The 1W FM LCD PLL No Tune Exciter/Transmitter Kit was ordered by phone, from the UK, direct from Broadcast Warehouse. It cost £89.95, including UK postage, and arrived promptly by recorded delivery. The unit is also available ready built. Note that this is a mono unit, for stereo it must be used in conjunction with a stereo encoder. Two pages of instructions were provided. These were:
The high quality double sided printed circuit board had a component placement silk screen on the top side, and top and bottom solder resists which assist in making good solder joints.
This kit marks a departure from the usual practice of kit manufacturers, in that it is a no-tune unit. The three adjustable inductors have been pre-set by Broadcast Warehouse, and require no further modification. The only other adjustable components are a variable capacitor to set the centre frequency of the PLL reference oscillator, and a variable resistor to set the audio gain. If no frequency counter is available, the reference variable capacitor VC can be left in its centre position (half meshed) in which case the output frequency should typically be ±1KHz from the displayed frequency.
The frequency is programmed by using two push buttons, each push increments or decrements the output frequency by 100KHz (0.1MHz). This ease of programming will appeal to intellectually challenged kit builders who find the prospect of setting dip switches to alter the output frequency too daunting. The unit boots with the last programmed frequency.
As the unit is no-tune, home constructors can have more confidence that providing they have correctly followed the assembly instructions, the resulting unit will have the same high specification (low levels of harmonics and spurious) as the prototype units.
As usual with Broadcast Warehouse equipment, no schematic was provided. Broadcast Warehouse have informed me that the reason schematics are not provided is not a disinclination for them to be circulated or published, but the fact that the engineers at Broadcast Warehouse do not work from schematics themselves, being sufficiently old school and immersed in their profession to carry the whole design in their heads. Nevertheless I suggest that a schematic remains an essential fault-finding tool, and should be provided.
The transmitter uses the standard PLL architecture.
The voltage controlled oscillator (VCO) is based on a Fairchild MPSH10 NPN bipolar transistor and operates at the output frequency. The centre frequency of the VCO is set by an adjustable inductor, which forms a tank circuit with a BB809 varicap diode. This varicap is used as both the audio modulator, and the VCO control element, as it is connected to the DC output of the PLL loop filter. This DC feedback path enables the output frequency to be locked to the frequency of a stable crystal reference oscillator. The audio input circuit is totally passive. Optional pre-emphasis is provided, by the fitting of a single capacitor in the audio input path. Two alternative capacitors are provided, one for European 50uS pre-emphasis and one for USA/Japan 75uS pre-emphasis. If the unit is to be used with a stereo encoder, the pre-emphasis must be disabled. This is achieved by shorting out the pre-emphasis capacitor C19 (not C11 as stated in the instructions).
The output of the VCO is connected to common base buffer stage based on a MPSH10. At the output of this buffer, a sample of the signal is passed to the fin input of the synthesiser IC. Next is an amplifier stage, also using a MPSH10. The DC emitter current of this stage is returned through one of the outputs of the PIC. As the PIC can sink up to 25mA, this enables the PIC to enable and disable the output of the exciter by shutting down this amplifier stage. The PIC derives this information from the lock detect (LD) pin on the synthesiser IC.
A bifilar 4:1 impedance transformer based on a small ferrite toroid is used to achieve a wideband match between this amplifier and the output stage, a trusty old 2N4427 operating in class C. Two adjustable inductors and 4 capacitors make up the output match for the 2N4427. The inductors have been preset by Broadcast Warehouse. The output matching network incorporates the harmonic filter. The 2N4427 is provided with a push-on aluminium heatsink. No output socket is provided, so a piece of co-axial cable must be soldered to the PCB.
The PLL is implemented with the popular Motorola MC145170 PLL frequency synthesiser IC (MC145170 data sheet (PDF format) - 382Kb). This synth chip has a three wire serial interface, consisting of data in, clock and enable. The frequency reference comes from a 8MHz crystal, trimmed by a variable capacitor. The lock detect (LD) output of the synth chip drives an input of the PIC. The PLL loop filter uses an LF351, which is a low noise, low distortion, J-FET input operational amplifier. Using of an active op-amp loop filter running from the +12V supply, allows the VCO tuning voltage to have a much wider range (5 to 10.5V for 87.5 to 108MHz in the unit I built), so the whole frequency band can be covered without having to make any adjustments to the VCO tank coil. Compare this with the Broadcast Warehouse 1W PLL, where the PLL loop filter is passive, and the range is limited to the +5V supply to the PLL chip. In this latter case, the tank coil of the VCO must be adjusted to ensure that the centre frequency of the VCO is within lock range of the PLL.
