12 VOLT TO 250 VOLT OR 300 VOLT, 40 WATT POWER CONVERTER.

This supply is a "beefing up" of the 12V to 170V converter. It
provides higher voltage, much more power, and better regulation.
It's main use would be in converting car battery voltage into
HT voltage to run a mobile tube based amplifier of respectable
power.


Specifications (measured on my prototype):

Input voltage range: 10.8 to 16.8 volts (top end can actually
   operate to about 20 volts as configured. By changing one
   part value, the range is 14 to 28 volts at even higher power.
   At 24 volts, 100 watts is available.
Power Output: 40 watts.
Power Consumption: About 2 watts at no load, rising to 47 to 51
   watts at full load. (10.8V @ 4.6A, 16.8V @ 2.85A). Current
   consumption is slightly higher at +70C and somewhat lower at
   -40C.
Output Voltage: 250 volts at 160 mA. Instructions are shown on
   the schematic how to convert this to 300V (at 133 mA) or 350
   volts (at 115 mA).
Output Ripple: Approximately 100 mV RMS at full load.
Line Regulation at full load: 10.8 to 16.8, output changes from
   248.4 to 250.1 volts.
Load Regulation at 13.6V line: no load to full load, output changes
   from 251.5 to 249.2 volts.
Operating Temperature Range: -40 to +70C.
Switching Frequency: about 22 kHz.
Delay on Power On: HT voltage is delayed approximately 25 seconds
   (to allow the tubes to warm up before HT is delivered).
Overload protection: None. Unit may be externally fused at 6 or 7
amps.


Description:

This is a voltage boost flyback converter. It uses an inductor
as the voltage boost element (no wierd transformers to find or
wind). There is 1 CMOS IC (4069UB - hex inverter) that is
configured as a schmitt trigger oscillator. The duty cycle of
the oscillator is controlled by one of the gates in the IC
package. This gate effectively forms a "comparator" as the
gate changes state at roughly half supply. The IC is run from
a temperature compensated zener regulator of approximately 9.8
volts.

The oscillator output is buffered thru another part of the
IC package, then into a 2 stage "totem pole" FET driver.

The power electronics are 4 paralleled sets of FET switch,
switching inductor, and "catch" diode. THe diode must be a fast
recovery type. The inductor is a 90 uH 3A inductor. It is made
from a ferrite "slug" 0.49 inch in diameter, 1.5 inches long.
44 turns of #22 gauge wire create the required 90 uH. If you
need more power, you can add more sections as shown in the
schematic. (Of course, if you need less power, you can remove
some of the sections). The available power is not exactly
additive: 1 section produces 15 watts, 2 section 25 watts, 3
sections 34 watts, 4 sections 40 watts, 5 sections 45 watts,
6 sections 50 watts. The power conversion works like this:
When the FETS are turned ON, the inductor current starts to
increase (in effect, it is "storing" current). The longer the
ON time, the more current can be stored. Then the FETs are turned
OFF, and the collapsing field in the inductor causes the voltage
to increase, and the catch diode to conduct. The stored current
is "discharged" into the 100uF 350V capacitor. (Which has an
additional 0.1uF capacitor in parallel to gather the "edges" that
the electrolytic doesn't like to handle).

There is an L-C filter section to reduce the switching ripple, which
is above the audio band anyway.


Setup Procedure:
The "catch" with a very simple circuit is device variability.
By using the CMOS hex inverter in an "analog mode", you are
subject to some level of device to device variability. (However,
the price is certainly right). This section shows you how to
verify correct operation and debug problems in construction.

When you first build up the circuit, make sure you have access
to points A, B, and C on the schematic. GROUND point B (disconnect
it from A first, obviously). Monitor point A with a scope.

When you first apply power, point A should remain LOW (near ground)
for about 25 seconds. Then you should observe a rectangular
waveform, spending most, but not all of the time HIGH. So far
so good.

If you never get a waveform, look at pin 12 of the IC. If it stays
high, the "delay" capacitor is not charging up. May be in backwards.

If that's normal, but you still get no waveform check the U1A, B, C
voltages (or monitor them with the scope), probably a part is
connected to an errant spot. Remember, this is a hex inverter IC;
if the input is over half supply voltage, the output should be
under half supply.

Now would be a real good time to make sure the IC has about
9.8V on pin 14.

If you ALWAYS get a waveform, but it goes from a low duty cycle
to a high duty cycle (or constantly high) at the 25 second point,
you may need to change the value of R4. (Not likely, but it is
possible). The proper setup at this point can be done using point
C. With C grounded, the duty cycle at point A should be 80 to 90%.
With C connected to pin 14 of the IC (power), the waveform should
disappear, and remain low. R4 and R7 control these points.

With C connected to half supply, the waveform ought to be about
50% (40 to 60% is OK). To make point C equal to half supply, connect
it to pin 10 of the IC (creating a zero gain "op amp").

With that latest connection in place, measure the frequency at
point A. Should be about 22 kHz (20 to 30 kHz is OK). If it is too
low, decrease the value of C10. If it's too high, increase the
value of C10.

You have just checked out the control circuit. Remove power and
restore the proper connections at points A, B, and C.

Now you can power it up MONITORING THE OUTPUT voltage, with no
load on the output. It will start out at about 12 volts for the
first 25 seconds, then increase to about 252 volts. If it goes
to about 20 volts instead, you put the zeners in backwards.

If it starts to get VERY HIGH, IMMEDIATELY TURN IT OFF AND FIND OUT
WHERE YOUR ZENER STRING (or R8 or R9) is disconnected.

Once you have the proper voltage, let it operate for a few minutes,
then turn it off. Wait for the big cap to discharge, then feel
around to make sure nothing is getting hot (it shouldn't.... even
under full load, the transistors only get 20 degrees (C) above
ambient.)

An abbreviated set of these instructions is shown on the schematic.