q of the VFO inductor. The higher the Q, the lower the phase noise of the oscillator. I was able to achieve a Q of 370 by epoxying two T-68-6 toroid cores together and using #18 guage wire for the inductor winding. The VFO circuit is based on an article by J Makhinson, N6NWP. Communications Quarterly, Spring 1999 page 9-17. Makhinson provides extensive details in the article about designing very low phase noise VFO oscillators. High main inductor voltage, loose coupling to the inductor and a low noise devise such at the J310 all contribute to the low phase noise of this oscillator circuit. The VFO circuit has it's own power supply which is never turned off and it is also compensated for freq. drift with a N150 neg. temp. capacitor. Drift is so insignificant that I can use the digital modes with out any problem. The following picture shows the VFO ready to put back into the enclosure.  
Not shown on the VFO diagram is a 14 db attenuator to reduce the output to -10 dBm, the required input into the HFO mixer. The crystal oscillators are biased for +7 dBm. The VFO is filtered with a 2 section LP filter to attenuate harmonics. When building the RIT/XIT circuit you may want a different tuning range. My RIT/XIT tunes 3 kHz below and 15 kHz above the centre freq. If you go to the DESIGN page and click on Tools, #6 has a tool that allows you to determine the correct bias resistors for different incremental tuning ranges. Overall phase noise is not a limiting factor in regards to the two tone dynamic range of this receiver because of the inherent low phase noise of crystal oscillators and the low phase noise VFO circuit I used. Frequency Drift Frequency drift can be an issue with analog VFO's so I decided to try to improve the frequency drift while changing the VFO to encorporate two ranges. Not that the drift was bad, +or- 20Hz once the VFO warmed up, but I was sure it could be improved. With no excuse not to address the drift situation I built a tempature sensor with a LM335 and intalled it in an old Coleman picnic cooler as a heat chamber. It includes a shelf and below the shelf is a bulb with an on/off switch, (I used a 40 Watt bulb for a slower temp. rise) and a fan. Wes describes the use of a temperature chamber (or warming oven) on page 4.5-4.6 and also 7.42 in EMRFD. A more detailed description is in the original QST article, Dec 1993 page 37. After several runs with the VFO in the chamber and running the formula's in EMRFD I determined that if I replaced an existing 50pF N150 with a 68pF N150 cap I should be able to nail the drift right on. I couldn't find one. No one stocks polystyrene caps around here either. I tried combining two caps in series, a 100pf N150 and a 220pF NPO and a couple other combinations but to no avail. I then remembered that I was given a box of parts when my friend VE7YQ passed away several years ago. I searched the attic and found the box and sorting though the stuff I found a box of ceramic capacitors. In it was a 68pF N150 in perfect condition, never used before and I suppose it was at least 50 years old. I popped it in the VFO, let it stabilize over night and the next morning I measured a +3.2 ppmC. Wow, I couldn't believe it so I ran another test later that same day the result was similar. I have since put the VFO back into the transciever. Monitoring the VFO ouput using my HP8640 counter, the VFO is stable to +or- 5Hz over a half hour period. Not bad for an analog VFO using a 70+ year old WWII tuning capacitor and other assorted old parts. 
