Video není dostupné.
Omlouváme se.
EEVblog
Vložit
- čas přidán 4. 09. 2024
- Why do many powered studio monitor speakers have a low level audible background noise?
This is an issue with many big and popular brand monitors like KRK, Yamaha, JBL, Alesis and many others, even top end brands like Focal.
Time for some investigation on the KRK Rokit 6 studio monitor speakers to track down where the noise is coming from.
NOTE: Not all monitors have this issue, many much higher end ones may not. It pays to check this first if you are going to spend decent money on studio monitors, as you may find it annoying.
Service manual and schematic for the Rokit G6:
eevblog.com/fil...
The black gunk of death: • EEVblog #1072 - KRK Ro...
Forum: www.eevblog.com...
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
Stuff I recommend:
kit.com/EEVblog/
Donate With Bitcoin & Other Crypto Currencies!
www.eevblog.co...
T-Shirts: teespring.com/s...
Likecoin - Coins for Likes: likecoin.pro/@...
💗 Likecoin - Coins for Likes: likecoin.pro/@...
NOTE: Some studio monitors will have an "auto-mute" function that is designed to make the speaker silent without an input signal, so the background noise isn't heard. This is a different feature to power-on mute. But the noise is still there in the amplifier and returns when you play music and goes out of mute mode.
Aye, try listening to audio books/spoken voice or any audio which is "on" but has lots of silence in it. You can hear the noise floor in the spaces.
I have a pair of Rocket 6 G2s and I've had to stop using them. I left them powered on for several years and they slowly developed a louder (but still faint) hum that you can detect from several feet away. I plan to replace some of the capacitors on the amplifier to higher quality ones, but I don't know if it's going to help. I would be very interested in a video where you attempt to lower the noise floor on this Rocket 6 by upgrading the components. :)
And before anyone says it, I've already verified my cables and power source.
Junction avalanche noise - Great for white noise source on a synth but I'm with Frank Zappa when it comes to amps. I'm one of the few unlucky people who bought a Marshall hybrid valve preamp/transistor power amp. What a piece of crap that turned out to be!
Sounds like it is ripple from the power supply to me. The very first thing I would do is stick a larger reservoir cap across the power to see if the noise goes down. It sounds like 100hz to me, which would be ripple from the full bridge in the power supply.
Yes, that's what I meant actually. I'm wondering if switching to an "audio grade" cap will sound better than a generic quality one. I have no experience with that sort of thing.
Did you check the woofer for snakes as a source of hiss? It is Australia after all!
I saw some odd looking wiring...
Don't be silly, snakes don't woof. But do drop bears hiss?
I had a quick look down my pants and yes all us Australian’s are hung like a horse, python next time thanks.
"Hung like a horse", is that judged by the colour or the shape?
the hiss noise is a little kid playing you
What a way to merge my passions in both electrical engineering and music production together.
An EE course usually includes knowledge in signal processing, so there you go.
I'm an EE undergrad interested in music tech too.
maybe build a synth
Yes, go ahead and build an eurorack modular synth! Sort modules by DIY projects and other functions. www.modulargrid.net/e/modules/browser?SearchName=&SearchVendor=&SearchFunction=&SearchSecondaryfunction=&SearchTe=&SearchTemethod=max&SearchBuildtype=d&SearchLifecycle=&SearchSet=&SearchMarketplace=&SearchIsmodeled=0&SearchShowothers=1&SearchShow1u=0&order=newest&direction=asc
@@haraldlindohf4032 maybe someday, but for now Serum has everything I'll need to make sounds for years to come
well now you have to do a video of actually fixing the noise
Yes!!!!
He can't. Or at least for cheap.
@@johnyang799 there must be a way
@@JustUploads Sure. Rebuild the amplification circuit. Yes.
It would be a bit of work, but honestly would make for some great content!
We from the VoIP department call that comfort noise, so you know the equipment is working.
Can we have the FINAL conclusion of where the noise comes from, and perhaps what can be done to alleviate it to some extent?
Nothing really, the amp is fixed. Best you could do is get a high end low thd+d power amp and passive monitors. But those are almost extinct.
Dave, really appreciate your audio-related vids. Thanks!
Glad you like them.
Most noise issues are due to bad PCB design and self noise.
It's especially a matter of designing the ground properly. I have worked on designs myself were that difference was 6dB or more.
That last one can easily be calculated with the gain aa Dave already mentioned. Don't forget that a tweeter easily has a sensitivity of around 90 dB/W. Maybe people also understand how difficult that's gonna be with super sensitive compression drivers.
A way to fix this for low power studio monitors is to add a resistor divider (called a L-PAD) after the power amp. Obviously you have to compensate for the gain loss in the pre-amp. The value of these resistors are critical because of the impedance. Wrong values will change the frequency response. These resistors barely add extra noise since the values are low (typically < 10 ohm)
That is some info that Dave could have told in the video, it would have been a little bit more interessting :)
PCB layout can affect mains related hum and buzz. Random, uncorrelated noise is typically design issues, thermal noise related either from the semiconductors or resistors. Bad gain structure compromises can make noise worse as well.
One possible way to reduce the amp noise, *IF* the power amp's residual noise reduces when its feedback is increased (value reduced) to reduce its gain....
Try reducing the gain of the power amplifier and then boost its input signal level by adding some gain previous to the amplifier. That is, if the amplifier input can then take a bit more input and the added gain doesn't bring back all the noise that was reduced in the amplifier.
That's not really the issue in this case. I shorted out the input resistor and there was nothing wrong with the layout. The noise was obviously inherent in the power amp chip.
EEVblog .Shorting out the input resistor is only isolating the power amp stage from the rest.
(the proper way btw, is shorting the input to ground)
PCB layout issues will still be noticeable, I don't understand why you think that just shorting the input would change that?
The power amp chip still has some circuitry that belongs to the signal ground and not the power ground.
Anyway, I was just giving a general explanation (so people can actually learn from it).
But I woudln't be surprised if it's bad PCB grounding in this case as well (not saying it is, i am only saying it wouldn't surprise me)
I have seen thousands of audio designs in my life and even the big brands make a big mess.
