Macintosh Portable: "pop" noise every 10s from Speaker & Headphones

[David Cook]

Good video. Don't give up.

Something is just not right about the noise floor on your oscilloscope. There is no way there is basically a 1V difference between your two probes. Could you email Rigol support and ask what the expected level should be for your scope in A-B math mode? I think we need to figure that out before we can diagnose the subtle noise issue in the Mac.

After you get it down to <100mV, then you really do want to compare all the test points I outlined earlier. For example, comparing A-B of pin 12 would tell you if there is a different in analog supply voltage between the two sound chips. I know the schematic shows them connected, but if there is hidden corrosion impeding the voltage reaching one of those pins, then it could affect the sound output.

 

[David Cook]

I've figured it out. It's the vertical scale you've set both your probes to. Try 1V or maybe even 500 mV per division.

 

[JDW]

I've figured it out. It's the vertical scale you've set both your probes to. Try 1V or maybe even 500 mV per division.

I have achieved success in setting up my scope as per your direction, which you can see in my video here:

You are seeing my function generator's output on the scope display.

My function generator was set to a 3.23kHz SINEWAVE with DC offset. Both scope probes were connected together and to the FuncGen output, and all GND's. Noise floor ≈ 32mV. Scope was floated (not earthed). (This video has no sound.)

I will use two scope channels to probe pin-11 and later pin-12 on both SONY chips after work this evening, doing everything in same manner as you saw in my previous video.

By the way, I was able to make that video of my scope's display by adding WIFI, as per this video, and then accessing it on my iMac, and then using QuickTime player to make a screen recording.

 

[David Cook]

I have achieved success in setting up my scope as per your direction, which you can see in my video here:

You are seeing my function generator's output on the scope display.

My function generator was set to a 3.23kHz SINEWAVE with DC offset. Both scope probes were connected together and to the FuncGen output, and all GND's. Noise floor ≈ 32mV. Scope was floated (not earthed). (This video has no sound.)

I will use two scope channels to probe pin-11 and later pin-12 on both SONY chips after work this evening, doing everything in same manner as you saw in my previous video.

By the way, I was able to make that video of my scope's display by adding WIFI, as per this video, and then accessing it on my iMac, and then using QuickTime player to make a screen recording.

Awesome! That is so much better.

I have no doubt that if we can hear the difference (we can) then we can measure the difference.

Here are the other steps from my earlier post. Basically, the methodology is to find pairs of test points up and down the sound chain. Hopefully, by comparing the places the difference between channels is distinct, and the places where it is not, we can determine the where the noise is getting in.

"3. Repeat on pin 13, 15, and 17. These are all power related lines and should be identical for both chips. If noise is present in the power to the sound chip, it stands to reason it would affect the sound output.

4. Repeat on pin 18. This is the data input to the sound chips. Let's hope the difference isn't here. Otherwise the issue is upstream.

5. Repeat on pin 11. Repeat C9-here-R90 (left) vs C10-here-R77 (right). Repeat L1 pin 1 (left) vs L1 pin 3 (right). These are all post sound chip areas whose purpose is to remove the DC voltage from the signal. If one of these components is bad or weakly connected, it could do a poor job removing circuit noise."

 

[JDW]

@ScutBoy
I found your Dec. 2020 post on the MLA, where you reported noise coming from your headphone jack. You said it was low priority for you, so I will assume you never tried to diagnose the noise. Even so, I'd appreciate knowing if the noise you hear from your headphone jack is similar or exactly the same as what you can hear in my video here.

@pocketscience
I found your Portable post here. If you connect headphones to your Portable and try to do the same things you see me do in my video here, do you hear the same noise, similar noise, or no noise at all? (I suggest an initial volume setting of 3, but also try 7 too.)

 

[JDW]

@David Cook
I didn't get much testing done this evening because I burned through all my testing time trying to get my scope to cooperate. Worked fine with the function generator earlier today, but with the actual Mac Portable motherboard readings, it's different. The actual motherboard signals are tiny compared to what my function generator produced.

