ATMega328 power consumption


1
2		Comparison of power consumtion of ATMega328P (Arduino chip) running at different voltages and frequencies
3
4		TL;DR: Running your battery powered bare arduino project at 3.3V and 1MHz is a great energy saving option, running it at 1.8V and 1MHz is even better, provided your peripherals can run at 1.8V
5
6		Measurement Methodology: I uploaded an empty sketch (infinite loop) to a bare ATMega328P and measured current flowing to the uC. uC is running in normal mode with all periperals running, no idle or power saving mode is used. External crystal is used for 16MHz and 24MHz, internal oscilator for 1MHz and 8MHz, watchdog oscilator for 128kHz
7
8		Support: If you found this information helpful, feel free to support me by visiting my store at
9		www.FemtoCow.com
10
11		Legend:	Good	Meh	Bad	out of specs
12
13
14
15		Current drawn at given Voltage and Frequency
16
17	U[V]	I_128kHz[mA]	I_1Mhz[mA]	I_8Mhz[mA]	I_16Mhz[mA]	I_24Mhz[mA]
18	5.5	10.7	11.5	16.5	20.8	25.0
19	5	6.57	7.45	11.6	15.3	19.4
20	4.5	3.47	4.00	7.70	11.4	14.8
21	3.8	0.66	1.32	4.40	7.70	10.3
22	3.3	0.33	0.85	3.52	6.30	6.50
23	3	0.31	0.75	3.10	5.60	5.73
24	2.7	0.26	0.66	2.74	4.95	5.07
25	2.2	0.19	0.53	2.13	-	-
26	1.8	0.15	0.43	1.60	-	-
27	1.5	0.13	0.33	0.89	-	-
28
29
30
31		Power consumption at given Voltage and Frequency
32
33	U[V]	I_128kHz[mW]	I_1Mhz[mW]	I_8Mhz[mW]	I_16Mhz[mW]	I_24Mhz[mW]
34	5.5	58.9	63.3	90.8	114.4	137.5
35	5	32.9	37.3	58.0	76.5	97.0
36	4.5	15.6	18.0	34.7	51.3	66.4
37	3.8	2.5	5.0	16.7	29.3	39.0
38	3.3	1.1	2.8	11.6	20.8	21.5
39	3	0.9	2.3	9.3	16.8	17.2
40	2.7	0.7	1.8	7.4	13.4	13.7
41	2.2	0.4	1.2	4.7	-	-
42	1.8	0.3	0.8	2.9	-	-
43	1.5	0.2	0.5	1.3	-	-
44
45
46
47
48		Relative power consumption at given Voltage and Frequency to standard 5V 16MHz
49
50	U[V]	I_128kHz[%]	I_1Mhz[%]	I_8Mhz[%]	I_16Mhz[%]	I_24Mhz[%]
51	5.5	77	83	119	150	180
52	5	43	49	76	100	127
53	4.5	20	24	45	67	87
54	3.8	3.3	6.6	22	38	51
55	3.3	1.4	3.7	15	27	28
56	3	1.2	2.9	12	22	22
57	2.7	0.9	2.3	10	17	18
58	2.2	0.5	1.5	6.1	-	-
59	1.8	0.4	1.0	3.8	-	-
60	1.5	0.3	0.6	1.7	-	-
61		running at 1MHz at 1.8V consumes only 1% of power compared to running at 16MHz at 5V
62
63
64
65
66		Relative power consumption at given Voltage and Frequency to 3.3V 1MHz
67
68	U[V]	I_128kHz[%]	I_1Mhz[%]	I_8Mhz[%]	I_16Mhz[%]	I_24Mhz[%]
69	5.5	2,098	2,255	3,235	4,078	4,902
70	5	1,171	1,328	2,068	2,727	3,458
71	4.5	557	642	1,235	1,829	2,366
72	3.8	89	179	596	1,043	1,390
73	3.3	39	100	414	741	765
74	3	33	80	332	599	613
75	2.7	25	64	264	476	488
76	2.2	15	42	167	-	-
77	1.8	10	28	103	-	-
78	1.5	7	18	48	-	-
79		running at 1MHz at 1.8V consumes only 1% of power compared to running at 16MHz at 5V
80
81
82
83
84		Mili Watts per MIPS at given Voltage and Frequency
85
86	U[V]	I_128kHz[mW]	I_1Mhz[mW]	I_8Mhz[mW]	I_16Mhz[mW]	I_24Mhz[mW]
87	5.5	460	63	11	7.2	5.7
88	5	257	37	7.3	4.8	4.0
89	4.5	122	18	4.3	3.2	2.8
90	3.8	20	5.0	2.1	1.8	1.6
91	3.3	8.5	2.8	1.5	1.3	0.9
92	3	7.3	2.3	1.2	1.1	0.7
93	2.7	5.5	1.8	0.9	0.8	0.6
94	2.2	3.2	1.2	0.6	-	-
95	1.8	2.1	0.8	0.4	-	-
96	1.5	1.5	0.5	0.2	-	-
97
98		Runnig uC at lower frequency saves power, but the calculation takes longer, consuming more power between sleeps. There is a sweet spot 1.8V@1MHz and 2.7V@8MHz