Oscilloscope Interpretation & Current Spikes: How harmful are they?

[Meatball]

So, how does one measure the Rsense value while it is in circuit?

 

[ryansoh3]

Double check all the scope settings are correct. Your DMM should read the average voltage, which for a PWM signal is not the normal 0.707 x Vpk (and Vpk = 1/2 x Vp-p for an ac only signal) that a sine wave uses, so there will be a small discrepancy. I'm not familiar with your scope, I'm only used to old analogue clunkers haha, so I can't tell from the readout what is going on. It does look like the spikes have stopped though, right?

Yup, thanks for the advice, I just realized that the scope was not calibrated to its test leads. (Until I figured it out, I was so frustrated since the scope displayed ten times the voltage in my DMM. I was about to go nuts! )

Regarding PWM, my dad has a Fluke that's advertised as a "true RMS meter." Would that get the accurate voltage?
I doubt my dad will let me use it anyway.

Unfortunately not. :cryyy: If you look closely, there are still some spikes through the driver output. No matter what combination of capacitors and inductors I'm using, the spikes just don't seem to go away.

I used the "max value" feature of the scope, and the spikes go up to 2.21A, but only very very briefly but repeatedly. I'll take a pic in a few hours once I get some sleep, but do you think the spikes will damage a M140 diode? Otherwise the mean current is 1.80A.

Sheesh that inductor is bigger than your driver. What's the magnitude of the spike we're trying to filter out?

Yes, I don't think that inductor will fit inside my host, so I guess I'll have to find a smaller one.

I can't really describe the magnitude of the spikes, but a video can!

(Please note the x-axis (time) modifications in the bottom that first starts out with 25ms.)
(Also, around 35 sec, I'm messing around with the controls that lets me see the "max/peak voltage" which is around 2.25V=2.25Amps, which may be too much for an M140.)

Please let me know if you want me to "do this or do that" with the scope.

I practically have no idea what I'm doing!

Ryan;

Although I am not familiar with that scope,

the voltage across the sense resistor looks too high.

You need someone familiar with the scope,

to measure voltage across the sense resistor, and translate that to peak current.

LarryDFW

Hmm.. How can you tell that the voltage across the Rsense is too high?

I don't have anyone around me who can use this scope properly, but I'll give it a shot.

You mean the sense resistors as in the ones I got replaced to lower the current from 2.35A to 1.8A, right?

I'll measure the voltage across them and let you know what I get.

Peak current? :thinking:



Thank you so much guys for helping me out! Y'all are seriously great.

Cheers! :beer:

(I think I'll hit the sack for a few hours. I've been playing with the scope for 6 and a half hours and I'm kinda getting sleepy. Cya in the morning! +9 GMT)

 

[ryansoh3]

So, how does one measure the Rsense value while it is in circuit?

I think Larry means the voltage across it--measuring the resistance while it's in a circuit can be very difficult indeed. :yh:

Cheers! :beer:

 

[Meatball]

Are those spikes in the video at ~2.5V?? Where was the voltage read from?

Right, so what I mean is: If we want the current through Rsense, we need both the voltage across it and the value of Rsense via ohm's law.

Do you know the value or your total sense resistance?

 

[ryansoh3]

Are those spikes in the video at ~2.5V?? Where was the voltage read from?

Right, so what I mean is: If we want the current through Rsense, we need both the voltage across it and the value of Rsense via ohm's law.

Do you know the value or your total sense resistance?

Yup, the spikes go up to 2.5 volts :gun:

The voltage is read across a dummy test load's 1 Ohm 10 watt resistor. Though Ohm's law, it can conveniently be translated to the current flowing across the test load.

I'm just wondering, if we know the current through Rsense, what advantages might we have? I'm not so sure how that value will help us, though the total sense resistance is 0.11 Ohm. I'll measure the voltage across them in a few hours.

Thanks for your reply!

Cheers! :beer:

 

[Meatball]

Larry maybe have to help me with this, but I believe what we need is the voltage drop across the Rsense, this is so that we can predict the both the voltage spike and the current spike that would occur once we know what the load is. A test load resistor will take on a constant spike current given some spike voltage, but if this is supposed to be a const. current driver, a laser diode or LED would take on a current spike (the important thing to watch) which is independent of its own voltage drop. So we need to use the Rsense on the board as a reference point from which to work.

We can pick filtering capacitors and inductor values based on the predicted magnitudes of each spike.

I'm no expert on SMPS, so anyone else's input should be as good as mine.

 

[ryansoh3]

Larry maybe have to help me with this, but I believe what we need is the voltage drop across the Rsense, this is so that we can predict the both the voltage spike and the current spike that would occur once we know what the load is. A test load resistor will take on a constant spike current given some spike voltage, but if this is supposed to be a const. current driver, a laser diode or LED would take on a current spike (the important thing to watch) which is independent of its own voltage drop. So we need to use the Rsense on the board as a reference point from which to work.

We can pick filtering capacitors and inductor values based on the predicted magnitudes of each spike.

I'm no expert on SMPS, so anyone else's input should be as good as mine.

Ah, I see. Thank you very much for your explanation.

Here is a video of what's going on across the sense resistor.

The mode is DC coupling, and please take note of the x-axis (time) variances I change to show you a better view.

The voltage seems very noisy and has lots of spikes. What cap + inductor combination should I use to filter the noise/spikes out? Where should I place the noise filter, on the output side?

Again, thanks for the info.

Cheers! :beer:

 

[Sigurthr]

Yes, noise filtering goes on the output side of the driver board. I still think you need a decoupling cap though, they can do wonders. Decoupling caps go directly (or, as close as feasible) across the power in and ground connections to the IC which needs decoupling. This cap removes tiny fluctuations in the supply voltage that are almost imperceivable, even with a scope, which cause unwanted oscillation on the IC's output. A small electrolytic cap anywhere from 1uF to 100uF directly across the Vcc and Gnd pins of the IC is all it takes. If the board is small enough, you can put it across the power in terminals where your wires are, but if the board has enough trace length between that and the IC, it won't work.

 

[ryansoh3]

Yes, noise filtering goes on the output side of the driver board. I still think you need a decoupling cap though, they can do wonders. Decoupling caps go directly (or, as close as feasible) across the power in and ground connections to the IC which needs decoupling. This cap removes tiny fluctuations in the supply voltage that are almost imperceivable, even with a scope, which cause unwanted oscillation on the IC's output. A small electrolytic cap anywhere from 1uF to 100uF directly across the Vcc and Gnd pins of the IC is all it takes. If the board is small enough, you can put it across the power in terminals where your wires are, but if the board has enough trace length between that and the IC, it won't work.

I see... do you mind pointing out where the IC's Vcc and ground are? I don't have a clue where they are.

I think the problem arises from the sense resistors too.

Here is a video of where I short out the sense resistor, and the medium power output is much much more smooth. (although higher) The scope is hooked up to a test load's resistor, which can be translated into current through the load.

Please note the x-axis (time) scale changes.

At 18 sec, I short out the sense resistors, which increases the current but make it very steady.

(In the video I removed the caps and inductor low pass filter on the output side.)

So.... my question is, how should I smooth out the noise/spikes?

Thank you very much for your help. I really appreciate it. :thanks:

Cheers! :beer:

 

[ryansoh3]

Arrghhhhh I'm on the verge of giving up trying to make this driver suitable for driving LD's.

Unfortunately, contrary to what formulist had suggested, I don't think the IC is an AX2003 because the pinouts don't match with that of the documentation, found here.

Tsteele has agreed to have a look at the driver--kinda my last resort.

Cheers! :beer: