engn4521/6521 embeddedwireless …users.cecs.anu.edu.au/~gerard.borg/anu/courses/engn4521/...3 radio...

ENGN4521/6521

Embedded Wireless

Radio Architecure Tutorial 3

V3.0

Copyright 2015 G.G. Borg College of Engineering and Computer Science. Australian National University

1

Contents

1 Introduction 3

2 Mathematical Preliminaries 3

3 Radio Components 4

3.1 Mixers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3.2 Hybrids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

4 Radio Architectures 14

4.1 Direct Conversion Radio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

4.2 Superheterodyne Radio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4.3 Low IF Radio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2

1 Introduction

In this short tutorial I explain the main types of radio archticture and the components that can be usedto build them.

1. Mathematical Preliminaries.

2. RF components used to make radios

• mixers

• hybrids

3. Radio architectures

• Simple radio receiver

• The direct conversion radio

• The superheterodyne radio

• THe low IF radio

4. RF filters

2 Mathematical Preliminaries

To understand how radios work, one needs a little electronics and the following trigonometric formulae.

cos (ω + ωo)t = cosωt cosωot− sinωt sinωot (1)

cos (ω − ωo)t = cosωt cosωot+ sinωt sinωot (2)

sin (ω + ωo)t = sinωt cosωot+ cosωt sinωot (3)

sin (ω − ωo)t = sinωt cosωot− cosωt sinωot (4)

cosωt cosωot =cos (ω − ωo)t+ cos (ω + ωo)t

2(5)

sinωt cosωot =sin (ω + ωo)t+ sin (ω − ωo)t

2(6)

3

3 Radio Components

3.1 Mixers

We have already introduced mixers and we know that they can be used to multiply two signals. Thesymbol for a mixer is shown in Fig. 1. A mixer has three ports: RF for radiofrequency, LO for LocalOscillator and IF for Intermediate Frequency. The inputs of a passive diode ring mixer (DRM) suchas that shown in Fig. 2, could either be the RF and the LO ports or the IF and the LO ports. Theoutputs would be respectively, the IF and the RF ports. The RF and LO are high frequency ports. TheIF is normally the lowest frequency port. The passive mixer can be used reciprocally for up or downconversion mixers and are the only mixers that work into the GHz range on the high frequency RF andLO ports. One important distinction between DRMs is whether or or not the IF is capable of operatingdown to DC. If this is the case then the mixer can be used in a direct conversion receiver, otherwise itmust have a finite IF frequency and could be used for example in the low IF radio.

Figure 1: Circuit symbol for a mixer

Operation of the DRM can be explained from Fig. 2. Consider the case of down conversion where theLO is multiplied by the RF1.

Consider the bottom left figure of Fig. 2 where the LO is positive at some instant of time. A positivevoltage appears as shown across the two secondaries of transformer T1. This voltage has equal excursionsin both secondaries. The two diodes at point A are reverse biased and therefore may be considered asopen circuits. They have been removed from the circuit. The diodes at B however are foward biased andact like a pair of equal value resistors across the secondary of the transformer T1. These resistors forma voltage divider. Since the voltages on the secondaries of T1 are equal, the centre of the resistors is atground potential. This is shown by the GND symbol connected by a dotted line in the figure. Note thatthis is not a direct connection to ground but is a fairly stiff one because the resistors have a low value.Since this GND is connected directly to the lower secondary of transformer T2, the IF signal at thisinstance of time is IF = +RF .

The right figure at the bottom of Fig. 2 shows what happens when the LO is negative. In this case,IF = −RF . Now we see how the mixer produces the multiplying or chopping effect of the LO. The samearguments above may also be applied if the IF is the input. The LO may either chop the RF (downconversion) or the IF (up conversion).

1In fact mixers do not strictly multiply signals by the local oscillator but instead the LO chops the signal. This means

that everytime the LO changes sign, it also changes the sign of the signal

4

LO RF

IF

LO RF

IF = +RF

LO RFA B A B

Diode Ring Mixer

Positive LO Negative LO

+

-

-

+

IF = -RF

T1T1

+ -T2 T2

Figure 2: Doubly balanced diode ring mixer

As a result of this chopping action and the non-linear characteristic of diodes, the mixer has a non-lineartransfer function as shown in Fig. 3. The correct operating point of the LO level is the minimum levelto attain the flat part of the characteristic. For the DRM shown in the figure it is about 7dBm. Theappendix shows the datasheet for the MINICIRCUITS ADEX-10L mixer which has a drive level of 4dBm.Take a look at the datasheet and see how much you can understand. Notice that the drive level is a veryimportant parameter and is the first thing mentioned on the datasheet.

5

Figure 3: Mixer transfer function

3.2 Hybrids

Hybrids (A.K.A splitter/combiners) are four port devices that are the solution for either splitting onesignal into two signals or combining two signals into one. This is very important to remember because itis not possible at RF to T-off two components located at the ends of long wires. Wheneversignals are to be split or combined, you need a hybrid.

Fig. 4 shows the circuit symbols for the two only types of hybrid. The symbol on the left is a 0/180o

hybrid and the symbol on the right is a 90o or quadrature hybrid.

Figure 4: Symbols for Hybrids.

Hybrids must always be terminated, VIZ. all ports must be connected to circuits whose input impedanceis Zo. When this is the case, Hybrids are designed to have the following properties.

6

1. A match on all ports. I.E. The input impedance on each port is Zo.

2. Each port has a corresponding port from which it is completely isolated.

3. Power entering one port is divided equally between two of the remaining ports. By reciprocity, ifappropriately phased equal amplitude signals enter two ports, then they emerge summed togetherfrom one other port2.

Under these conditions it may be shown that the two hydrids of Fig. 4 are the only possible.

The Transformer Hybrid

Fig. 5 shows the transformer hybrid. On the left (A) is the circuit for analysis and (B) is a practicalimplementation. The circuit for analysis employs internal resistors to achieve condition (1) for a match onall ports. This of course causes internal losses in the hybrid. The practical circuit is an obvious extensionthat uses a pair of 7 : 10 (actually it should be 1 :

√

(2)). autotransformers to do the match and therebyeliminates the need for the internal resistors.

V3

V2

V1

I1

I2

u

I-V4

I+

Zo

Zo1

2

3

4

(A) (B)

Zo

Zo

Figure 5: The 0/180o transformer hybrid (A) Basic circuit (B) Practical implementation that is modifiedto produce a match using autotransformers instead of internal resistors.