Did you short out the amplifier's feedback resistor ? 10K I think ? 12K Ohms actually from the schematic
Ooops... Maybe just reduce the value in case the amp isn't unity gain stable :)
Years ago, and not sure where I read this, but the article I read said that IC amplifiers have more noise to them than using discrete components. My old Mackie HR824's are super quiet. The only use IC's for the OP amp preamplifiers and output is power transistors. Well anyway, one day one of my HR824's went dead, so I went out and bought Mackie MR5MK3 monitors. They were low priced ones, but I just needed something to use while I fix the bad one. I did notice that they had a pretty audible noise floor. Opened one of them up, and sure enough, the power amp section used IC's. I was sure glad to get back to my 824's. Oh, before I forget, I also remember talking about it with my local ham radio group way back in the mid 70's. I'm old.
i think it might be a bit more complicated than a IC versus discrete components situation. im not as experienced but i found that i can never adhere to simple distinctions like that because i get bit in the backside with counter examples.
stings, man :/
Notice no roll-off in the room background noise, so the roll-off is likely in the tweeter output.
Good point
Eh, if the noise comes from other sources that aren't subject to the same signal path as the tweeter output (for instance, ADC noise?) then your conclusion isn't quite right
This is fully analog, what ADC noise?
I mean thing's gotta go through an ADC before it goes into a computer
@dm, okay, makes sense.
But still, the noise was different between the tweeter and the woofer with the same mic+adc+....
I could spend all day listening to Dave say "woofer".
One possible cause of the hiss is that these and many powered monitors use Class D power amps. So there is an inherent switching oscillator that is part of the power output stage. These are great sources of white noise.
Waiting on Mr. Carlson Lab input...
Given your conclusion that the power amp is the culprit,it would be interesting to connect the pre-amp/filter board to a separate amplifier and see if there is any improvement.
That is why I bought a pair of ELAC bookshelf speakers and the cheapest crown amp, no noise, sound just as good or better than most powered monitors
I bought secondhand Rotel RA-05 amp + a Dac and had a pair of ELACs so i know what u talking about :)
Anyway Studio monitors are a tool to make your mix sound correct when making music while HIFI-speakers often boost some frequencies to colour the sound and make it more plesant to listen to.
That hiss isn't a bug, it's a feature! It helps a audio tech quickly determine if the speaker is working and if the problem he's looking for is somewhere else. ;)
funny you say that, nut it has held true for me many times. listen, hear hiss, quickly jump and look upstream for the issue :))
Very interesting, glad to know I'm not crazy for hearing the noise coming through my monitors when nothing is playing. I can't hear it unless I'm really close to it, so thanks for doing this deep dive and tearing everything down with a very thorough explanation.
It's quite easy. Most noise in audio amplifiers comes from the preamp. In the usual preamp + power amp combinations the output of the preamp is attenuated by the volume control.
The volume control attenuates both the signal and the noise before it feeds into the power amp. This is true for separate preamp + power amp stages as well as integrated amplifiers.
But for active speakers like this studio monitor the preamp is not attenuated (let's say the volume control is at 100 all the time), and volume control is done by the user by lowering the signal at the preamp input.
That's why the noise level stays the same, because the output of the preamp has no attenuator and therefore always keeps the same gain level.
If you turn the volume control to 100 on any preamp + power amp combination without a signal at the input, the hiss is just as loud as those studio monitors.
It's just the volume control that turns the hiss down with the audio in that case, which does not apply to active speakers.
I'm in the process of setting up my powered monitors with a dac that effictively doubles as a pre, with a volume control. Are you saying that I would be able to reduce the noise if I keep the volume on my monitors near max, and reduce the volume at the pre stage?
But note that he got no change in the hiss and voltage measurements when he completely disconnected the power amp, or shorted the power amp's input pin. So much for your theories. EVERY active circuit has noise. It's just a matter of "how much is tolerable?" However, I agree with you that MOST noise comes from the first gain stage. In this case? No. He's got a relatively noise power amplifier, which is typical of these very old, very cheap, and very obsolete TDA power amplifiers from yesteryear.
This is a linear amp, not Class D switcher.
It would be really interesting to have a video done on how noise enters into an otherwise very quiet amplifier circuit, seems like there’s a lot of potential for noise that isn’t obvious.
I wonder if it could be improved by lowering the power amp gain (changing the feedback resistors) and making up for it by increasing the gain of the prior stages?
I enjoyed this departure from your normal content, it's really cool to see the troubleshooting approach in action even (or maybe especially) when there's no 'fault' - it puts a lot of things in a real world context that's not easy to see (for me at least :)) when doing a 'regular teardown'.
Klipsch speakers do this as well. Ear to the tweeter, with no audio playing ( connected still ) produces a white-noise
If an amplifier doesn't produce ANY noise, it means it's unpowered.
I work for the local cable company, and my job lately consists mostly of finding sources of Return Path Ingress (5-65MHz) in the network. Nothing better than coming home from work after a long day to watch dave do the same thing on a much smaller scale :D
First of all, you did not mention any of the "top brands", you mentioned affordable home studio monitors which are different animals all together. Genelec, PSI, Barefoot are some of the top brands. And two, what kind of AC line conditioner are you using? IF you run those KRK's through a Furman with L>I>F>T> that noise goes away because the problem is the filthy hydro power and no one who values their recordings at all would plug anything to do with recording into a wall socket! That will give you 3Db of noise that could be used for music or what ever you are recording. But I would think someone like you would know that, pretty surprised that you don't.
No mention of power supply rejection ratio of the power amp, nor the noise on the power rails... Figure 8 of the TDA2052 data sheet appears to show only -20dB to -40dB PSRR at 5kHz. Any chance we could get a noise measurement on the power rails?
Along the same line of thought, if the output stage is all NPN then the neg rail modulates the bottom emitter which makes a signal between it and the base. I have seen it with some opamps. I agree that is probably the power supply rejection of that cheap power module. Cool stuff.
I've fixed this issue with my studio monitors. I tried replacing the electrolytic caps, didn't work. I noticed when lifting the center tap of the power transformer from the board the noise went away. Turns out that the system wasn't star grounded. I fixed it by soldering the center tap to the pin of one of the bulk filtering capacitors.
I would try lifting then jumping the rh100/ch102 node to pin 4 direct, and/or lifting the 2 earth grounds on the tweeter output
@@DarthJe5us It's the transformer pin in the middle, between secondary coil 1 and 2.
