Variable bandwidth crystal filter

Some time ago we featured a design from Michael N2ZDB that used a Jones filter design from TenTec to produce a variable filter passband for his BITX40.

There has been more recent interest on the IOGroups BITX20 list in variable bandwidth filters in the µBITx.

Karl-Heinz DF9RU completed a build of a µBitx transceiver and found it to be an excellent learning platform. He has been toying with a variable CW filter.

Up until now audio filters have largely been adopted by members:  either active filters with operational amplifiers or DSP filters with microcontrollers.  Karl-Heinz acknowledges that these options represent easy solutions for integration into the µBITx.

Karl-Heinz has a CW transceiver HB-1B from Youkit. This transceiver has a quartz filter of variable bandwidth.  The bandwidth can be continuously changed using varicap diodes.  The circuit diagram of the HB-1B can be found here.

Karl-Heinz was impressed by the acoustic result of the simple circuitry of his HB-1B and wondered why this alternative had not previously been used?

TenTec has patented a filter design which describes the passband curves of this filter.  A German website  also documents results on the bandwidth of these filters, which match with data from the TenTec patent.

Thierry F1HSU suggests taking a look at Tasa’s site for a tunable quartz filter  : http://yu1lm.qrpradio.com/bp%20yu1lm.htm

Ashhar Farhan VU2ESE suggests that these filter designs are merely a variation of the min -loss cohn filters. As only the coupling capacitance is varied without varying the terminating impedance, we must expect high ripple at all settings except one. What does this mean? It means that the filter will exhibit ringing and phase delays.

A better option would be a smooth Butterworth response with minimum ringing at a fixed frequency.  You can vary the BFO for shifting the audio tone. A 400 hz bandwidth will be narrow enough and yet offer a brightness that we miss in more aggressive designs.

Wes wrote a paper on this on his website www.w7zoi.net.
Ted, KX4OM reminded us of the SSB6.1 transceiver, which employs a tuneable SSB filter with tuning diodes in place of the capacitors in a min-loss configuration. The rig is a basic SA612-based design which can be found here:

Allison KB1GMX suggests looking at this design as well.  It is not new and a bit tricky but works best with lower frequency crystals.

Reference

Monitoring your microphone level

Didier F5NPV notes that many options are possible to tune your AF level. Measurement tools like an ocsilloscope can tap directly at the output of the AF Amplifier.  Alternatively a spectrum analyzer can be very useful in monitoring spectrum quality.

It is also possible to use another RX, a SDR receiver or even WebSDR online to monitor your transmission in quite an accurate manner. When you establish the correct level  of gain required to generate a signal without distortion spikes you just need to calibrate a meter with the appropriate threshold value.

The µBITx does not have an AF level input monitoring system. It is simply impossible to know if the level applied from the microphone preamp is too high or too low.

in the picture above you will notice a little switch just below the vu-meter thatI use to monitor the AGC and AF input level.

Reference

100PPR encoder – mods to KD8CEC code to make it work better

Using a 100PPR encoder with the KD8CEC firmware may cause some issues as the firmware can’t keep up with the pulse train from the encoder.

Sascha Bonnet suggests some simple code modifications make a big difference. He replaced two commands (“millis” to “micros”) in the KD8CEC’s source code and my encoder worked.   Here’s his code for you to use as a template for editing the KD8CEC arduino sketch:

  1. int enc_read(void) {
  2.   int result = 0;
  3.   byte newState;
  4.   int enc_speed = 0;
  5.   unsigned long start_at = millis();
  6.   while (millis()  start_at < 50) { // check if the previous state was stable
  7.     newState = enc_state(); // Get current state  
  8.     if (newState != enc_prev_state)
  9.       delayMicroseconds(1);
  10.     if (enc_state() != newState || newState == enc_prev_state)
  11.       continue;
  12.     //these transitions point to the encoder being rotated anti-clockwise
  13.     if ((enc_prev_state == 0 && newState == 2) ||
  14.       (enc_prev_state == 2 && newState == 3) ||
  15.       (enc_prev_state == 3 && newState == 1) ||
  16.       (enc_prev_state == 1 && newState == 0)){
  17.         result–;
  18.       }
  19.     //these transitions point o the enccoder being rotated clockwise
  20.     if ((enc_prev_state == 0 && newState == 1) ||
  21.       (enc_prev_state == 1 && newState == 3) ||
  22.       (enc_prev_state == 3 && newState == 2) ||
  23.       (enc_prev_state == 2 && newState == 0)){
  24.         result++;
  25.       }
  26.     enc_prev_state = newState; // Record state for next pulse interpretation
  27.     enc_speed++;
  28.     delayMicroseconds(1);
  29.   }
Reference

Full assembly of a uBITx

Jonathan Kayne, KM4CFT, who is studying at Virginia Tech in the USA asks on the IO Groups BITX20 list whether it is possible to do a full assembly of the uBITX.   That is, he wants to start with a bare PCB and hand solder on the SMD components. He plans to use it in a school project.

