by 5B4AIY » 28 Sep 2015 08:42
Hi, Juhani,
Glad to be of assistance. Power supplies, OK, let's discuss what is required. There are two characteristics that govern the operation of a regulated power supply - the static regulation, and the dynamic regulation.
Static Regulation
This parameter determines how well the power supply maintains the desired output voltage under steady-state conditions. This factor is primarily set by the closed loop gain of the regulator amplifier. A typical method of measuring this is to use the TUNE mode of the transceiver and output a steady carrier into a dummy load and measure the change in voltage from the idle to the full-load state.
At the power supply terminals the change in output voltage should be a very small percentage of the preset output voltage, and, for example, a typical power supply might well be specified as less than 0.7V for a 16A change.
Dynamic Regulation
This parameter determines how well the power supply maintains its stated output voltage as the load current changes, and is largely determined by the transient response of the error amplifier. Obviously a fast transient response is desirable, but too wide a bandwidth and the power supply may have a tendency to oscillate with certain types of load. This parameter becomes important when keying a transceiver or linear amplifier. For example, when the key is up, the current drawn from the power supply is only the idle current and amplifier bias currents, typically of the order of about 500mA - 800mA, but as soon as you key down the output power is immediately at 100W, and the current drawn immediately increases to about 18A - 20A. If the transient response is lacking, the output voltage will momentarily drop and then recover. Equally, when you release the key the current immediately drops back to the idle current, and again, if the transient response is poor there may be considerable overshoot on the output voltage before it stabilises.
All of this also presupposes that the resistance of the power supply leads is low, and does not further add to the regulation problems. One way to overcome the inherent voltage drop in the power supply leads is to use a feature called remote sensing where an additional pair of wires connects from the load (amplifier) back to the power supply, and the voltage is then sensed directly at the amplifier, thus including the power leads inside the regulation loop and further improving the regulation. Unfortunately I do not know of any amateur power supplies that make use of this feature, but certainly industrial high-current power supplies are equipped to do this.
There are two types of power supplies commonly used for amateur transceivers and linear amplifiers, the linear power supply and the switch-mode power supply. Linear power supplies represent very mature and reliable technology. They use a large mains transformer, bridge rectifier, large filter capacitor, one or more series-pass transistors, and a linear DC amplifier that is the regulator. The output voltage is usually very well smoothed with very low AC ripple, typically of the order of 10mV or less, and excellent static and dynamic regulation. The drawback is that they are usually very bulky and heavy largely because of the size and weight of the mains transformer, and they are not very efficient, generating a fair amount of heat, and have large heat sinks which are often forced air cooled with a fan, and thus might well be quite noisy. One factor that does make them highly desirable for amateur radio applications is that they do not generate any RF noise, and are thus radio quiet.
The second type, switch-mode, use a bridge rectifier to convert the incoming AC line into a 300V - 400V DC voltage which then powers a high-frequency oscillator, using a switching frequency between a few hundred kHz and 1 or 2MHz. The transformer can now be a small ferrite-cored device. The regulator uses a pulse-width modulated switching technique to modulate the average output voltage. There is usually a significant amount of high-frequency ripple on the output voltage, typically about 100mV P-P, but this is fairly easy to eliminate in the transceiver or linear amplifier. The static regulation is somewhat worse than for a good linear supply, but still more than good enough for our purposes. The plus side is that these power supplies are physically much smaller and lighter than their linear counterparts, and much more efficient, often exhibiting full-load efficiencies of 90% - 95%. They are thus quiet, and cool, but the drawback is their tendency to produce RF interference. The switching frequency and its harmonics can be fed back into the mains wiring, which will then radiate it, and this can be quite difficult to satisfactorily eliminate. For amateur applications it pays to choose a switch-mode power supply that is specifically designed for this mode rather than trying to adapt a surplus industrial supply, even though these are often available quite cheaply on eBay for example. An amateur supply will often be constructed in a shielded metal case, with good quality input and output filters, and in some cases there will be a knob on the front panel to shift the frequency of the oscillator so that its spectral components do not interfere with the wanted signal.
A good linear power supply is, for example, the Yaesu FP-1030A, the Kenwood PS-50, or the Palstar PS-30M. One serious drawback of linear power supplies concerns their catastrophic failure mode whereby a series pass transistor fails short-circuit. I have had this happen twice in my professional career, with disastrous consequences. What happens is that the transistor fails short-circuit, and now the full unregulated supply voltage appears at the output terminals. To protect against this most power supplies have an over-voltage protection circuit built-in to either crowbar the supply or otherwise turn it off in this event.
Switch-mode power supplies such as the MyDel MP8230, and the MP30SWIV are now very mature and reliable, and I have found them to be radio-quiet as well. In the case of these types of power supply, a catastrophic failure typically results in an immediate and complete loss of DC output, thus protecting the load.
There are also some more modern switch-mode power supplies that, whilst slightly less efficient, using a resonant transformer technique which is inherently quieter in terms of RF interference than the usual switching technique. One such example is currently being sold by Elecraft for their KXPA100 add-on linear amplifier, and is the Power-Werx SS-30DV, switchable between 120V and 230V operation, and CE certified. (powerwerx.com) This is even smaller than the MyDel power supplies, and is rated at 25A continuous, and 28A for 5 minutes. I have used this power supply as well as the MyDel ones, and have not experienced any significant amounts of RF interference.
Hopefully this answers your questions.
73, Adrian, 5B4AIY