Author Topic: Small switching amplifier up in smoke. Twice.  (Read 2611 times)

0 Members and 1 Guest are viewing this topic.

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Small switching amplifier up in smoke. Twice.
« on: November 10, 2019, 06:07:51 pm »
Hi-

I’m working through the book “Learning the Art of Electronics.” The project I’m having trouble with is attached.

I carefully wired this and double-checked to make sure there were no mistakes. I omitted the decoupling capacitors (on the supply line) as the instructions stated— they said to install them afterwards and listen for the effect.

I never got to do this. I switched on the power, and heard some hiss noise coming from the speaker, as expected. I set the function generator for a sine wave, 300 hz, .5 volts amplitude, and switched it on. I heard the tone on the speaker for about 3 seconds, then silence and a small poof of smoke.

Thinking I might’ve made a mistake, I checked everything carefully, then used the 2nd amplifier that I had and tried it again, with the exact same result.

I’m not sure this is relevant, but the bookshelf speaker that I’m using has an auto-shutoff to protect it from being over-driven. Although the volume wasn’t particularly loud, I’ve seen it shut off in the past from a quiet volume because it didn’t like the home-brewed signal I was sending it. Keeping in mind that I don’t know what I’m talking about, I was thinking that maybe the speaker shutting off caused some kind of inductive spike that overloaded the amp?

Here’s a link to the amp’s data sheet:    https://www.ti.com/lit/ds/symlink/lm4667.pdf

Thanks so much for any help you can provide.

-John Adler
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Small switching amplifier up in smoke. Twice.
« Reply #1 on: November 10, 2019, 06:26:15 pm »
My telepathic ability is sort of weak today. Maybe you could post a schematic of what you did annotated with relevant data?
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 
The following users thanked this post: John0922

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #2 on: November 10, 2019, 06:32:11 pm »
Hi Soldar-

Thanks for your reply. The schematic is part of the project instructions that I posted.
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Small switching amplifier up in smoke. Twice.
« Reply #3 on: November 10, 2019, 06:38:14 pm »
With the 'gain' select pulled high, you are using low gain.  However, the + input should be tied to the incoming signal ground according to Figure 28 in the datasheet.
I don't know if it matters.

Why not just buy a cheap 8 Ohm speaker for your testing?  PC speakers are not usually 8 Ohms because they have some electronic amplification internal to the assembly.  I don't know if this is true for your speaker but the device wants to see 8 Ohms, not some magical high impedance from another amplifier.  The fact that your speaker has 'auto off' tells me that it isn't just an ordinary 8 Ohm speaker.

 
The following users thanked this post: John0922

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Small switching amplifier up in smoke. Twice.
« Reply #4 on: November 10, 2019, 06:43:20 pm »
If you did exactly what is in the schematic and connected it correctly my guess is it would not smoke.

So you are doing something wrong. Maybe the power supply. Maybe the way your measuring and testing equipment is connected. Maybe the input.  Maybe the output.

Really. You are obviously doing something wrong. Don't tell us what the book says you should be doing. Tell us what you are, in fact, doing.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 
The following users thanked this post: John0922

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Small switching amplifier up in smoke. Twice.
« Reply #5 on: November 10, 2019, 06:43:27 pm »
We can see from the documents attached what you were supposed to do but we don't have a picture of what you actually did.  I know you checked but that hasn't been working out well.  I really think the problem is with the speaker but a photo of the breadboard wouldn't hurt.

Personally, I wouldn't even think of running the chip without bypass capacitors as described in the datasheet.  They go to great lengths, employing high paid engineers, just to provide accurate instructions on how to use their product.  I think I would try it their way first.

High frequency oscillations come to mind.  You can't hear them, they're high frequency, but they can cause a lot of damage.
 
The following users thanked this post: John0922

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #6 on: November 10, 2019, 06:48:44 pm »
Hi rstofer-

Thanks for your help. The speaker I’m using is marked 8 ohms, although you’re probably correct in that there’s some circuitry inside that complicates the issue. I will pick up an ordinary 8 ohm speaker for my next test.

