LTC1867 (16 bit ADC)

NearSys

I'm experimenting with a LTC1867.· The datasheet makes it look a lot like the MAX186, an ADC I use frequently.· However, when I read values every second, there's a lot of funcuation in the values.· Besides, any one reading is not the right number.··The numbers look make me think I'm close to making it work properly.·

Has anyone used the LTC1867?

Thanks,
Paul·

NearSys

Here's the code I'm using.· I based it on the MAX186 (12-bit) ADC code I'm using 9the code for the MAX186 works perfectly).· There's two toher ADCs in this project and I set their CS lines high to keep them from conversing when I talk to the LTC1867 over the DATA and CLK lines

PreLaunch:
· HIGH CS_08
· HIGH CS_12
· HIGH CS_16
· RETURN
ADC_16:
· FOR Index = 1 TO Analog_16
··· LOOKUP Index,[noparse][[/noparse]%0000000,%1000010,%1100010,%1001010,%110110,%1010010,%1110010,%1011010,%1111010],ByteVariable
··· LOW CLK
··· LOW CS_16
··· SHIFTOUT DIO,CLK,1,[noparse][[/noparse]ByteVariable\7]
··· SHIFTIN DIO,CLK,MSBPOST,[noparse][[/noparse]WordVariable\16]
··· HIGH CS_16
··· DEBUG DEC WordVariable, CR
· NEXT
· PAUSE 1000
· GOTO ADC_16

The temperature sensor's voltage I'm trying to measure is 2.93 volts.· I verified there's 2.93 volts on pin 1 of the surface mount LTC1867.· The LTC1867 has a max voltage input of·4.096 volts.· So for a bit ADC with 65535 steps, I should get a result of 46948.· I ran the above code (which take readings once per·second) twice and got the following results (I've written it into two columns)

0······· 0
28928·18688·
34560·25088
38400·27392
42240·31232
45824·35072
50176·35584
49152·39680
50176·38144
50944·40448
53504·43264
52224·44800

Notice that I get a 0 for the first reading.· When I had no input voltage, I got the following results.

0
40192
17664
35840
29440
31488
39680
6400
12032
26624
53760
14080

I checked the Vref pin and measured 2.5 volts.· The REFComp pin was at 4.09 volts.· So I'm pretty sure the ADC is connected properly.·The datasheet for this ADC is at ·http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1155,C1001,C1158,P2496,D2141

I'm attaching two images.· The photo is a close up of the PCB I've designed.· The LTC is the surface mount chip in the SparkFun break out board.· The·screen capture is from·the PCB·CAD program I used to design·the PCB.· I've highlighted the·CS, DATA,and CLK traces.· Since the board is single sided, there are a couple of jumpers.· I don't see·any bad solder connections and besides the voltages I measure lead me to believe the·connections are good.

My guess is that I'm misreading something in the datasheet.

This chip would be great to include with the BASIC Stamp weather station, since it would give you so much resolution.·

Paul