The circuit of Fig. 5 is constructed using a trifilar wound transformer with the winding senses as shown.One can immediately see by the symmetry of the transformer that port V3 is isolated from V4. By

2Of course these signals must enter through isolated ports

7

application of the rules for transformers, namely that the flux linking each core is the same (same voltageu across each coil) and the currents into the coils sum to zero, we obtain,

V3 = Zo

I1 + I22

(7)

V3 = V1 − u = V2 + u =V1 + V2

2(8)

V4 = u =V1 − V2

2= Zo

I1 − I22

(9)

where the last expression follows from the currents in the transformer summing to zero. From these wederive the important condition that the inputs are terminated,

V1 = ZoI1, V2 = ZoI2. (10)

This shows that ports 1 and 2 are also isolated. The equations for the practical implementation ofFig. 5(B) differ only in the respect that,

V3 =V1 + V2√

2(11)

V4 =V1 − V2√

2(12)

which shows that the implementation of Fig. 5(B) also conserves power. V.I.Z it is lossless.

The Wilkinson Hybrid or Magic-T

Very often only three of the four ports are required. For example only three ports are required to combinetwo signals or to split one signal. For this one may internally terminate one port as shown in Fig. 6. Thissimpler circuit with only two windings is known as theWilkinson hybrid ormagic-T. Fig. 6(A) is internallyterminated in Zo for analysis and the practical implementation of Fig. 6(B) uses an autotransformer tomatch to Zo. Both implementations must have the internal 2Zo resistor. In this circuit ports 1 and 2 areisolated as in the case of the transformer hybrid.

Exercise. Using Fig. 6(A), derive relations between the voltages for the magic-T. Use thefollowing method from the top figure in 6(A).

1. Derive an expression for the output voltage in terms of the input voltages.

2. Derive an expression for i1 + i2 interms of V1 + V2

3. Derive an expression for i1 − i2 interms of V1 − V2

4. Show from these that V1/I1 = Zo and V2/I2 = Zo

8

2Zo

V+

V2

V1

I1

I2

2Zo Zo

Zo

i1

i2

(A) (B)

Figure 6: Wilkinson 0/180o hybrids.

In the transformer hybrid, transformer currents at opposite ends of transformer windings are summed atthe same node. Phase lags introduced by finite wavelength effects along the transformer windings cancompromise hybrid operation. As we have seen before, this makes high (microwave) frequency operationdifficult. We employed the term Ruthroff transmission line transformers to describe transformers thathave this property. The publication:Radio-frequency power combiner for cw and pulsed applications byG.G. Borg and T Jahreis provided in the appendix contains modelling and measurements that demon-strate graphically the deleterious effect of transmission line phase errors in the transformer hybrid.

We have also seen that if the circuit is designed so that only currents at the same end of the transmformerwindings are summed at a node, then the circuit can function well at high frequency. We first saw thiswith the 4:1 Guanella balun. For completeness, Fig. 7 shows a way to implement a transformer hybrid (ofGuanella type) which avoids summing currents at different phase lags and thereby eliminates the phaseshift error of the simple transformer hybrid of Fig. 5. Such a design could be made to work in microstripat microwave frequencies.

VA VB

VCVD

Figure 7: A transformer hybrid that is conducive to a transmission line implementation by virtue of itsGuanella winding style.

9

The circuit of Fig. 7 still needs output transformers to do the 1 :√

(2) match to Zo on all ports.

The Quadrature Hybrid

Quadrature hybrids are four port devices that produce two equal amplitude outputs that are 90o out ofphase. They are very useful in radio applications such as,

• 90o phase shifters for splitters in radios

• Hilbert transforms

• Directional couplers

• Phased antenna arrays

Unlike the transformer and magic-T hybrids which are intrinsically broadband, there is no device thatnaturally produces an amplitude independent 90o phase shift over a wide range of frequencies. In thissection we look at a simple example of a quadrature hybrid however which comes very close to this ideal.

Fig. 8 shows a practical quadrature hybrid that produces two outputs VB and VC . Port VD is isolatedfrom VA. Notice that this design involves a small inductor across the primary and the secondary of thetransformer. The design equations for this hybrid are ωcL = 2Zo and 2ωcCZo = 1 where 2πfc = ωc isthe centre frequency of operation: The band of operation straddles this frequency. The reactances of Land C are roughly around Zo. The transformer self-inductance, LT should be much larger than Zo atωc i.e. Zo ≪ ωcLT . Using the networking model, we can compute the performance of the hybrid for adesign centred at fc = 900MHz. From the design formulae and Zo = 50Ω, we obtain L = 18nH andC = 1.8pF . We choose the self-inductance of the transformer to be 2µH and coefficient of coupling κ = 1,though the performance is not very sensitive to these parameters.

10

Zo

ZoZo

L

C C

L

L

C C

L

(A) (B)

IAVA

VC

VB

VD

vC

il

ic

I

I

Figure 8: A lumped quadrature hybrid. (A) for analysis and (B) in practice

The output voltages VB , VC and VD for unit drive at VA are shown plotted versus frequency in Fig. 9.The phase in the second figure is the angle of VB with respect to VC . Notice that the only quantitieswhich turn out to be frequency dependent are the output signal amplitudes, |VB| and |VC |. They onlyhave the required equal amplitude at fc

3. Power conservation, V 2

B+ V 2

C= 1, of course applies at all

frequencies. The phase of VB with respect to VC is exactly −90o and VD is perfectly isolated from VA

at all frequencies. The input impedance Zin = VA/IA (not shown) is also exactly Zo = 50Ω at allfrequencies.

3Note that in some applications, signal processing may be used to compensate for this effect, in others we may have to

live with it

11

107

108

109

0

0.5

1

Frequency (MHz)

Output signals for input VA

VB

VC

VD

107

108

109

−90

−89.9999

−89.9999

−89.9998

Frequency (MHz)

degr

ees

Phase

Figure 9: quadrature hybrid performance

In the following we derive equations describing the quadrature hybrid. To do so we apply the designcriteria (ωcL = 2Zo and 2ωcCZo = 1) under the following assumptions,

1. The currents I in the transformer are equal and opposite in each turn.

2. The voltage drops across the primary and secondary of the transformer are equal.

The voltage across both Cs is the same and hence so too is the current ic, through them. The sameapplies to the Ls where the common current is il. The following equations relate the voltages VA, VB

and VC ,

jωLil + VB = ic/(jωC) + VC (13)

The current flowing from the source is

IA = I + il + ic (14)

where I is the current flowing in the transformer in the sense shown. The current flowing into Zo at VB

is

IB = I + il − ic (15)

12

and VB = IBZo. The current flowing into the Zo at VC is

IC = I − il + ic (16)

and VC = ICZo. The current flowing into Zo at VD is

ID = ic + il − I (17)

and VD = IDZo. Substituting for VB and VC in 13, yields

(2Zo + jωL) il = (2Zo + 1/(jωC)) ic (18)

Using our design equations, ωcL = 2Zo and 2ωcCZo = 1, we can reduce 18 to ic = jil. This means thatic and il are in quadrature.