He received several replies to his query.    Jerry KE7ER suggests:

Be aware it might not be as trivial as it looks.

The transformers are described here:
http://www.hfsignals.com/index.php/ubitx-circuit-description/

Search for “Coil Details”.

The KiCad files are not available, that was some sort of requirement
when they set up HFSignals to build this stuff. Perhaps to get a business loan they had to keep some part of the design private. The remainder of the design is open source, and building from scratch is encouraged. The uBitx is a two layer board, the bottom side is mostly ground plane. Perhaps just build “ugly style” or “Manhatten style” on copper clad circuit board, that way you have a solid ground plane under the entire design, which is highly recommended.

All quartz crystals in that IF filter should be matched by hand to within 100hz or so, that means building a crystal oscillator and having some way to accurately measure frequency. If your quartz crystals have different characteristics than what hfsignals uses, you will need some way to determine the passband of the IF filter and adjust the filter shape as described in Experimental Methods in RF Design (and/or search for Dishal on the web).

You will need some way to sniff and measure RF, an Antuino would be ideal, though you might get by with a diode RF probe and a Harbor Freight DVM. A good scope would be nice, perhaps 50mhz or more of bandwidth. The nanoVNA would be worth looking into, shows complex of 1 and 2 port networks, and thus is an education in itself.

If you are serious about studying “RF and Microwave”, all of the above is worth the investment.

And lastly, maybe you could consider getting a working uBITx from HF Signals, simply so you can  know for sure what the signals levels really should be when yours doesn’t work.  Many have built radios like the uBitx from scratch, but few find it easy.

Ashhar Farhan VU2ESE, the designer of the µBITx has indicated he can arrange for a blank PCB.   He notes, however, that there is very little education in it!   To build an actual ubitx all by yourself would be a better learning experience.

Although the µBITx is a double conversion design, it is actually quite easy to build, stage-wise.  Ashhar suggests building the IF amplifiers first.  You have to build six of these. After getting one going, the rest can be duplicated.  He suggests that each is built on a separate 2″x2″ copper board.

After the IF modules have been built, you can hook up a raduino from the Si5351 board from qrp-labs and an antuino.   Using the antuino as a signal generator, you can test and align the 45 Mhz filter and the LPF.  At that point, you are done with the RF parts and you can choose to go with any audio amplification system.

Ashhar encourages experimenters to build their µBITx one stage at a time, then test, measure and move on.    He suggests it is a great education!!

Reference

Digital interface kit for your uBITx

David N8DAH is now producing a very handy digital interface kit.  This is a Gordon Gibby KX4Z design, and it features:

  • On board 5v regulator that you can power from 12-13.8v or you can leave it off for direct 5v input.
  • Built in VOX PTT via reed relay
  • PTT LED
  • Input and Output gain control
  • Audio Isolation transformers 600:600ohm

The kits are available for:

US$15 unassembled or US$20 assembled

https://shop.kit-projects.com/index.php?route=product/product&path=59&product_id=120

Reference

An interesting i2c encoder board for managing multiple encoders

John Scherer came across an interesting, i2c addressable rotary encoder board that allows you to connect encoders in just about any configuration.

The board has pins for digital or ADC I/O, has 256 bytes of eeprom, pads for setting each encoders i2c address, and works from 3.3 – 5V DC.   For something like the µBitx, this could be an easy way to add additional physical I/O (i.e. knobs).

There is also a video that gives a pretty good overview of how they work.

John VK2ETA suggests that on a read of the manual it appears the 256 registers are at one I2C address, and are divided between control registers and EEPROM addresses (two banks of 128 bytes).

The current implementation has the current modes for the encoder:

A. Relative, where the counter keeps the number of steps since last read. Reading it resets the counter to zero. Useful for frequency tuning in our case.

B. Absolute
B1. Without limits.
B2. With user defined limits
B2.1. No wrap around, like a potentiometer, hard stops at the limits.
B2.2 With wrap around, goes back to low limit if high limit is exceeded and vice-versa.

The encoder push button events are:
– button down
– button up
– double-press with user selectable double-press interval.

Current consumption is listed at less than 2mA plus LED current if used.

The board mounted PIC controller’s program is open source so it can be customized as well.

Tom, WB6B believes they keep a counter in the PIC chip that is on each board, so each time you check the status of an encoder board it will tell you how many encoder pulses took place since the last time you polled the data. This should greatly relax any timing issues.

