LDO characterization
Laura GonellaPhysikalisches Institut Uni Bonn
2
Status
• Tests performed on the LDO mode of the Shunt-LDO regulator– Single device characterization
• Line regulation(*)
– T dependence• Load regulation(*)
– T dependence• Load transient
– Parallel operation: same Vin, different Vout• Line regulation• Load regulation• Load transient
(*) These results have already been shown and discussed. They are added here as useful material for the design review
4/20/2011
3
Single device characterization
• Both regulators on chip have been tested independently• Results are in good agreement• Shown here results from Reg2
4/20/2011
4
Line regulation
4/20/2011
• For a certain Iload, the Vout is stable as a function of Vin, once the output is regulated
• To have a regulated output up to Iload = 600mA the Vdrop has to be at least 200mV– Even ~270mA for Iload =
600mA and Vout = 1.2V• The lower Vout for Iload =
0.6A is due to the bad load regulation
5
Line regulation: T dependence
4/20/2011
• The line regulation is stable in a T range from +20°C to -20°C
6
Load regulation• For Iload<10mA, the output
stage of the amplifier A1 that controls the pass transistor is driven out of saturation which decreases the regulation loop gain. This explains the bad load regulation for Iload < 10mA
• Investigations are ongoing to explain the bad load regulation for Iload > 10mA– The test setup seems to be fine– Results on 4 chips from 2
different wafers agree, excluding process variation
– Simulations with corners and T are good
– Ongoing post layout simulations with extracted parasitics to extimate the on chip wiring resistance
4/20/2011
7
Load regulation
4/20/2011
Vdrop (V)
Rout (mΩ) @ Vout = 1.2V
Rout (mΩ) @ Vout = 1.5V
0.1 231 1550.2 164 1440.3 155 1510.4 155 1530.5 155 1550.6 155 156
8
Load regulation: T dependence
4/20/2011
• The load regulation is stable in a T range from +20°C to -20°C
9
Load transient
4/20/2011
• Load current pulse measured across a 100mΩ resistor
• Iload = 66mA → 190mA, ΔIload = 124mV
• Rise time = fall time = 200ns
• Pulse width = 10us
• ΔVout = 18.0mV• Rout = 145mΩ
• Vout = 1.2V, Vin = 1.9V– Need to have Vin high enough to make sure Vdrop is >200mV
during the transient
10
Load transient
4/20/2011
• Load current pulse measured across a 100mΩ resistor
• Iload = 98mA → 330mA, ΔIload = 232mV
• Rise time = fall time = 200ns
• Pulse width = 10us
• ΔVout = 29.2mV• Rout = 126mΩ
• Vout = 1.5V, Vin = 2.0V– Need to have Vin high enough to make sure Vdrop is >200mV
during the transient
11
Load transient
• This is what happens if Vin is not high enough• Same as previous slide, just with Vin = 1.9V
4/20/2011
12
Parallel operation
• Reg1 generates Vout = 1.5V• Reg2 generates Vout = 1.2V • The Vin is in common
– The Vdrop on the Vin lines (from supply to the chip pad) is slightly different so Vin1 ≠ Vin2
– Reg2 sees a Vdrop ≥ 0.4V
4/20/2011
13
Line regulation
• Iload1 = 380mA• Iload2 = 180mA
4/20/2011
14
Load regulation
• Iload2 = 180mA, Vout = 1.169V• Iload1 = 0 – 0.6A
4/20/2011
Vdrop (V)
Rout (mΩ)
0.1 1730.2 1450.3 1530.4 155
15
Load regulation
• Iload1 = 380mA, Vout = 1.446V• Iload2 = 0 – 0.6A
4/20/2011
Vdrop (V)
Rout (mΩ)
0.4 1530.5 155
16
Load transient
• Iload1 = 380mV• Iload2 = 58mA → 184mA, ΔIload = 126mV; rise time = fall
time = 200ns; pulse width = 10us
4/20/2011
ΔVout2 = 17.2mVRout2 = 137mΩ ΔVout1 =
6mVΔVin2 =
350mVΔVin1 =
320mV
Vout2
Vout1
Vin2
Vin1
17
Load transient
• Iload2 = 180mV• Iload1 = 114mA → 348mA, ΔIload = 234mV; rise time = fall
time = 200ns; pulse width = 10us
4/20/2011
ΔVout1 = 30.8mVRout1 = 132mΩ ΔVout2 =
4.8mVΔVin1 =
650mVΔVin2 =
590mV
Vout1
Vout2
Vin1
Vin2