LM3875 speaker pop when powered on

JDawg

Member
2010-05-15 5:16 pm
It only makes a pop when I turn it on. I'm not sure if it does it on both channels or just one of them so that's something I'll have to check later tonight.

It looks like everything is connected properly otherwise I don't think I would be getting good sound out of it. I'm using the star grounding system with the ground from the mains and two amp boards connected to a single point on the case. I did add the preamp buffer circuit found on the Decibel Dugeon website and I have a seperate power supply board to power the two buffer circuits. Would I have to connect the ground from the two buffer circuits and its power supply board to the same ground point as everything else?
 
I just checked for the popping noise again and it comes out of both speakers. I don't think its the switch because if I turn it on, then off, and on again after a few seconds there's no pop from the speakers.

I have the same issue as you, but it doesn't happen on every switch on or switch off. I'm using some pretty expensive Nikkai switches which are way overrated for the application. Doesn't seem to hurt the speakers, but one thing I haven't tried is placing a small ceramic cap across the switch.

Anand.
 
The LM3875 has 10mV of input offset. Multiply this times your configured gain and you may have quite a bit of output offset. This will cause an audible transient at power on as the output stage suddenly jumps to the offset voltage. This is a common problem in higher power amplifiers with no soft start circuitry. You can measure the output offset if you are so inclined by grounding the audio input, turning on the amplifier, and looking at the DC level of the output. My guess is it's a pretty large offset, which is definielty audible during turn on.

Another common issue is that the external caps are not charged at power on. If the amplifiers output stage is enabled while the caps are still charging, the output will swing around and create a pop sound (or some other sound...). JDawgs comment seems to point to this being a problem. You can test by turning it on for a while, then off and right back on. Then try turning it on for a while, then off for a while, then back on. If it clicks when you wait a while to turn it back on but not when you do it quickly, this would point to the cap charging issue.

Hope this helps... I couldn't advise you on any solutions unless I saw a schematic.
 

JDawg

Member
2010-05-15 5:16 pm
I actually tried turning it on and then off and then doing the same thing again but waiting a while before turning it on again. If I did it right away then the pop isn't there. If I wait a while to turn it on then I get the noise. So I think you're probably right about the capacitors. Its not a loud pop or anything but is there a way to fix this?
 
The LM3875 has 10mV of input offset. Multiply this times your configured gain and you may have quite a bit of output offset. This will cause an audible transient at power on as the output stage suddenly jumps to the offset voltage. This is a common problem in higher power amplifiers with no soft start circuitry. You can measure the output offset if you are so inclined by grounding the audio input, turning on the amplifier, and looking at the DC level of the output. My guess is it's a pretty large offset, which is definielty audible during turn on.

Just to do the maths: Datasheet specifies up to a microAmp of input offset current -- if you are direct coupling using a 100K/10K potentiometer you can have 100mv/10mV on the non-inverting input. Multiply this by typical gain of 26dB and you have some serious DC. You can rig a servo (see National's Application note 1192) to eliminate the offset.
 
Hi,
Try to use an inrush current limiter like the GE CL30 between the rectifiers and the filter capacitors. Add it to both negative/positive voltages. This will slow down the inrush current in the power up.

This may work, though it looks like the current limit is not less than 1A. I'm not sure that's enough. You can do the calculations by finding the time constant of the offending circuit (I think this would be the input cap) with 1A current limit and see if it's an audible ramp. Also, who knows what kind distortion these can cause...

I believe the problem is in the input caps. If you can slow down the charging of these caps, it should solve the problem. There are devices out there that do this, though the only one I know of is the MAX9892 which is a very small suface mount part (with bumps rather than leads). In Maxim's world this is used on the outputs of the amplifier, but I don't see why it couldn't be used on inputs.

You could make a discreet version of this with a FETs and an RC timer. Just put the FET from the input cap (amplifier side) to ground and make it turn off slowly. You could even put a pot in to have a variable turn off time. The idea is to hold the input cap at ground during turn on and slowly release it so there are no audible pops.

Remember I'm just guessing at the true nature of this pop... and again it's hard to say if the solution would work without a schematic. Good luck!
 
This may work, though it looks like the current limit is not less than 1A. I'm not sure that's enough. You can do the calculations by finding the time constant of the offending circuit (I think this would be the input cap) with 1A current limit and see if it's an audible ramp. Also, who knows what kind distortion these can cause...

