Vos and Ib - Specifications — Transcript

Full transcript

1. 0:00Hello and welcome to the TI Precision Lab
2. 0:02discussing Input Offset Voltage (Vos)
3. 0:05and Input Bias Current (Ib).
4. 0:08In this video, we'll discuss op amp Vos specifications
5. 0:11and Vos drift over temperature, as well as
6. 0:14input bias current specifications and input bias
7. 0:17current drift over temperature.
8. 0:19We'll also show the range of Vos and Ib
9. 0:21across many different TI op amps.
10. 0:25Let's start by defining offset voltage.
11. 0:28Offset voltage is the differential input voltage
12. 0:31that would have to be applied to force the op amps
13. 0:34output to zero volts.
14. 0:36Typical offset voltages range from millivolts
15. 0:39down to microvolts, depending on the op amp model.
16. 0:43Offset can be modeled as an internal DC source connected
17. 0:46to the input of the op amp.
18. 0:48Changing power supply voltage and common mode voltage
19. 0:52will affect input offset voltage.
20. 0:57Looking at the inside of an op amp,
21. 0:59we can see that the mismatch of transistors Q1
22. 1:02and Q2 in the differential input pair
23. 1:05is what causes the offset voltage.
24. 1:08In some cases, internal resistors
25. 1:11ROS1 and ROS2 are laser trimmed in order
26. 1:15to compensate for this mismatch and obtain very low
27. 1:18offset voltage.
28. 1:20In other cases, an internal digital correction circuit
29. 1:24is used to minimize offset voltage and offset drift.
30. 1:30This slide introduces op amp specifications for offset.
31. 1:34The top of the specification table
32. 1:37is the test conditions for all the parameters in the data
33. 1:39sheet.
34. 1:40In this example, the temperature is 25 degrees Celsius,
35. 1:44the load resistance is 10 kilo ohms,
36. 1:47the load is connected to mid supply,
37. 1:49and the common mode voltage is set to mid supply.
38. 1:53These conditions are true unless otherwise specified.
39. 1:57If you look at the offset voltage specs,
40. 1:59it lists some additional conditions.
41. 2:02The supply voltage is plus or minus 15 volts,
42. 2:05and the common mode voltage is zero volts.
43. 2:09Also note that we have a typical and a maximum specification.
44. 2:14The value listed in the typical specification
45. 2:16will cover plus or minus one standard deviation
46. 2:20R plus or minus sigma on a Gaussian distribution.
47. 2:24This means that 68% of the device population
48. 2:28will be less than the typical value.
49. 2:31So in this case, 68% of the devices
50. 2:34would have less than plus or minus 75 microvolts of Vos.
51. 2:39The maximum is a tested value, and so you will never
52. 2:42find a device with greater than the maximum Vos of plus or minus
53. 2:46150 microvolts.
54. 2:50We also have a Vos drift specification
55. 2:53that is measured in microvolts per degrees C,
56. 2:55describing how the Vos changes with temperature.
57. 2:59In this case, the typical drift is
58. 3:01given as 0.1 microvolts per degrees C.
59. 3:05The maximum drift is given as two microvolts per degrees C.
60. 3:12Most op amp SPICE models include the effects of offset voltage.
61. 3:16Several external conditions, such as power supply voltage
62. 3:20and common mode voltage, affect the offset voltage
63. 3:23on a real world device.
64. 3:25These effects are also included in the simulation model.
65. 3:30In order for the simulation result
66. 3:31to match the offset specifications in the data sheet
67. 3:34table, the same test conditions must
68. 3:37be applied to the amplifier.
69. 3:39In this example, the power supply is set to five volts,
70. 3:43the common mode voltage is set to mid supply R 2.5 volts,
71. 3:47and the load is connected to mid supply
72. 3:49in order to match the data sheet conditions.
73. 3:53The typical offset specification is 150 microvolts,
74. 3:56and the simulated offset is also 150 microvolts.
75. 4:00The goal of our models is to target typical op amp
76. 4:04performance.
77. 4:09The slope on offset voltage drift
78. 4:11can be either positive or negative.
79. 4:14This formula shows one possible definition for offset drift.
80. 4:19This formula will produce a positive or a negative drift,
81. 4:22depending on the slope of the curve.
82. 4:24Some other definitions use the absolute value,
83. 4:27so you will not have a negative offset.
84. 4:31This is the more common definition for drift
85. 4:33which is separated into two different regions, although more
86. 4:37than two regions could be used, if desired.
87. 4:41The idea with this definition is that you
88. 4:43get a more realistic view of what the expected error would
89. 4:46be than if you only considered the endpoints
90. 4:48over the entire region.
