Hello! I am used to the software and aerospace world, and only have an intermediate level of knowledge in electrical circuitry. I have been creating a few simple circuits to help improve my knowledge here.

The circuit I've been working on is a simple sine wave generator based on a Wien Bridge Oscillator and Op Amp, using a JFET for automatic gain control. This part of the circuit has been working correctly after choosing an Op Amp with a higher bandwidth product and slew rate. At low frequencies it outputs decent voltage, but by 35 kHz it only outputs about 60mV. I believe this is to be expected, so I thought I'd correct it with a non-inverting Op Amp to boost it to my 9V design point.

This non-inverting amplifier portion is where behavior is not quite what I am expecting. I am using an OPA552 because of its high bandwidth and slew rate, as well as decent current output ability. It appears to have very little gain at 35 kHz, however. Looking at the data sheets, I should be getting at least 50 dB of gain in an open-loop configuration. This seems to be enough, but the most I seem to be getting with both closed- and open-loop configurations is only about 23 dB (or G=15) to about 1V. The Gain Bandwidth Product is 8 MHz, so I would assume that I should at least be able to get G=200+ at 35 kHz.

If I pre-amplify the signal to about 1V and then feed it into the OPA552, I hardly get any gain at all. The output signal regardless of the resistor values is barely above the input of 1V. Am I missing something obvious in how Op Amps work? Or do I simply need an even higher speed Op Amp for these frequencies?

I've included a schematic of the non-inverting amplifier I am using. It's no different than the standard ground and feedback resistor design. I have tried resistor values ranging from (R_feedback = 1k, R_g = 6.8) to (R_feedback = 220k, R_g = 1k) and (R_feedback = 10k, R_g = 1k). None of these seem to provide any better behavior. I can post the Wien Bridge circuit if needed as well, but I'm not sure if that part of the circuit should matter. Not included in the schematic are the sets of decoupling capacitors (0.1uF and 10uF) bridging +V, -V to ground.

 

I should be getting at least 50 dB of gain in an open-loop configuration.

you probably don't want to use your opamp open-loop, however. their open loop bandwidth is typically less than 10Hz, and sometimes 1Hz.

So if you are designing a close-loop amplifier, open loop figures are of no use.

there are quite a few ways to compensate for lower swing, especially if you already have a means to control the gain. Alternatively, think of a vga like the 603 - effective and cheap.

or if you are comfortable with software, go dds.

 

Welcome to AAC!

What gain and input voltage aren't working as expected? Your schematic shows a 1V (peak to peak?) signal. You can't have a 50dB gain with a dual 9V supply.

 

I am inputting a sine wave that is 2 Vpp or 1 V peak amplitude. I'd like to boost that up to about 18 Vpp, 9 V peak (16 Vpp actually since it is not rail-to-rail). However, even at calculated gains greater than 200, the output hardly climbs above 1.1 V peak. If I don't pre-amplify and instead try to amplify the 60mV signal, I can still only boost it to about 1 V peak.

I realize I can't get out more than the supply to the Op Amp regardless of the gain; I'd be happy to get to the maximum swing of the Op Amp which is around 16 Vpp. I thought that the figure indicating 50 dB for open-loop is high enough that the closed-loop should be able to provide the 20 to 30 dB that I need. I suppose they are not related enough for this assumption to be true?

Also, which figures should I be looking at to size a closed-loop circuit if the open-loop figures are not useful?

 

I am inputting a sine wave that is 2 Vpp or 1 V peak amplitude. I'd like to boost that up to about 18 Vpp, 9 V peak.

so design an amplifier with a gain of 9x. With a 3Mhz gain bandwidth, you are good to a tad over 300Khz.

 

That's where my problem is coming in. The OPA552 lists its gain bandwidth at 8 MHz, yet I cannot boost it much beyond 1V at 35 kHz. For example, with R_feedback = 8.2k and R_g = 1k, I expect a gain of about 9.2. However, the signal I read at the output is hardly 1.1V.

 

What is RL?
Are you using a oscilloscope with a 50Ω input to look at the op amp output?
What is the output at lower frequencies?
Otherwise I don't understand why you output voltage is so low.

 

Per crutschow's commnet, what is the value or RL?

As an aside, I hope you are using adequate bypassing from your power supply connections to ground.

