In that example, more drive, does not mean more power in to a given load. It only refers to the ability to deliver more current in to a lower impedance load. In which case, it could then provide more power to that load than could a single such amplifier. This type of confusion over a simple circuit functioning seems to happen when intuition of how a circuit should function is substituted for a simple mathematical analysis which proves otherwise. The correct and simple mathematical analysis has been presented in this thread, over and over to no avail. That being the case, maybe, an intuitive approach may help.
In thought, apply a 12V output from a pair of identical paralleled amplifiers to a single speaker load. Being paralleled, the pair of amplifier outputs are always at the same voltage. So, as far as the speaker load is concerned, there is single virtual amplifier driving it. Then, by Ohm's Law, we find: 12V out / 6R load = 2A current.
So, 2 Amps total current flows through the speaker load, whether from a single amplifier, or from a paralleled pair amplifiers. Why, because the amplifier output voltage is the same in either case. That's all of the current flowing through the speaker, and so is divided between paralleled amplifiers. So, if there are 2 Amps of total current flowing through the speaker, then logically, since there are two amplifiers, each amplifier must be providing 1 Amp of the total, not each amplifier somehow maintaining delivery of 2 Amps. Which would then sum to 4 Amps through the speaker in complete violation of the basic physics dictated by Ohm's Law. Of course, if you reject Ohm's Law, no amount of explanation will help
Last edited:
This reminds me of people that put stuff in a microwave that doesn't belong there just to see the explosion. My personal favorite is automotive air bags... the microwave ends up round and the door flies off like it's propelled by a rocket.
Of course, this is done outside, with remote powering (an extension chord) and the person stands behind a barrier, like standing around the corner of a brick house.
I say to the OP go ahead and do it. Then you will learn for yourself what you wouldn't accept from a half dozen EEs with probably 200 years of combined experience between them... that it won't double the current. Then the light bulb in your head will finally come on. You will learn the lesson and you will remember it for the rest of your life. And, it will be fun.
I remember when I was very young (preschool) and I started messing around with electricity. My father bought me some whiz kid electrical experimental lab kit. I had been illuminating typical flashlight bulbs with D size batteries. I was out of D batteries and only had C batteries. In my mind, the C batteries wouldn't light the light as bright; but as most of here know they did. That was a powerful lesson that I never forgot. It's the exact same lesson the OP will learn if he goes ahead with his plan to parallel the amplifiers, which IMO is a terrible idea and may very well damage the units. Then maybe he will realize that the biggest gain to be made will be to bridge them. The irony here is that according to Ohm's Law, bridging them is what will double the current into the speaker.
Of course, this is done outside, with remote powering (an extension chord) and the person stands behind a barrier, like standing around the corner of a brick house.
I say to the OP go ahead and do it. Then you will learn for yourself what you wouldn't accept from a half dozen EEs with probably 200 years of combined experience between them... that it won't double the current. Then the light bulb in your head will finally come on. You will learn the lesson and you will remember it for the rest of your life. And, it will be fun.
I remember when I was very young (preschool) and I started messing around with electricity. My father bought me some whiz kid electrical experimental lab kit. I had been illuminating typical flashlight bulbs with D size batteries. I was out of D batteries and only had C batteries. In my mind, the C batteries wouldn't light the light as bright; but as most of here know they did. That was a powerful lesson that I never forgot. It's the exact same lesson the OP will learn if he goes ahead with his plan to parallel the amplifiers, which IMO is a terrible idea and may very well damage the units. Then maybe he will realize that the biggest gain to be made will be to bridge them. The irony here is that according to Ohm's Law, bridging them is what will double the current into the speaker.
In your example, the output impedance is (100R+Zout(opamp))/2. Often you put output resistors larger than the opamp's closed loop Zout in series with the output in opamp circuits, even with a single opamp, so the same circuit with a single opamp would have the output impedance of 100R+Zout(opamp). So indeed output impedance is cut in half.
