Was looking at this and going to set it up in a bit to look it over.

Also

I can suggest some experiments to help you understand why one opamp might be better than the other for this circuit.

And I mentioned I had some LM301s and somehow thought they were audio amps but apparently are general-purpose although do use class AB amp.

 

Was looking at this and going to set it up in a bit to look it over.

This one made me give some thought to what common mode input range really means. I had blindly insured enough headroom, but I couldn't do that with this circuit using the supplies for the other opamp.

I thought about it and decided that for this application, I wasn't really applying the supply outputs to the inverting input. Since I expected the voltages to be tracking relatively closely, I didn't need to worry about it. If I did overdrive the input, the output would experience phase inversion and I'd know why.

I mentioned I had some LM301s and somehow thought they were audio amps but apparently are general-purpose although do use class AB amp.

Even though they call it a general purpose opamp, it's a waste to use it in circuits where a lesser opamp would do. LM301 was an earlier design than uA741 and LM358 and it doesn't have internal frequency compensation. The guaranteed input voltage range is 3V from either supply; similar to the uA741.

 

it's a waste to use it

Just tested them in a simple inverting amp mode. Turns out the entire lot that I received are bad so I won't bother reordering then...

 

Just tested them in a simple inverting amp mode. Turns out the entire lot that I received are bad so I won't bother reordering then...

I take it that these were Ali Express specials? The ones I have are from the 70's; long before counterfeiting was a problem and opamps were relatively expensive back then.

I tend to use LM358 for almost everything. I only consider a "better" opamp when I need one.

 

Yep, AliX specials. Actually came in a tube and the uA741s that came with it were real and actually laser engraved and all tested perfectly. The 301s were stamp marked with Nat Semi Logo but sandblasted and the pins were already bent to fit a socket so obviously junk. it's always a crapshoot with AliX...

 

The 301s were stamp marked with Nat Semi Logo but sandblasted and the pins were already bent to fit a socket so obviously junk. it's always a crapshoot with AliX...

There are very few semiconductors I'd buy from AE. I bought 100 AO3400 and AO3401 and did some testing before I bought more. You can't fake threshold voltage and I was looking for some inexpensive MOSFETs with reasonable current capability.

Assembled products are a completely different consideration. I buy Arduino Uno/Mega and similar things without too much concern; though I do have a Uno with questionable drive capability. I just won't pay $30 for the real deal when I can get a legal workalike for $3-5. I also got a batch of 4 channel level shifters that weren't packaged properly; got a refund and kept the male header pins for future use. Haven't thrown out the suspect parts yet, but I won't use them in any of my projects unless I replace the MOSFETs. Unless I'm desperate, that won't be worth the time or effort it would take.

 

Be careful of the AO3400 - Alpha Omega set end of life - last time buy and not for redesign status. I have reels of them and complementary. They are nice to have around.

 

Be careful of the AO3400 - Alpha Omega set end of life - last time buy and not for redesign status.

That's good to know. I guess it's time to buy more. Is the AO3401 also being discontinued?

 

That's good to know. I guess it's time to buy more. Is the AO3401 also being discontinued?

Yes which follows... luckily I order directly from Taiwan and they do not have a sunset date. Plus I get them for about $60 per reel. The ones in the reel are perfect, I trust them more than AliX and ebay.

 

@SamR There's a saying from Phil Collins in a song that says "In learning you will teach, and in teaching you will learn". Well I learned, maybe remembered might be more accurate, something about opamps from these circuits.

We were taught to always consider the common mode input voltage range when using opamps. It turns out I've sometimes been applying that concept incorrectly.

When I made the positive supply for the opamps 10V, I was mistaken in the application of that rule. The inverting input to the inverting amplifier amplifier would never be much above ground, so the positive supply only needed to be 2V.

I was using a 2.4V zener to generate the positive supply for the inverting summing amplifier I designed to check tracking because I reasoned that the inverting input voltage would never be far from ground. It just occurred to me that that also applied to the tracking supply circuit. So I got it right once and wrong once.

This project has made me a better designer. Thanks for that.

 

Dennis I wondered why the +Vs was 10V when it was controlling to -20V? But the input from +V rail to the inverting amp is ~+7.3 so I assumed it was to accommodate that input? So I'm not sure that I understand that it only needs to be +2V. I've been working with some old circuits from a 1977 Electronics Today magazine article called Mr. Marston's 741 Cookbook. Interesting reading and the circuits appear to make sense but some don't work. Spent half a day testing chips only to reaffirm that they are good (had already done a simple go-no go test). Here is an example and I didn't go for the 100X amplification, only 10X so usually using a 100kΩ Rf and 10kΩ Ri. I can assure you, from the years I spent teaching, it is a good way to learn the material.

 

I wondered why the +Vs was 10V when it was controlling to -20V? But the input from +V rail to the inverting amp is ~+7.3 so I assumed it was to accommodate that input? So I'm not sure that I understand that it only needs to be +2V.

I'll let you think on that one for awhile.

Here is an example and I didn't go for the 100X amplification, only 10X so usually using a 100kΩ Rf and 10kΩ Ri.

What is the input and output?

 

I'll let you think on that one for awhile.

I assume it gets inverted before amplified. Since it's a negative value after inverting, then why does it need the +2V and not simply GND?