 

Hello! I am used to the software and aerospace world, and only have an intermediate level of knowledge in electrical circuitry. I have been creating a few simple circuits to help improve my knowledge here.

The circuit I've been working on is a simple sine wave generator based on a Wien Bridge Oscillator and Op Amp, using a JFET for automatic gain control. This part of the circuit has been working correctly after choosing an Op Amp with a higher bandwidth product and slew rate. At low frequencies it outputs decent voltage, but by 35 kHz it only outputs about 60mV. I believe this is to be expected, so I thought I'd correct it with a non-inverting Op Amp to boost it to my 9V design point.

This non-inverting amplifier portion is where behavior is not quite what I am expecting. I am using an OPA552 because of its high bandwidth and slew rate, as well as decent current output ability. It appears to have very little gain at 35 kHz, however. Looking at the data sheets, I should be getting at least 50 dB of gain in an open-loop configuration. This seems to be enough, but the most I seem to be getting with both closed- and open-loop configurations is only about 23 dB (or G=15) to about 1V. The Gain Bandwidth Product is 8 MHz, so I would assume that I should at least be able to get G=200+ at 35 kHz.

If I pre-amplify the signal to about 1V and then feed it into the OPA552, I hardly get any gain at all. The output signal regardless of the resistor values is barely above the input of 1V. Am I missing something obvious in how Op Amps work? Or do I simply need an even higher speed Op Amp for these frequencies?

I've included a schematic of the non-inverting amplifier I am using. It's no different than the standard ground and feedback resistor design. I have tried resistor values ranging from (R_feedback = 1k, R_g = 6.8) to (R_feedback = 220k, R_g = 1k) and (R_feedback = 10k, R_g = 1k). None of these seem to provide any better behavior. I can post the Wien Bridge circuit if needed as well, but I'm not sure if that part of the circuit should matter. Not included in the schematic are the sets of decoupling capacitors (0.1uF and 10uF) bridging +V, -V to ground.

Hello,

First check the value of RL as others have suggested.

What are you using to measure these quantities? Is it a scope, meter, etc. ?
Many meters will not respond well to 35kHz.

Also, you should try a gain of 5 to start with an get that working if you really intend to input 1v peak ac. With maybe 1k input resistor and 4k feedback and input to the non inverting terminal you should see a gain of about 5.

If you DO NOT see that gain, then something else is wrong, such as:
1. Power supply is not right.
2. Connections of the op amp are not right.
3. Measurement equipment does not work at the required frequency and/or levels.
4. Bad or 'fake' op amp.

You could also try a cheap LM358 to start with just to get going. Try a lower frequency like 1kHz and see if you can get SOMETHING working right
You can also try 1kHz with your present circuit and see if that makes any difference.

BTW a gain of 200 is probably too high at 35kHz. 100 might be more reasonable.

You could also show a few scope pics on there by uploading so we can see exactly what you are getting on the output.

Merry Christmas to everyone.

 

However, the signal I read at the output is hardly 1.1V.

could be a variety of other factors: fake chip, wrong value resistors, overloading, bad rails, ...

 

you probably don't want to use your opamp open-loop, however. their open loop bandwidth is typically less than 10Hz, and sometimes 1Hz.

No.

You are applying a condition of the maximum open loop bandwidth at very low frequencies to a medium gain, medium frequency application, and this is not correct. You are referring to the corner frequency in the open loop response plot, caused the the internal frequency compensation capacitor that makes the open loop plot look like a low pass filter because, well, it is a low pass filter at very high gains. But that does not apply to close-loop gains set below the open loop value for any particular frequency. The OP's estimation of his available gain at 35 kHz is correct.

ak

 

What are you using to measure these quantities? Is is scope, mete, etc. ?
Many meters will not respond well to 35kHz.

I do not have access to a scope at the moment, and have been using a multimeter. I had mistakenly assumed that since the frequency reading could go up to about 10 MHz, that the AC reading would also support this. Turns out after looking at the sheets that the particular multimeter can only read up to 400 Hz AC voltage!

I used my soundcard as a quick oscilloscope to test this theory, and everything seems to be working properly (after dividing down the voltage of course). Thanks for the help everyone, and sorry for the wild goose chase! I've definitely learned a thing or two regardless!