# Why is the MCP601 better suited than the LM833 for single supply applications?

Still not answered, please need some help. If you guys don't know the answer don't be afraid to admit it, but I want to generate discussion around potential solutions/calculations and not semantics like telling me what Vcm means when I defined it in my text, or telling me that the MCP601 isn't pin compatible with the LM833 which I obviously took into account for since my output on my 4th figure is working perfectly as expecting. All I ask is for people to read this please and not just nit pick on things that don't even add a spec of relevance to the conversation. Sorry for the rant, but it seems like people here don't have actual answers and only tunnel vision on things that make them feel superior to others.

TLDR: I tested two op-amps with similiar configurations (only adjusted Vcc to an appropriate range) and found very different results, I mentioned a couple hypotheses for why I believe this to be the case, but they don't hold any consistency when comparing the two op-amps, some help/guidance would be greatly appreciated.

Recently I posted a question about my LM833-N and saw that no one had an answer for me, so I am back again trying to figure out why this specific op-amp does not work for my circuit design. Here is my single power supply build as follows (w/ Vcc at 15V and Vee GND):

simulate this circuit – Schematic created using CircuitLab

I tested this circuit with the LM833-N firstly and got this as my scope output(CH2 output CH1 input):

I was confused about what's happening as I believe I connected everything properly (ignore the awful noise on my input signal since it comes from my trash SMPS), so I did some research and believe this to be a case of incompatible input common-mode range. By looking at the datasheet for the LM: http://www.ti.com/lit/ds/symlink/lm833-n.pdf

I found their values of VICR with test conditions of dual rail power supply w/ +- 15V to be within a range of 12V. This means that in order for my input signals to be compatible with the power rails at 15 and 0V, my Vcm must be no greater than +12V and no smaller than +3V. So to calculate Vcm I used a simple formula I found online of (Vin1 + Vin2)/2.

We can caluclate Vcm by analysing if the average peaks of the signals fall within our common-mode input range: For the non-inverting pin Vin1, the small signal is only 2V pk-pk as seen from our scope and is biased by Vcc in a 2:1 voltage divider leaving us with +8.5V max and +6.5V min. On Vin2, assuming op-amp laws that Vin1=Vin2, we should have the same +8.5V max and +6.5V min (please correct me if I am wrong here). Now using the same Vcm formula as above we get a range of +8.5V to 6.5V for our signal which is no greater than 12V and no less than 3V, so everything should check out here.

However, this is not the case as seen by my scope output and everything is working backwards in my eyes. I have rewired and deconstructed the circuit numerous times and the output remains the same. If anyone could shed light on this situation it would be greatly appreciated, thank you very much for taking the time to read this!

Here is a photo of my breadboard if anyone thinks there's a mistake in the connections somewhere, which I hope I didn't make any careless errors:

My next hypothesis was changing the op-amp to a MCP601, so I rewired my circuit by simply replacing the LM833-N with the other op-amp and changed Vcc to 5V (as that is the max rated Vcc suggested) and measured my new waveform (CH1 is my output and CH2 is my input signal):

From here we can see that everything worked 100% properly, so can someone explain to me why the LM833-N is not suited for this circuit? I did all the VICR and Vcm calculations and came to the conclusion that it should have worked, but clearly I made a mistake, but not sure where. If someone can shed some light on this discrepancy it would be much appreciated, thanks!

From the datasheet of the MCP601 I found that the VICR must be no greater than Vcc - (+1.2V) and no less than Vee + (-0.3V) so my input signal fits this criteria as my swing is from 3.5 to 1.5V which is within the VICR of 3.8 to 0V. My last guess for the discrepancy of these two op-amps would be the maximum output swing as I believe the LM833 is clipping before the capacitor decouples the DC, thereby showing a clipped signal in my output. This seems kinda reasonable to me in the sense that the max output swing for the LM833 is said to be 12V to 3V (same as our Vcm) and the signal before the capacitor would swing from (1)*Gain(2) + 7.5 = 9.5V to (-1)*2 +7.5 = 5.5V. And the swing for out MCP601 is +15 to -20 which is well above our output signal swing of (1)*2 +2.5 = 4.5 to (-1)*2 +2.5 = .5 This seems reasonable to assume that the LM833 doesn't work because my signal is fully clipped and the capacitor is only removing the DC bias for what is an already ruined signal and can't recover the output. However, according to these calculations, I can't notice anything wrong with the LM833 that the MCP601 does different.

If someone could please take the time to read this it would be super appreciated, thank you very much!

• Why is input dc blocking capacitor value is so small ? And, your op amp circuit has only AC gain of 2. Jun 27 '18 at 8:32
• C1and R1 and R2 form a high pass filter with a cutoff above 10kHz. There won't be much of your 60 Hz signal getting through to be amplified.
– JRE
Jun 27 '18 at 9:12
• That makes sense, will post an edit with a higher capacitance value, thank you! edit: done! output is still showing flatline though Jun 27 '18 at 20:09
• How did you connect the unused op-amp in your LM833 setup? Jun 29 '18 at 17:16
• MCP602 (a dual op-amp) is pin-compatible with LM833 (a dual op-amp). Jun 29 '18 at 17:35

Your main problem is that R1, R2, and C1 form a 32KHz high pass filter on the input. At 60Hz, you have about 6dB gain in the op amp stage, and about -54dB in the input filter, for an overall gain of -48dB. Replace C1 with one of those 470uF caps you clearly have in abundance, and I bet it will work like you expect it to.

• Alrighty so I tried what you said and swapped the 10nF cap with a 470uF and I seem to be getting the same problem, although my DC did get decoupled on the output so we are half way there. Scope output photo is edited in the main post, thanks! PS do you guys think something is wrong with my Vcm calculations? Jun 27 '18 at 20:05
• Vcm=Vcc/2 limits are peak limited. . Noise must be eliminated by low ESR cap on board and clipless probe method. (search here) Jun 29 '18 at 17:53
• Verify both inputs are Vcc/2 DC Then learn EMI proofing your measurements without long antenna wires. Jun 29 '18 at 18:00
• Okk thanks for the clarification of Vcm being peak limited, so I am glad I did both the calculations, since using peaks made a lot more sense and concurs with my LM833 analysis near the end. I can try EMI proofing my circuit but then why would adding the MCP601 in the same circuit without any proofing show linear results. Since my control is the wiring I would assume there is more to it than just noise, especially with a flatline output which isn't characterized with noise typically Jun 29 '18 at 19:33

LM833

$V_{CM}$ Input Common-Mode Range:+/- 12V w/ +/-15V supply

MCP601 Common Mode Input Range $V_{CMR}$ $V_{SS}$ – 0.3 to $V_{DD}$ – 1.2 V

The input of the MCP601 can come to within 300mV of ground when used signal sided, but it looks like the LM833 needs much more headroom.

This might not be a problem if you used an inverting configuration with the non-inverting terminal between the rails, but that LM833 needs a ton of headroom on the inputs, so your positive supply would need to be fairly large.

• Do my calculations for Vcm look correct to you? Since they tell me that the LM833 should have been compatible with my input and power rails. Also do you think its an issue with the output voltage swing since it seems that the LM833's swing is way below what my actual output is so I assume it would saturate the whole cycle of my signal on the output and when it gets decoupled by the cap and read by my scope the whole signal has already been lost. Input common-mode range shouldn't be an issue here since the input signal is biased appropriately to fit in between the +12V Vcc and 3V Vee Jul 2 '18 at 15:47