The synth chip is programmed by a Microchip Technology PIC16F84 FLASH/EEPROM (Electrically Erasable Programmable Read Only Memory) 8 bit microcontroller, also running off the reference clock. This chip runs code written by Broadcast Warehouse, which resides in the chips internal EEPROM. Programming information is derived from two momentary action push switches. This allows the frequency to be set in 100KHz steps, from 87.5MHz to 108MHz. Holding a switch down results in a rapid change of the programmed frequency. The synth is programmed on power-up and whenever one of the switches is altered. This chip incorporates 64 bytes of data EEPROM, this is used to store the programmed frequency. As this memory is non-volatile, the unit boots up to the last programmed frequency.
The PIC also drives a 16 character by 2 line LCD alpha-numeric display. This is used to display the output frequency, and the loop lock status. Note that this feature is not a frequency counter, it just displays the programming information that the PLL chip receives. For this reason if you manage to make such a hash of constructing the latter RF stages so that they oscillate independently, the display will not assist you in this matter. For confidence, the output of the unit should still be checked with a frequency counter or scanning receiver to check the output frequency is where you think it is.
A small voltage regulator provides a stabilised 5V supply to the synth and micro-controller chips. There is provision on the board to fit a 7812 or similar voltage regulator. This is covered in more detail in Web Surf's review of the Broadcast Warehouse 1 Watt PLL no-tune Exciter with LCD.
The only equipment required to test this unit is a +12 to +15V stabilised power supply capable of delivering 500mA and a small dummy load.
The double sided printed circuit board (PCB) supplied has plated through holes, making the removal of a wrongly placed component difficult, so extra care is required when fitting the components As extensive use of 0.125W resistors is employed, a small bit on the soldering iron is essential. Any potential constructors new to soldering are advised to practice on a scrap PCB with some cheap unwanted components before moving on to the real thing.
The ICs were supplied loose in the polythene bag with the rest of the components. Although there is a tendency amongst some to go over the top on ESD (electrostatic discharge) precautions, both the ICs are CMOS and therefore potentially susceptible to static damage. A bit of anti-static packaging would have been preferable. The LCD display was supplied in an anti-static bag. The PIC was supplied with a socket, making it possible for it to be changed in the event of firmware upgrades.
No veropins were supplied, I fitted four to make it easier to connect the DC supply and audio input connections. The values printed on some of the ceramic capacitors were extremely faint, needing a magnifying glass to read. There is a 4:1 transformer to be constructed. This is not difficult as the instructions are clear. I suggest using a little silicone sealant (e.g. Maplin YJ91Y) to glue the transformer to the PCB after testing is complete.
Four plastic pillars are provided to hold the LCD PCB off the main PCB, the idea being that the whole unit can be mounted on the front panel of a completed transmitter. I decided to do things slightly different, and joined the LCD display to the main PCB with a piece of 14 way ribbon cable (use 20 way and peel off the 6 spare ways). This allows the main PCB to be mounted at your convenience, whist still mounting the LCD display on the front panel. Also Broadcast Warehouse suggested a slight degradation of phase noise could be possible with the LCD positioned directly over the main PCB. I was not able to detect any difference in my measurements. I fitted the two push button switches and the audio gain variable resistor to the top of the PCB. As the push button switches are push to make, another pair of switches could be fitted in parallel with those mounted on the board, this extra pair mounted on the front panel.
The unit took a shade over two hours to build, and worked first time.
When the unit is powered up, the LCD briefly displays Broadcast Warehouse's web address, then the stored output frequency is displayed. The unit takes approximately 20 seconds to lock. This long lock time is a consequence of having a good low frequency audio response. Whilst the loop is locking, the LCD displays the message "Please wait", the RF output is suppressed by 55dBc during this period. When lock is achieved, the message changes to "Locked" and full RF output is available.
With a 13.8V supply, the output power and VCO control voltage were accurately measured and plotted below. The supply current varied from 217mA at 87.5MHz to 196mA at 108MHz.
As can be seen from the graph, the output power remained over 800mW for the whole frequency band. The plot of the VCO control voltage, taken from pin 6 of the loop integrator op-amp, demonstrates that the VCO control voltage is properly positioned within its valid range. If at the extremes of the frequency band, this voltage became too high (say above 11V) or too low (say below 2V), the loop would be in danger of losing lock with variations over time and temperature, or under high audio modulation. The centre of this range could be changed by altering the adjustable core of the VCO tank coil L1 (using a proper trim tool to avoid breaking the fragile ferrite core), but as the part supplied was ready adjusted, this was not necessary. The slope of the control voltage plot gives the tuning sensitivity of the VCO as approximately 4MHz/V.