I believe the problem is in the input caps. If you can slow down the charging of these caps, it should solve the problem. There are devices out there that do this, though the only one I know of is the MAX9892 which is a very small suface mount part (with bumps rather than leads). In Maxim's world this is used on the outputs of the amplifier, but I don't see why it couldn't be used on inputs.

You could make a discreet version of this with a FETs and an RC timer. Just put the FET from the input cap (amplifier side) to ground and make it turn off slowly. You could even put a pot in to have a variable turn off time. The idea is to hold the input cap at ground during turn on and slowly release it so there are no audible pops.

Remember I'm just guessing at the true nature of this pop... and again it's hard to say if the solution would work without a schematic. Good luck!

Audiophilia,

I think schematics and other details can be found here.

All the best,
Anand.
 
Hi,
Try to use an inrush current limiter like the GE CL30 between the rectifiers and the filter capacitors. Add it to both negative/positive voltages. This will slow down the inrush current in the power up.
CL30's are kind of expensive, about $5, and don't shut off after the unit is on, which would limit current flow on cannon shots on Tchaikowski, ACDC, etc. Peavey uses a 5 ohm positive temperature coefficient thermistor so slow their inrush, but I don't know a cheap place to buy one. My mixer I'm using for my records pops at turn-on. I've already installed a capacitor across the power switch to cut turn-off pop. I have to turn the mixer on before I turn on the power amp.
 
inrush limiter

Boy this edit feature disappears quickly. The Peavey 5 ohm 12A inrush limiter is a "SG26", Peavey pn 70240210. Mouser doesn't recognize the number. Peavey bridges the contacts of a relay with two of them in series, said relay picked up by the 12VDC control supply, so they are in use limiting the 120 VAC input until the amp is fully on. CS800S amp schematic, a bit bigger than a chip amp, but the PTC resistors mouser does list for $1.50 are 50 ma current limit, not too useful. I wish PC power supplies had them, i've got about five of those dead I could cut up for parts but no schematics for them.
 
Hi,
Mouser have it for $1.30. Also you can buy it in different current rating. The CL30 is 8 amps. Also to prevent the spikes in the powered up I use a GE GMOV V130LA10A across the transformer primary winding to prevent the spikes. The prices for the GMOV in mouser is 30 cents. I used both in every electronic equipment that I built or repair. They really worked.
 
Attached are two pictures of my LM3886 showing the LMR installed in the board.
 

Attachments

  • LM3886_pic1.JPG
    LM3886_pic1.JPG
    707.1 KB · Views: 181
  • LM3886_pic2.JPG
    LM3886_pic2.JPG
    636.9 KB · Views: 170
Indian,
I don't think a PTC is suitable to reduce the start up current of a transformer.
An incandescent metal (tungsten) filament light bulb is a sort of PTC.
It turns on and cripples the PSU if higher current is detected.
It's precisely this PTC behaviour that the mains light bulb tester makes use of to indicate and protect against wiring faults.

NTC does the job far better.
When continuous high current is demanded it's resistance falls (due to heating effect) and allows maximum voltage to the transformer soon after high demand.
Better still is to bypass the NTC, after start up and use the transformer direct to the mains (the way it is designed to be connected).
 
Hi,
Mouser have it for $1.30. Also you can buy it in different current rating. The CL30 is 8 amps. Also to prevent the spikes in the powered up I use a GE GMOV V130LA10A across the transformer primary winding to prevent the spikes. The prices for the GMOV in mouser is 30 cents. I used both in every electronic equipment that I built or repair. They really worked.
That price is the CL30. If your going to use those on a power amp, you can also write off bass drum hits in Polovetsian Dances and ZZ Top "Afterburner". I like installing the metal oxide surge suppressors across the power transformer input, but I salvage mine from dead PC power supplies for nothing (I'm not employed). The import power supply surge suppressors are blue, near the fuse, say "S14" or "S07", have a UL logo for power line use. They are useful for preventing lightning and motor turn-off pops or damage. Readers in the eastern hemisphere, make sure you use "240" or higher voltage MOS surge suppressor, with VDE rating or whatever for your country.
MOS suppressor across the power transfomer are not useful for eliminating turn-on pop. The Peavey surge suppressor is in series with the AC transformer, not across. (Actually the rectifier, there is no main Power transformer on this switcher amp).