Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Need help - What to do with 2 amplifier outputs into 1 speaker?

Status
Not open for further replies.

deweyusa

Member
Hi

I have a project where I'm trying to drive audio from my own source (an **broken link removed** from Nuvoton) into the speaker of a kid's toy that already has its own sound driver. When I hook up the output of my source, I notice a significant drop in the level of the toy's output, to a level so low that it's barely detectable. This is, I'm guessing a conflict between impedances of the two outputs. I can see the toy's audio output consists of a chip coupled to an SMT transistor, and the ISD2100's seems to be direct from the IC itself.

In fiddling with the circuits, I connected two low value resistors (47 ohms), one on each source's output, then fed that into an **broken link removed** I had lying around. It partially worked, but the amp's input level control (the pot in the diagram) has to be adjusted just to the point where it wants to start squealing (feedback?) to work, a condition that also has a lot of distortion.

Can someone please explain the interactions I'm seeing between the two sources (why one swamps out the other, even when not outputting anything) and/or recommend some designs that would work well for this application to make the two sound sources share the same speaker without distortion? Should I be using dual amplifiers, each with their own gain adjustments? Maybe an impedance matching transformer on the toy's output?

Thanks in advance for your help!
-Dewey
 

Nigel Goodwin

Super Moderator
Most Helpful Member
You can't connect two power amplifier outputs together, it's particularly bad in the case of the ISD2100 as it's a bridged output as well.

You need to feed the outputs of both through a 'mixer (as a minimum two resistors - say 10K each?, NOT 47 ohm), and feed it to a power amplifier.

I would also suggest having a pot on the output of each source, with the 10K resistors fed from the sliders, allowing independent level adjustment of each source.

As the ISD2100 is bridged, use just one of the outputs - and a low-pass filter as it's presumably class D.
 

deweyusa

Member
Thanks for your help Nigel. I don't have a lot of experience in this realm. I hope you don't mind me asking you a couple more:

I think what you're suggesting is effectively what I'm doing (except for the much lower value of resistance in my mixer). I did try higher values, but got nothing from the final amp when they were so high. Does that suggest that I possibly need a higher gain amp or a middle amp stage?


Also, I'm not entirely sure what to do about the fact that the toy's output is what I'd call, if I were talking about RF cable, "balanced" - neither side is ground - but the amp I'm using utilizes ground as one side of its input. I'm not sure if that is causing me issues.

Thanks again.
 

audioguru

Well-Known Member
Most Helpful Member
The resistors and volume control in the development amplifier kit have very low values that might not be able to be driven by your signal sources.

Each signal source might have bridged outputs with DC on them so use a coupling capacitor in series with one output wire to feed the mixer resistor and connect all the grounds together (the speaker outputs are not grounded).
 

4pyros

Well-Known Member
Most Helpful Member
Also, I'm not entirely sure what to do about the fact that the toy's output is what I'd call, if I were talking about RF cable, "balanced" - neither side is ground - but the amp I'm using utilizes ground as one side of its input. I'm not sure if that is causing me issues.
I would say just the opposite.
As Nigel pointed out the sound chip has a bridged output, nether side is referenced to ground.
The toy probably has a cheaper amp that is referenced to ground. You can check for continuity from the battery negative to the speaker wires in the toy, One or the other is probably connected to the battery negative.
As you found out you can not connect these together at the same speaker.
You will need some sort of coupling network.
Assuming the toy is run off batteries, how is the sound chip powered?
 

deweyusa

Member
4pyros, the toy IS battery powered, yes, using 3V (2 AA's). It appears that neither of the toy's outputs is grounded after all. I would have thought that if the cheaper method was to do that, that would be the case, but it appears not.

So in a bridged output, does the output voltage to the speaker swing into both the positive and negative at once?

audioguru, I think I understand what you're suggesting...since neither of the two outputs have a ground amongst them, that would mean connecting one of each to separate inputs of the mixer, and then just tying the remaining two output wires (not going to the mixer) together, but NOT to ground, right?

I don't know if it helps, but here is a photo of the output board of the toy:
P1190612.jpg


Thank you both for your help.
 

4pyros

Well-Known Member
Most Helpful Member
4pyros, the toy IS battery powered, yes, using 3V (2 AA's). It appears that neither of the toy's outputs is grounded after all. I would have thought that if the cheaper method was to do that, that would be the case, but it appears not.
It appears the toy is set up opposite then the why I thought.
One side of the speaker is hot all the time ( thru R9 ) and the COB controls a transistor switching ground.
It is probably making some horrible square wave sound.
Any way you can not connect grounds together.
You will need to isolate the return paths as well.

At this point maybe AG can come up with a mixer circuit with a seperate amp for you.
 

audioguru

Well-Known Member
Most Helpful Member
We do not know if this toy has an amplifier for sounds or instead has a simple squarewave driver for noises.

A bridged amplifier actually has two amplifier circuits. One drives one wire of the loudspeaker and the other amplifier inverts the signal and drives the other wire of the loudspeaker.
A bridged amplifier is used when the supply voltage is low (like in a car and in this toy) which causes a low amount of output power from an ordinary amplifier. A bridged amplifier almost doubles the voltage swing across the speaker which also almost doubles the speaker current which almost makes 4 times as much power as an ordinary amplifier.
The DC bias voltage is the same at the output of each of the amplifiers so a coupling capacitor is not needed to feed the speaker.

Take the output from one of the two amplifiers in the bridge that feeds one wire of the speaker. It has signal on it. Use a series coupling capacitor to block its DC and use a series mixing resistor to feed the input volume control of the final amplifier. Connect the grounds of the amplifiers and signal sources together. DO NOT connect together the other outputs of the bridged amplifiers.