91. 4:51In this example, you can see that the slope of the two
92. 4:53separate regions is much more severe than the drift
93. 4:56of the entire range.
94. 4:58Note that the absolute value is used in the formula,
95. 5:01so this formula will never give a negative result.
96. 5:07In this application example, we will
97. 5:09see how to calculate the output voltage
98. 5:11error from the offset voltage.
99. 5:14Consider offset voltage as a DC voltage source in series
100. 5:18with the non-inverting input of the op amp.
101. 5:21We have a 0.1 millivolt, or 100 microvolt offset, in this case.
102. 5:27The signal source is a very small input of 1 millivolt,
103. 5:30so the offset will generate a fairly significant error.
104. 5:34The gain for this part is configured
105. 5:36as 100 volts per volt, which can be calculated
106. 5:39as R2 over R1 plus 1.
107. 5:44The total output voltage is the series combination of the offset
108. 5:47and the input signal are 1 millivolt plus 0.1 millivolts
109. 5:53multiplied by the gain of 100, which gives us 110 millivolts.
110. 5:58The offset accounts for about 10% error.
111. 6:05Offset drift calculations can be done in a similar manner.
112. 6:08Notice that we have two sources, one for the initial offset,
113. 6:12and one for the offset drift.
114. 6:14The offset drift source will be zero
115. 6:17at 25 degrees C. As the temperature deviates
116. 6:21from 25 degrees C, the temperature difference will
117. 6:24be multiplied by the offset drift
118. 6:26to generate the additional offset voltage.
119. 6:29For example, at 25 degrees C we have 100 microvolts
120. 6:33of offset, which is just the room temperature offset
121. 6:36and no drift term.
122. 6:39At 125 degrees C, we have a total of 250 microvolts.
123. 6:43That is 100 microvolts from the initial offset,
124. 6:47and 150 microvolts from the drift term.
125. 6:52The table on the right illustrates
126. 6:53how the offset changes over temperature.
127. 6:57Keep in mind that the slope of the offset drift
128. 6:59can be either positive or negative,
129. 7:01so both cases are shown.
130. 7:04Drift is especially important in calibrated systems.
131. 7:08In calibrated systems, room temperature offset
132. 7:11is frequently measured and corrected for in software.
133. 7:14Temperature drift, however, is often difficult and expensive
134. 7:17to calibrate out, so devices with minimal drift
135. 7:20are preferable.
136. 7:25This chart shows a range of offset voltages,
137. 7:27from microvolt to millivolts, for different types of TI
138. 7:31amplifiers.
139. 7:33The first amplifier in the list, the OPA333,
140. 7:36includes a Zero Drift apology which
141. 7:39uses an internal digital calibration circuit to minimize
142. 7:42offset and offset drift.
143. 7:45Some precision bipolar amplifiers use laser
144. 7:47trimming to minimize offset.
145. 7:50Often you must trade off bandwidth
146. 7:51or other characteristics for low offset.
147. 7:54For example, the OPA835 is optimized for speed,
148. 7:59not for offset.
149. 8:01Also commodity or low cost amplifiers
150. 8:04are usually not optimized for low offset or low offset drift.
151. 8:12Let's now move on to input bias current, or Ib,
152. 8:15and input bias current drift.
153. 8:19Input bias current is the current flowing
154. 8:21into the inputs of an op amp.
155. 8:23These currents can be modeled as a current source connected
156. 8:26to each input as shown in this figure.
157. 8:29Ideally, the two input bias currents
158. 8:32would be equal to each other and would cancel.
159. 8:35In reality though, they are not equal,
160. 8:37and the difference of these currents
161. 8:39is defined as input offset current.
162. 8:42If the input offset current is low,
163. 8:45it's possible to match the impedances connected
164. 8:47to each input and cancel the offset developed from the input
165. 8:50bias currents.
166. 8:55In a bipolar amplifier, input bias current
167. 8:58is the current flowing into the base of each transistor
168. 9:01in the input pair.
169. 9:03Generally, the bias current for bipolar amplifiers
170. 9:06is larger than the bias current for MOSFET and JFET amplifiers.
171. 9:11Typical numbers are in the range of nanoamps.
172. 9:15You can see in the case of the LM741C,
173. 9:19the input offset current is about 200 nanoamps max,
174. 9:23and the input bias current is about 500 nanoamps max.
175. 9:31Some precision bipolar op amps use
176. 9:33a method called bias current cancellation in order
177. 9:36to minimize bias current.
178. 9:39This is done inside the op amp, so no external components
179. 9:42are required.
180. 9:44The amplifier simply behaves like a bipolar amplifier
181. 9:48with very low bias current.