Reference

Kit Projects selling outboard relay mod kit

David N8DAH from Kit-projects is marketing Gordon Gibby’s outboard relay mod as a kitset or for $5 more as a built-up board. This includes the circuit board, and all parts (including relays, diodes, inductors and caps).

Cost might seem high but David assures us he will be using quality parts.

The pre-orders were snapped up pretty quickly, but new stock will be coming shortly once the pre-orders have been shipped.

https://shop.kit-projects.com/index.php?route=product/product&path=59&product_id=119

Reference

SSM2167 board – improving compression control

Brad N8YG has been reading over posts concerning the ssm2167 compression pre-amplifier with quite a bit of interest. So he finally integrated the board into his ubitx v5.

He put a trimmer on the output of the board, and potentiometers on the compression, and noise gates. The problem that he ran into was that on the air, rotating the 100K compression potentiometer had little to no effect.  However he was getting compression that was clearly audible on the air.

He re-read the data sheet for the ssm2167 more thoroughly and found that the input voltage need to be quite low at -20 to -30 dBV.  A quick check on the electret microphone output using the oscilloscope, showed a much higher output from his microphone.

This led to him adding another pot to trim the input voltage.  He still needed the 5V to get to the electret so he wanted the trimmer after the 0.1 uF surface mount capacitor on the board.

Brad modded the board by removingthe 0.1 uf capacitor, and he made connections for the trimmer after the 0.1 uF capacitor.

The net result is a board with three additional potentiometers and one trimmer.  The trimmer adjusts the output voltage, and potentiometers control the input voltage to the board, compression and noise gate.

On the air.. with the compression potentiometer turned down the rig sounds like it has no compression… while turning up the potentiometer makes a profound difference in on the air sound.  The rig now has a LOT of punch and the power meter goes way up.   

Brad believes that what was happening was that the input level was too high forcing the chip into the limiting portion of the transfer curve.  Now that the chip is operating back in the linear portion, all is well!  

Reference

Temperature measurements on the fan shroud

Mark AJ6CU released a fan shroud for the “tall” Sunil case a short while back. See https://www.thingiverse.com/thing:3790450

He was curious to check out its effectiveness, against  just using the supplied heat sinks or using the heat sinks with a fan.  So here are the results….

Test Setup

  • ubitx v4 with stock PA transistors and the new larger heat sinks being shipped with the case
  • Noctua 1800RPM 80mm 12V fan (Blowing in)
  • Little Volt/Temp gauge that was mounted on the front panel (measurements in Fahrenheit)

Search for: ” DROK Car Voltage Temperature Display” on amazon if you want one!

Unfortunately Mark took the “shroud measurements” on a different day than the”no shroud measurements”.  So focus on the temperature gain over “ambient temperature, for a true comparison.

The first set of tests were for 5 minutes on 40m into a dummy load.
The second set was on FT8 CQ into a dummy load, until the FT8 watchdog timer terminated transmissions.

The Results

No Shroud            Ambient               After 5 Min Xmit       Temp Gain
Fan Off                   81F                          128F                               +47
Fan On                   76F                           94F                                +18

With Shroud       Ambient                After 5 Min Xmit       Temp Gain
Fan Off                   83F                          128F                               +45
Fan On                   81F                           98F                                +17

Operating for 5 minutes into a dummy load suggest that a fan is probably as good as a fan and a shroud.

On FT8 CQ (into Dummy load, until the watchdog kicked TX off):

WITH Shroud     Ambient                 After 5 Min Xmit      Temp Gain
Fan Off                   84F                          105F                               +21
Fan On                   82F                           88F                                +6

Of course the missing measurement is the FT8 CQ with just a Fan. Mark thought of this test *after* he had put everything back together. 🙁 .

Reference

TX IMD measurements for a v5 uBITx

Doug K4DSP just finished measuring the two-tone TX IMD on his v5 µBITx. These are 3rd order products, and they were measured relative to either tone. Since the ARRL states their measurements relative to PEP I have included that as well, for those who sleep better at night secure in the belief that their radio is clean 🙂

This is at 10W PEP using 700 and 1900 Hz tones. Doug’s radio puts out 10W from 80 through 20 meters, and falls off to 5W on 10M:

80m -25.5 dBc (-31.5 dB PEP)
60m -22.5 dBc (-28.5 dB PEP)
40m -22.0 dBc (-28.0 dB PEP)
20m -24.5 dBc (-30.5 dB PEP)
17m -21.0 dBc (-27.0 dB PEP)
12m -25.5 dBc (-31.5 dB PEP)
10m -22.0 dBc (-28.0 dB PEP)

As Doug’s flight instructor used to say after one of his landings, “I’ve seen worse, but I’ve seen a lot better.” This is probably not atypical of IRF510s.

Reference