What is the input and output?

Typically I use a 100mV or 1.00V input @ 1kHz. Looking at it, it is a hi-pass filter but the control frequency is ~15Hz so that is not it. And no output! Flatline. Which I why I spent most of the day testing chips. As just a simple inverting 10X amp without the decoupling caps on the input and output, and the extra resistors, the chips work fine using 1.00Vin and getting ~-10Vout.

 

I assume it gets inverted before amplified. Since it's a negative value after inverting, then why does it need the +2V and not simply GND?

Let's work through the process instead of assuming things that we don't need to make assumptions about.


To analyze opamp circuits without resorting to equations you've memorized, we just use the zero differential input voltage theorem.

For IC2B. If we have the non-inverting input connected to ground, the theorem says that the inverting terminal will also be at ground. If we have 5V at the node between R4,5,and 6, that gives a current of 0.5mA in R6. Since no current flows into or out of the inputs of an ideal opamp, that 0.5mA current goes through R7; giving a drop of 15V. That makes the output of IC2B -15V.

That gives a gain of -3. That's how the -Rf/Ri gain formula for an inverting opamp was derived.

Do you want to take a crack at analyzing IC2A?

EDIT: I'll let you think about why the positive supply can't be ground.

 

Do you want to take a crack at analyzing IC2A?

I think I need to do some more work on circuit analysis... I know from your example that the input @ the non-inverting pin would be -15 and the feedback on the inverting pin is the negative rail voltage. So the output of the op-amp is trying to adjust the negative rail voltage to match the input voltage. Still a bit fuzzy on the PNP math wise. Time for me to do some more bookwork.

 

I think I need to do some more work on circuit analysis... I know from your example that the input @ the non-inverting pin would be -15 and the feedback on the inverting pin is the negative rail voltage. So the output of the op-amp is trying to adjust the negative rail voltage to match the input voltage. Still a bit fuzzy on the PNP math wise. Time for me to do some more bookwork.

Let me know if you want me to explain this part.

 

Let me know if you want me to explain this part.

Please do. I see the generalities but short on specifics.

 

Please do. I see the generalities but short on specifics.

With -15V on the non-inverting input, the zero differential voltage theorem says the inverting input needs to have the same voltage. To do that, the opamp will apply about -15.7V to the base of the transistor which will turn it on and force the collector voltage to be -15V.

I say about -15.7V because the actual voltage will depend on the load on the negative voltage. If it's low, the junction voltage will be lower; if it's high, it will be higher.

Next question. What is the limit on the maximum current available from the negative supply? How could you increase it?

 

zero differential voltage theorem

That is what I'm missing. So far, all I'm seeing is how to use them and not how to design a circuit around them nor what they actually do in the integrated circuit. A bit of explanation about the long tail pair differential amplifier block and that it also has a voltage amplifier and push-pull output amplifier blocks is all that I've studied so far.

What is the limit on the maximum current available from the negative supply? How could you increase it?

That I do know and it would depend on the op-amp but typically less than ~20mA. To increase it you output it to the base of a transistor with sufficient CE ampacity to supply or sink the load.

 

So far, all I'm seeing is how to use them and not how to design a circuit around them nor what they actually do in the integrated circuit. A bit of explanation about the long tail pair differential amplifier block and that it also has a voltage amplifier and push-pull output amplifier blocks is all that I've studied so far.

Using the zero differential voltage theorem, Ohm's Law, and KVL make it unnecessary in most cases to understand what's inside of the black box. I wouldn't recommend trying to analyze every opamp at the transistor level. About the only time I look at the schematic for an opamp is when I'm trying to understand why the limits are what they are and/or how far I can push the limits.

EDIT: And voltage divider equation or standard gain equations.

For example, the LM358 states that the maximum positive input voltages are 2V from V+. Looking at the schematic (NatSemi in this case):

Shows that that parameter is conservative and it's going to be closer to 2 diode drops plus the current source voltage. The LM393 (TI) has a similar input (this is a simplified schematic):

And they specify the maximum positive input voltage to be 1.5V from V+.

EDIT: over the full temperature range, they specify 2V from V+.

I was able to operate the inverting amplifier with V+ as low as 1.25V (edit - didn't go lower because I was using an LM317 for the supply), but that isn't the way we design circuits. In general, conservative design practices don't require you to cherry pick components. There are times when companies do it, but they need to justify the cost of additional screening to find the cherries. HP did it because I got tubes of TTL that failed their cherry picking. They couldn't use the parts anywhere else because manufacturers marked the standard parts with HP part numbers and their original part numbers weren't shown.

Looking at the schematics also shows you why LM358 and LM393 input voltage ranges include the negative supply, but not the positive supply. I've seen some opamps with rail to rail inputs that use both NPN and PNP differential amplifiers on the input.

Unless you have need to understand the internals of an opamp, the black box approach is usually sufficient.

That I do know and it would depend on the op-amp but typically less than ~20mA. To increase it you output it to the base of a transistor with sufficient CE ampacity to supply or sink the load.

The current sink capability and the pass transistor beta limit the maximum current. The opamp will try to saturate the transistor if necessary and beta drops with lower collector emitter voltage and higher current, so we'd use a worst case value of around 15 for TIP30.

How could you increase current capacity?