The crystal. osc. boards and mixer are to the left, the BPF's in the middle and the transmitter box is folded towards the back lying on top of the receiver Mixer, IF and Audio boxes. The frequency counter is to the right with the VFO underneath. Below the frequency counter standing upright is a box housing the noise blanker TRF receiver. The transceiver is built so that each section can be either folded out and away from the main transceiver chassis or easily removed for testing and maintenance purposes. The VFO and Crystal oscillators are fed into a double balanced diode mixer and then filtered by 10 separate relay switched 2 section Series C BPF's (bandpass filters). Some may question the wisdom of using only a 2 section filters at this stage however, the receiver input BPF's (which are also shared with the transmitter) are 3 section filters and provide adequate stop band attenuation. After filtering, further amplification increases the LO power output where it is split into three separate outputs, +17 dBm for the receiver mixer, -5 dBm to the transmitter board (later to be amplified to +7 dBm for the transmitter mixer) and a low level output for the frequency counter. When I first built the LO system I did not pay attention to setting the VFO output level and for reasons I can't remember, the VFO output was 0 dBm. After connecting the LO to the receiver mixer I found a number of spurs that shouldn't have been there. Wes, W7ZOI came to my rescue explaining the importance of not driving the RF port of a balanced mixer with more than -10 dBm. After decreasing the VFO output to -10 dBm the spurs disappeared. The VFO is built into a box made of PC board material. The 10 crystals oscillators and LO mixer are built into the top half of a box made of PC material. The bottom half of the box contains the 10 BPF's and an amplifier to boost the LO output power to the desired level. 7. RF Filters Filters, play an important role in the development of high performance receivers and transmitters. The HBR-2000 contains over 30 separate filters. One test instrument that I found invaluable is the L/C Meter kit produced by Almost All Electronics Inc. This test instrument allows the experimenter to quickly determine inductor and capacitor values. When winding toroid coils, I calculated the number turns required for the desired inductance value and then after winding the coil I measured the value with the L/C meter. If required I then squeezed or expanded the turns until I reached the correct value. This process saved many hours of tuning filters after they were built. The L/C meter is also helpful in determining the value of poorly marked capacitors which seems to be all too common these days. You may find information regarding the L/C meter at: Picture below shows one of the ten BPF's shared between the receiver and transmitter. The filters are shielded from each other in addition to having shields between each sub-section of each filter. 
All the filter component values used in the HBR-2000 were derived using a program that my friend developed. See the Design Page on the main menu bar of this Web site for further details. I have found this program indispensable. It contains many other features besides filter design such as, a VFO component calculator, an impedance matching circuit calculator, mixer spurious image calculator, an active IC filters design and a Calculator page that has a number of other useful features such as, the reactance of capacitors and inductors, resistive attenuator values, an air coil calculator and receiver noise analysis program. It is easy to use and provides very good quality graphs showing attenuation and input return loss. Below is a copy of the graph produced by the RF Design Program for the 14 to 15 MHz BPF filter that precedes the receiver mixer in the HBR-2000 (filter output [to mixer] is actually on the left to show match to mixer!). The 14.5 MHz series trap matcher (diplexer) terminates the mixer RF port in 50 ohms over a broad frequency range improving the 3rd order input intercept of the receiver mixer. 
Notice that the insertion loss (blue line) of the filter increases more rapidly above the design frequency than below. This provides greater attenuation at VHF/UHF frequencies where local high power TV and FM stations are located. This diminishes the possibility of these out of band signals reaching the mixer and combining with VFO and HFO harmonics and mixer products that can produce unwanted spurious signals in the receiver. 14 MHz. RF BPF Response. 
This is a picture of the spectrum analyzer screen showing the characteristics of the 14 MHz
BPF filter. The left side of the screen indicates an attenuation at 7 MHz (40 meters) of
-58 db and the right side is 21 MHz (15 meters) with an attenuation of > -65 dB.
Horizontal divisions are 10 db.
For 160 and 80 meters I used Series L BPF's. The insertion loss of Series L BPF's increases more rapidly below the design frequency than above. This helps attenuate local strong AM broadcast stations. At my QTH there is an AM station at 980 kHz which is very strong. The filter I designed has an attenuation at 980 kHz of over 70 db. With a 160 half wave loop tuned for minimum SWR at 1.85 kHz, I inserted the 160 meter BPF between the antenna and my TS940s and tuned it to 980 kHz, I measured attenuation at 68 dB. I have no birdies on 160 or 80 meters caused by AM broadcast stations with the filters I designed for the HBR-2000. 
