Hello everyone,

I'm currently working on a project where I use a TI DSP to sample a ultrasound signal. Sadly, I have envountered a problem. The pulser I'm using sends out a pulse with an amplitude of 30V, while the input of the DSP only takes 3.3V. My question is therefor: How can I reduce the amplitude of the signal without changing the sinusoidal morphology of it? It might be worth mentioning that the frequency of the signal is 5MHz.

I've read that transformers, transisors, step down converters, operational amplifiers and simple voltage dividers can be used. Is one of these a good option or can anyone recommend a suitable method?

Thanks in advance!

 

I can not see why a resistor divider should not work in this case. But put a suited OPAMP between the divider and the ACD input. But another problem. Can you DSP sample with at least 10MHz samplerate?

 

A voltage divider using two resistors would be the simplest solution.
You may want to add a couple of diodes to clamp the voltage between Vcc and GND so that you don't blow your chip.

 

@t09afre
I use the DSP TMS320F28335 from Texas Instruments which can sample in 12.5 MHz, hence, over the nyquist frequence. Why should I use an amplifer after the resistor divider?

@MrChips
Do you have any schematic for such a circuit, with the diods that is.

 

 

5MHz/30V → resistor divider (3V) → Op Amp Voltage follower(3V) → 3.3V clamp → ADC

Op Amp Voltage follower(3V) - for isolation.

 

So, I should first use a resistor divider to get 3V over one of the resistors used (ex. 27kOhm and 3kOhm). Then I should isolate the circuit with the help of a OP amp follower. Last, I should use a 3.3V Zener diode and a resistor as a clamping circuit, which gives me the output over the Zener diode. Have I understod this correctly?

 

Seems right from here.

 

WHy do you need to preserve the sine wave? Are you going to digitize the transmitted signal? I would think that you would be digitizing the return signal, which will be low amplitude.

 

I need to preserve the waveform because the knowledge about the frequency components can help me find possible echoes.

Yes, both the transmitted and received signals will be digitized. This is done mainly because the pulser/receiver also "receives" the transmitted signal, and because I need to calculate the distance between both of them.

 

When I simulate this with pspice (OrCAD) I get a phase shift at the out signal. How can I avoid this?

 

A bigger problem might be this. Looking at your chip. It has only ADC unit capable of 12.5 MSPS. If you are sampling two ADC channals. These two has to share 12.5 MSPS between each other and end up with a half the sample rate.

 

Did you simulate as I mentioned?
You can separate each stage and to see the waveform, then you will find that where is the problem.

 

When I simulate this with pspice (OrCAD) I get a phase shift at the out signal. How can I avoid this?

Why is phase shift a problem?

 

@t06afre
I'm only using one analog input so that will not be a problem. But thanks for the information.

@ScottWang
Yes, I simulated as u mentioned and came to the conclusion that the zener diode was the problem. But now I'm thinking about skipping the diode because I'm already restricted by the feeding voltage of the op amp. If I feed it with 3.3v, the output will not exceed 3.3v. right?

@Ron H
If the signal is shifted then I need to sample more points making my sample array in my program much larger. This will lead to longer execution time.

I'm also wondering which op amp I should choose in othe words can all op amps handle 5MHz signals. Maybe LM6132 a good choice?

 

MSP430 Vcc=2.5V~5.5V, you can check your DSP.

If your DSP as MSP430 then you can choose a 3.6V Zener or a little higher.
If you want to get a 3.3V from Op Amp, then you better offer the Vcc=5V, because the output of Op Amp will reduce about 1.4V from Vcc, you can find out the infos from the datasheet of Op Amp.

 

MSP430 Vcc=2.5V~5.5V, you can check your DSP.

If your DSP as MSP430 then you can choose a 3.6V Zener or a little higher.
If you want to get a 3.3V from Op Amp, then you better offer the Vcc=5V, because the output of Op Amp will reduce about 1.4V from Vcc, you can find out the infos from the datasheet of Op Amp.

Scott, have you heard of rail-to-rail op amps?

 

Ron H:
Scott, have you heard of rail-to-rail op amps?

What do you mean, as this one?

LT1498 - 10MHz, 6V/μs, Dual Rail-to-Rail Input and Output Precision C-Load Op Amps

LT1498 datasheet.

 

Thanks for the answers.

Now I've encountered another problem. After the simulation I started connecting the different parts, one at a time, starting with the resistor divider. When I connected the oscilloscope over one of the resistors I got a good signal with lower amplitude, which is good. But when I connected the the DSP over the resistors the signal became distorted.Is it the input impedance that has something to do with it? If so, how should I modify my simple resistor divider so that I get a clean signal in to my dsp?

 

Thanks for the answers.

Now I've encountered another problem. After the simulation I started connecting the different parts, one at a time, starting with the resistor divider. When I connected the oscilloscope over one of the resistors I got a good signal with lower amplitude, which is good. But when I connected the the DSP over the resistors the signal became distorted.Is it the input impedance that has something to do with it? If so, how should I modify my simple resistor divider so that I get a clean signal in to my dsp?

What are your resistor values?