Body
Hello,

I am working on a power electronics project where I need to mix a 12V DC signal with a 1V AC signal and then amplify the output to achieve 20A while maintaining a bandwidth of 100 MHz. My current approach involves using operational amplifiers, but I am unsure about the best way to design the circuit to handle such high current without distorting the signal.

Current Setup:

• Input Signals: 12V DC and 1V AC mixed using an operational amplifier.
• Goal: Amplify the mixed signal to 20A output without distortion.
• Bandwidth Requirement: 100 MHz.

Final Application:

• In the final application, I will be injecting noise into the signal with a bandwidth up to 100 MHz.

Challenges:

1. Finding a suitable circuit design using operational amplifiers that can handle 20A output.
2. Avoiding the use of transistors due to potential oscillation issues.
3. Ensuring stability and minimizing noise in the amplified signal.

Questions:

1. What is the best circuit design to achieve a 20A output using operational amplifiers without distorting the signal?
2. Are there specific components or configurations recommended for such high current and bandwidth?
3. How can I ensure the stability of the circuit and minimize noise?

I have considered using multiple operational amplifiers in parallel, but I am concerned about the potential for voltage division and imbalance in current sharing. Any advice or suggestions would be greatly appreciated. Please feel free to contact me if you have any solutions or need more details.

 

You are asking for two opposing parameters, high current and high frequency. This will require an RF power amplifier and an RF radio transmitter licence in order to operate.

 

You are asking for two opposing parameters, high current and high frequency. This will require an RF power amplifier and an RF radio transmitter licence in order to operate.

Thank you for your response, MrChips.

To clarify, I have a noise frequency range that goes up to 80 MHz which I will add to my 12V DC signal. I am aware that reconciling high current with high bandwidth is challenging, and that is precisely the goal of my project.

The board on which I inject the 12V requires 20 amperes to function correctly, which is why I am looking for circuit designs with operational amplifiers to boost the power without distorting the signal.

Do you have any suggestions on specific approaches or circuit configurations that could help me achieve this performance?

 

1. Finding a suitable circuit design using operational amplifiers that can handle 20A output.
2. Avoiding the use of transistors due to potential oscillation issues.

How do you expect to get 20A without using transistors?
No op amp can deliver anywhere near that amount of current and few can operate at 100MHz.

Oscillations issues have nothing to do with with whether the circuit has transistors or not.

 

Welcome to AAC.

At that sort of current and frequency you'll be hard pushed to avoid nasty feedback effects whatever technology approach you choose. But I am not clear that your problem statement is correct. As I understand it, the mysterious 'board' you refer to takes 12v @ 20A to operate and you wish to inject wide-band noise into this power rail to, presumably, see the effects on said board. And the noise level is going to be around 1v ( peak, rms, or ???). What is this board & what makes you think it will be impacted? Any reasonably designed board will have bulk input capacitance and decoupling capacitors on all active parts, plus some fairly chunky power tracks which will exhibit a reasonable amount of inductance. I'd hazard a guess the bulk of your wide-band noise wont make it past the power connector and its effects will be negligible. So my starter for 10 would be to look at the wide-band admittance of the power rail to see what it might look like to a wide-band noise signal...

 

Amplify the mixed signal to 20A output without distortion.

To amplify a 100MHz signal with no distortion is practically impossible.
Why the 20A requirement? What load will be driven by this magic amplifier?

 

Every day people request help on a project but will not tell us what they are doing. So we try to help but can't because we cannot see the entire problem.

Is the bandwidth of the "signal" DC to 100mhz? Going to DC makes a big difference.
Is the signal 1V AC, p-p? rms? average?
I think the signal is a sign wave that can be any frequency form DC to 100mhz. (?) Not random noise.
I think you want a 12V 20A supply where you can add 1V of noise onto it. This type of test is done to amplifiers to measure Power Supply Rejection Ratio.

I worked on a project 10 years ago where we added 15V of signal to 120/200 Vac power line. I think the current was less than 20A. (15A) The bandwidth was DC to 100khz.

Several of us have made 100mhz amplifiers that do more far than 20A but they are fixed frequency. Not wide bandwidth. Maybe they could do 10:1 bandwidth.

There will be transistors.

 

How do you expect to get 20A without using transistors?
No op amp can deliver anywhere near that amount of current and few can operate at 100MHz.

Oscillations issues have nothing to do with with whether the circuit has transistors or not.

Yes, I understand that a single op-amp cannot deliver 20A. That's why I'm exploring the feasibility of using multiple op-amps in parallel, each generating 1A, to achieve the desired current.

The oscillations are not due to the use of transistors per se but are caused by the capacitive load at the input of the board where I inject my signal after the power stage.