Here's a scope screenshot showing what I get with the Portable powered off, and both probes set to 10X. In other words, that is with 5.3V active on the motherboard (powered, but shutdown, no audio at all playing in my headphones — dead silence)...

The Ground leads of both Probes are connect to the ground lead of capacitor C8 (470µF) because it's so easy to clip onto.

The yellow waveform above is 118mVp-p and 60Hz (AC wall socket frequency). Yellow is CH1, which is tied to pin-11 of UB15B, which is the LEFT Audio Channel.

(When the machine is powered on, the Left audio channel is much more quiet to my ears than the right channel, as you have heard yourself from my videos. Basically, I hear mostly a soft hiss in the Left channel, what I would consider to be normal.)

Anyway, nothing displays in the Math1 section, probably because pin-11 of UB15C (the Right Audio channel) doesn't appear at all on the left side of my scope. Or perhaps because there is no actual signal. Again, the probes are set to 10X.

When I then power-on the Portable, I can hear the whine in the Right Audio channel (scope Channel-2, colored in light blue), and that produces the following scope display...

Still nothing displayed in the Math1 section, probably because the two waveforms are so different. I pressed the AUTO button so the scope could auto adjust ranges, and the above is what I got.

Because the signals are rather small, and because I am seeing 60Hz too, I decided to then switch my probes to 1X and I changed the scope setting for each probe to reflect that. I then get the following scope display when the Portable is powered down...

The scope briefly put up an overlay just prior to the above screen that said NO SIGNAL, which is of course what I would expect when the machine is powered down, especially because I hear nothing at all through attached headphones.

After a little more pressing of the AUTO button, I was able to get the scope to give me the following in the powered-off condition, with my probes still both set to 1X...

Now that is the first scope display that looks like what I would expect! Again, there is no audio at all in the headphones when the Portable is powered off.

When powered on, I hear the whine in the Right channel, and that yields the following display on the scope...

Despite the spiking on the right (purple signal) above, there is in fact no pop every 10s or anything like that. Those spikes come and go randomly in Math1.

I am testing headless with no LCD attached, and no FloppyEMU, and I am not booting anything. The machine is simply powered on waiting for a disk. Even though the above screen shows 77.486mV peak-to-peak for the yellow Channel-1(left channel), those are faint ESL spikes because my probe ground leads are somewhat long (8cm). The actual signal (brighter yellow in the above) is much lower amplitude, as you can see.

Before testing anything else, we basically have to interpret these results first (basically my previous 2 scope screenshots). All it tells me is what my ears already tell me. I hear a whine in the Right ear when wearing headphones and the Portable is powered-on, and you can see an HF signal in the Right audio channel on the scope when the Portable is powered on. You don't see that HF signal (the whine) in the Left audio channel (nor can my left ear hear a whine), so of course Math1 will show you A-B differences between the two probes.

So is the A-B math really helping here?

It's Friday here in Japan and my weekend is packed with things to do, so I will be leaving the Portable here at the office and won't be able to test anything until Monday. But I will take into consideration any thoughts that are submitted to me before then.
Thanks!

 

[ScutBoy]

@ScutBoy
I found your Dec. 2020 post on the MLA, where you reported noise coming from your headphone jack. You said it was low priority for you, so I will assume you never tried to diagnose the noise. Even so, I'd appreciate knowing if the noise you hear from your headphone jack is similar or exactly the same as what you can hear in my video here.

@pocketscience
I found your Portable post here. If you connect headphones to your Portable and try to do the same things you see me do in my video here, do you hear the same noise, similar noise, or no noise at all? (I suggest an initial volume setting of 3, but also try 7 too.)

@JDW - I will pull the machine out this week and check. Feel free to remind me if I'm tardy in replying.

 

[SuperSVGA]

One thing you can experiment with as well with a debugger (you should at least have MicroBug included) is disabling the output on the Apple Sound Chip. Hit the programmer/interrupt button and type SM FB0802 1, then see if you notice a difference in the noise levels.