All the RF input filters are built inside a box made of PC board material. The box is subdivided into 10 sections, one for each filter with additional shields between each sub-filter section. I have measured stop band attenuation in excess of 96 db with the cover on the filter box. All of the filters in the HBR-2000 are switched with relays. Many commercial amateur radio transceivers use diodes to switch between filters. Diodes switches can introduce IMD in the presence of other very strong RF signals. European hams in particular are aware of this problem as 40 meters is also occupied by very strong AM shortwave stations. Relay Troubles: November 15, 2003 There is a problem associated with using relays to switch very low RF levels. I noticed over the period of several years that some of the relays used to switch the input BPF's did not close completely the moment a particular band was selected. For example, when I changed bands, the receiver antenna noise sometimes was lower than usual for 10 to 60 seconds and then all of a sudden the antenna noise would rise to the expected level. This was confirmed by looking at the 15 meter BPF with the spectrum analyzer and tracking osc (this band gave me the most trouble). As pointed out by Peter, G3RZP in Letters to the Editor, page 58 in QEX Sept/Oct this problem is associated with oxidation of the relay contacts. This solution is to run a constant DC current through the relay contacts when they are closed. I recently added resistor dividers to pass approx 7 mA DC through the input RF filter relay contacts when switched on. Initially this did not solve the problem so I increased the current to 100 ma DC and switched the relays on and off many times (50 to 100 times worked for me) to clean the contacts. After the contacts were clean, I reverted back to the lower current circuit. I no longer have problems with the relays. 8. Mixer, Post Mixer Amp and Noise Blanker Switch The first mixer circuit I employed was a double balanced diode circuit with a LO power of +7 dBm. Wanting to increase the dynamic range of the receiver I changed the mixer to a Mini Circuit TUF-1H which requires a LO power of +17 dBm. The change in mixer and increase in LO power did not increase the IP3 as expected. A double balanced diode mixer with an LO of +17 dBm should be able to produce an IP3 of approximately +23 to +25 dBm. I was only achieving +14 dBm. John Stevenson, KD6OZH, wrote an excellent article titled "Reducing IMD in High-Level Mixers" in May 2001 QEX, page 45. He examined how IMD was effected by the impedance match (SWR) looking into all the ports of high level double balanced diode mixers. He found that having a good match (low return loss) into the RF port was as important as the IF port. In order to achieve a low return loss at the RF port he employed a diplexer between the front end BPF and the mixer. When I inserted a diplexer between my 14 MHz BPF and the mixer the IP3 increased from 14 dBm to 25.5 dBm. I have since added diplexers to all the receiver input band-pass filters in the HBR-2000. The IMD, two tone dynamic range, (tones spaced 20 kHz apart), increased from 97db to 106 db with the new mixer and at 2KHz tone spacing it is now 103 dBm, well worth the effort. New Front End, December 2006. Since I recently added 6 meters capability to the HBR-2000, I decided that I wanted more sensitivity on that band in particular so I decided to build a "New Front End Circuit". The "New Front End Circuit" incorporates three relay selected 10 db attenuator pads, a relay selected RF PRE-AMP., the mixer, and two post mixer amplifiers with a 100 KHz wide BPF inserted between the amplifiers followed by the noise blanker diodes. Note there is no diplexer after the mixer. I found after making measurements that a diplexer was not needed. The high input and output return loss of the two post mixer amplifiers and the 6 db loss of the BPF provides sufficient isolation between the mixer output and the input to the xtal filters so that typical large variations in xtal filter input impedances are not reflected back to the mixer output port. The following picture shows the "New Front End Cirucit". Note the copper shield over the first two attenuator pads and the +17 dBm LO input to the mixer. +17 dBm is a lot of power in a receiver that has a input sensitivity of -138 dBm. It is a good idea to keep this level of RF power from getting into places it shouldn't be! Good shielding helps to accomplish this. The second blanker diode stage is also shielded to improve the isolation when the diodes are biased OFF. The RF PRE-AMP is located in the bottom left hand corner subsection, the two Post Mixer Amplifiers with the 100 kHz BPF is shown in the middle section.  
The RF PRE-AMPS are built on separate double sided PC boards. I used a dremel tool to etch pads for the component connections. The RF PRE-AMP transistor is a low noise UHF transistor (MRF581A) which has to be treated with care to prevent UHF osc. The 5 pf cap. at the junction of the base and 10 ohm resistor kill UHF osc. in the amplifier. Short component leads are very important with this amplifier. 