The frequency was set to 98MHz, and the supply voltage was increased to 15V. The table below shows the supply current and output power as the supply voltage was reduced from 15V.
|Supply Current (mA)||Output Power (mW)|
The unit lost lock at 7.5V, at which point the PIC shut down the RF output. Referring to the control voltage table above, we can see the reason for this, as at 98MHz the required VCO control voltage is about 7.5V. The output of the integrator op-amp cannot swing any higher than the supply voltage, so the unit must necessary loose lock when the supply voltage drops below the required VCO control voltage. As a consequence of this, referring once again to the control voltage table above, the more perceptive amongst you will note at an output frequency of 108MHz the unit will require a supply voltage in excess of 11V to stay locked.
The efficiency of this unit is good for a no tune output stage, especially for a 1W output power unit. If you take off the 32mA consumed by the logic, display and regulators, the DC to RF efficiency of the RF stages at 15V is 35%. With this level of current drain, battery operation is possible, bearing in mind the unit will shut down the RF output if low battery voltage causes the loop to loose lock. The instructions make clear that a regulated DC power supply of +12 to +15V is required for correct operation. Note this is different from Broadcast Warehouses' 1W tuned PLL unit, as in that unit the loop components are driven from an on-board +5V regulator. In this case the lower limit on the voltage supply is the class C output stage falling out of conduction.
The output device stays sufficiently cool even when operating at +15V supply.
At 13.8V supply, the harmonics were measured on a spectrum analyser at 88, 98 and 108MHz centre frequencies.
|Plot of Spectral Purity, 98MHz 0.88W, 500MHz span||Plot of Spectral Purity, 98MHz 0.88W, 200MHz span||Plot of Spectral Purity, 98MHz 0.88W, 4MHz span||Plot of Phase Noise, 98MHz 0.88W, 400KHz span|
As can be seen from the table and the spectrum analyser plots, the worst harmonic is at -62dBc, better than the -60dBc specification figure. No spurious outputs could be found at all, even using a high quality spectrum analyser and looking down to a staggering excellent -95dBc noise floor at 1MHz offset. This level of spurious performance is unprecedented in units of this nature, and attests to the careful design and layout that has obviously gone into this unit.
This amount of harmonic and spurious suppression is the best I have seen in units of this level of output power, and is perfectly good enough to be connected directly to an aerial.
Due to the small output power capability of the unit, coupled with the low supply voltage, bad VSWRs are unlikely to damage this unit. Audio response has not measured - yet. This will be the subject of further work. By ear the audio sounded fine, with no hums or whistles.
The unit can of course be directly connected to an aerial, or used to drive an RF amplifier for more output power. The vast majority of RF amplifiers currently have to be tuned to the operating frequency, although Broadcast Warehouse supply a range of broadband amplifiers (87.5MHz to 108MHz). The additional complexity in a broadband RF amplifier over a tuned amplifier is not that great, and in my opinion more amplifiers of this type should/will become available Here's my design for a 40W broadband amplifier. Presently broadband (87.5MHz to 108MHz) aerials for transmit use are even thinner on the ground, though once again Broadcast Warehouse supply some (e.g. the JAYBEAM 7050 folded dipole). I'd be interested to hear how much work is being done in this area. It's worth noting that only the exciter determines the close in spurious products of the complete RF chain (assuming subsequent amplifier stages are working properly), and any external RF amplifier will need its own harmonic filter to follow it to keep harmonic radiation down to an acceptable level.
The 1W FM LCD PLL No Tune Kit has the following advantages over the 1W PLL (Tuned) Kit
The 1W FM LCD PLL No Tune Kit has the following disadvantages over the 1W PLL (Tuned) Kit
*now no longer available? (June 2001)
This is an excellent, high quality unit using modern technology which will set the standard to which other units will be compared. It's relatively easy to build, needs no tuning, needs a minimum of test equipment to set up, and has an RF technical specification and results that will be hard to beat. I recommend that Broadcast Warehouse supply a schematic with future units. Contact Broadcast Warehouse, email@example.com
The review unit was purchased end July 1998. The PCB is marked "bw0001".
RF measurements were made 3 August 1998. This review was prepared 15 August 1998 and updated on 20 August 1998 after evaluating a revision 2 PIC, which Broadcast Warehouse started shipping week beginning 17 August 1998.
The differences between revision 2 and revision 1 are
Phase noise plot was added 22 August 1998.
Web Surfs' review of the Broadcast Warehouse 1 Watt PLL no-tune Exciter with LCD
[ How to be a Community Radio Station Home Page | Introduction to Community Radio Station Electronics ]