In an audio power amp, you generally do not have an output resistor in series with the load. So the output impedance of one power amp is Zout(power amp). In order to parallel them, you will usually want output series resistors on each amp to ensure good current sharing and compensate for any DC or gain offsets. So output impedance of the two power amps in parallel is (Zout(power amp)+ZSeries)/2. Zout(power amp) is often less than .2 ohms, so unless you have very small series output resistors, you will actually increase your output impedance with two amps in parallel.
Also, amplifiers that are paralleled have to match very closely. Critical components (like the feedback network resistors) have to match very closely (ideally 0.1%) and be very stable or else you're going to get a mess. TI has a tutorial on parallel/bridging the LM3886 and reading it would make the OP hopefully think about why paralleling amplifiers not specifically designed for it is not a good idea.
That stupid Ohm law is an obstacle.
Reducing amplifier output impedance will only reduce amplifier internal voltage loss. Amplifier in question is said to have 0.02 Ω output impedance. At 30 W/8 Ω output, current is 1.94 A rms. Amplifier internal voltage loss is I x R = 39 mV rms. Paralleling two amplifiers will halve output current per amplifier and reduce internal loss to 19 mV, or output voltage will be 19 mV higher. Now calculate how much will this whopping output voltage increase bring in output power terms. 🤣
Ah yes, those "Ah - Ha Moments" We all have had them!
Mine was Freshman year in High School Chemistry. Five years earlier I was told by an aunt never mix Ammonia and Bleach. That night I went into the bathroom, mixed them, remember smelling something really nasty, passed out, and hit my head on the bathtub... woke up in hospital...fast forward to High School Chemistry, learning basic stoichiometry and seeing that I had made Chloromine gas. Neato!
I say try it. If you destroy a couple amps or don't get the result you believe you would, well then, all good...
I will admit it, years ago I also was questioning the ability of Paralleling Amp outputs of a five channel amp to get two higher power channels to drive a pair of Magnepan SMGs. I saw paralleled output transistors in SS amps, paralleled output tubes in SE tube amps, heck, even paralleled LM3886 chips in chip amps......why wouldn't it work, right?
I used resistors and also tried diodes in series with the speaker...results? Crappy sound, and in one case, smoke....don't remember which was which, but one thing for sure was that five channel amp never quite sounded right after that experiment.
Last edited:
I think I was in grammar school when I got the bright idea to "improve" my reel to reel by hooking the record head directly to a solid state amplifier. I thought it would lower distortion and increase dynamic range. I even had the bias running through the amplifier which was a very bad idea. Primitive transistor amplifiers can't handle ultrasonic frequencies very well, and you will go off the SOA if you try to force it.
The amplifier emitted the magic smoke in short order. I quickly soldered the original wires back to the record head and surprisingly, it worked fine.
I knew just enough to be very dangerous. But I learned something I never forgot.
The amplifier emitted the magic smoke in short order. I quickly soldered the original wires back to the record head and surprisingly, it worked fine.
I knew just enough to be very dangerous. But I learned something I never forgot.
JMFahey
There's nothing remotely close to that name having been taken as far as one can see from the members list (the list of members is continuously expanding so the names are shifting in list position, but as of posting this post it's on page 822)
https://www.diyaudio.com/community/members/list/?page=822
it would be prudent of you to change your name instead of insinuating and fooling other members thinking you being an EE, you can ask for a name change in the name changing thread here:
https://www.diyaudio.com/community/threads/username.21170/
There's nothing remotely close to that name having been taken as far as one can see from the members list (the list of members is continuously expanding so the names are shifting in list position, but as of posting this post it's on page 822)
https://www.diyaudio.com/community/members/list/?page=822
it would be prudent of you to change your name instead of insinuating and fooling other members thinking you being an EE, you can ask for a name change in the name changing thread here:
https://www.diyaudio.com/community/threads/username.21170/
Here's that TI tutorial on bridge/parallel amplifiers.
https://www.ti.com/lit/an/snaa021b/snaa021b.pdf
And here's the precision resistors I would recommend, Vishay CMF Industrial series.
https://www.mouser.com/c/passive-components/resistors/film-resistors/metal-film-resistors-through-hole/?q=cmf&m=Vishay&series=CMF Industrial These are very stable with guaranteed temperature coefficient, and they have very good high frequency performance. Not cheap but they work great for critical precision applications in audio circuits.
https://www.ti.com/lit/an/snaa021b/snaa021b.pdf
And here's the precision resistors I would recommend, Vishay CMF Industrial series.
https://www.mouser.com/c/passive-components/resistors/film-resistors/metal-film-resistors-through-hole/?q=cmf&m=Vishay&series=CMF Industrial These are very stable with guaranteed temperature coefficient, and they have very good high frequency performance. Not cheap but they work great for critical precision applications in audio circuits.