Since,

VA = jωLil + VB = jωLil + Zo(I + il − ic) (19)

and using the design equations we obtain

VA = ZoI + (Zo + 2jZo − jZo)il = ZoI + Zo(1 + j)il (20)

We may write the expression for the input current as

IA = I + il + ic = I + il(1 + j) (21)

to obtain our first important result,

VA = ZoIA (22)

The input impedance of the hybrid is Zo.

Next we show that VD = 0. Using Kirchhoff’s law from VA to VD, equation 17 and the design formulae,we obtain

VA = VD + ic/(jωC) + jωLil = Zo(ic + il − I)− 2jZoic + 2jZoil (23)

Using ic = jil we obtain,VA = 3Zo(1 + j)il − ZoI (24)

Equation14 provides an alternative equation for VA,

VA = Zo(1 + j)il + ZoI (25)

We can substitute this into equation 24 to obtain 2ZoI = 2Zo(1+ j)il or I = (1+ j)il. Using equation17we obtain ID = 0 and VD = 0.

This demonstrates that port VD is isolated from port VA.

13

We use this result to obtain the relationship between VB, VC and VA. Following the path in the circuitfrom VA to VD = 0, using ic = jil and I = (1 + j)il again we obtain.

VA = −2jZoic + VC and VC = 2jZoil = 2Zoic (26)

VA = 2jZoil + VB and VB = −2jZoic = 2Zoil (27)

Finally we obtain,VC = jVB and VA = VB + VC (28)

and,

VB =VA

1 + jand VC =

VA

1− j(29)

From this equation we can confirm that power is conserved,

|VB |2 =|VA|22

and |VC |2 =|VA|22

(30)

4 Radio Architectures

Fig. 10 is a simple up and down converter. During up conversion, the mixer multiplies the LO bythe IF. During down conversion the mixer multiples the LO by the RF signal. In this case, the IFsignal is a baseband signal that is much lower in frequency than either the RF or LO frequency, VIZ.fIF ≪ (fRF , fLO).

Figure 10: Simple radio receiver

To model up conversion, use equation 5 with ωo = ωLO and ω = ωIF

cosωIF t cosωLOt =cos (ωIF − ωLO)t+ cos (ωIF + ωLO)t

2(31)

14

The mixer product, produces two output signals at f = fLO ± fIF . Since fIF ≪ fLO, these signals arevery close together and would normally be inseparable by a bandpass filter. Generally, the architectureof Fig. 10 is useless for most radio applications as a transmitter.

What about a receiver? Again consider equation 6 and let ω = ωRF .

cosωRF t cosωLOt =cos (ωRF − ωLO)t+ cos (ωRF + ωLO)t

2(32)

The two frequencies so produced are fLO ± fRF but this time fRF ≈ fLO. The output on the IF sideis fed to a IF BPF with an upper 3dB cutoff frequency, f3dB. If fIF < f3dB and f3dB ≪ fLO + fRF ,then only the IF will be passed to the baseband circuit. After the IF BPF we are left with any signalwith frequency, |fLO − fRF |. If indeed the wanted signal is the only signal to have entered the RF BPFat the input stage, then the down converter would work for the application. However if there were alsoa frequency somewhere in the range fRF − f3dB < f < fRF + f3dB, then it too would appear in our IFBPF passband as an interferer. The signals at fRF − fLO and fRF + fLO are direct interferers of eachother. They are referred to as images of each other. If fRF − fLO is our desired signal at the antennathen the unwanted signal, fRF + fLO is referred to as a birdie.

The circuit of Fig. 10 is suitable for double sideband suppressed carrier (DSBSC) modulation/demodulation.Recall that DSBSC modulation/demodulation just involves multiplication.

4.1 Direct Conversion Radio

Fig. 11 shows the direct conversion radio. The direct conversion radio, as the name suggests, is designedto convert a radio frequency directly to and from baseband. It differs from the simple radio of Fig. 10by having an additional mixer driven by a quadrature version of the local oscillator. It down converts anRF signal modelled by r(t) = A(t) exp j(ωRF t+ φ(t)) to the base band signal z(t) = A(t) exp jφ(t).

As a transmitter it is quite easy to understand. Consider the following message signal at radiofrequency,

r(t) = I(t) cosωRF t+Q(t) sinωRF t (33)

where I(t)+jQ(t) = A(t) exp jφ(t). From Fig. 11, the mixer pair simply multiplies the real and imaginarybaseband signals by cosωLOt and sinωLOt. However notice that if ωLO 6= ωRF then there will be afrequency offset in the baseband. The proces of working out this offset is a signal processing techniqueknown as carrier recovery orcarrier acquisition. Not all modulation schemes are susceptible to carrieroffsets, but in what follows we assume that ωLO = ωRF .

The case of reception (down conversion) is similar: multiply the message signal by the LO signal. Multi-plying equation 33 by cos(ωLOt) and sin(ωLOt) and removing the ωLO + ωRF terms from the output ofthe lowpass filters yields the desired result of A(t) cosφ(t) and A(t) sin φ(t).

15

Figure 11: The direct conversion radio

An existing example of a direct conversion radio is the Texas Instruments CC1101 low-power sub-1GHzRF transceiver shown in Fig. 12. In actuality the CC1101 down converts to a low IF in both channels.

16

Figure 12: The CC1101 low-power sub-1GHz RF transceiver

4.2 Superheterodyne Radio

A very common radio architecture is the superheterodyne radio shown in Fig. 13. The superhet looks likemultiple stages of the simple radio above - and it is. Recall that the problem with the simple radio isthe difficulty of designing a BPF sufficiently selective to separate the wanted signal from its image. For asimple radio operating from baseband to GHz that would be a problem. However this is not always thecase. In the superhet radio there are multiple stages of intermediate frequency (IF). The first IF frombaseband is chosen sufficiently low so that the images can be separated by the first IF filter. In the secondstage the second IF is again chosen sufficiently low that the second IF filter can separate the images ofthe first IF. This technique is continued right through to the last (RF) stage of the transmit chain. Thekey point is that by the time we reach the final stage, the final IF is so high that the image can be easilylocated outside the passband of the RF BPF at the antenna.