 

[JDW]

One thing you can experiment with as well with a debugger (you should at least have MicroBug included) is disabling the output on the Apple Sound Chip. Hit the programmer/interrupt button and type SM FB0802 1, then see if you notice a difference in the noise levels.

I did a search on Macintosh Garden for "microbug" but it yielded no results. If you meant to say "MacsBug" then please let me know what version I should use under System 6.0.7 and higher. I will then boot with your recommended version of MacsBug and give your command line and noise-check a try.

---

In the meantime, I did more testing tonight. Someone on FaceBook has a 5126 and they said they don't get a pop every 10 seconds. I asked their OS version and they said 6.0.7. So I downloaded the 1.44MB disk images from the Garden and tested using my FloppyEMU. I too got no pop sound. However, the default "Portable" control panel is version 1.2, which is made for the 5120 portable. It only offers a slider named Contrast, not Brightness. It is version 1.3 of that same control panel which changes the "Contrast" slider to a "Brightness" slider. And when I install version 1.3, then I get the pop. So I asked that fellow on FaceBook to try version 1.3 and am awaiting his reply. And yes, the version 1.3 Portable control panel is greatly needed. I find that when I boot without that version control panel, my backlight is shut off! Not sure if that is normal for the Portable, but it sure is strange.

Keep in mind that the pop every 10 seconds is a related but separate noise issue. All the other noise you heard in my videos remains, regardless of control panel presence or version.

I then did a lot of testing with games tonight, using the 32Ω stock speaker. Sure enough, much of the noise is filtered out. It's still there, but you'd need to press your ear against the speaker to hear it, even at higher volumes. I am talking about the background noises, not the pop. The pop is rather noticeable even at the default volume of 3, which is what set me on this investigation journey in the first place. It drives me mad, to be quite honest.

Sadly, the data pool of 5120 and 5126 Portable owners who have kindly made time to test their machines is incredibly tiny. So much so the data isn't that useful. So I would once again like to plead with my fellow Portable owners to please test their own noise levels, especially through the headphone jack, to share their experiences. I need to establish what is normal and what is not. Otherwise, I am just operating in the dark on this.

 

[SuperSVGA]

I did a search on Macintosh Garden for "microbug" but it yielded no results. If you meant to say "MacsBug" then please let me know what version I should use under System 6.0.7 and higher. I will then boot with your recommended version of MacsBug and give your command line and noise-check a try.

MicroBug is a much more basic version of MacsBug, I believe it is either built in to the ROM or the OS.

It looks like this:

 

[JDW]

@SuperSVGA

I booted my Portable into System 6.0.8 without the Portable control panel loaded. I pressed Interrupt, and at the prompt, I typed the following and pressed Return:

SM FB0802 1

Suddenly, all the noise stopped. And I should say that I can very faintly hear the same background noise in the left ear as the right. But it's extremely faint even at Volume 7. Most of that background noise (not the "pop") is in the right ear. In any case, all of that stopped in both ears. There was still a very faint hiss in the left ear, but so faint it's almost inaudible. But all the crazy noise like the whine and such you heard in my videos vanished.

I then typed G and pressed return to get back to the Desktop. The noise remained gone until I opened the sound control panel and adjusted the volume. Then all the noise came roaring back.

I then rebooted with the "Portable" control panel version 1.3 loaded. That is the one that has a brightness slider. And that is the version which triggers the "pop" sound every 10 seconds. I pressed the Interrupt Switch and typed in the same thing and pressed return. All the noise once again vanished EXCEPT the pop every 10 seconds. That pop remained even though all the other noise vanished.

Please let me know your thoughts in light of this. Thanks!

 

[nottomhanks]

Another thought, a friend of mine has built a small circuit board that reduces the "hum" from older pinball games. As a musician, I often deal with speakers and "ground loops" where a 60hz hum is anywhere from barely noticeable to extremely obnoxious.

