Digipot - linear (need log) - can charactersitic be changed in code?

[HankMcSpank]

I've a circuit where I need to use a dual digipot - the thing is I need one to be linear & one to be log - they don't make up hybrid - it's one or the other. I really don't have PCB space for two separate pots, so just wondering is there anyway this can be spoofed in code?

I'm using a linear pot as input...I'm thinking here along the lines of some clever maths on the AtoD sample (ie from the linear input pot) to then control the output pot in a log-esque way?

[Jerson]

Off the top of my head, you could just use a log lookup table for one of the pots that will return log values instead of linear ones.

[Acetronics2]

Off the top of my head, you could just use a log lookup table for one of the pots that will return log values instead of linear ones.

+1 ...

But care to very poor and variable resolution ...

Alain

[HankMcSpank]

+1 ...

But care to very poor and variable resolution ...

Alain

Oop, straight out the blocks, an epic fail on this one!

so to generate a lookup table, I need my digipot's 256 linear step converted to log.

So I use the Excel log function to log 256 number & then charted it within Excel, the problem is, I get a plot like the No4 trace in the first chart/diagram about halfway down this page (ie a reverse log characteristic), whereas what I want is a trace like no2 on that same diagram


http://www.geofex.com/article_folder...s/potscret.htm

How can I generate my LUT values for a 256 pot & have it look like trace 2?!!! (now I'm sure this is embarrasingly simple ....but I now realise that I shouldn't have tried so hard to pull Mary Harbottle during maths lessons)

[languer]

Try the antilog (i.e. 10^x)

[HankMcSpank]

Try the antilog (i.e. 10^x)

thanks but that wasn't it.

ok, I think I've cracked it, well actually Maxim did for me http://pdfserv.maxim-ic.com/en/an/AN3996.pdf (there's an exe embedded in their Apllication note which yields the attenuation per step - I was able to dick about in Excel until I got my values to match theirs, so for a 256 step pot, you have this formula

Attenuation = 20*log(256 -actual tap_position) / (255)

which yields the cuve in excel I want....albeit in DBs, but I rebased it to actual steps., giving me something like this...