17

Figure 13: The superheterodyne radio

The superheterodyne receiver is clearly a very selective design and, though it is nearly 100 years old, isstill widely used. One modern day example is the SA626 FM-IF receiver chip shown in a test schematicin Fig. 14.

18

Figure 14: The SA626 FM-IF complete wideband FM receiver

4.3 Low IF Radio

The low IF radio achieves an improvement over the direct converion radio without the complexity of thesuperhet. The production of a single IF in the baseband, where the modulation passband straddles thecentre frequency of the IF, only requires a single ADC/DAC rather than the two required by the directconversion receiver. The low IF radio may be viewed as a single sideband radio that uses a quadraturehybrid in the baseband to do the Hilbert Transform. Note that the quadrature hybrid at the RF end ofthe radio could be swapped with the 0o hybrid in the LO circuit.

19

Low noiseamplifier

mixersPoweramplifier 90o hybrid 90o hybrid LPF

LO

0o hybrid0o hybrid

Antenna

Figure 15: Low IF radio

Exercise. Derive the output signal to the antenna given an input IF signal of the form,z(t) = A(t) exp j(ωIF t+ φ(t)). Try to determine whether the passband of the RF signal lies inthe lower or upper sideband at radiofrequency

20

INTERNET http://www.minicircuits.com

P.O. Box 350166, Brooklyn, New York 11235-0003 (718) 934-4500 Fax (718) 332-4661

Distribution Centers NORTH AMERICA 800-654-7949 • 417-335-5935 • Fax 417-335-5945 • EUROPE 44-1252-832600 • Fax 44-1252-837010Mini-Circuits ®

ISO 9001 CERTIFIED

ADEX-10L

Typical Performance Data

REV. ORM78421ED-9623/2ADEX-10LDJ/TD/CP011011

Features• EXcellent L-R isolation, 60 dB typ.• low conversion loss, 7.2 dB typ.• flat conversion loss ±0.2 dB typ. over entire band• good VSWR, 1.5:1 typ. for LO&RF, 1.8:1 typ. for IF• good performance to 1500 MHz• solder plated leads for excellent solderability• low profile package, US Patent 6133525

Outline Drawing

RFMHz

LOMHz

Conversion Loss(dB)LO

+4dBm

Isolation L-R(dB)LO

+4dBmLO

+4dBm

Isolation L-I(dB)

VSWR RF port(:1)

VSWR LO port(:1)LO

+4dBmLO

+4dBm

FFFFFrequencrequencrequencrequencrequenc y Mixy Mixy Mixy Mixy MixerererererLevel 4 (LO Power +4 dBm) 10 to 1000 MHz

Surface Mount

A B C D E F G H.280 .310 .220 .100 .112 .055 .100 .0307.11 7.87 5.58 2.54 2.84 1.40 2.54 0.76

J K L wt.–– .065 .300 grams–– 1.65 7.62 .40

Outline Dimensions ( )inchinchinchinchinchmmmmmmmmmm

Maximum RatingsOperating Temperature -40°C to 85°C

Storage Temperature -55°C to 100°C

RF Power 50mW

IF Current 40mA

Pin ConfigurationLO 6

RF 3

IF 2

GROUND 1,4,5

Applications• cellular• PCN

EVALUATION BOARD P/N: TB-03

Frequency

CASE STYLE: CD542PRICE: $2.95 ea.

QTY.: (10-49)

Electrical SpecificationsLO-RF ISOLATION

(dB)LO-IF ISOLATION

(dB)

TotalRangeMax.

CONVERSION LOSS*(dB)

Max.

Mid-Bandm_

x

FREQUENCY(MHz)

IFLO/RF

fL- fU

LTyp. Min.σσσσσ

IP3@centerbandTyp.

(dBm)MTyp. Min.

UTyp. Min.

10-1000 DC-800 7.2 0.1 8.2 8.8 75 55 60 40 47 37 40 26 33 20 24 13 16 1.2

1dB Compr.: +1 dBm typ.E= [IP3(dBm)-LO Power(dBm)]/10*Conversion loss increases 0.8 dB when IF is above 150 MHz.

EFACTOR

LTyp. Min.

MTyp. Min.

UTyp. Min.

NEW!NEW!

L = low range [fL to 10 f

L] M = mid range [10 f

L to f

U/2] U = upper range [f

U/2 to f

U]

m= mid band [2fL to f

U/2]

10.00 40.00 7.30 82.88 58.83 1.18 1.5425.00 55.00 7.23 82.79 51.06 1.13 1.5455.00 85.00 7.27 80.30 44.57 1.12 1.5370.00 100.00 7.31 78.35 42.47 1.14 1.53

100.00 130.00 7.37 75.43 39.36 1.15 1.51

172.00 202.00 7.31 68.52 34.38 1.21 1.48244.00 274.00 7.21 64.68 31.33 1.25 1.46316.00 346.00 7.20 61.44 29.83 1.29 1.44352.00 382.00 7.13 60.51 29.38 1.28 1.43424.00 454.00 7.19 61.30 28.92 1.28 1.43

460.00 490.00 7.21 61.56 28.63 1.27 1.42532.00 562.00 7.21 59.88 28.24 1.27 1.39604.00 634.00 7.46 57.30 27.79 1.29 1.40640.00 670.00 7.49 55.44 27.54 1.30 1.40712.00 742.00 7.58 52.02 26.70 1.34 1.40

748.00 778.00 7.46 51.61 25.74 1.38 1.40820.00 850.00 7.38 51.53 23.84 1.38 1.39856.00 886.00 7.34 52.51 22.81 1.42 1.39928.00 958.00 7.43 51.02 21.76 1.48 1.35

1000.00 1030.00 7.65 47.97 21.23 1.57 1.27

ADEX-10LCONVERSION LOSS

6.0

6.5

7.0

7.5

8.0

8.5

9.0

10 100 190 280 370 460 550 640 730 820 910 1000

FREQUENCY (MHz)

CO

NV

ER

SIO

N L

OS

S (

dB)