I noticed quite a few unpopulated cap positions on PCB, probably bean counters removing "unnecessary" filter caps
Simple answer. Because the built-in amps are always on MAX. All ACTIVE monitors do this. The volume is usually adjusted in the studio before the signal hits the speakers. You can replicate this same noise floor on high-quality PASSIVE HiFi speakers by turning your separate amplifier up to maximum and keeping it there all the time while making all your volume adjustments before the input to the amp. It'll be noisy as FxcK. It's just bad gain structure.
I will now watch the video.
(love your channel by the way)
As per your study, maybe not always the HISS is coming from Preamplifiers or Corrections (as you name it FrontEnd). First, yes, ribon cable is really not appropriate for analog signals. Replace with coax. Then, the OpAmps. They used very cheap OpAmps... Maybe in the same range of price they could place LM833, with considerably much better results. Try to extract all of them, place sockets and LM833. See results, make nice measurements. Then replace with LT1678 and come back with some measurements. And then replace the ordinary carbon resistors with metal oxide. And replace the ElCeapo Aluminum capacitors with Tantalum drops of 10u. At least, I am doing this exercise when I optimize commercial audio stuff. Cheers mate!
As long as the hiss is proportional to output volume, its not too much of a concern. That is, if you're using them at high volume, you're probably standing back a distance from them and the sound will swamp the noise. That might seem like an obvious point, but take for example the EV ETX series of PA speakers - I've found they have a constant, quite loud hiss that is independent of any of the volume controls. So you have to be using them at theater volume levels to make the sound/noise ratio acceptable, but I wan't to use them at lower volumes a lot of the time, and it's very disappointing to have all that hiss there. There's a lot of DSP going on in them, maybe the output circuitry is "full on" all the time.
The hiss is not proportional, it's a fixed base noise floor level. Totally insignificant when playing content.
Just tested my Yamaha HS80M speakers. It's probably identical to what you're experiencing: base level hiss independent of the input level gain. So in these cases, the output gain is fixed regardless of input level, theoretically you could reduce output noise by lowering the gain via RH103-RH104? Just hypothetically if you were always using them in close quarters at low volumes.
Yeah, I would also expect the noise to decrease if you lowered the gain on the power amplifier. Only way to be able to tell for certain is to test it, though.
Karreth that's if the noise is from before volume pot. after volume pot, noise volume is fixed.
Dave has a video about op amp noise, it's super awesome. You can use the knowledge, and build a amp with nearly no noise :D square root rule.
You could build a mute standby timer with some circuitry! That would be awesome! So that it mutes the amplifier when there's no signal present for a little bit of time, and re-enables the amplifier when signal is detected. I would love to see that.
My 1997 Mackie HR824 Gen 1 are absolutely full quieting. No hiss, no buzzing, no nothing except music. This is with pre-amp mixing connected.
Me too, the Mackie make no discernable noise from 3 feet away.
These things are not high grade monitors, what they are are a cobbled together car stereo grade output with a active seperation of input frequencies. Those amplifier modules are designed for mobile usage with minimal bias etc to save power, THD and all that audiophile "nonsense" (like low noise) are not parameters of the design. You ask yourself why do all these E-E have all that "expensive equipment" when you can get the same things on Ebay for a lot less. Dave is currently in the grip of the first stages of being an audiophile.
so, for a newbie, of what are you talking about when you say
"when you can get the same things on Ebay for a lot less. "
any tips and suggestions?
I recently bought a headphone amp that was really noisy no matter what I did. Same kind of thing. The output amplifier had a fixed, somewhat high gain, so the input noise voltage was always amplified by the same amount regardless of the volume setting. As luck would have it, I had another headphone amp that had more/less the same components, but the volume set the gain of the output amp. On that one the noise was inaudible when the volume was down, but was pretty comparable when turned up, but at least it was less obnoxious when running at reasonable volume levels.
i suggest to decrease the gain of output amplifier to something like 10... to change 12k resistor for 6k.. it will improve gain staging too
@Ruslan
Yes, gain balancing would help, but one can't lower the gain that much. The TDA2052 requires gain to be at least 30dB (32x) for stability.
so its better to put lm3886 there:)
The buzz could be a ground loop. I had the same trouble with my Presonus active monitors. For some reason I had less trouble using passives and a separate amp.
Use an audio isolation transformer between your signal source and the speaker amp's input. This breaks the loop and usually fixes the ground loop problems.
Not much to improve by having a pre amplifier, because the TDA2052 also has a 30Db minimum gain requirement. All these kind of integrated power amplifiers seem to be in the same "noise category", even OPAxxx power op amps, hard to find something 20Db better, this is why all studio monitors are the same. You would have to do a discrete power amplifier with a good low noise op amp to get better.
Idea - what if you cool down whole board? I saw even webpage of guy, who put peltier cell into DSLR camera to reduce noise. Would be interesting to see how noise of semiconductors changes with temperature.
I had a pair of Genelec (8040A's), and they had a very faint hum. Then I down graded to M-Audio, and they have a much more noticeable hum. So apparently it is something that can be improved if you give the engineers more money to work with. The currently available verson of that speaker is the 8040B, which costs $1,200/each (US).
In other words, you get what you pay for.
Stringer: This isn't a very informative statement. My M-Audios are tri-amped, with 220Amps...they're "fun", because who doesn't want 8" woofers and 200+ amps per speaker? In stark contrast, there are speakers like the Avantone MixCubes, with single 5.25" drivers and 60amps...I guarantee the amps in the Avantones are way cleaner. Different products, different uses, different compromises. The M-Audios were $750/pair, and the Avantones are $450/pair... So where does that put us on the "you get what you pay for scale?" :)
fredy gump, you said yourself that "it is something that can be improved if you give the engineers more money to work with." While some engineers are more talented than others, and some brands will spend less on engineering and more on advertising and other things that do not add value, with all other factors being equal, you said it yourself albeit in different words. You pay more money for engineering, you get more (or better) product. In other words, you get what you pay for.
I can't help you with your sudden case of amnesia, but am amused that you's shoot your own point down just to be antisocial.
