Author Topic: Odd Glitch In Waveform  (Read 1948 times)

0 Members and 1 Guest are viewing this topic.

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Odd Glitch In Waveform
« on: February 10, 2020, 06:34:47 am »
I made this circuit to experiment with amplifiers, then I tested it with my function generator to see if I could use it as a mic amplifier and I see this weird glitch in the output. Not sure if that's normal or if I messed something up. My boards pretty messy so I also drew a diagram in MSPaint cause I suck at SPICE. Also excuse the mess on the desk, I don't have that much space so it gets cluttered quick.
its your boi, dj ethanol
 

Offline Nerull

  • Frequent Contributor
  • **
  • Posts: 694
Re: Odd Glitch In Waveform
« Reply #1 on: February 10, 2020, 07:12:28 am »
Looks like clipping and crossover distortion. Clipping is caused by not having enough headroom in your supply rails.

741 opamps are old and pretty garbage and prone to crossover distortion.

You can try a load resistor between the output and the negative supply rail.

https://youtu.be/VgodYtiD_F0
« Last Edit: February 10, 2020, 07:15:25 am by Nerull »
 
The following users thanked this post: codingwithethanol

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Re: Odd Glitch In Waveform
« Reply #2 on: February 10, 2020, 07:16:58 am »
@Nerull Yeah I understand it clipping cause i was playing with the power supply and noticed that I was changing where it flattened out, but idk what crossover distortion is so i'm gonna google it. I have a few other opamp types so i'm gonna try them out.

Edit: Thanks for the link I love this guy! Alot of the stuff he does kinda goes over my head but ill take a look at this one

Edit: https://youtu.be/VgodYtiD_F0?t=552 Ok this seems pretty much like black magic but ill try it out
« Last Edit: February 10, 2020, 07:25:42 am by codingwithethanol »
its your boi, dj ethanol
 

Online magic

  • Super Contributor
  • ***
  • Posts: 7458
  • Country: pl
Re: Odd Glitch In Waveform
« Reply #3 on: February 10, 2020, 10:36:06 am »
Check what's your virtual ground voltage and how it varies in time. 1MΩ seems rather high for rail splitting.

I don't know about µA741 crossover distortion, I thought it wasn't supposed to be that bad. But if you bought these chips from an auction site, chances are good that they are LM358 with fake markings.
 
The following users thanked this post: codingwithethanol

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Odd Glitch In Waveform
« Reply #4 on: February 10, 2020, 04:13:04 pm »
Check what's your virtual ground voltage and how it varies in time. 1MΩ seems rather high for rail splitting.

I don't know about µA741 crossover distortion, I thought it wasn't supposed to be that bad. But if you bought these chips from an auction site, chances are good that they are LM358 with fake markings.

Lower those 1M resistors to 1K and see what happens.

Considering their use in aged Analog Computers, the crossover distortion of 741 Op Amps can be managed.  These old computers use +-15V supplies with +-10V signal swings and I have never seen clipping or crossover distortion.
 
The following users thanked this post: codingwithethanol

Online mikerj

  • Super Contributor
  • ***
  • Posts: 3384
  • Country: gb
Re: Odd Glitch In Waveform
« Reply #5 on: February 10, 2020, 08:07:37 pm »
The audiophools will be soon be seeking that classic '741' sound.
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Odd Glitch In Waveform
« Reply #6 on: February 10, 2020, 08:27:24 pm »
The audiophools will be soon be seeking that classic '741' sound.

But only those in the metal can.  They're actually hard to get!
 
The following users thanked this post: mikerj

Online magic

  • Super Contributor
  • ***
  • Posts: 7458
  • Country: pl
Re: Odd Glitch In Waveform
« Reply #7 on: February 10, 2020, 09:32:35 pm »
National actually released their LME49xxx line of audio opamps in metal cans, quite recently :scared:
 

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 4135
  • Country: 00
Re: Odd Glitch In Waveform
« Reply #8 on: February 10, 2020, 11:17:14 pm »
I caught exactly the same issue with fake Chinese operational amplifier. It worked with no distortion on high impedance load (1 MOhm), but the same distortion appears when I tried to used it with 50 or 100 ohm dummy load (according to the datasheed it should work ok with 50 ohm). Also I found that opamp chip doesn't respond on output control pin. I tested this pin with multimeter and it looks like this pin is not connected at all. Also I found that opamp bandwidth at least 1000 times smaller than specification. So, I think this is a cheap low speed opamp remarked as a high speed one...
« Last Edit: February 10, 2020, 11:18:53 pm by radiolistener »
 
The following users thanked this post: codingwithethanol

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: Odd Glitch In Waveform
« Reply #9 on: February 11, 2020, 03:03:45 am »
What voltage are your supply rails?

The LM741 is specified for ±10V minimum supply rail voltages (i.e. your supply with the splinter would have to be at 20V). Looking at the oscillogram I'm going to guess that you were feeding it something much lower. The clipping, dual slope non-linearity and crossover distortion all point to that.

You can try a load resistor between the output and the negative supply rail.

That trick wouldn't work here, the 741 has a proper class AB totem pole output unlike something like the LM358 where exactly that trick is often called for to compensate for the weaknesses of the lop-sided class B output stage. I also suspect you're thinking about the LM358 when you talk about a proneness to crossover distortion; there's no bias applied to the output transistor bases on the LM358 to offset the VBE drops on the output transistors - it operates in pure class B and relies on external negative feedback to get rid off the problem. The LM741 has a classic VBE multiplier used to bias the output transistor bases into the active region in classic class AB operation.