LO +1dBm LO +4 dBm LO +7dBm

at IF Freq. of 30 MHz

ADEX-10LISOLATION L-R

30

40

50

60

70

80

90

10 100 190 280 370 460 550 640 730 820 910 1000

FREQUENCY (MHz)

ISO

LA

TIO

N (

dB)

LO +1dBm LO +4dBm LO +7 dBm

at LO DRIVE of +1/+4/+7 dBm

ADEX-10LVSWR

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2.0

10 100 190 280 370 460 550 640 730 820 910 1000

FREQUENCY (MHz)

VS

WR

#LO VSWR #RF VSWR

at LO DRIVE of +4 dBm

This datasheet has been downloaded from:

www.DatasheetCatalog.com

Datasheets for electronic components.

http://www.datasheetcatalog.com

L = 10-100 MHz M = 100-500 MHz U = 500-1000 MHz

FREQ. RANGE(MHz)

ISOLATION(dB)

INSERTION LOSS (dB)ABOVE 3.0 dB

PHASE UNBALANCE

(Degrees)

AMPLITUDEUNBALANCE

(dB)

fL-fU

L M U L M U L M U L M U

Typ. Min. Typ. Min. Typ. Min. Typ. Max. Typ. Max. Typ. Max. Max. Max. Max. Max. Max. Max.

10-1000 25 20 23 16 19 14 0.3 0.5 0.4 0.9 0.8 1.5 1.0 3.0 5.0 0.15 0.2 0.4

ISO 9001 ISO 14001 CERTIFIED

Mini-Circuits®

P.O. Box 350166, Brooklyn, New York 11235-0003 (718) 934-4500 Fax (718) 332-4661 For detailed performance specs & shopping online see Mini-Circuits web site

The Design Engineers Search Engine Provides ACTUAL Data Instantly From MINI-CIRCUITS At: www.minicircuits.com

RF/IF MICROWAVE COMPONENTS

minicircuits.comALL NEW

A B C D E F G

.272 .310 .220 .100 .162 .055 .100

6.91 7.87 5.59 2.54 4.11 1.40 2.54

H J K L wt

.030 .026 .065 .300 grams

0.76 0.66 1.65 7.62 0.25

ADP-2-4+ADP-2-4

2 Way-0° 50Ω 10 to 1000 MHz

Power Splitter/CombinerSurface Mount

Typical Performance Data

Electrical Specifi cations

Maximum RatingsOperating Temperature -40°C to 85°C


Power Input (as a splitter) 1W max.

Internal Dissipation 0.125W max.

Outline Drawing

Outline Dimensions ( )inchmm

electrical schematic

PCB Land Pattern

Suggested Layout, Tolerance to be within ± .002

REV. DM102713ED-7386/3ADP-2-4HY/TD/CP070424

CASE STYLE: CD636PRICE: $11.95 ea. QTY. (10-49)

Frequency(MHz)

Insertion Loss(dB)

AmplitudeUnbalance

(dB)

Isolation(dB)

PhaseUnbalance

(deg.)

VSWRS

VSWR1

VSWR2

S-1 S-2

Features• low insertion loss, 0.4 dB typ.• excellent amplitude unbalance, 0.10 dB typ.• very good phase unbalance, 0.5 deg. typ.• aqueous washable• protected under U.S. Patent 6,133,525

Applications• instrumentation • cellular

Pin ConnectionsSUM PORT 1

PORT 1 3

PORT 2 4

GROUND 6

Externally connect together & isolate 2,5

Demo Board MCL P/N: TB-208Suggested PCB Layout (PL-116)

ADP-2-4

INSERTION LOSS

3.0

3.2

3.4

3.6

3.8

4.0

4.2

4.4

0 200 400 600 800 1000

FREQUENCY (MHz)

INS

ER

TIO

N L

OS

S (

dB

)

S-1(dB) S-2(dB)

ADP-2-4

ISOLATION

10

15

20

25

30

35

40

0 200 400 600 800 1000

FREQUENCY (MHz)

ISO

LA

TIO

N (

dB

)

ADP-2-4

VSWR

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

0 200 400 600 800 1000

FREQUENCY (MHz)

VS

WR

#S-VSWR #1-VSWR #2-VSWR

+ RoHS compliant in accordance with EU Directive (2002/95/EC)

The +Suffi x identifi es RoHS Compliance. See our web site for RoHS Compliance methodologies and qualifi cations.

10.00 3.27 3.26 0.01 29.66 0.03 1.12 1.27 1.27 50.00 3.26 3.26 0.00 27.85 0.05 1.12 1.24 1.24 100.00 3.22 3.22 0.00 26.35 0.13 1.15 1.25 1.25 200.00 3.30 3.30 0.00 24.51 0.21 1.22 1.27 1.26 250.00 3.35 3.34 0.01 23.57 0.21 1.25 1.28 1.27 350.00 3.38 3.36 0.02 22.19 0.32 1.32 1.31 1.30 400.00 3.53 3.52 0.01 21.78 0.36 1.36 1.33 1.31 500.00 3.56 3.53 0.03 20.95 0.44 1.42 1.36 1.33 550.00 3.58 3.55 0.03 20.65 0.50 1.44 1.37 1.34 650.00 3.66 3.60 0.05 20.35 0.54 1.48 1.39 1.35 700.00 3.69 3.63 0.07 20.39 0.62 1.49 1.40 1.35 800.00 3.87 3.79 0.08 20.86 0.70 1.48 1.40 1.35 850.00 3.72 3.62 0.10 21.23 0.76 1.46 1.39 1.34 950.00 3.94 3.82 0.12 23.04 0.78 1.39 1.38 1.31 1000.00 4.01 3.88 0.13 24.55 0.90 1.34 1.37 1.30

Radio-frequency ower combiner for cw and pulsed applications

G. G. Borg and T. Jahreisa) Plasma Research Laboratory, Research School of Physical Sciences and Engineering, Australian National University, Canberra ACT 0200, Australia

(Received 20 September 1993; accepted for publication 3 November 1993)

We describe the design and operation of a circuit to combine the output signals of two high-power radio transmitters into a single load. A working unit is realized using a commonly available ferrite and cheap coaxial cables. Performance tests indicate satisfactory operation from 3 to 50 MHz with less than 1 dB total power loss while guaranteeing adequate isolation between the transmitters. The power combiner can also be constructed specifically to withstand high line voltages for short time periods thus making it suitable for high-power pulsed applications such as plasma formation and heating.