Okay, I agree with you...you get what you pay for...and that means nothing to an engineer. Engineers don't make the "best" thing they can possibly make...they do their best to meet all the design requirements that the customer gave them! If the requirement is to make a $200 monitor speaker with extremely low noise, that is what you'll get. And you may be surprised that some expensive monitors actually sound worse...and they sound worse in a very particular way...on purpose! There is a reason for it, but that's a whole 'nother subject.
The power amp is coupled as a inverting opamp with a gain of 1+ RH104 / RH103: 12k / 456 ohms = 27, that's a lot of gain, what I learned in Audio design class, was never leave feedback resistor unbypassed, have not read the data sheet the amp might not like it, but a small feedback cap over RH104 to cut down on the power amp, frequency response could be a good idea, I know that there is some parasitic capacitance in the PCB, but Add 330pF over the 12k resistor, would cut down the frequency response to 40Khz 1/(2 x PI x 12k x 330p ) = 40,2Khz This could lead to instability of would properly do with a good read of the data sheet before doing this :)
I figured out bypassing the feedback resistor by trial. Old TDA2030As didn't need it but the newer drop-in replacements (LM1875) would straight up oscillate in the ultrasonics without it.
You might just be able to change RH104 and RH103 to different values the tweeter really does need a low frequency rool off at 34Hz (CH104 is 10µF) 1 / (2 x PI x 456 x 10µ) I would try dividing the resistors by two, 6,2Kohm and 220ohm would cut down the thermal noise of RH104
Cool I was just thinking out loud :) My other respond might be worth a try :)
Can't get the feedback resistor too small since there's a series resistor on the output of the opamp feeding the power amp.
Given that dave said it's probably not the input noise characteristic, there is probably noise being put into the amp, such as through resistors or oscillation.
I think the noise is inherent in the amplifier output stage, as they do not have an output refereed noise specs in the datasheet like some other higher end amps.
"And the [hiss], that was planted in my brain... still remains
Within the sound of silence"
I have some personal findings on high frequency hiss coming as Common Mode noise from the power supply rectifier diodes. It is the recovery spikes - microsecond or less - that by themselves are not audible, but when they get into the amplifier, they cause issues by overloading a stage which has high native gain moderated by feedback. But the feedback has a propagation delay, so it does not attenuate all frequencies alike and some not at all. In fact, the feedback may approach turning the stage into an oscillator.. Or you could relate it to Intermodulation Distortion. Anyway, I would look for ways to add some (lossy) ferrites to certain places, like the emitter or source leads of a few transistors.
all class D amplifiers have a great deal of HF switchmode noise in the output!
that is part of the design, but clearly some are noisier than others ...
TDA2050 has only 45 dB power supply voltage rejection.. LM3886 has 120dB typical, and no less than 80dB. That means TDA2050 cannot suppress 100 Hz hum from unstabilized power supplies well enough..
The number in the title is funny, because the LT1118 was the only linear regulator on the market that I could find, match the specs I needed, and could SINK current. Nice. :)
My Rokit 5s had a noise problem due to a nearby plasma TV. With the TV on there was a noticeable buzz in the woofers. Wasn't a ground loop or mains power. The buzz got louder the closer the speakers were to the TV. Eventually worked out an aluminium foil "shield" wrapping the speaker and attached to ground removed 99% of the buzz.
I fixed the speakers by disassembling them and lining the inside with adhesive copper foil then attaching a ground wire to the amplifier board using lugs and screws. The buzz is now completely gone.
Where did you attach that ground and what did you ground it back out to? Does an external ground make sense with what you are saying or did you just ground it back to the power supply?
@@DavidTOBerry They're powered speakers with 3-pin kettle cable for 240V. There was already a ground lug running to the amplifier board so I ran another wire and used a spiky washer and a self-tapping screw to get good contact with the copper foil. The copper foil is grounded to the mains via the 240V plug, same as the amplifier board.
I knew it wasn't a ground loop because the hum got larger simply by proximity to the TV. I could pickup the speaker and walk towards the TV and it hummed louder. So I knew it was RF interference and shielding the amp with grounded copper foil proved it.
The TV was plasma and used a lot of watts and I think very high voltages. I've long since replaced that TV with a low-power LCD and probably the problem doesn't exist with the new TV, but there's no harm leaving the foil in there, it doesn't make it worse.
@@nhand42 Thanks Nathan. Much appreciated!
Often minor HF instability reveals itself as hiss/noise A 100 PF(?) in // with feedback resistor may help. It is unusual for a power amp to generate that much noise , even with a sensitive driver connected. Though most of those studio monitors are horrible - just like these....
Clouse loop gain has to be upper 30dB, in this monitors 28dB. This is stability
@@vitalley The closed loop gain of the amp assuming half a volt sensitivity and 100W out is around 34db (with 8ohm load for 100W but voltage gains are same of course). Not sure I understand what gain has that to do with HF oscillation? What is happening in the audio band may have nothing at all to do with the HF. That also describes almost every audio amp ever made? Many are motor boating and whistling and hissing for various reasons and some are wide band and have a -3db point above 100khz where it is difficult to ensure phase is kept in the stable region.
I'm probably missing something obvious - I often do - so forgive me if I missed your point. Regards.
Looking at amplifier noise numbers, those 2uV are specified for A-weighted, 20Hz-20KHz. If your meter has a 100kHz bandwidth, then you should divide the output voltage by a factor of roughly 3 (sqrt(5) for the badwidth, and around 1.4 (3dB) for the A-weighting). That makes the measured value much more plausible from the amp specs.
In order for your meter readings to make sense with the amplifier data sheet noise spec (20Hz-20KHz), it needs to be reading through a 20KHz LP filter. Your 246uV measured noise is mostly ultrasonic.
If the amplifier is operating at the datasheet "typical" value of 3uV Ein with 20KHz bandwidth, output noise should be 82uV. Your meter bandwidth would then be estimated as 20KHz * (246/82)^2 or about 180KHz.
When I make audio noise measurements, I use my Tek scope in high-res sampling mode, at a sample rate of 50Ks/sec. That gives a 25KHz bandwidth, then I make the small correction to come up with the 20KHz BW figures.
Remember those huge (and doubtless prohibitively expensive) extremely high value resistors in the Keithley 617? That's what it takes to eliminate this kind of noise.
