Can PSOC4 be applied in thermal couple temperature device? Because of there is only 2 OP. We need to do the sample with differenctial mode.
Due to there is no VDC in PSOC4, so we need to use one IDAC and one OP for reference power. And then the there is no enough OP to amply differencial samples signals.
The SAR ADC has differential mode, however its input buffer has only a gain of 1, gain
control as in PSOC 3/5 is not supported.
If you need more gain, then athe only way to get it would be external IA or external OpAmps
with precision R network to set gain, and its precision (and OpAmp G) control CMRR.
Some reference material attached.
I have a lot more ref material, if you send me email address I can supply.
Thank you for your reply. And your means that the PSOC4 cannot do the application of thermal couple one chip solution?
And I have tried to change the sampling mode from differenctial mode to single mode with AMUX. But the common mode interfering noise is too big and the sample data is not good.
And the customer is interesting our one chip solution because of the low BOM cost and simple circuit. Maybe the application is not the target filed for PSOC4.
Take a look at this ap note -
This was done using single ended measurements, with a DelSig. Using SAR,
in order to get high resolution, you would have to do signal averaging. Which also
would aid CMR performance.
Signal averaging can reduce noise if its uncorrelated. If there is high correlation then
you need a diff amp, which can be done with 2 OpAmps, with an external precision
R network. Or a combination of both.
So I think its possible to do it with PSOC 4, the primary consideration is SAR
resolution and with averaging you should be able to get a reasonable accuracy.
The only drawback to averaging is averaging latency, which in 99% of control type applications
not a problem.
With respect to your CM noise you are experiencing -
1) If power supply related noise take a carefull look at grounding.
2) Use Polymer Tantalums vs standard tants, they have an order of magnitude
better Z vs f performance.
2) Band limit where possible, R-C filters, IIR and FIR filters in code.
Some additional info -
I have tried added a 358(two channel op) for PSOC4 application. But I found the analog bus cannot be routed internal in creator.Can you release your email address to me. And I will send the project configuration to you and help me to route it.
I have sent a email to you. And I attached my project to you. Please help me to route it. Now it seems that there is a conflict when using AMUX and IDAC.
You have 2 OpAmps external, any thought to using 4200 part ?
Right now those OpAmp R networks have to be .1% to keep
common mode in spec (actually you would need to do a more
thorough analysis), and coupled with the cost of the 2 OAs might
offset cost of using 4200. Also 4200 enables the SAR input buffer
which allows G up to 16.
Just a thought.
Still looking at routing, seems like you are meeting all restrictions
and comitted poin issues, not obvious.
Question, when you choose ch2 (of ch0 - ch4) you essentially short the
SAR diff inputs together, is that what you intended ?
I would post a CASE while we look at this (if you have not already done so),
I too am having no luck with routing. I have dropped paths, etc.., still something
“Create a MyCase”
Scratch my comments about using the 42xx part, it does not
have an input buffer on the SAR either.
That being said keep in mind you need good matching of the R networks
on the external OpAmps.
Why PSoC4, because you needed a 32 bit processor to measure temperature?
AN2226 covers thermocouple measurements in PSoC1 with near zero external parts. There are also PSoC3,5 versions with the same technique. The differential measurement is sequential, using correlated double sampling to eliminate offset and 1/f noise.
---- Dennis (Cypress apps)
Hi Dana & Dennis,
Due to PSOC3&5 cost is too high. So the customer cannot accept it. And our competitor is samsung MCU.
And I thought PSOC4 could easy cover the application(two thermal coulple & one segment LCD display).
Now, we found that there was many limited in PSOC4. And we are trying to use CY8C27X43 to do that. But now the sampling speed is too slow. And the AN2226 is too simple and no any compensation.