Since both your signal sources are designed to drive speakers then fairly low values can be used for the mixing resistors like 330 ohms but then the value of the coupling capacitors must be pretty high like 3.3uF to pass low frequencies.
 

4pyros

Well-Known Member
Most Helpful Member
Connect the grounds of the amplifiers and signal sources together.
AG; I didn't think the OP could tie the grounds together because the speaker driver in the toy is a low side driver and the speaker is hot all the time.
 

audioguru

Well-Known Member
Most Helpful Member
You think that Q1 simply puts negative-going pulses into the positive (hot) speaker? The grounds can still be connected together when coupling capacitors block the DC because C1 couples the positive supply to ground..
 

4pyros

Well-Known Member
Most Helpful Member
You think that Q1 simply puts negative-going pulses into the positive (hot) speaker?
Well yes sort of, Q1 clearly is switching B- or ground so hooking up to that without a pullup or dummy load for speaker will not give you any voltage in reference to ground.
Am I right?
 

deweyusa

Member
Excellent observation on the fact that the ground is switched! That explains a lot of the problems I've been experiencing I think.

After much experimenting until late last night, I could not get a proper combination of mixer input resistances and amplifier gain (adjustable input on the LM4876 -based kit) that mixed both signals with fair volume, low distortion, and no feedback. It seems that there is no perfect input resistance balance point in between the two possible. The ISD2100 can do fine, but with the mixer adjusted in its favor, it seems to swamp out the toy's ability to drive its own output, even when the resistance on the toy side is fairly high. Adjusted in the toy's favor, the amp usually provides lots of feedback in the absence of an input signal.

I'm not 100% sure how to deal with the amplification of the negative-going pulses from the toy. You're both thinking that by putting a series resistance between the negative-pulsed output from the toy and Vcc, I can then mix the signal developed across it with the ISD2100's then feed the amp with that? If I understood correctly, that sounds like it could work, but I wonder if I may need a pre-amp for the voltage developed on the pullup resistor before I try to put it through the mixer.

Either that or two separate amplifiers that are then tied together at the speaker (which I think I can do, as long as the grounds are tied together).

Thanks for the help so far!

-Dewey
 

4pyros

Well-Known Member
Most Helpful Member
You can try this;
Put a voltage divider net work (two resistors) across the outputs of each source. Use a value somewhere between 10 and 100 ohms for each of the resisters in the network. That should provide enough of a load for the amps to work properly and provide a signal output between the resistors equal to half the supply voltage as referenced to ground. Because they both run on 3 volts they should be close to the same level. Use a coupling cap about 4.7uf from the middle of both resistor networks to your amp positive input and tie all the grounds together. Hopefully half the output voltages from the sources is enough to drive your LM4876 at a decent volume. If one source is still louder than the other then you can adjust the resistor values of one or the other resistor divider networks.
 

deweyusa

Member
Okay...after much more experimenting, I tried putting an op amp circuit on the toy's end consisting of a 1W kit built around the TDA7052. Somehow, in the process, I managed to release some magic smoke, and the magic stopped (blown driver, perhaps). I'm not sure if it was a simple short, or if I hooked up a very strange configuration.

This still has me very curious though, as I'd really like to figure out how to combine these two signals effectively. (I'll obviously need another toy of the same kind). I'm' guessing other toys use the same method to produce sound. It seems that the toy is expecting 8 ohms: the speaker. If I were to put two 4 ohm resistors in series across its outputs (with one side being the positive terminal), could I then amplify and buffer that signal in preparation for the mixing with the other one? I hope to find that the case.

-Dewey
 

Nigel Goodwin

Super Moderator
Most Helpful Member
As we keep telling you, you DON'T use both speaker terminals for a bridged amplifier, you also don't need an opamp to provide enough drive for a power amp, the details I posted in post #2 explain what you need.
 

deweyusa

Member
I understood that. Thanks for your suggestion, but I tried the method you listed in post #2, but it attenuated the toy's amplifier beyond the point of producing distinguishable sound output when also connected to a simple resistor-based mixer having the ISD2100 on one leg (both resistors = 10K). As I mentioned in my follow-up post, there was NO resistance value in the mixer that provided a happy medium between the ISD2100 and the toy's output such that both were loud enough to be usable. It was either one or the other. The only way to get the toy's sound source to work acceptably was to put a lower resistance between it's output and Vcc. Around 22 ohms worked the best.
 
Last edited:

audioguru

Well-Known Member
Most Helpful Member
It looks liker 4pyros was correct in his post #7 that the output of the toy's amplifier is simply an open collector that uses the speaker as its load to B+. A 22 ohm resistor to replace its speaker will be fine.

As I said in my post #4, "The resistors and volume control in the development amplifier kit have very low values that might not be able to be driven by your signal sources."
Therefore you cannot use 10k mixing resistors because they attenuate the signals too much. Try using 330 ohms mixing resistors in series with 3.3uF coupling capacitors to feed the development amplifier.
 

deweyusa

Member
Will do! (I just have to wait for another one of the toys to replace the one I blew out). Just to clarify what you wrote in your previous post about this: I drew the circuit here. Is this what you were recommending? Right off the bat, I'm noticing something I have a question on. As an open collector class D amp, the toy is essentially sending a pulsed low/high impedance signal, so it needs Vcc on the other end of the speaker to pull current. The ISD2100, being bridged, is sending (I think) a waveform that is driven alternatively high and low. Does the fact that one is an open collector and one a totem pole type of configuration create a conflict of some kind?

Is this what you are suggesting doing?
toy2.png
 
Status
Not open for further replies.

Latest threads

Top