OPA857EVM-978: Setting the Module Into Test Mode and Conducting Frequency Tests

Hi James,

Sorry we missed your first post. Please close it and use this one for further questions.

Use the EVM user's guide in conjunction with the datasheet to help evaluate how this part works.

You have the right setup. Is the scope AC coupled? Did you try an AC signal on the Test_In pin? If you think the part may be damaged, you can put on a new one too.

Best regards,

Sean

Hi Sean,

I have looked and looked and quadruple checked the userguide/datasheet for both the OPA857 module AND the OPA857.

Here are the specifications I have set to determine the VIN1 and the VIN2:

VCC (banana plug) = 3.3V

Test_In = 2.18 V (DC only)

OUT = 1.83V

OUTN = 1.83V

VOUTN = 0V

VOUTP = 0V

VCC (banana plug) = 3.3V

Test_In = 1.77 V (DC only)

OUT = 1.33V

OUTN = 1.83V

VOUTN = 0V

VOUTP = 0V

Calcuations (as per the OPA857 Module User Guide):

VIN1 - VIN2 = 0.41V

(VIN1+VIN2)/2 = 1.975V

Now, I set the DC bias to 1.98V and the Vpp to 0.41 and set up the circuit as specified by Figure 3 in OPA857 Module User Guide, connected the DC bias using a breadboard and connecting a sine signal at 0.41Vpp. I then verified that the signal was appropriate and could see a 0.41Vpp sine signal on a 1.98V bias via an oscilloscope. Taking the output of the Vpp = 0.41 with DC bias at 1.98V and connecting this to the Test_In SMA connector. Now, I took the VOUTN SMA output and connected to an oscilloscope.

There is ZERO output that's readable, just a static line at 0V (the oscilloscope is in AC coupled mode). Do I need to take the differential output of BOTH VOUTN and VOUTP and connect to the oscilloscope to see the signal when in Test Mode?

There isn't any clear indication on any user guide or datasheet on measuring the output signal.

Regards,

James

Hi James,

Looking at Figure 16 of the Module User guide, can you check to see if there is a signal at OUT or OUTN? If there is not, check if either output is grounded. If there is, what frequency is your sine signal? It should be at least 100kHz to get through the transformer T1. Also note the resistor divider formed by R8/(R6+R7). It is there to give a 50 ohm source impedance from the amplifier while requiring a lower output current from the amplifier. However, in many applications you should just replace R6 and R7 with 25 ohms and remove R8.

-Sean

Hi Sean,

My previous reply states that I get voltages from the OUT and OUTN pin directly from the amplifier.

The issue is that there is NO output from the SMA outputs which are connected to the OUT and OUTN. I've tried multiple sine frequencies which are all above 100kHz and there is no visible output signal from the SMA outputs.

I'm expecting a nice output similar to the input which I've checked. However, the output is flat at 0 Volts.

Why would I need to adjust the resistors on the board? I thought the point of the evaluation module was to test out the OPA857 straight from the box?

James

Hi James,

Can you measure a voltage at both of the inputs to the transformer? Not just the output of the amplifier? If there is a voltage there, is there still no voltage at either output of the transformer? Are you taking into account the 56.2/(2*487) attenuation caused by the resistor divider? (This is there to minimize amplifier load, while still providing 50 ohms impedance matching for high frequency measurement. This makes the amplifier look good, but is not practical for use as-is for amplification.)

The default configuration of this EVM is to convert the differential output of the amplifier voltage into a single ended voltage on VOUTP. VOUTN is tied to ground by R14, so the transformer pushes all the voltage into VOUTP. Some people will remove R14 to get a truly differential voltage between the two outputs to drive a differential input ADC. I do not have the board so it is difficult for me to determine the issue. At what node does the signal disappear?

-Sean

Hi Sean,

Sorry for the delay in my replies I have been writing my thesis.