If it's white noise in an audio amplifier, 99.44% that's thermal noise through a resistor. Since it's still there with the preamp disconnected, it's gotta be the feedback resistor. Heavily over-spec'd resistors help cut that down, which is why you'll often find 1/2W or higher rated metal film resistors throughout boutique amplifiers, including the preamp signal path where the current is naff-all. Some passive component choices in higher-end amps aren't entirely due to juju or woo.
Tried using 1/4W Wun Hunlo carbon film resistors in a few dirt boxes (guitar distortion pedals) a while ago. They sounded even worse than you'd expect.
I thought thermal noise increased with the value of the resistance
daycounter.com/Calculators/Thermal-Noise-Calculator.phtml
You made me curious so I checked out my British handmade Quested S7 active monitors, which the company I worked for also sold a few times to Abbey Road Studios. Here is the datasheet info:
POWER AMPLIFIER
LF output power: > 120W RMS continuous (note 1)
HF output power: > 70W RMS continuous (note 1)
THD: < 0.03% at levels up to 1 dB below clip, 20 Hz-20 kHz. typ 0.005% @ 1kHz.
Hum & Noise: -100 dB referred to clip
NOTE 1: The continuous rating is for a period not exceeding 5 minutes with unrestricted airflow
(100 mm clearance) around the amplifier heatsinks and an ambient temperature not
exceeding 30ºC.
POWER REQUIREMENTS
Voltage: Set by internal plugs: nominal 115V or 230V @ 50-60 Hz AC
230V setting: min 160V; restricted output power: max 255V
115V setting: min 80V; restricted output power: max 127V
CONSUMPTION
Quiescent: 18W
Typical use: 75W (average music)
Max: 140W (average music)
Note: 0 dBu = 775 mV into open circuit
Dimensions: 244(w) x 347(h) x 302(d) mm
Weight: 11.7 kgs
Drivers: Bass cone - 1 x 165 mm (6
1
⁄2”)
High Frequency soft dome - 1 x 28 mm (1
1
⁄8”)
Maximum SPL: 115 dB (c) continuous pink noise @ 1 m (121 dB (c) per pair RMS music)
Frequency Response: 65 Hz-22 kHz ± 2 dB
Connector: XLR and
1
⁄4 inch Jack combo socket
Impedance: 10k ohm electronically balanced with RF filters
Wiring: Pin 1 ground Sleeve
Pin 2 hot Tip
Pin 3 cold Ring
Sensitivity: -12 dBu to +6 dBu for 96 dB SPL @ 1 m set by 10 position rotary switch on rear
Subsonic: -3dB @ 30 Hz, 24dB/oct
Ultrasonic: -3dB @ 75kHz, 4dB/oct
Crossover: 1k19
LF EQ: -2/-4 @ 65 Hz
HF EQ: +/-2dB @ 10 kHz
INDICATORS
Power on: Front mounted green LED indicates pow
Clip: Red LED 2 dB before amplifier clip.
At 7:53, you state that the blue AC meter has a BW > 100 kHz, which is fine. But note that En is specified at 3uV when measured over a BW of 20 kHz. To get your meter readings to jibe better with the the datasheet (and your ears), you should have added a lowpass filter to the input of the meter to limit its BW to 20 kHz. Without such limiting, your RMS meter will measure and add more HF noise that your calculations and the datasheet do not assume! However, since all of the measurements were done using the wide BW meter, then the amp's datasheet numbers as well as your thermal noise calculations (11:16) have to be expanded out to this assumed meter BW of 100 kHz (+/-).
I'm pretty sure that the source of your noise is En and the thermal noise of (RH9//RH10//RH100 = 1Kohm). You can't just add these two noise sources. You have to do the square root of the sum of the squares. When you do that, then En dwarfs the resistor noise, so it comes down to just En. First you expand En to a BW of 100 kHz, which you do by multiplying it by sqrt(100kHz/20kHz) = 2.24. En is now 3*2.24 = 6.7uV. This noise gets multiplied by the gain of the amplifier, (RH4/RH3 + 1)=27, giving a noise value of 6.7*27 = 181uV.
This is getting close to the measured value, but now you have to consider the potential error sources in the calculations. The biggest one is the BW of the blue meter. If it is really 180 kHz (Dave said "100kHz-plus bandwidth"), then the calculation bumps up to 245 mV. Then there is the unspecified error of the meter when operating at this frequency. The user manual claims accuracy out to only 5kHz. What happens to signals at 100 kHz? 200 kHz? Let's face it, this meter is being used far beyond its stated calibrated capabilities. Note to Dave: you really need to measure the freq response of this meter out to the frequencies you intend to test with it. You certainly have the calibrated instruments to do it.
Another error source is En itself. It has a very wide tolerance to it (3uV typ, 10uV max). If this particular amplifier's noise is 4uV, a very real possibility, then you also get a 245 uV calculated result.
One last thought: much of the measurements shown here are way too high only because who can hear anything beyond 20 kHz? Why include the noise that is out there beyond 20 kHz? To get a true measure of the voltage that counts, namely "hiss" that we can hear, the measurements should be reduced to about 250uV * sqrt(20kHz/100kHz) = 120uVrms (20kHz-20Hz).
*BOTTOM LINE: the source of the hiss in this video is the power amplifier itself.* These very old TDA amplifiers were well-known in their times for being noise generators. They are obsolete today for good reason!
Is there any possible fix? Some have correctly said to lower the gain of the power amplifier, which would lower the output power by 3dB, which is perceptible. But then you have to raise the gain of the preamp by 2 to get the same total gain you had before. Maybe the design has too much gain as it is? Is there enough headroom in the preamp to do this? I'd disconnect the speakers (to save my ears) and replace them with dummy load resistors, then attach a signal source and start driving the system to its limits. Adjust the gains of the preamp and the power amp such that they both begin to clip at the same time. See how much headroom you have in each subsystem and get them "balanced" this way. Ideally the designers did that in the first place. Such a mod could be an entire video in itself.
