Been in electronics a long time actually but sometimes my memory fades on certain items in circuit design (what design? I write software these days) -- and something that I need a refresher on is the purpose of the small value power resistors on the outputs of these amplifiers when running in parallel.
Some sort of isolation? (but what?) can't be AC and at one tenth of an ohm wouldn't have much effect on DC either. ?
Is it a protection circuit? If an output is shorted to ground does it initiate some sort of internal shutdown and at the same time provide a buffer or something?
What is going on with these? --and thanks
Some sort of isolation? (but what?) can't be AC and at one tenth of an ohm wouldn't have much effect on DC either. ?
Is it a protection circuit? If an output is shorted to ground does it initiate some sort of internal shutdown and at the same time provide a buffer or something?
What is going on with these? --and thanks
The resistors are there because the amplifiers are in effect pure voltage sources at their outputs. Their output impedance is very low - of the order of milliohms.
If you connect two voltage sources together and their voltages are even slightly different, large currents will flow between them. So the resistors are there to try to limit the currents that flow between the amps because ideally we want all the current to flow into the speaker, not into the other paralleled amplifier.
Any clearer now?
Sort of the "path of least resistance thing". If one source is lower, lookout.
But because the voltage difference between the 2 is inevitably not that much,
only a tenth of an ohm will suffice.
As a buffer really.
Right? (close enough?)
But because the voltage difference between the 2 is inevitably not that much,
only a tenth of an ohm will suffice.
As a buffer really.
Right? (close enough?)
Think of the two amps without any speaker connected - maybe that will help. If there's a gain mismatch, one amp 'sees' a load of 0.2ohm into the other one. That 0.2ohm load is more than an order of magnitude lower than the speaker impedance - even a 1% gain mismatch at an output voltage of 30V is 300mV - giving 1.5A into the lower gain amplifier.
Gotcha.
You know my head has been out of analog audio for some time and playing around with the chip amps will be a good tutorial.
I'll get out the test gear and get this module aligned better for parallel operation.
Funny, I haven't even HEARD an LM3886 before -- bought this module for the convenience of quick audio and value. thanks.
You know my head has been out of analog audio for some time and playing around with the chip amps will be a good tutorial.
I'll get out the test gear and get this module aligned better for parallel operation.
Funny, I haven't even HEARD an LM3886 before -- bought this module for the convenience of quick audio and value. thanks.
you may have "heard" this or one of it's brethern in a TV or table top radio or in computer speakers or a multitude of other worse than useless implementations.
Now you tell me.
So, the LM3886 on a scale of 1 - 10?
Are you talking the about the chip itself or as you say, the implementation?
Lousy speakers etc.
So, the LM3886 on a scale of 1 - 10?
Are you talking the about the chip itself or as you say, the implementation?
Lousy speakers etc.
I'll put in my tuppence worth here. Its mostly about implementation, not the chip itself.
I started playing with the LM3886 shortly after RS components introduced it into its UK catalogue. I don't recall exactly when this was, but sometime in the mid 1990s. First up, I built the application circuit straight out of the datasheet and found it sounded terrible - a really, really harsh top end.
It took me a while to figure out why the sound was so lousy - in those early days, the recommended application circuit didn't look like fig1 in the current (dated 2003) datasheet, it looked more like the AC test circuit shown on p5. Notice that circuit has a 220pF between the inputs of the chip. I found when I removed that capacitor, amazingly all that harshness at the top end disappeared. From then on, I was hooked on chip amps.
The only problem I found with the sound of the LM3886 was when it ran into clipping - the protection circuits would activate and the waveform looked a mess. It also didn't sound great. But keep the beast out of clipping and it sounded fine. In the end I went for the TDA7294 in my final design because this part didn't have the messy clipping behaviour I found in the LM3886.
The greatest advantage of chip amps I found is the ease of building an active speaker. The sound quality improvement from ditching the passive crossover and putting the amps right up next to the drive units has to be heard to be believed. So much so that I pretty much have given up listening to passive speakers now. An active speaker driven by LM3886s to me has always sounded better than the same speaker with a passive XO driven by an amplifier of any hi-end credentials, provided the implementation is done right and its not allowed to clip.