It is important to use one DPDT relay for each 10 db attenuator, making all leads as short as possible to preserve the 50 ohm impedance and reduce stray inductance. I chose a modified Low Noise Norton, single ended, unbalanced circuit for all the amplifiers in the New Front End for it's excellent input to output reverse isolation (Return Loss)and it's simplicity. I measured the Return Loss, at both the input and output in excess of 20 db and that is with the opposite end either terminated in 50 ohms, shorted to ground or open. The RF PRE-AMP is biased for 20 mA and achieves an Output 3rd order intercept (OIP3) of +33 dBm and the gain is flat to within +/- 1 db from 1.8 MHz to 50 MHz with the UHF MRF581A transistor. Using a 2N3866, (I didn't have any more MRF581A's) the first post mixer amplifier is biased for 30 mA with an OIP3 of +38 dBm and the 2nd post mixer amplifier is biased at 35 mA and has an OIP3 of +40.7 dBm. A number of other transistors are suitable for the post mixer amplifier circuit such as the 2N5109, 2N3553, 2N5943, BFR-94 and for a SM transistor try the NE46134. If you want a low noise transistor and are unable to obtain a MRF581A try to obtain a 2N3866 as this transistor has a low base spreading resistance. Low base spreading resistance helps to achieve a low noise figure with high collector current. The 100 KHz BW, 8.83 KHz BPF located between the two post mixer amplifiers delays IF incoming noise pulses about 4 us. This allows time for the noise banker diodes to be biased off before noise pulses arrive at the diodes. A 100 kHz BPF has a delay of approx. 4 us. If you are using an IF freq. of 9 MHz, monolithic filters with a BW of 20 KHz are available as a substitute however do not use a BW of less than 20 KHz as narrower BW also stretches the noise pulses making them wider and more difficult to eliminate. To achieve 80 db isolation when the noise blanking diodes are biased off be sure to connect the 1 K resistor, 1000 pf capacitor, the 4.7 uH inductor and the 0.01 capacitor to the respective centre taps of T4 and T5 with very short leads. The blanker receives its blanking pulses from a TRF receiver that tunes from 7.5 to 10 MHz. The blanker is very effective and does not suffer from the usual IMD products typically found in commercial receiver blankers. Noise pulses from power line leakage or RFI from dirty commutator brushes in kitchen appliances and lawn mowers etc. that register S-8 to 9 on my S meter are totally eliminated with this blanker circuit. The circuit shows a relay between the LO input BNC connector and the Mixer LO port. This reed relay switches the + 17 dBm LO from the LO mixer port to a 50 ohm resistor on transmit. By turning off the LO power to the mixer I was able to improve my QSK system such that I can hear between characters when I am sending CW at 30 WPM. See the link below for a copy of an article published in the March/April issue of QEX 2006; "Perfecting A QSK System" by VE7CA. QEX Mar-06, Perf. QSK (Copyright ARRL. All rights reserved, used with permission of the ARRL.) The New Front End performs better than I had expected. The Input Third Order Intercept and Two Tone Dynamic Range increased over the old front end design and the sensitivity when the RF PRE-AMP is selected is -139 dBm +/- 1dB. The increased sensitivity has proven especially helpful on 6 meters. Even on 15 to 10 meters when my yagi is pointed away from the city where the noise floor is very low the increased sensitivity does make the difference between being able to copy a very weak station or not. 9. CW Audio Filter There are times when a little extra filtering can make the difference between copying a station or not. Even though my narrowest IF crystal filter is 250 Hz wide, I've had experience using a narrow bandwidth audio filter and found it quite effective especially on a noisy band like 160 meters. The problem with the narrow bandwidth audio filter is that it was a sharp filter design and it rings excessively. I asked my friend to incorporate the ability to design audio filters that have a Butterworth shape hoping that it would not ring as much as a peaked filter design. If you go to the the Design tab in the menu above and then to Active Filters, you can use the program to design a filter to your liking. Adjusting the bandwidth, centre frequency and trying different capacitor values, I designed an audio filter employing 3 op amps with a BW of 144 Hz centred at the frequency I like listening to, 650 Hz. The specified resistor values are close to standard values and I was able to find the ones I needed by measuring enough resistors to find ones that were within 1% of the design values. The circuit called for 6 capacitors each 0.01 uF and they also had to be accurate to within 1%. I took my capacitance measuring meter with me to a local electronics store and asked permission to go through a bag of capacitors to find the right values. After building the audio filter on a breadboard, I was able to determine that the filter shape matched very closely to what the program predicted. The filter is inserted between the audio pre-amp stage and the panel mounted volume control and switched in or out with transistor switches. The Butterworth design still rings a little but not nearly as much as the peaked audio filter. I find that the audio filter often helps when trying to copying a weak CW signal. The following picture shows the finished version of the audio filter built inside a box made from PC board material. Notice the point to point wiring, "ugly construction technique". I turn the IC's and transistors upside down and wire the components directly to the IC and transistor leads. This type of construction lends itself to making changes to a circuit which as experimenters we seem to do quite often. Also it easy to trace a circuit as all the components and wires are on top of the board. In hindsight, I wish I had not used any etched PC boards in the HBR-2000. To make changes to a particular stage built on an etched PC board, I have to remove the board from the PC box which is a lot of trouble. With ugly construction technique, everything is on top of the board which makes it a simple matter to remove a component(s) and/or insert new ones. 