Do you think it is possible to design a circuit with multiple op-amps in parallel to collectively provide 20A?

 

I can't help but wonder about these types of questions. They strike me as questions that reflect some very specific perceived need. I recall a young lad I tutored years ago for a few weeks, he had learned something about electronics from the internet and his approach to learning was much more like looking up specific questions and finding answers.

He'd ask me a question and I'd begin to talk about the subject, how things are related to one another, but he'd get impatient and just want the specific answer to his specific question, he was too rushed in his mind to want to learn how to answer it himself in a general sense.

This is IMHO a real problem with the web, young minds use it to get answers rather than learn how to calculate an answer themselves. So this requirement about 20A might well stem from some other thing where the 20A is mentioned but is never intended or expected to be used in the way the questioner thinks.

 

Yes, I understand that a single op-amp cannot deliver 20A. That's why I'm exploring the feasibility of using multiple op-amps in parallel, each generating 1A, to achieve the desired current.

The oscillations are not due to the use of transistors per se but are caused by the capacitive load at the input of the board where I inject my signal after the power stage.

Do you think it is possible to design a circuit with multiple op-amps in parallel to collectively provide 20A?

Yes but whether that design would generate more problems than it solves, that's the important question here. If you connected 20 op amps together in parallel I'd expect issues, what are the chances that every op amp has exactly the same output voltage at all times? what will happen if they do not?

It's like asking if one could solder a hundred LEDs together to make a bright light, yes one could but its better to just get an already made bright bulb!

Some of the people here are very very experienced, and their questions are relevant, the question you are asking suggests some assumptions have been made that they can see are bad assumptions.

 

Welcome to AAC.

At that sort of current and frequency you'll be hard pushed to avoid nasty feedback effects whatever technology approach you choose. But I am not clear that your problem statement is correct. As I understand it, the mysterious 'board' you refer to takes 12v @ 20A to operate and you wish to inject wide-band noise into this power rail to, presumably, see the effects on said board. And the noise level is going to be around 1v ( peak, rms, or ???). What is this board & what makes you think it will be impacted? Any reasonably designed board will have bulk input capacitance and decoupling capacitors on all active parts, plus some fairly chunky power tracks which will exhibit a reasonable amount of inductance. I'd hazard a guess the bulk of your wide-band noise wont make it past the power connector and its effects will be negligible. So my starter for 10 would be to look at the wide-band admittance of the power rail to see what it might look like to a wide-band noise signal...

I am really glad to be able to exchange ideas with you and others. I didn't expect to receive feedback so quickly on my call for help. I'll try to explain my situation better.

No, my noise level is on the order of 10^3 µVrms/√Hz. Regarding the oscillations, I am basing my concerns on an application note [ https://www.analog.com/en/resources/app-notes/an-159.html (Figure 13. Driver Board Sums AC and DC Voltages to Provide Several Amps at Frequencies to 10MHz.) quote : (One major caveat needs to be noted when connecting this circuit to the DUT: input capacitance for the regulator should not be used. The first reason is that the circuit is not optimized for driving capacitive loads and may oscillate.) my goal is to do the same and grow ) ] and the advice from my electronics tutors. The specific case I am working on is very sensitive, and what might be negligible in other scenarios is quite significant for this particular board.

I didn't fully understand what you meant by, "So my starter for 10 would be to look at the wide-band admittance of the power rail to see what it might look like to a wide-band noise signal..."

I apologize if I am not being clear enough; I am doing my best. Electronics is fascinating, but I am still at the beginning, and the topic I am working on is quite complex from my point of view.

Thank you again for your assistance!

 

What is the "board" you refer to, the one that at 12v input consumes 20A? What is it, what does it do? can we read about it? does it have a specification we can look at?

 

To amplify a 100MHz signal with no distortion is practically impossible.
Why the 20A requirement? What load will be driven by this magic amplifier?

Thank you for your remark.

I take note of the statement "To amplify a 100MHz signal with no distortion is practically impossible." My objective is to try and complete this project, so even though I know a "magic" amplifier does not exist, I am wondering if using multiple amplifiers in parallel, each providing 1A, could work.

Regarding the 20A requirement, these are the specifications I have been given to try and meet. I understand it is a challenging task, but I am looking for any possible solutions or approaches to make it work, even if it is practically impossible.

Thank you for your help!

 

Every day people request help on a project but will not tell us what they are doing. So we try to help but can't because we cannot see the entire problem.

Is the bandwidth of the "signal" DC to 100mhz? Going to DC makes a big difference.
Is the signal 1V AC, p-p? rms? average?
I think the signal is a sign wave that can be any frequency form DC to 100mhz. (?) Not random noise.
I think you want a 12V 20A supply where you can add 1V of noise onto it. This type of test is done to amplifiers to measure Power Supply Rejection Ratio.

