For early portable radios

In the 1960s, the radio became portable, thanks to the appearance of tubes operating with plate voltages below 90V and filaments heated at 1.5V. The 1.5V round battery would fulfill this latter function. The anode voltage source was a 67V dry battery. This latter source no longer exists today. It must therefore be reconstituted respecting the voltage and, of course, the dimensions.

67V Batteries for Vintage Radios - An Example of Substitution.

Observations

• Consumption of a few µA at no load
• Overall efficiency of 90%.
• Good stability of the output voltage which depends practically only on the stability of the source.
• The operating frequency depends mainly on the value of the inductor of the transformer primary: the values of the capacitors have very little influence on it.
• The operating frequency with the prototype is in the order of 6 kHz for a 6 V power supply (inductor value 5.15 mH, without taking into account the DC component..)
• The frequency increases with the voltage.
• The frequency varies little depending on the load.
• The output voltage is determined by the input voltage and the turns ratio.
• The rectified AC voltage is added to the power supply voltage.
• V_s = V_p (N+1) with N turns ratio (sec/prim).
  In this case: V_s = 9 V * (6.3+1)= 66 V, which is very close to the measured values; without taking into account the voltage drops of the transistor and the diode

An example of the realization of this assembly

(photos Jacques Flamand)

Additional Information

(by Jacques Flamand). Some additional information; Synthesis of other answers that I had saved March-June 2003.

Here is the list of components, all recovered including the transistor...
T1: BC 327
Tr: ferrite pot 18/11 quality 3E1 without air gap (recovered from electronic starter - Philips)
Prim: 30 turns
Sec: 165 turns
C1: 0.1 uF (within 50% tolerance)
C2: 1 nF (within 50% tolerance)
C3-4 : 1 uF
R1: 10 Ω (transistor base) allows measurement of the waveform with an oscilloscope.
R2 or (L1): 100 Ω (filtering) advantageously replaced by an inductor of +-100 turns on a ferrite core
C5: Coupling capacitor ~1nF (HF decoupling) can be replaced by 4.7-10 nF wired as short as possible at the HT output
D1: BYV95C (recovered from electronic starter) or other diode (1N4001, BY....)

Test results:
Vp: 12.9V Ip: 145 mA
Vs: 83.2 V Is: 21 mA Vs:83.9 V for Is :10.5 mA
efficiency varies from 90.8 to 93.4 % depending on the load

The operating frequency depends mainly on the value of the inductor of the transformer primary; the values of the capacitors have very little influence.
- The operating frequency with the prototype is in the order of 6 kHz for a 6 V power supply (inductor value 5.15 mH, without taking into account the DC component..)
- The frequency increases with the voltage
- the frequency varies little depending on the load
- The output voltage is determined by the input voltage and the turns ratio.
The rectified AC voltage is added to the power supply voltage.
Vs = Vp (N+1) with N turns ratio (sec/prim).
In this case: Vs = 12.9 V * (5.5+1)= 83.85 V, which is very close to the measured values; although the voltage drops of the transistor and the diode are not taken into account....

It is normal for the circuit not to oscillate without a load; a load resistor is needed; the resistance of a voltmeter with low Ri is sometimes sufficient to start it.
This is an advantage because this circuit consumes practically nothing without a load (except for leakage); see the first message.
For simulation, a load resistor is therefore needed, let's say 10 kO should be sufficient.
When powered on without a resistor, the circuit oscillates for a fraction of a second, just to charge the capacitors. Then complete calm until partial discharge (supply voltage) and restart; therefore interesting no-load consumption, which is not the case for other circuits in general.

Here are tests carried out by « Vieille radio » (Old Radio):
For my first tests, I only wound the thirty turns of the primary over the existing winding and I obtain a voltage of 130V for a 13.2V supply to the circuit

To know certain limits of this circuit, I successively
replaced the rectifier diode with different types, this in order to
improve the already good initial efficiency.
What was my surprise to see this small converter accept
practically any diode with sufficient voltage rating,
I started from the "1N4007" as a reference, here is a
table giving the current consumption of the assembly
converter and load (130v on 10K) with a 12V battery
(13.2V exactly).

1N4007 - 149.2 mA
1N5062 - 150.9 mA
BY399 - 147.2 mA
BY255 - 148.7 mA
BYW98 - 153.1 mA
BYX10 - 148.9 mA
BY133 - 148.7 mA
BY127 - 155.8 mA
ESM513 - 149.7 mA
...and to finish, hold on tight!
OA214 - 148.5 mA
...well, I'll continue, I'll surely find a weakness somewhere?

Gérard.

Some additional thoughts:
- Parasitics, perhaps you have not observed the circuit diagram given at the beginning of the thread; the battery + inverter assembly is to be mounted in a metal case. The negative (common LV and HV) is connected to the ground of this case.
The positive output of the HT passes through the "feed-through capacitor C5". Unless there is a specific configuration at the receiver level (see the receiver diagram), the ground of the case and the receiver chassis will be common.
- The waveform generated by this converter is not rectangular and is therefore less likely to generate harmonics. The signal resembles a half-sine wave like half-wave rectification.
- I placed a "flying" assembly without any metal case or HF decoupling next to a transistor receiver with the ferrite loop placed in the worst orientation. Results: in medium wave, no whistles; in long wave, interference with certain stations; good EMC wiring should be effective.
Professional or other receivers are equipped with HT converters intended to power displays; such as the Kenwood R5000, and this without any problem.

Yes Jacques, in my case the frequency is around 9KHz
The "flying" assembly on a board, therefore unshielded, is less than
5cm from a small shortwave receiver without any interference.
and indeed the shape of the signal has no sharp edges, that's why
it is not necessary to use fast diodes.
I compare this schematic not to a switching converter, but
rather to resonance, and I have gone so far as to use germanium transistors
of the type AC128, 2N526, OC80, etc...
This circuit is very flexible and accommodates a wide range of
components.
I think you had a good idea to present this schematic
Gérard.

Clarifications:

- It is obvious that current ICs allow this type of assembly, however their operating frequency and the shape of the generated waves make me fear the presence of unwanted whistles in reception; moreover, it is not easy to find all the components.
- The proposed assembly is intended to be simple and relatively efficient and tolerant with regard to the use of components (long live recovery).
- Regarding the choice of the transistor, care must be taken that it can withstand at its base at least the output current of the power supply; this is not guaranteed with our old germanium transistors.

Jacques Flamand


The Spice simulation predicts a voltage rise as follows.

---

An Off-the-Shelf 3.5-25V to 60V Converter