LM741 internal circuit


925738-0

LM358 internal circuit

925742-1
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 
The following users thanked this post: codingwithethanol

Online magic

  • Super Contributor
  • ***
  • Posts: 7458
  • Country: pl
Re: Odd Glitch In Waveform
« Reply #10 on: February 11, 2020, 08:37:09 am »
I caught exactly the same issue with fake Chinese operational amplifier.
I recently got an AliExpress coupon and spent it on some cheapest jellybean opamps for shit and giggles.

Long story short, they were all LM358 inside or RC4558 for the "audio" parts. You just can't buy that stuff from there. And lots of people import them and resell in Europe, so domestic auction site sources should be suspect too.
 

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Re: Odd Glitch In Waveform
« Reply #11 on: February 12, 2020, 06:45:35 am »
Thank you guys for all the feedback! So I took into consideration some people in the thread said:

I first tried the original suggestion by @Nerull, I added a 1k and 100k pot in series as a dummy load, but I didn't see anything similar to what was shown in w2aew's video. The glitches sort of "flattened out" and then the output started clipping.

@rstofer and @Cerebus mentioned the 741 being used with at least ±10V, so I adjusted my PSU to 20V so the split would be -10V +10V.
@rstofer mentioned that the resistors for the voltage dividers needed to be lower so I used 1k as suggested.

These modifications didn't affect the output in any way, so finally, since I purchased these off of Ebay from a dubious chinese reseller, i'm going to take into consideration what @magic said and assume these amps are just janked clone ripoffs. I have some other amps which I want to try out, I actually got some ceramic chips from an old bag of mixmatch parts I got a while ago, so I'll try them next.







its your boi, dj ethanol
 

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Re: Odd Glitch In Waveform
« Reply #12 on: February 12, 2020, 07:06:16 am »
Success! I skipped the ceramic chip and tried an RC4558, which the datasheet says is basically a dual 741 and it works perfectly, as seen in the attached picture. Gonna try to use it as a mic amp next.
its your boi, dj ethanol
 

Online magic

  • Super Contributor
  • ***
  • Posts: 7458
  • Country: pl
Re: Odd Glitch In Waveform
« Reply #13 on: February 12, 2020, 07:56:58 am »
Dual 741 was 1458 and 4558 is a slightly different animal.

4558 is what goes into all those NE5532's on auction sites and fetches 100% positive feedback so it's clearly good enough for audio :D
 

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Re: Odd Glitch In Waveform
« Reply #14 on: February 12, 2020, 08:07:10 am »
@magic Just rigged up electret mic and cranked the gain and it works beautifully! I have an offtopic question if you don't mind, I fed the opamp output to a logic level mosfet to try to power a speaker but the fet cooked for some reason, also I tried hooking the opamp output directly to the speaker but i just got a wierd clicking noise. After that I tested the speaker with my function generator to make sure it wasn't faulty and it's fine so how would you go about this? End goal of this whole thing is to make a mini megaphone, I have these cute little speakers I ripped out of an old dell thing (See picture) that I wanted to use in a project so why not this.
its your boi, dj ethanol
 

Online magic

  • Super Contributor
  • ***
  • Posts: 7458
  • Country: pl
Re: Odd Glitch In Waveform
« Reply #15 on: February 12, 2020, 08:17:26 am »
Post the schematic. You can't make an amplifier with just one MOSFET, even Nelson Pass can't do it :P

Fundamental question about every speaker: what impedance, what power.
 

Offline codingwithethanolTopic starter

  • Regular Contributor
  • *
  • Posts: 73
  • Country: us
  • always tilted
Re: Odd Glitch In Waveform
« Reply #16 on: February 12, 2020, 08:25:55 am »
@magic
Schematic is same as before except electret mic is input not function generator. Also the feedback resistor is 100k now. Output of the amp varies from under a volt to clipping depending on what the mic is picking up. Speaker is 12ohm 5W, I set my function generator to 1k sine 1V p-p and the tone was low but audible so 5V-10V should be earpiercing.
its your boi, dj ethanol
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17429
  • Country: us
  • DavidH
Re: Odd Glitch In Waveform
« Reply #17 on: February 13, 2020, 01:13:45 am »
Dual 741 was 1458 and 4558 is a slightly different animal.

4558 is what goes into all those NE5532's on auction sites and fetches 100% positive feedback so it's clearly good enough for audio :D

The 4558 has "741 type" performance but is otherwise very different.  The quad RC4136 is the same way.  The basic difference is that they use a PNP differential input stage to get a wide differential input voltage range, like a 324/358, while the 741/301 type of operational amplifier uses a PNP cascaded with the NPN differential stage.  The reason this works is that the PNP transistors built on these NPN processes have a high base-emitter breakdown voltage.

The major difference is that the input bias currents are reversed which causes problems in some circuits.

That trick wouldn't work here, the 741 has a proper class AB totem pole output unlike something like the LM358 where exactly that trick is often called for to compensate for the weaknesses of the lop-sided class B output stage. I also suspect you're thinking about the LM358 when you talk about a proneness to crossover distortion; there's no bias applied to the output transistor bases on the LM358 to offset the VBE drops on the output transistors - it operates in pure class B and relies on external negative feedback to get rid off the problem. The LM741 has a classic VBE multiplier used to bias the output transistor bases into the active region in classic class AB operation.

It still works to force class-A operation but that is of more limited use when the output stage is already class-AB.
 

Offline nigelwright7557

  • Frequent Contributor
  • **
  • Posts: 706
  • Country: gb
    • Electronic controls
Re: Odd Glitch In Waveform
« Reply #18 on: February 13, 2020, 01:47:45 am »
The very old 741 has all sorts of problems.
Poor bandwidth, not rail to rail,crossover distortion, lower input impedance than JFET device.

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf