How to connect multiple amplifiers

[Rattanee]

Hi folks,

The past weeks I've been cursing the paint off my ceiling trying to figure out what's wrong with my designs. I've even had a thread about a TDA2009 powered amp overheating and being noisy. Well I still can't figure the problem out, but I'm fairly convinced it has to do with me running the amps off of the same DC rail and probably/possibly the way I've been mixing the L and R channels to obtain a third for the subwoofer amp. Since I need this ASAP, I've ditched the TDA 2009 and used the innards of a cheapo PC speaker I had on hand to use as the stereo section, and it works fine aside from very light noise from the PC (if anyone has an idea on how to filter this, that's welcome too though it'll be okay as it is if it's too much a hassle).

Thusly I'd like your advice on the following. I'd need to run a stereo amp and a mono amp off of the same DC rail, and use a single stereo input to drive them in a 2.1 setup. I've attached a blank pic with connector symbols named appropriately. Could anyone suggest a way to wire this up?

Basically I have two amplifiers built based on the datasheet application schematics. I'd like to run these off of a PC's 12v powersupply rail, and mix a stereo signal to the stereo input of one amp, while tapping them both for a combined mono signal for the mono amp. I'm looking for suggestions on how to achieve this. I'd like to keep this as simple as possible, preferrably with all passive components. Any help is appreciated, THANKS!

GND is pin 2 on all connectors!

PS: I've also attached a pic of how I have them wired up now, which actually almost works, except the LA4460 keeps giving a rythmic mid-low frequency thumping and of course distorts as hell, the latter certainly because I've not wired a volume control in for the LA4460. Basically power comes to the stereo amp from the PC power rail, and is tapped right after the filter cap on the board, to go to the LA4460. Signals go straight to the stereo amp, where each channel is tapped through a 4k7 resistor, and grounds are joined and tapped through a 200k resistor, thus going to the LA4460.

 

[BobK]

All of the grounds should be connected together, with no resistor.

Bob

 

[Rattanee]

All the grounds are now tied together on the circuit boards, I removed the resistor, yet I'm still getting the rythmic popping.

 

[Rattanee]

Okay so I wired in a 100k pot for volume, and the thumping is gone. However I'm still getting a considerable amount of distortion at higher volumes from the mono amp. How can I get rid of this? (I know the amp should be able to do much more volume-wise)

 

[Rattanee]

Could the passive mixing be what introduces distortion? I started adjusting volume on all channels, and noted that a light distortion in the lower-midrange is always present even at low volumes on -all- channels (that meaning the stereo amp as well so it's surely not a design fault of the amps.)

 

[Rattanee]

Okay I removed the amps yet again, changed the old lytics in the stereo amp just to be safe, and the light distortion was still there. At which point I plugged my phone into the audio jack, and music plays crystal clear from the phone. Apparently the integrated in the PC just blows chunks...

Hopefully a PCI sound card will solve that issue.

The mono amp is still clearly distorting however... Any tips on why this could be ?

 

[Rattanee]

Okay so I put the scope to the output of the LA4460. Here's the deal.
The stereo amp has a cheapo TA8217 in it, which is supposedly 2.5 watts per channel going into the proper 4 ohm load. The LA4460 should be 12W at 4ohm, it's driving an 8 ohm speaker so I suppose that drops to around 8 watts. Still, it should be louder than the stereo section. However, the stereo amp is only at halfway, not distorting when I start to turn the volume up on the LA4460, which goes to about half the volume of the stereo section before it starts to distort. The distortion is outright clipping on both low and high halves of the output. Is this all I should expect from this IC being driven into 8 ohms with 12VDC power, or am I missing something?

The LA4460's layout is an exact replica of what is in the datasheet here: http://pdf1.alldatasheet.com/datasheet-pdf/view/40092/SANYO/LA4460.html

What is weird is that at 12VDC supply voltage, I'd expect the voltage swing to be around 6V in each direction, but at best I'm getting 4.5V? Actually changing the V/DIV, I can confirm that it's a hint less then 4V... which seems very low. Is it not ?

 

[BobK]

Using darlingtons in the output, you can only get about 2V from the rails, so 4V sounds about right. Your 12W figure cannot be right. Given that the most swing you can get is 6V p-p that is 4.6V RMS and V^2/R is 21.3/4 = 5W. With an 8 Ohm speaker this falls to 2.5 W. With your 4V swing, that is only 1W into an 8 Ohm speaker.

Bob

 

[Rattanee]

The datasheet lists it as a 12w chip. Weird.
But anyways, in my quest for an amp chip that can be fed 12v and produce some volume as well, I found a TDA7240A in my amp chip drawer, and quickly put it together on a breadboard. According to the datasheet this thing should produce 20watts into 4 ohms with 18vdc. According to the tables, at 8 ohms, fed 12vdc, it should be 8 watts. The maximum voltage swing I'm getting is about 10 volts. It sounds better, but not by much.

Is there a way to crank out more volume from a 12VDC rail, or am I stuck getting a separate powersupply if I want to go any louder?
How the heck do those tiny 2.1 PC speaker systems have such (comparatively) loud woofers?

 

[Rattanee]

I also added a 800nf cap to filter the high frequencies from the input, this way I could crank a little more out of it and it stays clean. Is there a way to calculate how a 'tone' capacitor's value affects frequency cut?

 

[Rattanee]

And the question of the day... I've thought about tapping the -5V of the PC PSU to use as the ground for the amps, thus obtaining a 18v rail that -should- help the volume issue, however, were I to do that, I'm not sure If I could connect it to the PC's audio output without some sort of interfacing needed? (As the audio out would be 5VDC above the amp ground)

 

[(*steve*)]

In theory that would work.

In practice you may find that the -5V rail is rated for a very low current. If you exceed that current you may cause yourself some problems.

In addition the *real* ground rail is almost certainly connected to mains earth. That means your amplifier will have some potential in the ground side of the input which could result in shorting out the power supply when you connect a signal to it.

 

[Rattanee]

That's exactly what I was thinking of. I'd not be really worried about overloading the -5V rail, this isn't that much power, but indeed signal ground is connected to the AC ground as well in the PC. I wonder if I could get around that with an op-amp stage or in any other way, but that is too much effort to be honest. With the cap on the mono amp's input I was able to crank it up enough without distortion that it sounds good enough to stay.

What's odd, is I have a Sega Megalo arcade cabinet (big 50" projection sitdown cab) that has a quite sophisticated stereo system in it, which is powered by a mere 13 VDC, and that can get insanely loud if I crank it up with no distortion whatsoever, but that has 8 ohm speakers everywhere? I thought about just copying that but it's quite complex due to all the EQs and whatnot in it. I'm going to have to reverse engineer that and see how it's getting that kind of output power. (And no we're not talking +-13VDC, just 13VDC for each amp.)

 

[BobK]

I looked at the datasheet for the LA4460. A couple of observations:

It is a bridged amp, so it can produce 4x the power I computed before.

It has a gain of 51db which is a voltage gain of 345. This is way too much if you are giving it line level inputs, which explains why you get so much noise and distortion.

Bob

 

[Rattanee]

I have a 250K trimmer in front of it for volume control, which I've set as high as it will go without the output signal getting clipped.

Does that mean there's a way to get it go louder, or is this still as loud as it will go?

 

[BobK]

The clipping point is as loud as it can go without serious distortion. What kind of speakers are you using? Perhaps more efficient speakers would help.

Also, when you measure 4V how are you doing that? Is that between the two speaker outputs, or from one to ground? It should be going to nearly 8V between the two outputs.

Bob

 

[Rattanee]

Yeah a 4 ohm speaker would help but not in the budget for this project, already went overboard with the expenses.

I put the scope to the output in AC coupled mode and the voltage swing was 4 volts either way, so a total of 8 volts. Using the TDA ic it sounds okay though, that's getting me a 10V output so it's a tad louder.

 

[BobK]

How are you connecting the scope, across the two speaker outputs or from one to ground. If across the two speaker outputs, you should be seeing + and - 8V.

Bob

 

[Rattanee]

Between ground and either speaker output pin.

 

[BobK]

And the speaker is connected between the two speaker outputs, right?

Bob