op amp comparison LMC6484 vs LM324

Thread Starter

dlatch

Joined May 15, 2016
91
These two quad op amps have the same pinout so, I can compare them in the same circuit via an IC socket.

Both work fine, the pwm outputs are good. However, there is an anomaly with the 6484 that I want to understand.

The control voltage at the bottom comparator inputs is distorted, like, "modulated" by the triangle wave (I guess)

I have drawn in approximately what it looks like on the scope with both the tri wave and the control voltage shown. The tri wave is clean and the control voltage basically mirrors it.

The LM324 does not exhibit this. With the 324 the control voltage is, as expected, a clean flat variable DC voltage.

What could cause this?

comparator-input-anomaly.JPG
 

ericgibbs

Joined Jan 29, 2010
18,766
hi dl,
Do you still get that cross over dead band, when the 'C' and 'D' inputs are disconnected.?
ie: is the 'B' output OK when not driving C/D.

E
 

Thread Starter

dlatch

Joined May 15, 2016
91
I'd have to "trouble shoot" some more to be sure my answer would be correct. I'll do that later today.

I have mostly built this circuit using the 324 and have never seen this anomaly with, or without, the output comparators connected.

So it seems to be caused by the technology difference between the 324 and the CMOS R-R 6484.

My illustration of the wave form is intended to show how the control voltage roughly follows the tri wave. The two control voltages (at the two comparators look slightly different (one feeds + the other feed -) but, they both appear perturbed by the tri wave.))

As mentioned in post 1, the pwm outputs work well even with the distorted control voltages. I mostly want to understand where it comes from. And if it can be cleared up without adding greatly to the complexity and part count.

(I am a novice, "trouble shooting" may be too strong a word for what I do)
 

Thread Starter

dlatch

Joined May 15, 2016
91
Mistake in the above thread: for what it's worth, the quad in question is the LM 6134 (not the LMC6484)

the question remains: why is the control voltage "modulated" by the tri wave? If I disconnect the control voltage, the tri wave shows up on both inputs.
 

OBW0549

Joined Mar 2, 2015
3,566
I've used the LM324 and LM6134 (as well as the LMC6484) in a variety of circuits and have never observed this "cross over" effect you're describing.
 

Thread Starter

dlatch

Joined May 15, 2016
91
I double checked again yesterday and the issue remains. A few micro farads on the control voltage does smooth it out but, I need the control voltage on the Pfet (upper comparator) to swing faster than that. Besides I want to understand where this anomaly is coming from.

The circuit above is my completed project, it was the easiest picture to post.

The test circuit is the same but with no termination on the output comparators and I am using a single 5k pot to sweep a control voltage from the same single ended 13.8 supply, one comparator at a time. The test circuit is soldered up, not a solderless bread board, and the bypass caps are as shown above.

Would simulation show this issue? I will attempt that but, it might take days since I have never done that. If it did show this issue, it would confirm that it has to do with the technology difference of the two chips. (although the 6134 is bipolar not cmos.)

What circuitry is between the inputs on the 6134? As observed yesterday, the tri wave reflects right through with no other terminations.
 

ericgibbs

Joined Jan 29, 2010
18,766
hi dl,
Looking thru the LM6134 d/s I don't see a reason for the effect you are seeing on the scope.?
This circuit is from that d/s.

E
AA1 28-Nov-18 13.39.gif
 

Thread Starter

dlatch

Joined May 15, 2016
91
I just set this up again for fresh look and yes the other 6134's I have all exhibit this anomaly: with 2 amps for the tri wave, you jump that tri wave into either input on either remaining amp and the tri wave shows up on the other input and perturbs the control voltage when you insert that. (the pwm output still works fine sooo...not really an issue?)

For any still reading, I located some LMC6482 (cmos dual version) and I tested with that as the comparators and - small good news - they do NOT exhibit this issue. They do NOT pass the tri wave through, the control voltage is unmolested and, thankfully, the pwm works well too.

So, figure 30 above not withstanding...this novice is here to tell ya...the LM6134 does this. I would still love to know why even though the chip works regardless.
 

ebp

Joined Feb 8, 2018
2,332
looks to me like the base-emitter junctions of Q5 and Q6 in the circuit at #7 will behave as inverse parallel diodes between the two inputs (inboard ends of resistors) when the amp is used with large differential input (which would be prevented in normal operation as an amplifier)

[EDIT] from the datasheet:
If the input signal exceeds the slew rate of the input stage and the differential input voltage rises above a diode drop, the excess signal bypasses the normal input transistors, (Q1–Q4), and is routed in correct phase through the two additional transistors, (Q5, Q6), directly into the current mirrors.
 
Last edited:

Thread Starter

dlatch

Joined May 15, 2016
91
Thank you, ebp. I skimmed that section of the DS but my limited knowledge failed to lock on. Large differential input. indeed, about 2/3 supply all on one input!

I was needing a better high rail op amp than the 324 since my motor controller here is high side. Looks like the LMC6484 will be it.

OP amp instead of comparators? Well at motor control frequencies this is the simplest solution.
 

OBW0549

Joined Mar 2, 2015
3,566
looks to me like the base-emitter junctions of Q5 and Q6 in the circuit at #7 will behave as inverse parallel diodes between the two inputs (inboard ends of resistors) when the amp is used with large differential input (which would be prevented in normal operation as an amplifier)
I could kick myself for not catching that!

When using an op amp as a comparator, it's important to check the op amp's internal schematic to make sure it will tolerate a large differential input voltage. Some op amps use a pair of diodes across the inputs, connected in reverse parallel, for protection; these can spell trouble in a circuit like this.

Always check.
 
Top