 

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #7 on: November 10, 2019, 07:09:13 pm »
Hi rstofer-

I looked at figure 28 that you referred to on the data sheet. It shows the input ground tied to + in through a capacitor as you stated, but it doesn’t show the input ground tied to the supply ground, as the instructions in my book imply. This was a good catch on your part, and the next time through I will wire it like the data sheet.

Thanks!
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Small switching amplifier up in smoke. Twice.
« Reply #8 on: November 10, 2019, 07:45:38 pm »
However, the + input should be tied to the incoming signal ground according to Figure 28 in the datasheet.
Even though it says it is "single ended", the way it is shown it is differential input. If you connect the ring to gnd then it would be single ended. It should work in both cases but need to be careful that you do not do the old mistake of grounding the power supply through the circuit or the scope or the .. whatever. It has been mentioned and warned a million times in many threads.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 
The following users thanked this post: John0922

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #9 on: November 10, 2019, 07:47:42 pm »
Regarding my including a photo of the breadboard, I broke it down already but will definitely include one next time.

Thanks again for all the suggestions.
 

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #10 on: November 10, 2019, 07:50:13 pm »
Fortunately I’ve avoided the grounding mistake thus far! Thanks for the heads up, though.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Small switching amplifier up in smoke. Twice.
« Reply #11 on: November 10, 2019, 08:07:23 pm »
Yikes, they actually suggest powering it up without decoupling?!

I wonder what their bench supply and lead lengths were.  Possibly that can be done with short leads and a stable supply.  At some point I doubt it's going to survive.  It may be that you've found that point...

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: John0922

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Small switching amplifier up in smoke. Twice.
« Reply #12 on: November 10, 2019, 08:10:13 pm »
We never discussed the power supply.  So, I'll make some unwarranted assumptions...

If your supply has current limit, look at page 4 of the datasheet.  The quiescent current (no input, with 8 Ohm output) is just 9 mA.  So, set the supply for 10 mA if you can.  This way there is probably no chance of the magic scope escaping.  See how it goes before you apply a signal.  Feel around for heat - you won't find any at this current level but it won't hurt to look (feel).

What you don't want is a power supply with no current limit and capable of 3A or 5A.  What should happen is, if the current exceeds the 10 mA settiing, the output voltage starts to roll back.  On a flat out short circuit, the voltage should drop to zero.

The other thing you might do is put your DMM, in current mode, in series with the power supply.  Assuming you don't have a low level ammeter in the PS itself.  This is a place where an analog meter is handy.  If I see the needle start to swing wildly, I can shut off the PS.  The slow update rate of many DMMs means I won't know there is a problem nearly as soon.

My Simpson 260 VOM (analog) has a 10 mA and 100 mA scale.  I might select the 100 mA scale and if the needle swings much past 10 mA it's time to hit the switch.  Later on, I can verify the quiescent current on the 10 mA scale.

One of the modern bench supplies from Rigol or Siglent would be highly recommended.  I bought the Rigol DP832 a couple of years ago and it has been terrific. I have actually saved some chips by setting the current limit fairly low.  The chips drew too much current (address conflict) and the PS rolled the output voltage back to about 2V.  All chips survived!

There are many other options for power but something with current limit would be very handy for this project.

Just some random thoughts...

 
The following users thanked this post: John0922

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #13 on: November 10, 2019, 08:26:20 pm »
Hi Tim-

Thanks for your comments. I’m using a Dr. Meter power supply— not particularly stable I would imagine, and lead lengths are fairly long. So maybe that’s my problem. Next time through I will start with the decoupling capacitors on, and then maybe give a listen with them off, knowing that might blow the amp.

It will be a few more weeks until I can try this again. I have to order these amps from a special place in Canada because they are tiny surface mount parts— this place sells them mounted to a breakout board.
 

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #14 on: November 10, 2019, 09:27:56 pm »
Rstofer-

Thanks for your suggestions. I am able to set a current limit on my Dr Meter power supply, so that’s something I will definitely do. I don’t have an analog ammeter like you (actually I have Dave’s EEVblog Ucurrent Gold ammeter) but monitoring the current is a great idea.

Thanks again.
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Small switching amplifier up in smoke. Twice.
« Reply #15 on: November 10, 2019, 09:43:43 pm »
Setting the current limit is described in the User Manual so that's a workable feature.