My Specifications of the Circuit:

VCC = 3.7 V

DC Bias = 1.98 V

AC Signal = 100 mVpp at 1 MHz

J4 = OPEN CIRCUIT

J5 = TEST SETTING

J6 = 5k SETTING

J7 = SHORT CIRCUIT

The DC and AC signals are coupled as per the figure 3 in the datasheet for the EVM for OPA857 (http://www.ti.com/lit/ug/tidubx7/tidubx7.pdf)

I tested the outputs of the transformer and are summarised below:

Transformer Output 1: Measured at R14

Fig 1: Here we can see the input signal (in blue) and the transformer signal from the R14 output (in yellow)

Fig 2: The transformer output signal is scaled and measured to have approx 11.6 mVpp - quite noisy signal

Transformer Output 2: Measured at R15

Fig 3: Here we can see the input signal (in blue) and the transformer signal from the R15 output (in yellow)

Fig 4: Now, our output signal at R15 in slightly lower with a 8.6 mVpp and equally noisy output.

I have attached a picture of the circuit I am using, please advise if there are any issues with the jumpers or what not.

Fig 5: The Circuit being used

Questions:

1) Shouldn't the signal be amplified by some factor (or at least remain the same) for a 1 MHz signal?

2) What is the purpose of this transformer in the circuit?

3) I am measuring everything relative to the GND signal, is this correct?

4) Should I measure the signal from the transformer with a 50 Ohm impedance on the oscilloscope, when I do this - there is an even smaller signal (barely noticeable).

What node is the blue waveform coming from? You have to find where the signal chain is broken.

The blue waveform is the input signal at the TEST_IN node.

Hello James,

You are correct in measuring your signals relative to ground. The purpose of the transformer is to convert your psuedo-differential signal to a single-ended signal that can be measured from J3 alone. Based off your previous responses, it does seem that the OPA857 is functioning correctly with the OUT signal varying based off the TEST_IN signal you are applying. Could you please confirm that your are getting a varied voltage at the inputs of the transformer as well? I've highlighted the nodes below:

It seems that the transformer may be the issue here. Could you also measure the outputs straight from the transformer (pins 4 and 6 of transformer)? 

Best,

Hasan Babiker

Hello Hansan,

I will check the transformer nodes tomorrow and let you know the signal(s) I get.

I don't understand how you can say the OPA857 is functioning correctly when there is a clear attenuation of my input signal (blue) where it comes out as a noisy and barely noticeable signal (yellow). Blue is the INPUT at TEST_IN input to the OPA857, and yellow is the OUTPUT at the J3 and J2 - how is this an amplifier when the signal isn't amplified but attenuated instead?  

James

Hey James,

My assumption of the amplifiers functionality was not based off your scope shots but from your measurements in an earlier post where you showed that the OUT signal from the amplifier varied between 1.83V and 1.33V.

Best,

Hasan Babiker

Hello James,

Were you able to make measurments of the transformer nodes?

Best,

Hasan Babiker

Thanks for following up, I've been on holiday.

I took the measurements the next day, they showed no output at those nodes. I also too measurements directly from the transistor with DC coupling on the oscilloscope and found no signal.

At this point I've basically given up getting this to work and went on a holiday because otherwise I would have lost my mind.

I also tried using the device OUT OF TEST MODE and with a photodiode.

The photodiode showed no output, and there is a 3.3V voltage across the IN pin and ground, probably coming from the supply. I don't believe that is suppose to happen.

Hey James,

Hope you are enjoying your holiday.

Did you mean the amplifier or transformer when you mentioned the transistor? Are there any nodes on the EVM where you can see an AC signal? Are you getting anything out of the outputs of the amplifier itself?

Best,

Hasan Babiker

Hey James,

Just wondering if there were any updates on this topic? Were you able to get the EVM working?

Best,

Hasan Babiker

Hey James,

Glad you were able to work out were the source of the issue came from. If you do find out that the EVM was in fact part of the problem in the future, feel free to open up another thread and we can certainly look into again.

Best,

Hasan Babiker