I. INTRODUCTION

A power combiner is a broadband four-port device that combines the output signals of two transmitters into two loads. When the transmitter signals are cophased (have the same amplitude and phase) at the input ports of the combiner, all their power appears at one output port to which the main load is connected and none at the second output port to’which a dummy or reject load has to be connected. The combiner has the important property that its input impedance is equal to the characteristic impedance of the system under all conditions. Correspondingly, the input ports do not cross talk and the transmitters are fully isolated from each other. The power combiner finds application in plasma formation and heating experiments where a set of transmitters are combined in pairs by a pyramid of combiners into a single antenna and in broadcast applications where two identical transmitters are required to feed a single aerial.

Commercial power combiners are expensive and the operating principle of a workable device for high powers is not widely known. In this paper we publish a solution to the high-power combiner problem which requires minimal labor and material costs. We also discuss the theory of the combiner so that the user can make modifications to suit his own application.

II. OPERATING PRINCIPLE

The basic circuit of a small signal, “hybrid” combiner is shown in Fig. 1. In practice, the windings are made trifilar and wound on a toroidal ferrite core so that the response of the device is determined by transformer effects at low frequencies and transmission line effects at high frequencies.* Simple circuit theory shows that the output signals Vc, and Ve, are related to the inputs V1 and V, by

VI+ v2 VOl” 2

FJI - V, - and Vo2= 2 . (1)

Provided the reactance of the core is much larger than the terminating impedance R, the input impedance at each

input port is equal to 2R. It is clear that for cophased inputs all the signal appears at Vo, and none at V,,. The main load is therefore connected at the Vol port and the reject load at the Vo, port. For equal output terminating impedances, no signal appears at an unused input port when the other is excited. The role of the reject load is therefore to guarantee isolation between the two sources.

This simple circuit is useful without further modifica- tion for small signal applications where power transfer efficiency and component heating are unimportant. If this small signal hybrid is intended to operate in a system where source and load have the same characteristic impedance Zo, then the input can be most simply terminated in 2, by inserting either a resistor of value 2, across each output or a resistor of value 2Zo across each input. A problem encountered in adapting the circuit of Fig. 1 to high powers therefore is that a pair of 1:2 impedance transformers or baluns are required. A second problem is that, as the size and number of ferrite cores and windings are increased to avoid component heating at high power, finite wavelength effects begin to, limit performance. This is not modeled by the circuit theory leading to Eqs. ( 1) .

To understand the origin of this finite wavelength lim- itation we constructed the circuit of Fig. 2. This is the same device as in Fig. 1 except that the three windings are re- placed by 15 turns of two 6.4 m lengths of RG-58 coaxial cable wound on a 0.1 lo-m-diam plastic cylinder. The inner conductors form two of the windings and the braids are connected together to form the third winding. The measured impedance Z, of a single winding as a function of

Vl

VOl

V2

“Fachhochschule Regensburg, Fachbereich Elektrotechnik, D-8400 Re- gensburg, Germany.

FIG. 1. Circuit diagram of the sim$e transmission line transformer small signal hybrid combiner.

Rev. Sci. Instrum. 65 (2), February 1994 0034-6746/94/65(2)/449/4/$8.00 @ 1994 American Institute of Physics 449

Downloaded 24 Oct 2007 to 150.203.179.78. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp

FIG. 2. A coaxial cable version of the combiner of Fig. 1 for testing the effects of finite viavelength.

frequency with the other windings open circuit is shown in Fig. 3 (a). The following equations of transmission line theory may be used to describe the combiner:’

v, = ( v,,- F7&os 81 +j1;z, sin PI,

v,- vo2= vo, cos Pl +j1;zc sin 01, (2)

xwq.. .,.X.--m -- -

(cj SW

:; 4 l\, ---

Im(Zc) i,---..‘.’

+....+.+~

72 E

6

lx”, a) Re(Zc)

Ilao 8

5 xx)

g

P

(, 1 _.___.... -~ -.-. -- -+ _._,. 4 ”

y’. \

I I I I

L_I-.“..“^ . ..I......_. ^- L_I-.“..“^ . ..I......_. ^- - .._. L _-.. - L - .._. L _-.. - L

1 10 25

Frequency (MHz)

I2=1$0S /31 +j v,, ( ) zo

sin jjl,

v- - vo~=j(l3+I4)z~,

where the quantity Z. is the characteristic impedance and j3 the phase constant of the coaxial cable. The other symbols are defined in Fig. 2. Because Z, is finite, currents I:, and I4 flow on the outside of the coax braids and set a limit to low-frequency operation. As noted above, correct termi- nation at the inputs requires that R = Z,J2 = 25 Cl. In Fig. 3(b) are shown as a function of frequency the ratios P’,,/ V, and Vo2/V, for cophased inputs VI and V,. Also shown is V2/ VI, where V2 is the signal appearing at input port 2 terminated in 50 Cl when port 1 is excited by signal VI. In Figs. 3 (c) and 3(d) are shown respectively the impedances seen at each of input ports 1 and 2 when the other is terminated in 50 CI. The symbols are the measured! values and the curves are obtained by solving Eqs. (2) using the measured values of Z,.

The agreement between theory and experiment is quite satisfactory but the combiner performance is poor. The problem with the basic combiner of Fig. ,2 is that although the input signals are added to give Vo, after undergoing identical propagation phase delays along each coaxial cable, they are subtracted after different delays to yield V,,.

.tSll-.- Ai-- ‘ 1 IO

Frequency (MHz)

FIG. 3. Test results for the combiner of Fig. 2. (a) The.measured core impedance. The curves are best fit. (b) The ratios of output voltage to input voltage ( V&F’,) and reject port voltage to input voltage ( V,,/V,) for cophased inputs. The ratio of detected voltage at input port 2 to that at port 1 with input applied only at port 1 (cross talk), ( VJV,). The points are measured and the curves are calculated from Eqs. (2). (c), (d) Input impedance at respectively ports 1 and 2 when the other is terminated in 50 a. The points are measured and the curves are calculated from Eqs. (2).

450 Rev. Sci. Instrum., Vol. 65, No. 2, February 1994 rf power combiner


(4 (b)

FIG. 4. A transmission line schematic of the high-power combiner. (a) Input 1:2 baluns. (b) Power combiner stage. (c) Output 1:l: baluns.

