Forums - /etc... Data Converters Forum - /etc... Data Converters

choice of ADC

This question is answered

Posted by amj

on Jan 26 2011 07:07 AM

I'm using a apd5100 pressure sensor to monitor arterial line blood pressure. Since its an integrated sensor, i figured i could directly connect it to a 16 bit ADC.However I ain't sure of the choice of ADC for this application...as 1. the sampling rate 2. architecture-SAR/delta sigma 3. Input configuration range

Kindly help me out.

All Replies

Posted by Michael W

on Jan 26 2011 09:23 AM

Expert3285 points

Hello,

Unfortunately, we do not have details on this device. Do you have the datasheet, company/product page link, or any other details on this sensor? If so, then we will gladly look into some suggestions for an ADC.

Regards,

Michael Wheeler

Regards,

Michael W

Posted by amj

on Jan 26 2011 12:40 PM

Intellectual380 points

i apologise for the error...its an ADP5100 pressure sensor from panasonic...i decided to go for this one coz I couldnt fing any analog/digital output pressure sensors with a range of -700 to +700mmHg....if you have other suggestions, pls let know....Also, kindly answer my query with regards to the ADC.

Posted by Tom Hendrick

on Jan 26 2011 21:27 PM

Verified Answer

Verified by Tom Hendrick

Guru86190 points

Hi amj,

Looks like the ADP5x00 will get you +/-100kPA. That should give you about .45 to 4.5V output. Depending on your sample rate and resolution requirements, there are plenty of data converters from Texas Instruments to choose from.

A one channel low power option might be to take the OPA365 configured as a simple voltage follower between the sensor and an ADS8326. The output of the amp goes to the (+) ADC input, the (-) adc input goes to ground and you have a simple low power single supply option with 16-bits of resolution. There are other options in the ADS8326 family that might be more appropriate based on what your end goals are - the ADS8326 datasheet lists six or seven ADC options on the first page if memory serves me.

Let us know how you make out, this sounds like a fun project!

Regards,

Tom

Posted by amj

on Jan 27 2011 01:45 AM

Intellectual380 points

Thanks a tonne :)

the datasheet of the ADS8326 states that Fclock=24*Fsample...why so? isnt clock frequency=no. of bits*sampling freq ? How many equivalent DCLOCKS would that be considering my Fsample=250kHz

Posted by amj

on Jan 27 2011 01:47 AM

Intellectual380 points

ultimately my ADC output should go to a Spartan 3AN FPGA...

Posted by Tom Hendrick

on Jan 27 2011 11:50 AM

Guru86190 points

amj,

The maximum SCLK speed the ADS8326 device can tolerate is 6MHz. For 250KSPS, you would normally have the /CS tied low for 22 DCLKs to transfer data and high for two additional DCLK cycles between conversions - see the top of page 8 for details.

Regards,

Tom

Posted by amj

on Jan 28 2011 00:50 AM

Intellectual380 points

Thankx Tom!

any way i can obtain the spice model of those ICs..

Posted by Michael W

on Jan 28 2011 09:27 AM

Expert3285 points

Hi Amj,

The product folder for OPA365 contains the SPICE model for it, which you can find via the following link:

http://focus.ti.com/docs/prod/folders/print/opa365.html#toolssoftware

As for the ADS8326, there are no available SPICE models for it. However, there are 2 EVMs available, if you'd like to use those with your pressure sensor. For the EVM alone, that you could wire into your FPGA, we have the ADS8326EVM. If you would like a "ready to go" setup, that includes a controller and software, we have the ADS8326EVM-PDK.

Regards,

Michael

Regards,

Michael W

Posted by amj

on Mar 07 2011 02:50 AM

Intellectual380 points

Before trying this circuit with the sensor, i want to build the circuit by simulating the output signal characteristics of ADP5100 sensor...how do i do so and give it to OPA365?

Posted by Michael W

on Mar 10 2011 10:50 AM

Expert3285 points

Hello Amj,

You may try simulating the input using TINA-TI.

Regards,

Michael

Regards,

Michael W

Posted by amj

on Mar 15 2011 01:30 AM

Intellectual380 points

any specific reason as to why u suggested this ADC? and is a voltage follower circuit necessary?

Also, if I'm simulating the output signal of the sensor, can I do so by connecting a 5k pot across 5V and gnd and give its output directly to any ADC which then communicates with an MSP430.

Posted by Tom Hendrick

on Mar 15 2011 08:06 AM

Guru86190 points

Hi amj,

No specific reason for the ADS8326 other than its 'ease of use'. Typically a SAR adc wants to see a low impedance input. You may be able to get decent results with the 5K pot depending on your sample rate. In another post to this forum, you are looking for an 8-bit device in a DIP package. Here are a few:

