A Triggered Monostable Blocking

Oscillator
Used in legacy Channel Repeaters

Carlos Gil Soriano

BE-CO-HT
carlos.gil.soriano@cern.ch

August 21, 2012

Abstract
Along this document, a complete description of the previous genera-
tion of repetitors is offered. It allows the designer to better understand
the requirementes for good interoperability between the different types
of repetitors.

History of changes

This document version has been checked by:

This document version has been approved by:

Date Pages Changes

August 30, 2011 All Initial submission
September 21, 2011 All Deleted pulse definition chapter.
Moved to Standard Blocking
Definition Document [1]
August 21, 2012 All Minor changes in the title page.
System Description and Purpose This documents exposes how to de-
sign a Monostable Blocking Oscillator used in old Channel Repetitor boards.
A blocking oscillator offers an easy, and low-cost way of obtaining a sharp
pulse of fixed width.
By following the instructions given here, an optimal Monostable Blocking
Oscillator can be built.
Contents
1 Triggered Monostable Blocking Oscillator 1
1.1 References and tips . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Topologies evaluated . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Analysis of the chosen topology . . . . . . . . . . . . . . . . . 3
1.3.1 Switching on . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.2 On state . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.3 Switch off . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.4 Off state: Recovery time . . . . . . . . . . . . . . . . . 6
1 Triggered Monostable Blocking Oscillator
1.1 References and tips
A blocking oscillator is usually employed in synchronization applications
due to its simplicity, the little number of elements required -one BJT, a
transformer, few diodes, resistances and capacitors- and the sharp slope of
the rising edge it provides. Because of these reasons, a monostable blocking
oscillator is selected for the pulse conversion over other possibilities –such
as flyback converters, for instance.
An intuitive introduction to pulse converters can be found in [2], where the
blocking oscillator is sketched pretty simplisticly. Millman’s book, [3], offers
an easy and straightforward view on the pulse top state of the circuit.
The more mathematical insight of the switching state is found in Linvill’s
classical literature [4] and [5], which warns the reader of the aproximations
taken out for the sake of simplicity. It should be noted that, for a proper
switching analysis, some values of the model -such as base resistance- are
usually difficult to find in the datasheet of the manufacturers, and must be
inferred –a Ning-Tang method for the base resistance, i.e. Apart from this,
it is a good advice to carefully checking these values in the SPICE models
provided by the manufacturers. Some manufacturer’s base resistance param-
eter model corresponds to the intrinsic value of it and not to the intrinsic
plus extrinsic one, as it is requiered for a good matching with the mathe-
matical analysis. This will produce a misleading simulation which will turn
out into an unexpected outcome for the designer. Thus, we encourage not
to give 100% confidence to the simulation results due to this innacuracies
with respect to the aproximate model used in the analysis.
A comprehensive study of the switching state based on an extension of
Linvill’s aproximation is done in [6]. However, we have considered it as
overcomplicated compared to the more reasonable original Linvill’s aproxi-
mation.
Last but not least, Norman’s guide for the design of monostable blocking
oscillator [7] constituted the invaluable help which serves as the reference of
this design.

1.2 Topologies evaluated

Three blocking oscillators circuits were considered for the Pulse Converter
Units in CTDAH. The first two are positively collector-emitter feedbacked
and the later has a base-collector feedback.

Circuit 1
This circuit can be found in [5]. The main drawbacks it presents lay on
the charge it produces the output load and a worse triggering option

1
 compared to Circuit 2.

Figure 1: Linvill’s circuit, taken from [5]

Circuit 2
It is the circuit chosen for the design. With the inclusion of a resistance
in the positive feedback loop it is really simple to accomodate the pulse
width to the designer needs.

Figure 2: Norman’s circuit, taken from [7]

Circuit 3
However it is the most intuitive among all the designs, the triggering
is not as independent from the input as Circuit 2.

2
 Figure 3: Millman’s circuit, taken from [3]

1.3 Analysis of the chosen topology

The analysis follows the flow pointed out in [7]. The complete mathematical
resolution of the switching state from [4] is included. Furthermore, the fixed-
point methond algorithm is added to clarify how the normalized natural
frequency of the circuit is gotten.

Figure 4: States and equivalent circuits

1.3.1 Switching on
When it is switching both on and off, an equivalent circuit is shown in the
figure below. Emitter resistance and capacitance are omitted for simplicity.
The critical value while switching lies on obtaining the natural frequency of
the circuit. This value is closely related with the transformer turns-ratio
and, given the aproximations of the model, an optimun turns-ratio can be

3
calculated. If this turns-ratio is used, the fastest rise time will be achived
an no ringing will be obtained, ideally.

