Hello Everyone,
I was presented with a design challenge several months ago. I thought I had a pretty good solution in mind but things are not working out as I originally imagined. I was asked for an embedded solution that would provide 20 bipolar square waves. The user needs to control amplitude and frequency. The amplitude must be maintained in the range of +/- 1V in 0.1V steps while the frequency is in the range of 0-1Hz in 0.1 Hz step. I setup a PSOC3 to generate 20 50% duty cycle PWM on independent pins no problem. Then I imagined that I would use clock division to provide frequency stepping (I believe I can live with the bit/divider resolutions for now). So far so good. Now what I had imagined was that I would use a differential OP AMP configuration http://en.wikipedia.org/wiki/Operational_amplifier_applications. I have been experimenting with a TLC277CP http://www.alldatasheet.com/datasheet-pdf/pdf/28882/TI/TLC277CP.htmlfor this concept with a constant +5V as an input (V1 PIN2) , the PWM as (V2 PIN3), and -5V as ground (PIN4 Only) and +5V (Vdd PIN5) . Taking R1, R2, Rf, and Rg with 1:2 ratios (taken as 1K and 2K in my circuit) to get a +/- 2.5 V output. The idea was to step this down further before feeding into a digital 1024 step (10bit) variable resistor made by Maxim (MAX5483 with SPI interface or perhaps the up/down interface). For now I have been using a function generator (unipolar square wave) to feed into my differential op amp circuit for testing. I can not get the bipolar square wave out of the circuit for the life of me. (Checking I/O on an oscope). Thoughts on what I may be doing wrong would be highly appreciated. My friend thinks that it may be related to differential signals vs. single ended but I honestly do not know. I made a simulation based on the equation provided by the Wikipedia page and it seems like it should work (please see attached plot). My sense is that there is something in the hardware or implementation that I do not fully understand.
I was presented with a design challenge several months ago. I thought I had a pretty good solution in mind but things are not working out as I originally imagined. I was asked for an embedded solution that would provide 20 bipolar square waves. The user needs to control amplitude and frequency. The amplitude must be maintained in the range of +/- 1V in 0.1V steps while the frequency is in the range of 0-1Hz in 0.1 Hz step. I setup a PSOC3 to generate 20 50% duty cycle PWM on independent pins no problem. Then I imagined that I would use clock division to provide frequency stepping (I believe I can live with the bit/divider resolutions for now). So far so good. Now what I had imagined was that I would use a differential OP AMP configuration http://en.wikipedia.org/wiki/Operational_amplifier_applications. I have been experimenting with a TLC277CP http://www.alldatasheet.com/datasheet-pdf/pdf/28882/TI/TLC277CP.htmlfor this concept with a constant +5V as an input (V1 PIN2) , the PWM as (V2 PIN3), and -5V as ground (PIN4 Only) and +5V (Vdd PIN5) . Taking R1, R2, Rf, and Rg with 1:2 ratios (taken as 1K and 2K in my circuit) to get a +/- 2.5 V output. The idea was to step this down further before feeding into a digital 1024 step (10bit) variable resistor made by Maxim (MAX5483 with SPI interface or perhaps the up/down interface). For now I have been using a function generator (unipolar square wave) to feed into my differential op amp circuit for testing. I can not get the bipolar square wave out of the circuit for the life of me. (Checking I/O on an oscope). Thoughts on what I may be doing wrong would be highly appreciated. My friend thinks that it may be related to differential signals vs. single ended but I honestly do not know. I made a simulation based on the equation provided by the Wikipedia page and it seems like it should work (please see attached plot). My sense is that there is something in the hardware or implementation that I do not fully understand.
Attachments
-
6.5 KB Views: 21