Function generator question

[bitman]

I "invested" a while ago in a cheap function generator FG085 - I do not expect it to be perfect or anything - it's just a tool for me to use while I get my head straight.

So I wanted to go a bit deeper with capacitors and actually watch them "in action" while using that as a chance to calculate what I'm seeing to see if I can get it to match the capacitor specification. Anyway - that's not the question. However, to be sure here's a "diagram" of what I built to do that: http://tinyurl.com/ycyhravt

On the simulator things are of course nice and clean. However, on my scope only the cap line looks like. There's a lot of noise (??) on the top of the square wave. My question is if that signal noise is due to my cheap generator or something else.  I've attached a screen-dump to this post to illustrate.

Do note I have a variable resistor in front of the 1k on my bread-board as I realized I needed to be able to adjust the resistance as I changed caps. It's not the actual values I'm asking about, but if you cannot make the results you see on the scope match the diagram, well that's why

[bitman]

I know it's a cheap function generator, and noise is a known problem. But I'm seeing about .5v of "noise" which seems rather excessive.  For now, I don't think spending $200+ on a "gadget" I'll be using rarely, and it does seem I can get this work albeit I shouldn't count on it being too precise (albeit the frequency generated seems good enough to trust).

[BreakingOhmsLaw]

 - What are you using to measure this signal? An oscilloscope probe, or one of those BNC/Banana-plug adapters? Let's see your measurement setup.
 - Does the FG085 Square look noisy too, when you don't attach a load to it? Try a BNC lead straight to the scope.
 - I see your function generator board runs on a wall socket power supply. Your supply might be inadequate, check the rating.

[rstofer]

I have the same signal generator.  Unterminated, I get about 300 mV of noise and about 250 mV with a 50 Ohm terminator at the scope input.  This is not a precision SG 

My Siglent SDG2082X puts out nearly no noise, the trace is about as narrow as the DS1054Z will produce.

[bitman]

- What are you using to measure this signal? An oscilloscope probe, or one of those BNC/Banana-plug adapters? Let's see your measurement setup.
 - Does the FG085 Square look noisy too, when you don't attach a load to it? Try a BNC lead straight to the scope.
 - I see your function generator board runs on a wall socket power supply. Your supply might be inadequate, check the rating.

I use 10K oscilloscope probes. The ones that come with the scope.
When I plug the scope straight onto the function generator I see the same noise. On every type for wave form. I noticed, that if I change the scope to use "dots" I have two parallel lines - it's not just a wide fuzzy line. From reading/browsing around noise on this "model" is expected. It just felt more than normal to me.

Hmmm - interesting, I never looked at the adapter. It came with the unit when it arrived. I'll have to figure out what it is supposed to be at first, but I'll check. That said, the generator is pretty stable.

[bitman]

I have the same signal generator.  Unterminated, I get about 300 mV of noise and about 250 mV with a 50 Ohm terminator at the scope input.  This is not a precision SG 

My Siglent SDG2082X puts out nearly no noise, the trace is about as narrow as the DS1054Z will produce.

I didn't expect it to be anywhere close to perfect. I was wondering if there was something I could adjust etc.
As to your Siglent, that's WAY out of my price league ($600+).  As my current newbie level it's probably overkill too. That said, a few months ago I had no clue what I would use a function generator for

[SebG]

Hey. The waveform that the rigol oscilloscope shows is a collection of hundreds or thousands of waves superimposed so it appears as noise on the flat part of the square wave but what is probably happening is the whole wave shifts up and down.  Try capturing only a single waveform to see if the top and bottom of the square wave are flat or noisy.  The shifting of the wave can be either noise being picked up by the wires or there is a poor ground connection.  Try zooming out when viewing the square wave only until the square wave will be to narrow to see and you will only see the top and bottom profile of the square wave then you might see the possibly 60 Hz signal (the square wave oscillating up and down 60 times per second) (i'm assuming your AC is 60Hz)

[Kleinstein]

The "noise" looks excessive, especially if this is also with square wave output.

I somewhat doubt that this is true noise. More likely it is some kind of other frequency added to the signal. This could be something like ripple of it's supply (120 Hz or something in the 100 kHz range from a SMPS) or something like a processor clock. So it might be worth looking for the "noise" frequency. Depending on the sources there might be chances that one could improve on this.

[alsetalokin4017]

Using a 60 dollar function generator (MHS-5200A) :

[danadak]

Looking at schematic does not look like there is much bypassing on the LM6172
used to amp the DAC output or the X9C102 digital pot. I would have thought they
would have a 1 uF or better tant and a .1 uF ceramic disc on the supply rails
physically close to the part.

Set up math in scope to do a HP filter, and see if you can establish its related
to fgen clock noise for starters. Even easier would be do an external HP filter
and see what scope triggers on.

Also probe supply rail of these parts in fgen.


Regards, Dana.

[StillTrying]

On the simulator things are of course nice and clean. However, on my scope only the cap line looks like. There's a lot of noise (??) on the top of the square wave. My question is if that signal noise is due to my cheap generator or something else.  I've attached a screen-dump to this post to illustrate.

On the scope pic you've got the 20MHz BW limit on the cap. trace but not the square trace, which will make the square look a bit more noisy.

if I change the scope to use "dots" I have two parallel lines - it's not just a wide fuzzy line.

Scope in peak-detect mode ?

[danadak]

One approach is to do peak detect probing of fgen, supply rails of
each part, and have a .1 uF ceramic with short leads, and look at
supply rail with and w/o cap applied.


Regards, Dana.

[rstofer]

Here a quick snapshot of the noise on my FG085

Looks like switching noise as it's quite fast.

[bitman]

Using a 60 dollar function generator (MHS-5200A) :

WHERE? I see listings at more than twice that amount.

EDIT - never mind, found one on ebay. Thank you.

[ebclr]

Your scope has a nice test generator on the right bottom corner,  just use it, to see a better wave and check from where the noise come from.

Noise is the less desirable thing  and most present in any circuit don't be afraid of some noise just look and tell this is noise this is signal, later on you can make advanced research on noise, It's not a good idea to fight noise while learning, just assume " ok, that is noise"

[ebclr]

In fact, this generator price is 60 USD with shipping from China included, product only price is 44 USD

[welpester]

Did you find out noise source?

-welpester

[Vtile]

Hmm.. What is the meaning of square wave with that test circuit.. I'm on the believe it should use sine wave, then you can see also the ESR and capacitance by "just" measuring the time difference of the signals and some complex (or geometric) calculus, which can be copied from link below. 

http://www.tek.com/document/application-note/capacitance-and-inductance-measurements-using-oscilloscope-and-function-ge