Figure 5: Eq. circuit B: switching

Kirchoff ’s equations
The electrical relationship of the circuit are:
E1
+ IE − niL = 0 (1)
RL
E1 IE
+ − IE − (n − 1)E1 Cs s = 0 (2)
rb 1 + wso
IE
iL − + (n − 1)E1 Cs s = 0 (3)
1 + wso

Fundamental equation
Operating the previous equations yields:
s 2 s GL + gb gb
( ) + · 2
− =0 (4)
w0 w0 (n − 1) Cc w0 Cc w0 (n − 1)

That can be greatly simplified by normalizing the natural frequency

and performing the following changes of variable:
s gb gb + GL
x= n−1=∆ =k =k (5)
w0 C c w0 C c w0
l k
x2 + x(1 +2
)− =0 (6)
∆ ∆
However, as ∆ is a design parameter we can rewrite the equation to:

∆2 (x2 + x) − k∆ + xl = 0 (7)

4
 If a value of the normalized natural frequency of the circuit, x, is lower
than xmax two possible optimal turns-ratios will exist. If it is higher
than xmax no real turns-ratios exist. We can find an optimal value of
∆ that corresponds to xmax , by obtaining a double root of the previous
equation. Hence:
k 2 − 4(x3max + x2max )l = 0 (8)
We can get the normalized natural frequency by applying iterations
by a fixed-point method, given that it converges –because we expect
a value close to 1
k
xmaxi = q i = 1, 2, 3 . . . (9)
2 l(xmaxi−1 + 1)

Results
The expected rise-time is:
2.3
trise = (10)
xmax w0
for an optimal turns-ratio is:
k
∆opt = nopt = ∆opt + 1 (11)
2(x2max + xmax )
Subsequently, getting a good rise-time depends mainly on choosing the
optimal turns-ratio value and using a fast switching bipolar transistor
in the design.
TIP: as stated in [7], the triggering signal must be active for 3trise , so
as to effectively switching the cirucit state.

1.3.2 On state
The on-state is related with the value of the magnetizing inductance in the
collector and the positive feedback resistor. With this two parameters the
designer is able to chose the pulse width.

Kirchoff ’s equations
From the collector we can get the following equation:
Vp Vp Vp t
ic = + + (12)
n2 (RL + rEBt ) n2 RL L
ic
During the on-state we can define hF Bt = iE , thus the previous equa-
tions can be rewritten as:
Vp hF Bt Vp Vp Vp t
= 2 + 2 + (13)
n(RL + rEBt ) n (RL + rEBt ) n RL L

5
 Figure 6: Eq. circuit A: on-state

Results
When the transistor is close to leave its saturation state, we can change
all its parameters for the large signal model ones. This yields to ob-
taining the width of the pulse:

L(hF B − 1) L
Tp = − (14)
n2 [RE + rE + (1 + hF B rB )] n2 RL

As it is noted in [7], adding RE helps to desensitizing the circuit from

rE

1.3.3 Switch off

The switch-off state is governed by the same equations of the switch-on
state.

1.3.4 Off state: Recovery time

Off-state is reached when the circuit completely removes all the current from
the magnetizing inductor through the snubber consisting of the diode and
discharge resistor. Once the current is completely removed, a new triggering
can be faced by the monostable oscillator.

Kirchoff ’s equations
At the end of the on state, the magnetizing current is:

VP TP
IL = (15)
L

6
 Figure 7: Eq. circuit C: off-state

The magnetizing inductance current must equal the current flowing

out of the two multipoles formed by the diode and the resistor and
the diode and the zenner diode. The multipole consisting of the diode
and the resistor helps to avoid overdamping while switching off. In the
second case, the current decreases linearly until the diode acts as an
open circuit. From this time on, the only multipole draining current
from the inductor is the one with the resistor. The discharge equation
is:
V2 Rc −
t−t3
iL = [1 + (t − t3 )]e L/(2Rc ) t ≥ t3 (16)
Rc L
The critical damping happens with a snubber resistor value of:
s
1 L
Rc = (17)
2 Cs

where Cs is the shunt capacitance formed by the addition of the

collector capacitance of the bipolar transistor, the transformer self ca-
pacitance and the wiring capacitance.
The resistance of Rsnubber should be less than Rc , so as to avoid over-
damping:
Rsnubber < Rc (18)

Implicit equation
The residual current in the inductance, IR , will be:
V2 Rc t −t
− 4 3
IR = [1 + (t4 − t3 )]e L/(2Rc ) t = t4 (19)
Rc L

7
 that cannot be resolved, but it is bounded by:
t −t3
V2 − L/R4
IR < e snubber (20)
Rsnubber

Thus, operating along the two recovery stages as in [7]:

V1 Tp L V2
TR = + (log − 1) (21)
V2 Rsnubber Rsnubber IR

Results
First, we must set a threshold for IR .
Then, the resistance of Rsnubber that makes TRmin is:

V2
Rsnubber = (22)
IR
but must comply with:
Rsnubber < Rc (23)
for avoid overdamping.

8
References
[1] C. Gil Soriano. Standard Blocking Output Signal Definition for CTDAH
board, September 2011. http://www.ohwr.org/documents/109.

[2] W.A. Stanton. Pulse technology. Wiley, 1964.

[3] J. Millman and H. Taub. Pulse, digital, and switching waveforms: de-
vices and circuits for their generation and processing. McGraw-Hill,
1981.

[4] J.G. Linvill and J.F. Gibbons. Transistors and active circuits. McGraw-
Hill electrical and electronic engineering series. McGraw-Hill, 1961.

[5] J.G. Linvill and R.H. Mattson. Junction transistor blocking oscillators.
Proceedings of the IRE, 43(11):1632 –1639, nov. 1955.

[6] J. McDonald. Circuit models to predict switching performance of

nanosecond blocking oscillators. Circuit Theory, IEEE Transactions on,
11(4):442 – 448, dec 1964.

[7] P. Norman and E.J.E. Smith. The design of transistor blocking oscilla-
tors. Proceedings of the IEE - Part B: Electronic and Communication
Engineering, 106(18):1251 –1259, may 1959.