182. 9:50Bias current cancellation is done by measuring the input bias
183. 9:53current, and summing in equal but opposite currents, which
184. 9:57cancel the bias current.
185. 9:59This effectively takes an amplifier
186. 10:01with hundreds of nanoamps of bias current
187. 10:04down to single nanoamps of bias current.
188. 10:08You can see from the specification table
189. 10:10in this example that the input bias current of the OPA277
190. 10:14is plus or minus one nanoamp maximum.
191. 10:18In the previous example, the bias current
192. 10:21had to flow into the base of a transistor,
193. 10:23so the bias current could only have one polarity.
194. 10:27In this case, however, the bias current
195. 10:29can have either polarity, since the bias current cancellation
196. 10:33circuit is not perfect, and it's not
197. 10:35known whether the polarity of the residual current
198. 10:38will be positive or negative.
199. 10:43In the case of MOSFET or JFET op amps,
200. 10:46the input bias current is primarily
201. 10:48due to the leakage of the input ESD protection diodes.
202. 10:53The gate of the input MOSFET transistors
203. 10:56has extremely low leakage, so it doesn't contribute
204. 10:59significant bias current.
205. 11:01You can see in this example that the OPA369 has 50 picoamps max
206. 11:07of input bias current.
207. 11:13One thing to remember with low bias current amplifiers
208. 11:16is the effect of Ib over temperature.
209. 11:19In MOSFET amplifiers, the bias current
210. 11:22can double every 10 degrees C. You can see in the example
211. 11:26on the left, with the OPA350, that the input bias
212. 11:30current increases significantly at temperatures above 25 degrees
213. 11:34C. If you only considered the room temperature value of Ib,
214. 11:39and then operated the amplifier at elevated temperature,
215. 11:42you would have significant errors.
216. 11:45Notice that the vertical axis of the plot
217. 11:47uses a logarithmic scale.
218. 11:51With a bipolar amplifier, the initial input bias current
219. 11:55at room temperature is often large enough such
220. 11:58that the relative change in input
221. 11:59bias current over temperature is minimized.
222. 12:03You can see in the example on the right with the OPA277
223. 12:08that the input bias current starts
224. 12:09to increase at temperatures above 75 degrees C. Note,
225. 12:14however, that the vertical axis uses a linear scale.
226. 12:20This bias current calculation is very similar to what
227. 12:23was done for offset voltage.
228. 12:26First, we model the bias currents
229. 12:28as two current sources connected to the op amp inputs.
230. 12:32Note that the input bias current connected
231. 12:34to the non-inverting input is flowing back
232. 12:36into the input signal source.
233. 12:39And since there is no source resistance,
234. 12:41that bias current source does not add any error voltage.
235. 12:45If there was a source resistance connected
236. 12:47to the non-inverting input, the bias current
237. 12:50would generate an error voltage.
238. 12:53The error in this configuration arises
239. 12:55entirely from the input bias current source
240. 12:58on the inverting input, which flows into the feedback
241. 13:01network of R1 and R2.
242. 13:04If we perform a nodal analysis, we
243. 13:07can see that the output voltage caused by Ib
244. 13:10is equal to Ib multiplied by Rf.
245. 13:15We can then calculate the output caused by Ib and the output
246. 13:18from the input signal.
247. 13:21Using superposition, we can add the output signal
248. 13:23from the bias current equal to 20 millivolts and the output
249. 13:27signal from the input source equal to 100 millivolts,
250. 13:31since they are independent.
251. 13:33In this example, the total output voltage
252. 13:36equals 120 millivolts, and the error introduced by the input
253. 13:40bias current is about 20%.
254. 13:43Please keep in mind that this was an error calculation using
255. 13:46high temperature Ib values.
256. 13:49If this calculation was done at room temperature,
257. 13:52the error would have been significantly smaller.
258. 13:57This table gives a range of input bias currents
259. 13:59for different TI op amps.
260. 14:02Values can range from FEM to amps for specialized CMOS
261. 14:05amplifiers all the way up to hundreds
262. 14:07of nanoamps for high speed and commodity op amps.
263. 14:12Note that bipolar amplifiers will always
264. 14:14have higher input bias currents than CMOS amplifiers.
265. 14:18Also, bipolar amplifiers with bias current cancellation
266. 14:22circuitry, such as the OPA277, will have lower input bias
267. 14:27current than bipolar op amps without cancellation,
268. 14:31such as the OPA211.
269. 14:36That concludes this video.
270. 14:37Thank you for watching.
271. 14:39Please try the quiz to check your understanding
272. 14:41of this video's content.