Calculating noise is a real black art involving a lot of squares and square roots, combined with recognizing Thevenin equivalent circuits. I understand a little bit about it, but I make no claims to being an ace at it!
me at 0:00 : it's the power amplifier
me at 23:28 : it's the power amplifier
What the hell kind of "studio monitor" comes with volume controls ? And some HF amplification too ? Seems like a misnomer for simple overpriced active speaker. Although its guts DO look pretty good, even though the design is more of "classic" approach. But they did skimp on op-amps a little, also socketed them as if they expected users to change them.
By the way, after reading a tons of datasheets I concluded that best stereo PDIP-8 (socked here) op-amp is LM4562 (aka LME49860), best SOIC-8 are OPA1612 and LMP8672. Best headphone drivers are TPA6120 and MAX97220 (not to be confused with MAX9722). THS4551, OPA2365 and OPA2277 are also pretty good but for mono you would want OPA637. As for power amps: TAS5342, TPA3250-TPA3255, LM3886.
It's fairly typical these days to have a volume control and shelf adjustments for HF and LF built into the speaker for basic compensation. The market these are aimed at would not have sophisticated correction or room treatments so the shelf filters may be the only correction the user has.
+dwindeyer: It is typical but "monitor" speakers and headphones are supposed to be calibrated for flat response curve and hardwired for power/volume that is balanced for maximum dynamic range at maximum input signal while maintaining minimal distortion+noise (which is why they are fully "active" and self-contained, mono-only), aren't they ? One may argue that it is for compensation of room's imperfections but it's too simplistic for such task, it's naive to think that any knob can account for your room not being studio-grade. And if you have to compensate, you better off doing it in source's software or in-between source and "monitor".
They have a clearly defined bypass point on all the controls, so I'm not sure exactly what your objection is. The choice to use the controls or not is there, if you want full bypass, set everything to zero.
+dwindeyer: "Bypass" or simply zero-setting ? Because each of those filters would have op-amp and potentiometer and each of those is another layer of unnecessary distortion and noise. But, most importantly, it's the sign that manufacturer might not take responsibility for this being "drop-in" product, ready for use out of the box and conformant to "monitor" expectations, hence my comment about it looking like overpriced active speakers. Although, active crossover and multiple power-amps IS some serious business. Certainly better than most small "active" "HiFi" stereo crap you can get.
I've got similar monitors and I really don't need so much gain on the main amplifiers. Is there a relatively easy resistor to change out there to lower the gain? Lower gain would probably mean less noise I assume.
Because they are crap? Buy some genuinely industry standard Genelecs if you really have to have powered monitors. I use single ended valve amps with monitor speakers. BTW, Yamaha NS-10s are used in music studios as they are awful - if it sounds good on them it'll sound good on anything. Though they have good timing.
Had some good like adding snubber caps (x4) to the bridge rectifier in custom studio monitors back in the late 90's & early 2000s, as well as making sure any analog circuitry decoupling/power-bypass was sufficient. Had thought most of the noise floor was johnson/thermal, until the improvements dropped the noisefloor below isolated recording studio ambient, though that may be because we were running 600 ohm balanced, so didn;t have the regular 10k-100k input impedances involved
Use an net filter for the power and you get rid of that noice
So its from the amps .. thanks ! I was wondering if this noise is not a ground loop , but its not that loud ...
Both of my Roland CM 30 monitors have a humming sound with no connections & all the volumes set to zero. I did replace the speakers that don’t flub out with the volumes half way up. Same ohms, size & 60 watt instead of 30 watt. They sound quite a bit better , but now I can hear the humming more.
Absolutely great video - thanks for taking us into that detailed step-by-step journey, and narrowing it down! :)
The equivalent input noise spec for the power amp chip would be with the input shorted. Leaving the input open presents a high impedance to the power amp chip's bipolar input stage, making the BJTs' current noise the dominant noise source. Try shorting the inputs and you'll be measuring the EIN voltage.
Also, notice that the EIN spec says "A curve", which means that their measurement was done with an "A weighted" filter on the input of their analyzer (basically a band reject filter intended to replicate the human ear's partiality to the 2 - 4 kHz range). If your measuring instrument is measuring at a wide bandwidth and doesn't have an "A weighted" filter, then the readings will be higher.
I did short the input resistor in the video, the noise didn't change. The datasheet also has 20KHz bandwidth noise figures, I mentioned that in the video too.
How did you get verified with only 3 subs?
@@notaname8140 he must be the dedicated technician of the electrical company providing supply for YT servers, hated but respected LOL
sooo ... you'll follow through and chase down the problem with the power amp section next? or that's that?
KRK is not a good reference point to use. KRK use a glue on the PCB's that goes conductive over time and causes all kinds of chaos, it is best to avoid using them overall. It would be interesting to see a side by side comparison with this power amp chip used in the KRK and a LM3886, the LM3886 is used in quite a lot of HF drive designs.
i have to admit that i would like a full course of you on a quest to build the best audio amplifier possible.. :) .. you are the best on the net..
Hack the thing. Attach an audio detector that'll switch on or switch off the power amp board.
I've mucked around with the TDA2030A for my active speakers. I found that the noise is very voltage dependent. I dropped mine from 12V to 9V now the noise is OK. Perhaps this is also a thing with the TDA2052.
I bet the noise is from grounding layout. The high current driver returns should go directly to the psu smoothing cap center taps, while the low current signal returns should come together at a star further away from the caps. Must avoid high current return leaking in to the low level signal returns. Also mains ground should join at the low current star point.
One issue is that many of these monitors have very sensitive drivers. If you look at specs on most pro audio drivers, they are much more efficient than home theatre speakers.
It's a box with ok components,crammed in with an ok amp,like Class D,so what would we expect?I use passive monitors,as in 3way,with no controls,and a separate AB class amp..We are dealing with price/performance points and the convenience of not having to use a power amp.We already know that the more stuff you cram inside a box together,close to each other,the more noise,interference,and crosstalk you get..That was figured out even with computers,not using the onboard audio,and using external interface audio...
Did you try cooling the amplifier with some canned air to see if that affected the noise?
heat can amplify the buzzing noise.
Hello, Dave could you please test just the output stage of the IC ? If the datasheet is correct you can choose which stage is off by voltage level on the mute control input, in some state only the input is muted (as you have thought originally) with other voltage level output stage is off too if I understood it well. So it would be interesting to know which stage in the power amp is responsible for that noise.