I started playing with the LM3886 shortly after RS components introduced it into its UK catalogue. I don't recall exactly when this was, but sometime in the mid 1990s. First up, I built the application circuit straight out of the datasheet and found it sounded terrible - a really, really harsh top end.
It took me a while to figure out why the sound was so lousy - in those early days, the recommended application circuit didn't look like fig1 in the current (dated 2003) datasheet, it looked more like the AC test circuit shown on p5. Notice that circuit has a 220pF between the inputs of the chip. I found when I removed that capacitor, amazingly all that harshness at the top end disappeared. From then on, I was hooked on chip amps.
The only problem I found with the sound of the LM3886 was when it ran into clipping - the protection circuits would activate and the waveform looked a mess. It also didn't sound great. But keep the beast out of clipping and it sounded fine. In the end I went for the TDA7294 in my final design because this part didn't have the messy clipping behaviour I found in the LM3886.
The greatest advantage of chip amps I found is the ease of building an active speaker. The sound quality improvement from ditching the passive crossover and putting the amps right up next to the drive units has to be heard to be believed. So much so that I pretty much have given up listening to passive speakers now. An active speaker driven by LM3886s to me has always sounded better than the same speaker with a passive XO driven by an amplifier of any hi-end credentials, provided the implementation is done right and its not allowed to clip.
That's EXACTLY why I'm interested in these.
To use with an active system.
Audio is so full of contradiction, that I get distracted, and the past few days I've been thinking of building one of these 4" full range speakers systems (Fostex 127e in a vented box -- I'm sure you've seen the "hoopla").
I have a subwoofer and listen at low volume levels anymore, so it may be worth trying -- if I like it I may skip my active plans for now.
But I don't know -- the original plan was to use the sub and an MTM speaker which setup the need for (1) mixed mono sub amp with 2 LM3886 in parallel (1)
stereo LM3886 for the low/mid (probably not enough power) and (1) stereo LM3886 for the tweeters.
I have a USB preamp already built and thought I'd add the crossover to the board with maybe a simple jumper system on the board itself for crossover points.
I've been wanting to do something that in theory is a vast improvement over conventional passive crossovers and "a stereo amp".
Have you ever heard the 4" full range speakers in a bigger box?
I must admit when they are used in near field monitors they sound incredible.
To use with an active system.
Audio is so full of contradiction, that I get distracted, and the past few days I've been thinking of building one of these 4" full range speakers systems (Fostex 127e in a vented box -- I'm sure you've seen the "hoopla").
I have a subwoofer and listen at low volume levels anymore, so it may be worth trying -- if I like it I may skip my active plans for now.
But I don't know -- the original plan was to use the sub and an MTM speaker which setup the need for (1) mixed mono sub amp with 2 LM3886 in parallel (1)
stereo LM3886 for the low/mid (probably not enough power) and (1) stereo LM3886 for the tweeters.
I have a USB preamp already built and thought I'd add the crossover to the board with maybe a simple jumper system on the board itself for crossover points.
I've been wanting to do something that in theory is a vast improvement over conventional passive crossovers and "a stereo amp".
Have you ever heard the 4" full range speakers in a bigger box?
I must admit when they are used in near field monitors they sound incredible.
output resistors..
I found this old thread and because the title fully meets my question I saw no need to start a new thread.
I´m planning to design a LM3886 amplifier PCB based on the following schematics below. According to TI´s application notes the output resistors should be 0.1R, 3W, 1%.
The designer of the schematic found these hard to source so he used three 0.5R, 1W, 1% paralleled resistors (R6, R15, R16 and R11, R13, R14).
Over at Decibel Duncan´s webpage I found a different reason why not to use the 0.1R, 3W, 1% resistors in series.
I have obviously never heard the difference, but I'd like to get some opinions about the wirewound resistors vs carbon film or even metal film resistors in this part particular of the circuit.