Part Number	Status	Resolution (Bits)	Sample Rate (max)	# Input Channels	Architecture	Input Configuration Range	Interface	Serial SPI Interface	INL (+/-) (max) (%)	INL (max) (+/-LSB)	DNL (max) (+/-LSB)	Analog Input BW (MHz)	SNR (dB)	SFDR (dB)	SINAD (dB)	ENOB (Bits)	Power Consumption (typ) (mW)	Reference Mode	Analog Voltage AV/DD (min) (V)	Analog Voltage AV/DD (max) (V)	Digital Supply (min) (V)	Digital Supply (max) (V)	Rating	Pin/Package	Approx. Price (US$)	Description
Pict Pict TLV0838	ACTIVE	8	37.9kSPS	8	SAR	Vref	Serial SPI Interface		0.2	0.5	0.5						0.66	Ext	2.7	3.6	2.7	3.6	Catalog	20PDIP, 20SOIC, 20TSSOP	1.45 \| 1ku	8-Bit 37.9 kSPS ADC Serial Out, Ratiometric Op or Vcc Ref, TTL/MOS Input & Output Compatible, 8 Ch.
TLV0834	ACTIVE	8	41kSPS	4	SAR	Vref	Serial SPI Interface		0.2	0.5	0.5						0.66	Ext	2.7	3.6	2.7	3.6	Catalog	14PDIP, 14SOIC, 16TSSOP	1.45 \| 1ku	8-Bit 41 kSPS ADC Serial Out, Ratiometric Op or Vcc Ref, TTL/MOS Input & Output Compatible, 4 Ch.
TLV0832	ACTIVE	8	44.7kSPS	2	SAR	Vref	Serial SPI Interface		0.2	0.5	0.5						5	Ext	2.7	3.6	2.7	3.6	Catalog	8PDIP, 8SOIC	1.40 \| 1ku	8-Bit 44.7 kSPS ADC Serial Out, Muxed Twin Ch. w/SE or Differential Option, 2 Ch.
TLV0831	ACTIVE	8	49kSPS	1	SAR	+3.6V(Vref=VDD)	Serial SPI Interface		0.2	0.5	0.5						0.66	Ext	2.7	3.6	2.7	3.6	Catalog	8PDIP, 8SOIC	1.75 \| 1ku	8-Bit 49 kSPS ADC Serial Out, Differential input, Configurable as SE input, 1 Ch.
TLC548	ACTIVE	8	45.5kSPS	1	SAR	Vref	Serial SPI Interface		0.2	0.5	0.5						9	Ext	3	6	3	6	Catalog	8PDIP, 8SOIC	1.20 \| 1ku	8-Bit, 45.5 kSPS ADC Serial Out, Low Power, Compatible to TLC540/545/1540, Single Ch.
TLC542	ACTIVE	8	25kSPS	11	SAR	Vref	Serial SPI Interface		0.2	0.5	0.5						6	Ext	4.75	5.25	4.75	5.25	Catalog	20PDIP, 20PLCC, 20SOIC	1.50 \| 1ku	8-Bit, 25kSPS ADC Serial-Out, On-Chip 11-Ch. Analog MUX
TLC540	ACTIVE	8	75kSPS	11	SAR	Vref	Serial SPI Interface		0.2	0.5	0.5						6	Ext	4.75	5.5	4.75	5.5	Catalog	20PDIP, 20PLCC, 20SOIC	1.85 \| 1ku	8-Bit, 75 kSPS ADC Serial-Out, On-Chip 12-Ch. Analog Mux, 11 Ch.
TLC0838	ACTIVE	8	20kSPS	8	SAR	Vref	Serial SPI Interface		0.2	0.4	0.4						3	Ext	4.5	5.5	4.5	5.5	Catalog	20PDIP, 20SOIC, 20TSSOP	1.45 \| 1ku	8-Bit, 20 kSPS ADC Serial Out, uProcessor Periph./Standalone, Rem. Op w/Ser. Data Link, Mux option
TLC0834	ACTIVE	8	20kSPS	4	SAR	Vref	Serial SPI Interface		0.2	0.4	0.4						3	Ext	4.5	5.5	4.5	5.5	Catalog	14PDIP, 14SOIC	1.45 \| 1ku	8-Bit, 20 kSPS ADC Serial Out, uProcessor Periph./Standalone, Rem. Op w/Ser. Data Link, Mux option
TLC0832	ACTIVE	8	22kSPS	2	SAR	Vref	Serial SPI Interface		0.2	0.4	0.4						12.5	Ext	4.5	5.5	4.5	5.5	Catalog	8PDIP, 8SOIC	1.40 \| 1ku	8-Bit, 22 kSPS ADC Serial Out, uProcessor Periph./Standalone, Mux option w/SE or differential, 2 Ch.
TLC0831	ACTIVE	8	31kSPS	1	SAR	Vref	Serial SPI Interface		0.2	0.4	0.4						3	Ext	4.5	5.5	4.5	5.5	Catalog	8PDIP, 8SOIC	1.40 \| 1ku	8-Bit, 31 kSPS ADC Serial Out, Microprocessor Peripheral/Standalone, Single Channels

Regards,

Tom

Posted by amj

on Mar 15 2011 12:52 PM

Intellectual380 points

thanks :) do i need to give an anti aliasing filter?

Posted by Rick Downs - WEBENCH Design Center

on Mar 16 2011 16:16 PM

Expert5510 points

You always need some kind of band-limiting filter before a data converter. Sampling a signal always results in some kind of aliasing, you just need to be sure to keep the aliased signals below the noise floor of the converter. Besides, it's good practice to only use the bandwidth you really need to keep noise low, so limiting the bandwidth of the input signal is the right thing to do.

If you're sampling at 250kSPS, and your pressure signal is a fairly low frequency, then a simple RC filter may be all you need, or you can use FilterPro to design a second-order active filter at whatever your maximum signal frequency might be (100Hz, 500Hz?).