THS4532: Reference design of converting single-end 2V span signal to +- 1V pp differential signal.

Hi Nick 

Thank you for the reply.

the diagram illustrated 1V peak sine wave input Vin with Vcm=1.5,  but our analog signal is 0 to 2V single-end Vin with Vcm=0.5V (actual Vin is 0.5V-2.5V).

How to shift our Vin to get Vout_diff= +-1V pp?  Should I set Vin- = 1.5V? 

Currently, I set Vin- = 0.5V,  then I always get Vout_diff > 0V with Vcm=1.5V.  

Thank you

Tony

Hello Tony,

Our team will get back to you.  As I understand it, you are looking for +/- 1Vpp Vout, centered around 0V (GND)?

Best,

Alec

Hi Alec

Yes, Vout_dff = +/- 1Vpp is required by ADC (LTC2295),  its datasheet indicates that as below.

Currently, I have 500mV reference connected to Vin- and analog signal (0.5V to 2.5V) connected to Vin+,  then I can only get ADC reading between #2000 to #3FFF (corresponding Vout_diff = 0V to 1V and analog signal 0.5V to 1.5V).

What I need is when analog signal = 0.5V, I can read ADC #0000, when analog signal = 2.5V, I can read ADC #3FFF.

this is current schematic.

Thank you

Tony

Hi Tony,

Thank you for your clarification. When using a single-supply voltage operation, the output common-mode voltage can not be set to 0V (GND) and therefore, the differential output signal can not be centered around ground. When operating in a 3V single-supply voltage, the output common-mode voltage range is approximately 0.75V to 2.2V typical. Are you able to have a dual-supply voltage in this application?

Thanks,

Nick

Hi Nick

The output common mode voltages were not set to GND.  In fact, Vocm1 and Vocm2 are connected to VCMA=VCMB =1.5V reference voltage.

In the schematic, CS (connected to Vin+) = 0.5V to 2.5V analog single-end input (2V span),  C_500mV (connected to Vin-) = 0.5V single-end reference voltage which has to be substracted from the analog input.

Now, Vin+ always > Vin-,  that is the reason why I can only get reading when single-end input = 0.5 to 1.5V (1V span).

I am thinking of connecting a 1.5V reference voltage to Vin- instead of 0.5V,  but not sure if that will work fine for the 2V span analog input(0.5V to 2.5V).

Would you please help to do a simulation and see what are output of THS4532?   For example, Let Vin = 0.5V to 2.5V,  connect Vin- to 1.5V (prior 1K resistor).  

 Thank you

Tony

Hi Nick

Thank you for the simulation.

Would you please indicate which output (in color) is Vin, which is Vout_diff, which is VF1 and VF2?

It seems pink color signal is Vout_diff,  gray color signal is Vin,  but what I don't understand is why the first simulation Vout_diff is centered at 1.5V, while second simulation Vout_diff is centered at GND? 

Also in your simulation circuit, there is a transformer between VF1-VF2 and Vout_diff. buy we don't have such a transformer in the real circuit (see the schematic).  so in reality, Can ADC get +/- 1V without the transformer? 

Another thing in question is that VF1 and VF2 are both centered at 1.5V in simulation 2, while in simulation 1, VF1 is centered at 0.7V, VF2 is centered at 2.2V.  both simulations have the same Vin = 0.5V to 2.5V,  why so much difference?

Thank you

Tony

Hi Tony,

Apologies for not including a legend in the simulation snapshot that I provided for you. However, you correct on which signals are which.

I included a VCVS in the original simulation so that I could use a single measurement probe in TINA-TI. However, having this does not affect the signal as I set the amplification of the VCVS to 1. Here is the simulation without the VCVS and the signal is still the same.

The reason for the difference in where the +/-1V output differential signal is a result of the reference voltage that is applied on the Vin- signal path. In the first simulation that I sent, the reference voltage was 0V so that output was centered around Vocm, which was set at 1.5V. Adding the reference voltage bias, caused the offset of 1.5V which is why the most recent simulation is centered at 0V. 

This is also why VF1 and VF2 changed from the original simulation to the most recent simulation. Something I did not catch with the first simulation that I sent was, the Vin signal has a DC offset of 1.5V in order to achieve a range of 0.5V to 2.5V. Vout+ and Vout-  will also try to be centered around 1.5V because Vocm was set to that value. Because of the DC offset of the Vin signal being the same as Vocm, it was causing a distorted signal. 

Thanks,

Nick

Hi Nick

Thank you for the latest simulation.

The closer passive component equivalent to VCVS in reality is a transformer, however, we can not use a transformer because our analog signal to be digitized is a DC signal, therefore, the third simulation provides confidence of getting +/- 1V that ADC needed.

Would you please help to do another simulation (no VCVS) with Vin- = 0.5V which is the existing circuit, so we can verify with ADC reading.

Thank you

Tony

Hi Nick

Thank you so much for the simulation that helps a lot.

Best Regards

Tony