Also, what does such a paralleled resistor circuit do to the sound compared the series resistor? ...or does that not have any influence at all? Just trying to understand, so I can make the best choices for my PCB design🙂
I found this old thread and because the title fully meets my question I saw no need to start a new thread.
I´m planning to design a LM3886 amplifier PCB based on the following schematics below. According to TI´s application notes the output resistors should be 0.1R, 3W, 1%.
The designer of the schematic found these hard to source so he used three 0.5R, 1W, 1% paralleled resistors (R6, R15, R16 and R11, R13, R14).
Over at Decibel Duncan´s webpage I found a different reason why not to use the 0.1R, 3W, 1% resistors in series.
I have obviously never heard the difference, but I'd like to get some opinions about the wirewound resistors vs carbon film or even metal film resistors in this part particular of the circuit.
Also, what does such a paralleled resistor circuit do to the sound compared the series resistor? ...or does that not have any influence at all? Just trying to understand, so I can make the best choices for my PCB design🙂
Carbon film are noisy and offer no advantage over metal film, and the two are usually constructed the same way (helical resistive path on a ceramic substrate). Wirewound usually have the lowest noise, but unless specially wound (Ayrton-Perry winding - Wikipedia, the free encyclopedia), exhibit higher inductance.
See Selecting resistors for preamp, amplifier and other high-end audio applications for some numbers.
It is my understanding that these output resistors are there to minimize the effect of different output offsets between chips, so my main concern would be noise. I have used IRC GS3 0R1 3W resistors in my LM4780 amps, I suggest you check out the datasheet http://www.ti.com/lit/ds/symlink/lm4780.pdf basically two LM3886 on one chip.
If you want to go all out on the output resistors, maybe look for a 0.1R precision shunt resistor, basically a strip of metal if I understand correctly.
See Selecting resistors for preamp, amplifier and other high-end audio applications for some numbers.
It is my understanding that these output resistors are there to minimize the effect of different output offsets between chips, so my main concern would be noise. I have used IRC GS3 0R1 3W resistors in my LM4780 amps, I suggest you check out the datasheet http://www.ti.com/lit/ds/symlink/lm4780.pdf basically two LM3886 on one chip.
If you want to go all out on the output resistors, maybe look for a 0.1R precision shunt resistor, basically a strip of metal if I understand correctly.
I use 4 cm of 0.50mm diameter Constantan wire.
Which by the way is directly solderable and zero Temperature Coefficient (that's where the "constant" part of the name comes from).
Also known as "Kanthal".
Which by the way is directly solderable and zero Temperature Coefficient (that's where the "constant" part of the name comes from).
Also known as "Kanthal".
@Rodeodave
Thanks for the link. Interesting material. So wirewound and foil resistors both introduce less noise than other types. The foil type have the advantage of having a temperature coefficient of resistance (TCR) about 0 PPM, but these are much more expensive. Perhaps best to use these in the part of the circuit where the gain is determined or in or around the signal path before the signal gets amplified.
@AndrewT
I know how to calculate the resistance of a parallel resistor circuit. I would still like to know what a parallel resistor circuit does to the signal compared to a series resistor. In terms of signal integrety and/or soundwise.
@JMFahey
Thanks, I'll look into it.
Thanks for the link. Interesting material. So wirewound and foil resistors both introduce less noise than other types. The foil type have the advantage of having a temperature coefficient of resistance (TCR) about 0 PPM, but these are much more expensive. Perhaps best to use these in the part of the circuit where the gain is determined or in or around the signal path before the signal gets amplified.
@AndrewT
I know how to calculate the resistance of a parallel resistor circuit. I would still like to know what a parallel resistor circuit does to the signal compared to a series resistor. In terms of signal integrety and/or soundwise.
@JMFahey
Thanks, I'll look into it.
One of the best sounding resistor in that role is Caddock MP915/930 current sensing resistors, you can find them in all low values from Mouser.
A small heatsink should be used for safety (Mouser 532-577202B00).
I will have to change my PCB design radically when using resistors with such packages 🙂
Thanks.
- Status
- Not open for further replies.