Sometimes, by changing the location of the ground, certain audio artifacts can be reduced and/or almost eliminated.

So, my theory is that if the audio circuitry can use a "different" ground in the 5126 that has NOTHING to do with redraws/backlight, the popping/clicking/excess noise might be resolved, or at least diminished in some way.

Just an idea. Are there other ground sources we can experiment with and see?

 

[JDW]

Are there other ground sources we can experiment with and see?

There are grounds all over the motherboard, so the best way I can answer that question is with the...

1. Schematic here (keeping in mind I have a 5126, and the schematic is for the 5120).

2. 5126 board Photo below (by Bruce Rayne of Branchus Creations, not my board)

 

[JDW]

...disabling the output on the Apple Sound Chip. Hit the programmer/interrupt button and type SM FB0802 1

So that command line I tested last night is disabling only U14A, while leaving U15B & U15C enabled, correct?

I can only assume so because as I said last night, the background noise all went away when I tested your suggested command line, but the pop sound every 10 seconds remained audible (only when I have the v1.3 "Portable" Control Panel active at boot time, of course).

I purchased other capacitor types because someone on FaceBook said he thought the use of polymer Tantalum caps on C9/10/11 to be a problem, but I've not swapped out those caps yet because the pop every 10 seconds is totally gone when I boot System 6.0.7 or 6.0.8 without the v1.3 "Portable" control panel. If the caps caused the pop, then it's only logical to conclude the pop would never go away. It does go away when the v1.3 control panel is not used, so the caps cannot be related to the pop at all. As to whether the cap type has an impact on the other background noise, that's something I've not experimented with yet.

Anyway...

@SuperSVGA , please let me know your thoughts in light of this. And please know I really do appreciate all your kind help to date!

 

[Breezy]

There are grounds all over the motherboard, so the best way I can answer that question is with the...

1. Schematic here (keeping in mind I have a 5126, and the schematic is for the 5120).

2. 5126 board Photo below (by Bruce Rayne of Branchus Creations, not my board)

View attachment 15968

This is interesting because I, too, thought this seemed like a grounding issue. James, I would start by measuring the resistance between ground pin-1 on each sound chip to each other and also to other ground locations on the board. The intent will be to see if there is a difference in potential between locations. If you find a point with better ground, you could run a temporary wire clip between the other ground point and pin1 to determine if the sound response manifests differently. Measuring from pin-1 could also show you where resistance from corrosion or loose solder joints could be present. Because the content of the display causes the sound behavior to change so much, it makes me think that when the current draw increases or decreases, it is inducing or reducing voltage into the ground plane.

 

[SuperSVGA]

So that command line I tested last night is disabling only U14A, while leaving U15B & U15C enabled, correct?

The registers aren't really documented anywhere. I believe that bit switches between PWM and analog output on the ASC, but as you see the analog is not connected to anything at all.

There was also a thought I had on the ripple from the 12V regulator. The design guide for the LT1070 recommends low ESR for the output capacitors (C3 and C6 in this case) to reduce ripple, but I don't remember if I ever tested this.

I believe this is the VBL code that generates the "click" when it detects sound has been enabled:

        move.w  #61,(vblCount,A0) 
        movea.l (VIABase),A1 
        move.b  (A1),D0 
        andi.b  #%10000000,D1 
        move.b  ($E,A0),D1  
        eor.b   D0,D1 
        beq.b   .Exit
        move.b  D0,($E,A0)
        beq.b   .ASCEnabled
        movea.l (ASCBase),A1
        clr.b   (ascMode,A1) 
        rts
.ASCEnabled:
        movea.l (ASCBase),A0
        lea     (ASCRegBase,A0),A0
        cmpi.b  #1,(ascMode,A0)
        beq.b   .Exit
        clr.b   (ascClockRate,A0)
        tst.b   (ascType)
        bne.b   .L1
        clr.b   (ascChipControl,A0)
        bra.b   .L2
.L1:
        move.b  #1,(ascChipControl,A0)
.L2:
        move.b  #$FE,($830,A0)
        clr.b   ($831,A0)
        move.b  #$FE,($832,A0)
        clr.b   ($833,A0)
        move.b  #$FE,($834,A0)
        clr.b   ($835,A0)
        move.b  #$FE,($836,A0)
        clr.b   ($837,A0)
        move.b  #1,(ascMode,A0)
        bset.b  #ClearFIFO,(ascFifoControl,A0)
        bclr.b  #ClearFIFO,(ascFifoControl,A0)
.Exit:
        rts