Made total sense. I have the KRK 10/3's and the noise floor in those is a real shocker have thought about changing out the mid and hi amps for something far less noise inherent but finding a match is difficult. The only tri-amped monitors I have heard of in this price point are the "Keli 8"v2" ...they are VERY quiet. ;)
i have rokit 5's and yes there is a hiss (i was also under the assumption that they all do this) but i can tell where yer speakers have been repaired. the noise they make is different than any of the ones ive had. more buzz and less hiss than a fresh set.
to clarify for any haters readign this (yea last time i commented on my rokit speakers dave himself had to come shut the guy up) i can confirm from use that my speakers are acting the same way as have all my rokits ive owned. my point is a fresh set makes a slightly different sound than the set he recorded. a bit more buzzy than my own set. and im going to assume that difference in noise is a difference in the health of the caps or something inside
I think ST have superseded the TDA2052 with the TDA7296? Notably the new amplifier spec emphasises "VERY LOW NOISE" in the feature list - I wonder if they've solved the problem. It has a few extra pins with a new Standby function and separate pins for output stage power, but it looks as though it would be straightforward to swap in with some pin remapping. Challenge for you!
Had this problem with my KRKs and got a ground delete end for each plug which cleared it right up. The end converts the 3 prong USA style plug to a 2 prong. I understand not having it grounded could fry it in an electrical storm but it’s worth the risk since it completely removed all hiss
I did the same to my Yamaha hs50 monitors and they have been running with the grounds lifted for 5 years now without any problems.
ALL powered monitors have a self generated noise. Some have more (KRK, Dynaudio, pretty loud) some have less (Neumann, really quiet).
The reason for this is very simple: A powered monitor has its amp set to max.
However you should get a bit lower self noise (not completely silent though) by lowering the volume on the monitor. So maybe some of the noise actually is not only your monitors self generated noise but also from electrical interference? (As with that obvious hum, that's nothing you should expect from a monitor and most likely caused by interference in your room?)
The numbers you would be looking for is stated as “self generated noise”, but for some reason not many monitor manufactures specify this. (My guess: they don't want to because it might influence the customers negatively.)
Personally I think testers in media should talk more about this-however in most situations this self noise isn't really an issue because of the actual listening distance. In my case it is an issue because I have a pretty short listening distance to my monitors.
I'd say it's by design, a tradeoff between cost and proper monitoring arrangement. Making it "totally noise free" would make the cost explode.
The brands you mentioned woud never be considered "tp brands" by anyone who knows monitors, but are affordable home studio near feild monitors that are definatly better than tryint to use stereo speskers but lack the headroom to reproduce transient spikes etc. And an important piece of gear would be something like a Furman AC line conditioner that gets rid of dirty electricity giving an average of 3db extra headroom for your music. No one who is serious would plug monitors or any gear into a wall socket! Then you can turn your monitors up and hear pretty much nothing at all. The noise is comming from the electricity that has been used for dishwashers, power tools and anything that uses electricity unless you have a direct isolated line from your hydro company.
Oh...never NEVER put your ears up to a tweeter if you plan on mixing any music, or even listening to it. That's likely when something will go wrong and a loud crackle will damage your ear.
Oh, the calculator sitting there on the bench at the start - that was my calculator all the way through secondary school. Fond memories, and I wonder if I still have it somewhere..
I have both these (8 years old, haven't been affected by the black gunk of death yet), and a pair of JBL305 (which has class D amps), and the hiss from the JBL305 is really awful (but no low freq humm), still audible when playing music at a normal volume. I was really disappointed with them because so called audiophiles recommended them. I dealt with it by inserting a 2 watt 40ohm resistor inline with the tweeters, then used eq to boost it back to more or less flat response, sounds great now. Might do the same with both the woofer and tweeters in the RP6s next.
Sorry but the JBL305 is a cheap speaker and probably a response from JBL to other cheap "made in china" stuff. I have never seen or heard it but from the specs I can imagine it would be wise to avoid them.
wow good on ya man. i think i might investigate something similar. i just dont wanna deal with phase problems. how is your phaseyness with your modifications? did you manage to match it right again?
Wow, really, audible at listening levels? By most accounts the JBL305 are fairly highly regarded, but I have heard about the noise, didn't think it was that bad.
I don't know sorry, never measured it for phase issues, if there are any I haven't noticed. To be honest I used a lot of eq even before adding the resistors...
Hi Dave, yes, and this was at 2-3 meters listening distance, although depended on the music really (i.e. heavy metal masks it, classical doesn't). Someone else here said removing the ground pin stopped the hiss, so I'm thinking of trying that as well, but really can't see how it would work... If it does, I would really appreciate if you could investigate that :)
Just picked up on this, I never been keen on powered studio monitors! Because speakers can be used as microphones along with their wires, Very much like aerials! Maybe shielded speakers would help
I have rp8 g4 and no matter what socket they plugged in round the house they both have the same level of white noise no matter what gain you set, connecting this to interface does not help.
Faulty units? White noise is annoyingly loud tbh. u can hear it from 2-3m away
In my case noise was coming from monitors plugged in to the same socket as PC and display monitors. Plugged them in to separate power socket - dead quiet. Other devices on the same socket were somehow "contaminating" the current going to the monitors. But it's a probably different cause in here.
Wasn't it obvious that the mute function on the TDA would also kill the output stage?
BTW, thank you Dave for doing videos like this. Very very informative for people like myself who are far more mechanical than electrical but wish they were. :-)
No worries.
I bought a little songbird radio. It had a 7. something dc wallwart to power the thing. First thing I noticed was the blue back light would brighting and dim with the music. So I added a cap at input. Fixed.. Then decided to play it at night with very low volume.Now I have noticeable humm which is very loud at night when everything is quite . Replaced the 7 volt wallwart with one with a little better filter. Humm was a little better but still aggravating at night. At this point I am thinking about drop kicking it. The next thing was add 5 AA battery's. That seems like alot for a little radio but that was what the comportment had designed in the radio. Still humm but better. The only solution I could find was to add a 18650 battery in place of the 7volt walwart and 5AA batteries . So I went from over 7 volt Dc wallwart to 7.5 with 5 batteries to a 3.8 volt 18650.Fixed the humm but lost some playing time. It seems like something in the radio was causing to much power going to speaker. I know this fix is not the right way,,,at least I can use it at night. I might try a few other PS around 4 volts when I get a chance. Great video Dave. Love the videos and see if you can get that snake out of your speaker.That would make a very interesting video.