I worked on a project 10 years ago where we added 15V of signal to 120/200 Vac power line. I think the current was less than 20A. (15A) The bandwidth was DC to 100khz.

Several of us have made 100mhz amplifiers that do more far than 20A but they are fixed frequency. Not wide bandwidth. Maybe they could do 10:1 bandwidth.

There will be transistors.

Thank you for your understanding and for trying to help despite not having all the details. I appreciate your patience.

Unfortunately, I am limited in how much I can share about the specifics of the board I am working on, which is why my initial question focused on the feasibility of using multiple amplifiers in parallel to increase current without creating a voltage divider effect at the output.

Here is what I can share:

I am working on a project where I need to add white noise with a frequency range up to 100 MHz and a level around 10^3 µVrms/√Hz to a 12V power supply.
The goal is to achieve a 12V output with 20A while maintaining this noise profile.
Currently, I am using a sinusoidal signal for testing purposes because I am not yet a specialist in noise generation.
I am also trying to meet or lower the specification requirements, but achieving 20A output would be ideal.
I don't have a particular issue with transistors, but I have been asked to avoid them. An application note I read advised against their use for this specific application. They included a design similar to what I am aiming for but with less power. Hence, I am exploring the hypothesis of using multiple amplifiers in parallel, but I am open to any schematic suggestions that I could simulate.
Thank you again for your assistance!

 

Very well, use a vacuum tube to perform the amplification. Your ball parking around 12v * 20A = 240W.

Here you go, 250W at 500 MHz 4CX250B/7203

 

I can't help but wonder about these types of questions. They strike me as questions that reflect some very specific perceived need. I recall a young lad I tutored years ago for a few weeks, he had learned something about electronics from the internet and his approach to learning was much more like looking up specific questions and finding answers.

He'd ask me a question and I'd begin to talk about the subject, how things are related to one another, but he'd get impatient and just want the specific answer to his specific question, he was too rushed in his mind to want to learn how to answer it himself in a general sense.

This is IMHO a real problem with the web, young minds use it to get answers rather than learn how to calculate an answer themselves. So this requirement about 20A might well stem from some other thing where the 20A is mentioned but is never intended or expected to be used in the way the questioner thinks.

I understand your reasoning, the perspective. I have been working on this project for several months, trying to understand as much as possible to defend my thesis. The specifications I mentioned are the ones I need to meet. I have been stuck on this point for many weeks, which is why I reached out for help. The people supervising me recognize that the project is quite complex, at least from their perspective and mine, and they are limited in what they can teach me.

I am trying to determine if such a power stage exists or if there are any leads I can follow. Answering these questions forces me to face my understanding of my subject, which is not yet perfect as I am still learning. I am very motivated to see this project through because it is genuinely fascinating.

If you have time to guide me in the directions I can explore, it would be really kind of you.

Thank you!

 

Yes but whether that design would generate more problems than it solves, that's the important question here. If you connected 20 op amps together in parallel I'd expect issues, what are the chances that every op amp has exactly the same output voltage at all times? what will happen if they do not?

It's like asking if one could solder a hundred LEDs together to make a bright light, yes one could but its better to just get an already made bright bulb!

Some of the people here are very very experienced, and their questions are relevant, the question you are asking suggests some assumptions have been made that they can see are bad assumptions.

"what are the chances that every op amp has exactly the same output voltage at all times? what will happen if they do not?"

That's exactly what I'm trying to understand. I'm learning electronics, which is why I'm opening this discussion with those who have the knowledge. I realize it might seem like a bad assumption, which is why I'm trying to be as clear as possible with what I've understood so far.

If I could find a component that acts like an already-made bright bulb, I would gladly use it. However, given my specifications, it's not very easy to find such a solution. I'm here to learn and explore possible avenues to meet my project requirements.

 

If I could find a component that acts like an already-made bright bulb, I would gladly use it. However, given my specifications, it's not very easy to find such a solution. I'm here to learn and explore possible avenues to meet my project requirements.

See my earlier post: https://forum.allaboutcircuits.com/...ut-distorting-the-signal.201654/#post-1919293

 

Hi Vincent,

What level is this, undergrad, masters, PhD? I know how it feels to suddenly realize that you're out of your depth because what you need to know is peripheral to your research question but fundamental to the experiment you want to perform...

What can you share with us about this mysterious board? Understand it might be sensitive but we need something to go on...

 

Very well, use a vacuum tube to perform the amplification. Your ball parking around 12v * 20A = 240W.

Here you go, 250W at 500 MHz 4CX250B/7203

Gosh, I haven't used one of those since 1975! Used to have a pair in push-pull at 144MHz pumping out about 400W...