It is clear that this will adversely affect electrical performance at high frequency. The solution to this problem is to reinvent the combiner of Fig. 2 so that it adds and sub- tracts Vi and V2 with equal phase delays. The solution we adopt is that of Guanella2 in which transmission lines are connected in a parallel arrangement so that in-phase voltages are summed at all circuit nodes.

III. THE HIGH-POWER COMBINER

In Fig. 4 we show the complete circuit of the power combiner. All cable runs 1-12 are wound on separate cores. In Fig. 4(b) is the combiner stage of the device (lines 7-10) which consists of four equal length transmis-

TABLE I. The general balun.

input I- - I

Symmetric T Outputs

1 7-

; or ’ ! I I 7 ’ I

-r M lrnmoedance

Transformatiw Cable Impedance fgb7’-7z Transf,

1 I:1 22 2 I:4 (l/z)*zz

3 4:9 (z/3)*22

4 9:16 (3/4)*22

id &ml ((N-l)/N)*Zz

sion lines. For a 50 fl system the characteristic impedance of these lines is also 50 fi so that each parallel pair has an impedance of 100 Sz in push-pull. The output lines 11-12 in Fig. 4(c) are 1: 1 baluns consisting of 50 0 lines which convert the push-pull signals back to single ended. It is clear that the input signals are added at the Vol port and subtracted at the Vo2 port with equal phase delays as required. For cophased inputs, all the input power is trans- ferred to the Vol port to which the main 50 Cl load is connected. The reject load is connected at port Vo2.

The combiner stage requires a pushLpul1 or symmetric version of the input signals. We therefore require a 1:2 balun with single ended input and push-pull output. Al- though there are many designs of I:2 baluns that are used as impedance transformers,3 a version has to be chosen which does not suffer from the same phase delay problems as the small signal hybrid of Fig. 2.

In Fig. 4(a), lines l-3 (and 4-6) is shown a 1:2 balun satisfying these requirements. The impedance ratio of this balun is in fact 4:9 and is a member of a general class of devices whose properties are summarized in Table I. To the best of the authors’ knowledge, this class of devices is not well known except for the N= 1 and N=2 cases. The baluns in Fig. 4(a) correspond to the case with N=3. From Table I, the impedance of the transmission lines of the 1:2 balun is therefore about 67 0 for a 50 R system.

-60 0.1 1 .o 10.0 100.0

Frequency (MHz)

FIG. 5. Input VSWR (at either port), total power loss, ratio of detected voltage at input port 2 to that at port 1 with input applied only at port

1( VJV,), and the ratio of reject port voltage to output voltage ( V&V,,) for cophased inputs.

Rev. Sci. Instrum., Vol. 65, No. 2, February 1994 rf power combiner 451


A version of the device of Fig. 4 has been constructed to combine a pair of 4-26 MHz tunable broadcast transmitters rated at 30 kW PEP into the tuning circuit of an antenna used for pulsed plasma formation and heating. The input baluns of Fig. 4(a) were constructed by winding four turns of four RG-58 inner conductors on each of three separate cores. The cores each consisted of a stack of two AMIDON FT-240-43 nickel-zinc ferrite toroids. Bach of the four RG-58 inner conductors were cross connected to give an impedance of - 70 a on each of the lines, l-6. The combiner of Fig. 4(b), lines 7-10, was constructed using four similar cores each with four turns of 50 fi coaxial cable RG-213. The V,, cable of line 12 also consisted of four turns of RG-213 wound on a similar core. The output Vc, of line 11 was made using semirigid 50 0 coaxial cable RG-231 wound into a large air-core coil of ten turns.

IV. TEST RESULTS

In Fig. 5 are shown test measurements of input VSWR, V&V,,, V,/V,, and total power loss in dB (in- cluding both reflection and insertion losses). The overall performance is comparable to similar commercial devices. The large power loss of about 1 dB, however, compared with an insertion loss of typically 0.2 dB for a commercial device is attributed mainly to the input 1:2 baluns which have not been optimized. The subject of insertion loss in transmission line transformers is dealt with by Sevick.3 The present device could be operated at about 5% duty cycle for 30 kW output power with adequate isolation of the transmitters.

The ferrites in the cores of lines 2 and 5 of the input 1~2 baluns undergo the most heating because they have the largest voltage drops across them. From the results of Fig. 5, however, it is clear that the number of ferrites and windings can still be increased to meet higher power and duty

cycle requirements without sacrificing the frequency response. Forced cooling may also be introduced for cw applications. A second version of the device also tested suc- cessfully in which the coils of all coaxial cables were wound on air-core formers eliminating the need for ferrites. In this case two requirements had to be satisfied. First, the coil lengths had to be chosen long enough to give acceptable coil reactances at low frequencies or otherwise the electrical performance would be affected. On the othe.r hand, quarter-wave dipole resonances at high frequenciels were observed to occur as a result of standing waves on the outsides of driven return conductors. These resonances de- grade the input impedance and power transfer efficiency of the combiner and set an upper frequency limit to acceptable operation. The inclusion of ferrites significantly im- proves this situation because the cable lengths required to produce a large reactance at low frequencies are significantly reduced. Lowering the cable length also lowers the insertion loss.

V. DISCUSSKIN

We have described the design and operation of a simple and cheap radio frequency power combiner. The ideas used in this article can be adapted to suit either pulsed or cw applications over a wide range of frequencies.

ACKNOWLEDGMENTS

One of the authors (GGB) would like to acknowledge the receipt of a Queen Elizabeth II fellowship. The authors are grateful to Messieurs Robert Davies and Frank Uhlig for their involvement in the construction of the combiner..

’ C. L. Ruthroff, Proc. IRE 47, 1337 ( 1959). ‘G. Guanella, Brown-Boveri Rev. 31, 327 (1944). 3J. Sevick, Transmission Line Transformers, American Radio Relay League Inc., Newington, 1990.

452 Rev. Sci. Instrum., Vol. 65, No. 2, February 1994 rf power combiner


Surface Mount

Monolithic Ampliier

Page 1 of 4

Notes

A. Performance and quality attributes and conditions not expressly stated in this specification document are intended to be excluded and do not form a part of this specification document. B. Electrical specifications and performance data contained in this specification document are based on Mini-Circuit’s applicable established test performance criteria and measurement instructions. C. The parts covered by this specification document are subject to Mini-Circuits standard limited warranty and terms and conditions (collectively, “Standard Terms”); Purchasers of this part are entitled to the rights and benefits contained therein. For a full statement of the Standard Terms and the exclusive rights and remedies thereunder, please visit Mini-Circuits’ website at www.minicircuits.com/MCLStore/terms.jsp

Mini-Circuits®

www.minicircuits.com P.O. Box 350166, Brooklyn, NY 11235-0003 (718) 934-4500 [email protected]

Product Features

Typical Applications• Cellular• PCN instrumentation

simpliied schematic and pin description

Function Pin Number Description

RF IN 1RF input pin. This pin requires the use of an external DC blocking capacitor chosen for the frequency of operation.