It seems to be writing the "click" sound in the area of sound memory just after the standard 4 wave tables, so I'm not sure what this undocumented section is.

We can of course import that as raw data in Audacity:

It is quite quiet though, you'll really have to turn the volume up.

There are a few more locations that seem to write those same 8 bytes there, so there seem to be other methods that generate that sound. I may try and see if I can get a good trigger point on the sound to get a trace on the instructions leading up to it.

 

[JDW]

There was also a thought I had on the ripple from the 12V regulator. The design guide for the LT1070 recommends low ESR for the output capacitors (C3 and C6 in this case) to reduce ripple, but I don't remember if I ever tested this.

Meaning these two electrolytic caps...



As shown on the 5120 schematic...

And shown as C2 and C3 on the LT1070 datasheet...

It seems to be writing the "click" sound in the area of sound memory just after the standard 4 wave tables, so I'm not sure what this undocumented section is.

We can of course import that as raw data in Audacity:
View attachment 15989
View attachment 15990
It is quite quiet though, you'll really have to turn the volume up.

Yes, that is what I am hearing and is what I have been calling a "pop" sound.

 

[YMK]

This VBL task basically just turns the sound off when not in use, and then turns it on when something is using it.

I believe this is the VBL code that generates the "click" when it detects sound has been enabled:

It looks like the click code is a hack that prevents the audio circuit from going to sleep.

 

[JDW]

It looks like the click code is a hack that prevents the audio circuit from going to sleep.

The pop sound does NOT occur every 10 seconds when I boot into System 6.0.7 or 6.0.8 while also (1) loading "Portable" control panel version 1.2 (which is made for the 5120 and therefore lacks a "Brightness" slider, using Contrast instead), or (2) not loading any version "Portable" control panel at all.

I downloaded the 1.44MB System 7.1 installer floppy pair from Macintosh Garden and booted from one of the disks. As you know, System 7 uses different INITs than System 6. Sadly, I get the pop every 10 seconds while booted from that System 7.1 floppy. So it's not a problem exclusive to the v.1.3 "Portable" control panel that is used on System 6. And while some may say, "Just don't use System 7 or the v1.3 Portable control panel under System 6." The problem with that advice is that my backlight often is NOT LIT when I power on the machine. And I cannot activate the backlight with the v1.2 "Portable" control panel or with no control panel at all. Frustrating!!!

Of course, there's also all that loud background noise, mostly in the right channel too. But that seems to be a separate and distinct problem compared to the pop every 10 seconds. Like I said earlier, when I press INTERRUPT and use the "SM FB0802 1" command line, all the background noise vanishes, but the pop/tick every 10 seconds remains.

Lastly, despite all my begging on FaceBook and even here, only a scant few Portable owners have made time to test their machines to kindly provide feedback about their audio noise. The data pool is too tiny to raise any conclusions. What I really want is at least 2-3 more Portable owners (5126 especially -- all recapped and working well) to please test what I am testing to see if they too get the pop/tick sound every 10 seconds. The matter changes entirely if it is a problem common to ALL 5126 machines, as opposed to simply something that James Wages alone is having. I'm really quite desperate to receive such testing feedback!

 

[JDW]

I would start by measuring the resistance between ground pin-1 on each sound chip to each other and also to other ground locations on the board. The intent will be to see if there is a difference in potential between locations.

Thank you for the advice.
I just measured resistance between the two pin-1's on both SONY sound chips and individually to the Ground plane. No problems detected.