Both product approaches have been very successful in the marketplace. Below is a more detailed analysis of success drivers:
1. Texas Instruments’ AFE44xx family is suited to the pulse oximeter application, which inevitably results with an optimized and compact solution. Integrated additional features (like LED driver and fault diagnostics) make the AFE44xx an “all-in-one” solution for that photometric application. TI’s approach, however, also makes the AFE44xx family very application-specific. The drawback is that AFE44xx is not suitable or could be redundant for many other sensors and applications. On the other hand, Systemcom’s AFE family is very modular and therefore far more versatile and applicable to a wide range of current output sensors and applications.
2. Unlike Systemcom’s SC-I-AFE-180F110, TI’s AFE44xx features a fully-differential transimpendance amplifier (TIA) which eliminates the need for a single-ended to differential converter. This provides better overall simplicity in design but probably at the expense of linearity. The fully-differential TIA inherently biases the photodiode to the voltage of 0V which increases the photodiode’s capacitance. SC-I-AFE-180F110 is designed to bias the photodiode with higher reverse voltage, thus reducing its capacitance and consequently its response time. Such a sensor biasing solution makes SC-I-AFE-180F110 suitable for interfacing other current output sensors, e.g., electrochemical sensors.
3. The four-stage-PGA in SC-I-AFE-180F110 offers a gain of up to 1296 and a transresistance of 64kΩ by TIA. Such a combination of gain and transresistance allows significantly smaller resistance values in TIA and consequentially faster response time. The AFE44xx family provides 1MΩ transresistance by TIA, combined with a gain up to 4 in a single-stage-PGA. The PGA used in the SC-I-AFE-180F110 is based on the switched capacitor architecture, which in general exhibits lower noise performance when compared to solutions based on resistive amplifiers. Both products use the averaging technique to reduce noise and improve SNR.
4. AFE44xx provides limited ambient (light) cancellation (up to 10uA), which involves a procedure run by a microcontroller and DAC to perform this cancellation. Such a solution requires more time to complete and suffers from residual offset (because of 10 discrete steps of DAC). The relative measurement in SC-I-AFE-180F110 offers ambient light suppression performed by the analog circuitry itself regardless of the amount of ambient light. In other words, this allows high gain amplification of the desired signal relative to the ambient light. Furthermore, the ambient light suppression doesn’t rely on the procedure run by a microcontroller or digital post-processing, so it is therefore simpler and faster. The amount of ambient light to be suppressed depends only on the selected current range (up to 1mA) and is not performed by DAC; this reduces the residual offset.
5. The pulse oximeter application allows small low-gain amplifier and ambient light suppression to be done by a microcontroller instead of the analog circuitry; this ultimately simplifies the design and reduces power consumption. In the end, it all comes down to the requirement for specific application and the sensor used. Versatility means a wide spectrum of sensors and possible applications but it also requires an effort to implement it for a specific application. In contrast, a dedicated application means a more optimal, off-the-shelf solution.
Temperature range of AFE4490 is -40 to +85C and the lower cost AFE4400 is 0 – 70C.
AFE4490 programmable gain range is from 10k to 1000k in seven steps for the TIA and a further gain of 1 to 4 in five steps. Furthermore, the TIA has programmable feedback capacitors (5 to 250pF in six steps), which allow for tailoring of the bandwidth.
The ADC in AFE4490 is 22bits versus 13bit ADC for SC-I-AFE-180F110 (that is what the product brief says. Pity there is no full datasheet on-line to enable a full comparison).
AFE4490 splits the analogue and digital supplies to allow interface to a range of micro I/O voltages over the SPI. The SC-I-AFE-180F110 states that it has a single supply requirement of 1.6 to 1.8V. How to interface that? Probably need level translators.
Broad application ICs can end up being a jack of all trades and a master of none. AFE44xx is primarily for pulse oximeter, but is successfully used in other applications where the integration of both the Tx for the LED and the Rx photodiode portions make sense. For interfacing to other types of sensors then different AFEs are required; LMP91000 / 2, LMP91050 / 1, LMP91200, LMP90100 etc.
This whole piece looks like an advertisement for this little unknown start-up company from Croatia packaged as a product "evaluation" by an unknown consulting company. If anyone believes that they can compete with TI in anything he/she needs his/her head examined. Why would EETimes lend its name and credibility to something like this is a mystery.
hmmm - I found it useful and interesting.
Are all four of you above working for TI? TI is a GREAT company - we agree but -- do I detect a pro-TI bias?
I actually checked on "little un-known start-up company" - it has been in business for 20+ years and its customers included Intel, Digital, Bosch, etc....
@sranje, we are not talking about some trivial digital stuff this is serious analog, you cannot beat TI or ADI for performance, These folks have invested plenty to get to this level, be it proprietary process development or design...
Well, neither TI and ADI apparently agree with you -- they both see the threat from fast growing analog specialists in China, for example...
One can also look into who is that "unknown" Petrov Group.... In 2012 they published a seminal 400+ pages strategy analysis of TI - this time post-National acquisition. They posted extensive in-depth excerpts from that report on their Web page.
They have also published numerous strategy reports on Linear, Maxim, National, Intersil, TI, etc.
Prove please and what this upcoming companies doses that TI or ADI or other well established companies cannot do if they found its important and will make money. They can incorporate the many of the best system in their one products if needed but not do so because of specific needed (referring to Systemcom solution to TI).
If we are talking about treat no need to mention those Chinese specialist even a number of US small analog companies alone can become treat to the giants in specific applications. Thanks.
There are several other dimmensions of the product that was not considered, what makes this article, more marcom than technical. Doing the product focused to specific applications, it is a strong trend in analog. Why offer more than the application requires ????? That´s why TI is at US$ 14 billion level and this startup will take some time to double the current revenue.
I am intrigued by the article and by the discussion it has stirred up.
I obtained the datasheet for Systemcom’s product (at its web, through registration). Article’s point was that there are approaches where there are opportunities for small companies in competition with very capable and competent giants….
I noticed that AFE44xx with the gain of 4 and 1Mohm gain of TIA gives 4Mohm in total. SC-I-AFE-180F110 on the other hand provides 64kohm (in TIA) and 1296 (PGA) which gives 80Mohms in total. It means bigger and, as it seems, more programmable gain for SC-I-AFE-180F110, giving quite adaptable solution, while TI's is sharply focused to the specific application.
In summary, one would highly welcome both approaches from the customers' point of view -- it offers them vendor choices and flexibility. Also, note, the SC-I-AFE-180F110 datasheet states that the chip I/O accepts 3.3V logic levels.