You are correct. But in the real world, there is no gain at all.
Let's see how much you gain. Let's say your amplifier's Vmax (in your example) is 15 volts RMS. It was calculated (from the damping factor somewhere in this thread) that paralleling amplifiers would reduce the voltage loss by about 17 mV. That's a net gain of 10E-3dB, or .01 dB. Nothing.
log(15.017/15)x20=.01dB
From the TI paper. The advantage of the parallel solution is total Pdmax is divided equally among each of the amplifiers in the parallel solution. By dividing up the total power dissipation among two or more ICs, lower impedance loads can be driven for much higher power solutions.
There showing how to use an amp designed for a max 8 ohm load to drive a 4 ohm load.
There showing how to use an amp designed for a max 8 ohm load to drive a 4 ohm load.
(1) Yes, increasing the capacitance to the power supply will solve the issue of transients.... but that would require redoing the amp. My idea is to do it externally.
(2) BTW, it's not what I "want"... I'm just curious.
Because they won't fit into the Z-Rack. Note the first post... where I noted I wanted to fit the amps in the Z-Rack.
BTW, I'm not a "big shot"... I am what I am... I mentioned my background because people were accusing me of being a hack.
Why is it that so many are so rude?
What does a microwave have to do with this... are you all trying to ridicule me with non sequiturs?
Honestly, I brought up the point of driving a low impedance load into transient peaks... but almost no one is paying attention to it. You all seem to be stuck with steady state considerations.
It turns out that a better solution is to simply increase the capacitance in the power supply to begin with. But given an amplifier already designed and with no desire (or room in the rack) to redesign the power supplies I wondered about paralleling instead of bridging.
There is enough literature about it to make it an interesting read... and with some considerations (which I have addressed) you can prevent the amplifiers from immolating themselves.
I do wish you'd READ the posts.
And, I couldn't care how many eons of soldering and schematic experience you have... I also have experience... part of which was smoking a $250K not for flight, conformal coated board back in '85. Do you know what a 69 vdc, 3A source will do when wired into a 5VDC input? Did you know that a 1/4 watt diodes will "emit" a 1 inch, white jet of plasma? Since I was working as an engineer, I got away with it because I found out the root cause: the connectors in the wiring harness were not wired per the drawings and the QA guy ( whose job was it to accept such equipment) had not properly documented this. I just had to deal with lots of paperwork in the next six months.
So, IMHO, you haven't learned much until you smoke something in a fantastic manner.
Thanks... if you read some posts back, we were discussing some power resistors from Mouser. I'm looking at getting 100 watt, 1% resistors.
Jesus H. Christ! Will you give it up?
I signed up for this forum eons ago... tonyE was taken then. It is what is is.
I am NOT INSINUATING ANYTHING. YOU ARE INSINUATING. And you sure are spending a lot of time on accusing me of things I have never even thought about... If you had spent half the time browsing the web for "parallel amplifiers" you might have learned something useful instead.
What would be prudent would be for you to stop with the ad hominem attacks and stick to the facts.
Come to think about it... you too are ignored... It's incredible how many people come out of the wood being rude. After 15 years or so in these forums, I had to "ignore" three people. Nuts!
I think it has been established that the only benefit of paralleling amplifiers is to drive low impedance loads where current drive, not voltage gain, is necessary.
I've read that TI doc... that's why I started this thread by specifically using a stereo amplifier and paralleling its two channels. The idea being that the two channels in one amplifier would be as closely matched as you're gonna get.
- Status
- Not open for further replies.