Tap     Log     Attenuation
1	0	0.0000
2	0	-0.0341
3	0	-0.0684
4	0	-0.1028
5	1	-0.1373
6	1	-0.1720
7	1	-0.2068
8	1	-0.2418
9	1	-0.2769
10	1	-0.3121
11	2	-0.3475
12	2	-0.3830
13	2	-0.4187
14	2	-0.4545
15	2	-0.4905
16	2	-0.5266
17	3	-0.5628
18	3	-0.5993
19	3	-0.6358
20	3	-0.6726
21	3	-0.7094
22	3	-0.7465
23	4	-0.7837
24	4	-0.8210
25	4	-0.8586
26	4	-0.8962
27	4	-0.9341
28	5	-0.9721
29	5	-1.0103
30	5	-1.0486
31	5	-1.0872
32	5	-1.1258
33	5	-1.1647
34	6	-1.2037
35	6	-1.2430
36	6	-1.2823
37	6	-1.3219
38	6	-1.3617
39	7	-1.4016
40	7	-1.4417
41	7	-1.4820
42	7	-1.5225
43	7	-1.5632
44	7	-1.6041
45	8	-1.6452
46	8	-1.6864
47	8	-1.7279
48	8	-1.7695
49	8	-1.8114
50	9	-1.8535
51	9	-1.8957
52	9	-1.9382
53	9	-1.9809
54	9	-2.0238
55	10	-2.0669
56	10	-2.1102
57	10	-2.1537
58	10	-2.1975
59	10	-2.2415
60	11	-2.2857
61	11	-2.3301
62	11	-2.3748
63	11	-2.4197
64	11	-2.4648
65	12	-2.5101
66	12	-2.5557
67	12	-2.6016
68	12	-2.6476
69	13	-2.6940
70	13	-2.7405
71	13	-2.7874
72	13	-2.8344
73	13	-2.8818
74	14	-2.9294
75	14	-2.9772
76	14	-3.0254
77	14	-3.0737
78	15	-3.1224
79	15	-3.1713
80	15	-3.2206
81	15	-3.2700
82	15	-3.3198
83	16	-3.3699
84	16	-3.4202
85	16	-3.4709
86	16	-3.5218
87	17	-3.5731
88	17	-3.6246
89	17	-3.6765
90	17	-3.7286
91	18	-3.7811
92	18	-3.8339
93	18	-3.8871
94	18	-3.9405
95	19	-3.9943
96	19	-4.0484
97	19	-4.1029
98	19	-4.1577
99	20	-4.2128
100	20	-4.2683
101	20	-4.3242
102	20	-4.3804
103	21	-4.4370
104	21	-4.4939
105	21	-4.5513
106	21	-4.6090
107	22	-4.6671
108	22	-4.7256
109	22	-4.7845
110	23	-4.8437
111	23	-4.9034
112	23	-4.9636
113	23	-5.0241
114	24	-5.0850
115	24	-5.1464
116	24	-5.2082
117	25	-5.2705
118	25	-5.3332
119	25	-5.3964
120	25	-5.4600
121	26	-5.5241
122	26	-5.5887
123	26	-5.6538
124	27	-5.7193
125	27	-5.7854
126	27	-5.8519
127	28	-5.9190
128	28	-5.9866
129	28	-6.0547
130	28	-6.1234
131	29	-6.1926
132	29	-6.2624
133	29	-6.3327
134	30	-6.4036
135	30	-6.4751
136	30	-6.5472
137	31	-6.6199
138	31	-6.6932
139	31	-6.7671
140	32	-6.8416
141	32	-6.9168
142	33	-6.9927
143	33	-7.0692
144	33	-7.1464
145	34	-7.2243
146	34	-7.3029
147	34	-7.3823
148	35	-7.4623
149	35	-7.5431
150	35	-7.6247
151	36	-7.7070
152	36	-7.7901
153	37	-7.8741
154	37	-7.9588
155	37	-8.0444
156	38	-8.1308
157	38	-8.2181
158	39	-8.3063
159	39	-8.3954
160	39	-8.4854
161	40	-8.5763
162	40	-8.6682
163	41	-8.7611
164	41	-8.8550
165	42	-8.9500
166	42	-9.0460
167	43	-9.1430
168	43	-9.2412
169	43	-9.3404
170	44	-9.4408
171	44	-9.5424
172	45	-9.6452
173	45	-9.7492
174	46	-9.8545
175	46	-9.9611
176	47	-10.0690
177	47	-10.1783
178	48	-10.2889
179	48	-10.4010
180	49	-10.5145
181	49	-10.6296
182	50	-10.7462
183	51	-10.8643
184	51	-10.9842
185	52	-11.1056
186	52	-11.2288
187	53	-11.3538
188	53	-11.4806
189	54	-11.6093
190	55	-11.7399
191	55	-11.8725
192	56	-12.0072
193	57	-12.1440
194	57	-12.2830
195	58	-12.4242
196	58	-12.5678
197	59	-12.7138
198	60	-12.8622
199	61	-13.0133
200	61	-13.1670
201	62	-13.3235
202	63	-13.4829
203	64	-13.6453
204	64	-13.8107
205	65	-13.9794
206	66	-14.1514
207	67	-14.3269
208	68	-14.5060
209	68	-14.6888
210	69	-14.8756
211	70	-15.0666
212	71	-15.2618
213	72	-15.4614
214	73	-15.6658
215	74	-15.8751
216	75	-16.0896
217	76	-16.3095
218	77	-16.5351
219	78	-16.7668
220	79	-17.0048
221	80	-17.2494
222	81	-17.5012
223	83	-17.7605
224	84	-18.0278
225	85	-18.3036
226	87	-18.5884
227	88	-18.8828
228	89	-19.1876
229	91	-19.5035
230	92	-19.8313
231	94	-20.1720
232	96	-20.5266
233	97	-20.8962
234	99	-21.2823
235	101	-21.6864
236	103	-22.1102
237	105	-22.5557
238	107	-23.0254
239	109	-23.5218
240	112	-24.0484
241	115	-24.6090
242	117	-25.2082
243	120	-25.8519
244	124	-26.5472
245	127	-27.3029
246	131	-28.1308
247	135	-29.0460
248	140	-30.0690
249	145	-31.2288
250	152	-32.5678
251	159	-34.1514
252	168	-36.0896
253	180	-38.5884
254	196	-42.1102
255	224	-48.1308
256	256	Infinity

[Acetronics2]

Hi,

Just confirms #3 ...

try to add a linear, or best 2nd order, regression between each step ...

simple, you said simple ???

Alain