10. Front Panel Layout Front panel layout plays a big role in how a transceiver feels when operating. The placement of knobs and switches should all be placed so that minimal hand movement is required when different bands, modes, and filters are selected. I use my rig for listening as well as operating events like hunting DX or contesting. I have arranged the most commonly used knobs and switches close to the tuning knob allowing one hand operation of the HBR-2000. For example, the filter selection buttons are just to the left of the main tuning knob, with the mode selection buttons just below. The RIT/XIT switch and tuning buttons are just to the right of the main tuning knob. With this arrangement I do not have to take my hand off the main tuning knob to change selectivity. I just use my left little finger to select which band width I want. I designed the front panel by making a paper template of the panel. I then laid all the knobs, switches, the S meter, frequency display and band select and mode/filter push button units on the template. I moved everything around until I felt good about the layout. However I did not make a final decision until I had thought about it a lot. I thought about the layout during the day, dreamed about it at night and time after time I sat in front of the paper layout and visualized how it would feel to operate. When I came up with the final panel layout I traced around the knobs etc. with a pencil. I then taped the paper template to the aluminium front panel and used the template to centre punch the holes for the different controls and corners of the digital display etc. 
After cutting and drilling the holes I sanded the panel with very fine sandpaper and spray painted with the colour of my choice. I then used press-on Letraset for the control labels. After applying the labels I sprayed the panel with clear lacquer. It is time consuming to make a front panel that is both eye appealing and ergonomical functional however it is well worth the effort. 11. Transmitter Portion (under construction), a few pictures.  100 Watt Amplifier Low Pass Filters.  12. Test Equipment The following picture shows some of the test equipment I built during the development of the HBR-2000. At the bottom is a step attenuator, the box below the scope houses the Spectrum Analyzer based on the W7ZOI/K7TAU design (August 1998 QST) and above the scope is a the Power Meter based on W7ZOI/W7PUA design (June 2001 QST) with digital read out addition by K3NHI (May/June 2002 QEX) 
13. Keeping Records Whether you decide to build a simple 2 stage transmitter or a full fledged transceiver, keeping complete and accurate notes and records plays a very important part in the end result. I used a three ring binder complete with an index and dividers for keeping circuit diagrams, notes, hand drawn layouts, pictures etc. for each bulding block in my transceiver. All the parts are labeled. When making measurements I note the DC and RF voltages at key points on the circuit diagrams. In some cases a colour photo of the block is included. If I decided to experiment with a particular circuit and try different component values and measure the results I used spiral bound note books for the records. After I decided on the final design of a paricular circuit or building block, I then made a copy of the final circuit diagram complete with the measurement results and inserted them in the main 3 ring binder. By keeping good records I can go back and see what changes I made and what the results were if I want to use a particular circuit again. You will be glad you made good notes if your project quits and you want to find out why, ESPECIALLY if it is a complex and full featured 10 band HF transceiver. Conclusion The receiver section of the HBR-2000 is an absolute joy to use. There are no unwanted birdies or images, the audio is cyrstal clear, the IF filters are sharp, plus the AGC has to be heard to be appreciated. Since completing the transmitter portion of the HBR-2000 I have sold my old TS-940s and the HBR-2000 is my main transceiver. The many, many hours spent in planning and building have been well worth while. Sincere thanks to a special friend and Elmer, who wishes to remain anonymous, who gave unselfishly of his time to answer my unending questions and Wes, W7ZOI who offered much encourgement and technical advise. Published Articles (Copyright ARRL. All rights reserved, used with permission of the ARRL.) QEX Mar-06, Perf. QSK QST Mar-06, HBR-2000 |