RF-OUT and DC-IN 3

RF output and bias pin. DC voltage is present on this pin; therefore a DC blocking capacitor is necessary for proper operation. An RF choke is needed to feed DC bias without loss of RF signal due to the bias connection, as shown in “Recommended Application Circuit”.

GND 2,4Connections to ground. Use via holes as shown in “Suggested Layout for PCB Design” to reduce ground path inductance for best performance.

General DescriptionMAR-8ASM+ (RoHS compliant) is a wideband amplifier offering high dynamic range. It has repeatable per-formance from lot to lot. It is enclosed in a Micro-X package. MAR-8ASM+ uses Darlington configuration and is fabricated using InGaP HBT technology.

MAR-8ASM+

GROUND

RF IN

RF-OUT and DC-IN

REV. FM136025MAR-8ASM+120215

DC-1 GHz

CASE STYLE: WW107PRICE: $1.17 ea. QTY. (20)

• Exact footprint substitute for MAR-8SM and MSA-0886a,b

Benefits: • lower device voltage, 3.7 typ.

• lower power dissipation in the MMIC

• may eliminate need for choke (RFC)

• High gain, 31.5 dB at 0.1GHz, reduces component count• High power output, +12.5 dBm typ.• Internally Matched to 50 Ohms• Low noise• Improved stability• Protection against power supply transients

GND GND

RF-OUT and DC-IN

RF IN

1

2

3

4

Notes:a. Suitability for model replacement within a particular system must be determined by and is solely the responsibility of the customer based on, among other things, electrical performance criteria, stimulus conditions, application, compatibility with other components and environmental conditions and stresses.b. The Avago MSA-0885 part number is used for identification and comparison purposes only.

+RoHS CompliantThe +Suffix identifies RoHS Compliance. See our web site

for RoHS Compliance methodologies and qualifications

Monolithic InGaP HBT MMIC Ampliier

Notes


Mini-Circuits®

www.minicircuits.com P.O. Box 350166, Brooklyn, NY 11235-0003 (718) 934-4500 [email protected] Page 2 of 4

Electrical Speciications at 25°C and 36mA, unless noted

MAR-8ASM+

Absolute Maximum Ratings

Parameter Ratings

Operating Temperature -40°C to 85°C


Operating Current 65mA

Power Dissipation 250mW

Input Power 13dBm

Note: Permanent damage may occur if any of these limits are exceeded. These ratings are not intended for continuous normal operation.1Case is defined as ground leads.2Full temperature range.

Parameter Min. Typ. Max. Units

Frequency Range* DC 1 GHz

Gain f=0.1 GHz 31.5 dB

f=1 GHz 202 25

Input Return Loss f=DC to 1 GHz 15.5 dB

Output Return Loss f=DC to 1 GHz 11 dB

Output Power @ 1 dB compression f=1 GHz +12.5 dBm

Output IP3 f=1 GHz +25 dBm

Noise Figure f=1 GHz 3.1 dB

Recommended Device Operating Current 36 mA

Device Operating Voltage 3.7 V

Device Voltage Variation vs. Temperature at 36 mA +1.2 mV/°C

Device Voltage Variation vs. Current at 25°C 11.3 mV/mA

Thermal Resistance, junction-to-case1 140 °C/W

*Guaranteed specification DC-1 GHz. Low frequency cut off determined by external coupling capacitors.


Notes


Mini-Circuits®


Product Marking

Recommended Application Circuit

4

2

3

1

C block

IN

C block

Ibias

OUTV d

R F C (Optional)

C bypas s

V ccR bias (R equired)

Test Board includes case, connectors, and components (in bold) soldered to PCB

MAR-8ASM+

8A

Additional Detailed Technical InformationAdditional information is available on our web site. To access this information enter the model number on our web site home page.

1

2

3

4

index over pin 1

Markings in addition to model number designation may appear for internal quality control purposes.

R BIAS1

Vcc Bias Resistor Value2

7 88.7

8 118

9 143

10 174

11 200

12 226

13 255

14 280

15 3091 When being used as a substitute for MAR-8SM or MSA-0866, the bias resistor values must be changed to the values in this table.2 1% Resistor values (ohms) for optimum bias.

Performance data, graphs, s-parameter data set (.zip ile)

Case Style: WW107

Plastic micro-x, .085 body diameter, lead finish: tin/silver/nickel

Tape & Reel: F4

7” Reels with 20, 50, 100, 200, 500, 1K devices

13” Reels with 2K, 4K devices

Suggested Layout for PCB Design: PL-253

Evaluation Board: TB-411-8A+

Environmental Ratings: ENV08T3


Notes


Mini-Circuits®


No. Test Required Condition Standard Quantity

1 Visual InspectionLow Power MicroscopeMagnification 40x

MIP-IN-0003(MCT spec)

45 units

2 Electrical Test Room TemperatureSCD(MCL spec)

45 units

3 SAM AnalysisLess than 10% growth in term of delamination

J-Std-020C(Jedec Standard)

45 units

4Moisture SensitivityLevel 1

Bake at 125°C for 24 hoursSoak at 85°C/85%RH for 168 hoursReflow 3 cycles at 260°C peak

J-Std-020C(Jedec Standard)

45 units

VisualInspection

Electrical Test SAM Analysis

Reflow 3 cycles,260°C

Soak85°C/85RH168 hours

Bake at 125°C,24 hours

VisualInspection

Electrical Test SAM Analysis

Start

MSL Test Flow Chart

MAR-8ASM+

Human Body Model (HBM): Class 2 (2000 v to < 4000 v) in accordance with ANSI/ESD STM 5.1 - 2001

Machine Model (MM): Class M1 ( <100 v) in accordance with ANSI/ESD STM 5.2 - 1999

MSL RatingMoisture Sensitivity: MSL1 in accordance with IPC/JEDECJ-STD-020C

ESD Rating

engn4521/6521 embeddedwireless …users.cecs.anu.edu.au/~gerard.borg/anu/courses/engn4521/...3 radio...

Documents