1. 4580D is the plain DIP version, the selected 'D' version of that would be called 4580DD.
2. So they are driving a tweeter directly from the power amp? That's a stupid idea for at least two reasons:
a) Dome tweeters tend to be ridiculously sensitive (100-ish dB/W, give or take), making every bit of amplifier hiss (and preceding crossover hiss) heard. Output noise from the power amp would be expected to be expected to be ~34 µV(A) or ~50 µV unweighted, which actually is quite civil and would be dead quiet in most any passive speaker, but here you're basically sitting a foot away from something with the sensitivity of a big horn (PA) speaker, which is about the equivalent of looking at your screen with a magnifying glass. You could try replacing those 12Ks with metal films if they aren't already.
b) You are missing out on the positive effects of current driving, which can cut down on distortion levels quite dramatically (particularly in anything not using conductive voice coil formers). Yes, it means you have to deal with the frequency response peak at resonance, but in a high-order active crossover that's not too much of an issue. Unfortunately, it would still take the tools of a speaker designer to tackle this issue properly.
3. I am still suspecting that a great many active monitors suffer from subpar gain staging in general, or more precisely they are not running their crossovers as hot as they probably should be - I see no indications of any padding between crossover and power amp here. When a dynamic range in the order of 106-110 dB is required, you have to put some thought into gain staging or else results are going to be less than ideal.
In this case you can already tell that tweeter XO levels must be quite low, given that the amplifier would be able to deliver over 35 W into 4 ohms but obviously the woofer is going to limit at maybe the 2 W level (tops). That's about 2.8 Vrms out, or a measly 100 mVrms coming out of the crossover! (Plus an estimated 1-2 µV worth of noise, adding roughly 3 dB to output noise level.) Good for distortion levels, mind you, but easily leaving 14 dB of dynamic range untapped. It's slightly less depressing-looking in the woofer section, with maybe 500 mVrms coming out of the crossover, but even there I'd want 6-10 dB higher levels.
4. These KRKs appear to have their share of issues with XLR pin 1 routing, this would be a good thing to investigate while the speaker is ripped apart. AES48-2005 says that pin 1 *must* connect to chassis ground and nothing else (chassis ground then typically connecting to signal ground and safety earth in the power supply section). I suspect ease of production mandated that pin 1 actually connects to input stage signal ground, meaning any circulating ground loop currents would have to travel from there to the power supply and then to safety earth, thereby contaminating signal ground and wrecking input common-mode rejection.
Could you have skipped all the testing by just shorting the input of the power amp (and I suppose preamp’s) to ground? Just moving along the chain to see if the noise ever stopped?
Thanks, this video makes me feel better. Just got a pair of Adam Audio T5Vs and was hearing this low hiss from the tweeter.
Perhaps it's very difficult to get rid of that from active monitors at cheaper prices. Considering the competition between them, if they could, someone would.
IC amplifiers especially those TDA ones are always a big compromise. They are used in cheap mini-hifi's. This should really have something better!
humm is inherent in any amplifier design for audio ... most higher end consumer speakers mute or filter out the noise when the signal is low enough but for monitors you dont want that
monitors are not designed to just plug in to your home theater system to leave on all day
but full disclosure i have never worked in the audio industry (unless you count humming ceramic caps :P )
Yes, many have auto-mute
I had the exact same "problem" with my new Adam Audio T7V. At first, I thought something was wrong with it, cause the noise seemed too loud (while not playing anything), I could hear it from like arm's length, but it turns I just have too sensitive hearing 😂 I actually went back into a store and tried like 10 different studio monitors, from cheaper to more expensive ones, and yes, they all have it, while the guy from the shop was looking at me like something was wrong with me 😂
Apple would be like: you are using it wrong!
T7V user here aswell. Would love to see a fix for this issue.
@@JustUploads I was going crazy at one point 🤪 The speakers themselves are awesome considering the price, but I was disappointed by this. I'm going to do a review of it and address this issue, maybe they'll have a better look at it.
Next time whip out your phone, start any of the billion FFT apps, hold it close to the speaker and show him the visibly rising noise floor when he turns it on... ;)
Of course something is wrong with you - you have intact hearing. Clearly, you don't have enough experience. The experienced guys don't hear that at all. ;)
Studio monitors offer a lot for the money. The amplifiers are cheap and simple though, compared to state of the art discrete amps. A high end transistor amp has no noise. It is dead silent. The cost of putting two or three of such circuits into a monitor is prohibitive for most studios. One monitor would amount to about 5k or more, I estimate, and only large monitors have enough space, anyway. My Focal SM9 monitors ain't cheap and are one of the best around, but they hiss in the tweeter. The circuits are not bad, but still don't match a high end amp. Power supply, preamp, poweramp, everything is simpler. I would love to see a monitor with an amp like the Benchmark AHB2 built into it!
If I remember the times 30 years ago the TDA2030 was available and one feature was the hissing noise while the TDA2020, a more expensive amp had no or much less noise. Perhaps it is a "feature of quality" of the TDA2052.
Solution is simple. Turn the audio up!
yeah ive turned down like top of the line monitors in the music shop because of the hiss. my research deals in hiss. so you can bet it was a show stopper. i believe they want to keep the power amp constantly loud in order to profile its nonlinearities. now i just wish i had more desk space.
What research are you doing?
whole bunch of stuff around making shit survive intermodulation distortion :) your interview with the guy from the signal path got me onto it. this was a freaking excellent video. the amount of times ive stared at that tweeter with anger & wonder about why its effectively noiser than these free headphones i got with my phone.
The same thing happens to passive speakers on a seperate amp, they too produce the faint hiss from the tweeters.
I've noticed this before, same brand ,the noise was LOUD and sounded like data lines 60 hz, and almost everything else. Lots of what sounded like 5k range square or sinusoidal. We shut down equipment that might have been making noise (that we could shut down) in the place to see if any of it went away... it didn't.
Thank you for this video. Now i'm sure that nothing is wrong with my JBL LSR 308.