The PS Ammeter will read in mA so you should get a pretty good indication of current, even at low values.  I might still insert a DMM in series if I had any doubt about the PS Ammeter accuracy.
 
The following users thanked this post: John0922

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20363
  • Country: gb
  • 0999
Re: Small switching amplifier up in smoke. Twice.
« Reply #16 on: November 10, 2019, 09:45:38 pm »
Regarding my including a photo of the breadboard, I broke it down already but will definitely include one next time.

Thanks again for all the suggestions.
Layout is pretty important for a class D amplifier. A breadboard is completely unsuitable. It needs to be on a proper PCB. The data sheet you linked to shows a suitable PCB layout.

How long was the speaker cable? I believe these filterless class D amplifiers are designed for the speakers to be close to the IC, not on a long length of cable, which can produce voltage spikes and current surges due to parasitic inductance and capacitance.
 
The following users thanked this post: John0922

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Small switching amplifier up in smoke. Twice.
« Reply #17 on: November 10, 2019, 09:59:54 pm »
I believe the data application figure 28 contains a mistake and the only difference between fig. 28 and 29 should be the connection of the Gain Select pin and everything else should be the same.
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 
The following users thanked this post: John0922

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #18 on: November 10, 2019, 10:03:26 pm »
Hi Zero999-

Thanks for your suggestions. Since I’m building this just as a temporary project, hopefully I’ll get away with a breadboard for the few minutes that I test it, knowing that the performance would be better on a PCB.

The original speaker cables were 3 feet long, but since I’m replacing the speaker on the next go-round, I’ll be sure to keep the new cables short.

Thanks again.
 

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #19 on: November 10, 2019, 10:05:29 pm »
I believe the data application figure 28 contains a mistake and the only difference between fig. 28 and 29 should be the connection of the Gain Select pin and everything else should be the same.

Interesting, so I guess I wired it correctly then.
 

Offline Audioguru again

  • Regular Contributor
  • *
  • Posts: 115
  • Country: ca
Re: Small switching amplifier up in smoke. Twice.
« Reply #20 on: November 10, 2019, 10:27:11 pm »
I agree that the "Lab Assignment" is completely wrong to say it should be wired on a breadboard (the datasheet talks about using short but wide pcb traces) and wrong again to try the circuit without the important power supply bypass capacitor mounted as close as is possible to the IC.

I am Canadian. The "special place" you bought the mounted ICs in Canada might be using cheap fake ICs from China. There are many Chinese stores in Canada.
 

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #21 on: November 10, 2019, 10:36:17 pm »
Thanks AudioGuru. I’ll do everything I can to keep the leads short.

I’m guessing the place I’m buying the amp from is trustworthy because it was referred by the book author himself.
 

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Small switching amplifier up in smoke. Twice.
« Reply #22 on: November 11, 2019, 05:17:02 am »
If you want an easier to source micro class D amplifier IC try the NS4150, NS4158, or CS8138. These also have better efficiency and higher max output power. SO-8 breakout boards are easily found on aliexpress and soldering SO-8 is trivial.

LM4667 is a bit overpriced for what it does which makes it a potential high target for counterfeiters.

I am Canadian. The "special place" you bought the mounted ICs in Canada might be using cheap fake ICs from China. There are many Chinese stores in Canada.
Why use "fake" ICs from China when you can get real, Chinese made ICs that perform better and cost less?
« Last Edit: November 11, 2019, 05:22:08 am by OwO »
Email: OwOwOwOwO123@outlook.com
 
The following users thanked this post: John0922

Offline John0922Topic starter

  • Contributor
  • Posts: 44
Re: Small switching amplifier up in smoke. Twice.
« Reply #23 on: November 16, 2019, 08:11:25 pm »
Hi-

I received the new amp today and set it up, incorporating the changes suggested above. It worked fine. I tried it with the wiring in the datasheet and the wiring from the book. Both worked equally well.

Thanks again for all the help.

-John
 

Offline soldar

  • Super Contributor
  • ***
  • Posts: 3595
  • Country: es
Re: Small switching amplifier up in smoke. Twice.
« Reply #24 on: November 16, 2019, 09:25:37 pm »
Yay!
All my posts are made with 100% recycled electrons and bare traces of grey matter.
 
The following users thanked this post: John0922


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf