ADC Propeller Problem

Discovery

I was successful in soldering the ADC124S021 ADC onto the fine pads of my printed circuit board. It took some real effort to get the VSSOP fine spacing leads of the micro package soldered to the board.

Everything checks out at the propeller but the ADC will not convert an analog input voltage. After reviewing the Texas Instruments data sheet, I noticed that the power supply lead must be bypassed with two monolithic capacitors of 0.1 ufd and 1ufd placed within 1cm of the power pin. I have bypass caps on the power lead but they are about 3cm from the chip.

Could this be the problem? Does anyone have experience placing and operating this micro dot ADC on a PCB?

Are there equivalent ADCs in larger packages such as DIP or SOP?

I bought the monolithic caps but I found it impossible to hand solder these grains of sand on the PCB.

Discovery

JonnyMac

I like the MCP320x series; there is a 2-channel version in an 8-pin PDIP and SOIC package. You can find code in ObEx for it.

Duane Degn

Parallax sells the MCP3208 in a DIP package. I also like these 2 channel 10-bit chips.

I've had a few applications where the 10-bit chip worked better than the 12-bit chip (I think the 10-bit chip may read high impedance sources easier than the 12-bit chip).

Domanik

This is the same part used for an ADC on the Activity Board. Schematics are available on line as well as the code to initialize and read input voltage value to the ADC. The code is in the simple library and can be downloaded:

SimpleIDE/Learn/Simple Libraries/Convert/libadcDCpropab/Documentation adcDCpropab Library.html

I don't think the caps will make much difference with your problem-- they usually help to reduce power supply noise and you aren't getting any kind of read back.

This would be my check list:
Pin to Pin shorts: measure with a meter
double check pin connections from prop to ADC
does the device get warm, it might be in backwards
look at the digital inputs with a scope-does the SW wiggle pins
will init send clocks to the device on the clock pin and data on the data pins
are the input levels at full high/low (two outputs might be connected together)
should CS be high or low / is it set properly
if you put 3.3v on the Ainputs what do you expect for a binary output
where does your reference voltage come from-is it connected
Good luck,
Dom..

Discovery

I have experience with the ADC on the Propeller Activity Board.

No pin to pin shorts.
PC layout is correct.
Did not touch the chip for temperature.
Unfortunately, my Tektronics four channel digital scope crapped out as I was setting up to take signal measurements.
The instrument is at Tek for service.
Ohmmeter checks show no shorts pin to pin.
Ohmmeter shows traces are correct to the propeller.

I like the idea of switching to the 10-bit MCP3208 DIP...what is involved in making the ADC call using "C" coding?

Discovery

ValeT

You just have to include the library, initialize the ADC pins, and call adc_volts( pin# ). Below is example code for each step:

#include "adcDCpropab.h" //include the library

adc_init( 21, 20, 19, 18 ); //initialize adc pins

adc_volts( pin#( 0 - 3 ) ); //get input

Duane Degn

Discovery wrote: »

I like the idea of switching to the 10-bit MCP3208 DIP

Just to clarify, the MCP320x are 12-bit chips the MCP300x are 10-bit chips. As one would expect, the 10-chips usually cost less than their 12-bit counterparts.

Discovery

Vale T
The code you post is the same code that I use in SimpleIDC to run the ADC124S021 on the Activity Board. Are you saying that the code call is the same for the MCP320x and MCP300x ADCs?

Discovery

ValeT

I am not sure, but it's worth a shot since it is really easy to test.

Discovery

Easy enough.

I will order the ADC from Digikey tomorrow.

Of all the complex devices on my new PCB...the ADC is the only device that is not functioning.

Discovery

Duane Degn

Discovery wrote: »

Vale T
The code you post is the same code that I use in SimpleIDC to run the ADC124S021 on the Activity Board. Are you saying that the code call is the same for the MCP320x and MCP300x ADCs?

It's not the same. There are several different objects available for the MCP3xxxx chips in the OBEX.

Before I knew there was an object available for the ADC124S021 chip, I tried using an object written for the MCP3204 chip and it did not work correctly.

I was using Spin code but I bet you'll have the same problem trying to use ADC124S021 C code MCP3xxxx chips.

I do agree, this sort of thing is easy to test.

Do you plan to power the chip with 5V or 3.3V? Technically 3.3V logic is below the threshold listed in the datasheet for the MCP320x chips when powering the chips with 5V. I personally haven't had trouble using 3.3V logic with a MCP320x running at 5V.

Another thing to keep in mind if you're powering the ADC with 5V is to have a current limiting resistor between the chip's output and the Prop's I/O pin used as an input.

Discovery

Yes...5V is the Vdd feed for the ADC location on the board.

Is there a better choice of ADC that has SimpleIDE code in the library? I have not visited the OBEX.

I got the impression that C code is not on the venue at OBEX.

Okay on the limiting resistor at the Prop's I/O.

By the way...I got a big surprise today when I applied power to the fully interconnected system under control by the propeller. Since there is no operating program in the chip on power up...unlike the Basic Stamp design, the outputs of the Propeller drift to a high state and turned on all the motors and relays. Not a good idea for this system. I found that by pulling the propeller outputs down to ground using 360 ohm resistors...it was safe to energize the system then down load the program to the propeller. Everything worked properly.

The ADC is a key element in this system and I really could use an ADC C program pulled from the SimpleIDE library. Is that a possibility?

Discovery

Domanik

Last week I bought three MCP3021's from Digkey in a SOT23-5 package. They are now working. If you are interested in that part I can post my C code. I found it was straight forward to get it running and the code is solid. I like the idea of an I2C part that only needs 2 wires and uses only 2 prop pins. Everything is done with addresses so you don't need chip select and you can have up to 127 parts on one bus. The MCP3021 is 10 bit and the PN determines the address:
MCP3021A6T has an address 4E (48 + 6). The range is 48 to 4F.

If you're interested let me know.

TI might have another part number that uses the same timing as the ADC124S021 but comes in a larger package. A parametric check for 4 chan. Serial 12 bit ADC might turn something up. That might be easier than writing a new driver.
Dom..

Discovery

I cannot use the SOP...soldering becomes an issue to avoid.

It appears that the ADC124S021 is only produced in the VSSOP package.

Discovery