The DIY 2.1 amplifier. Should I build it? Would you buy it?

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

The 221 ohm resistor for the VCC/2 op amp should be outside the feedback loop and could be as small as 50 ohms without any problems I think. The smaller the value the less signal voltage will appear your active ground.

I've seen a virtual ground driver circuit drawn both ways (with the feedback before and after the series resistor). I think the reason for having the feedback after the series resistor is that the op-amp will compensate for any voltage droop caused by current draw through the series resistor. Which do you think is better? I will yield to your expertise.

The point about the pot is that if the pot is driven by a low impedance the impedance looking at the wiper will have a maximum impedance of half the pots value. If the circuit after the pot has an impedance say 20 or more times the impedance at the wiper, different settings of the pot should not impact the frequency response of the filter.

Ok - now I understand you. I thought I had designed the circuit to avoid any interactions between the pot and the filter. On the master volume pot (R1), U1A and U1B form stereo buffers to prevent any filter interaction (U2/U3). On the subwoofer volume pot (R2), U4B is a buffer designed to prevent any filter interaction. In case anyone is wondering, the series resistors (R13/R14) are there prevent crosstalk between the left and right channels when the pot is at maximum volume.

By putting the pot after the mono summing op amp you could loss a pot since the level you are adjusting is mono. Probably a 5k pot min and 20 or more times that impedance means jfet input op amps would be advisable. You have fixed all the real issues and put me down for a couple of boards. PS don't use a ground plane, and don't be afraid the use jumpers to keep
It at a 2 layer board.

I chose to keep the solution as a dual pot because that increases purchasing power and helps reduce the cost. For example, the person who builds the design can easily source two of the same pot form eBay, etc. to help save money. A single pot + a dual pot usually doesn't save any money as versus two dual pots. A lesser reason is that it prevented me from having to create two separate symbols in the DesignSpark software....

I do have to ask, though - why "no ground plane"??? I've been doing audio layouts in PC's and consumer electronics for 15 years, and I'll admit that careful attention must be paid to the grounding, but I certainly wouldn't call out a blanket "no ground plane" rule. My experience has shown that planes are preferred to shield grounding. Star grounding can be beneficial, but it can be complicated to get right. Basically, a star ground can yield better performance than a plane, but it's so much easier to screw up and yields worse performance when it's bad. Do you have some specific concerns here? The design could be done as a one layer board, especially if I were to use jumpers. The pricing for a two layer is cheap enough to not worry about trying to make it a one layer.

[fdieck]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

I've seen a virtual ground driver circuit drawn both ways (with the feedback before and after the series resistor). I think the reason for having the feedback after the series resistor is that the op-amp will compensate for any voltage droop caused by current draw through the series resistor. Which do you think is better? I will yield to your expertise.




Ok - now I understand you. I thought I had designed the circuit to avoid any interactions between the pot and the filter. On the master volume pot (R1), U1A and U1B form stereo buffers to prevent any filter interaction (U2/U3). On the subwoofer volume pot (R2), U4B is a buffer designed to prevent any filter interaction. In case anyone is wondering, the series resistors (R13/R14) are there prevent crosstalk between the left and right channels when the pot is at maximum volume.



I chose to keep the solution as a dual pot because that increases purchasing power and helps reduce the cost. For example, the person who builds the design can easily source two of the same pot form eBay, etc. to help save money. A single pot + a dual pot usually doesn't save any money as versus two dual pots. A lesser reason is that it prevented me from having to create two separate symbols in the DesignSpark software....

I do have to ask, though - why "no ground plane"??? I've been doing audio layouts in PC's and consumer electronics for 15 years, and I'll admit that careful attention must be paid to the grounding, but I certainly wouldn't call out a blanket "no ground plane" rule. My experience has shown that planes are preferred to shield grounding. Star grounding can be beneficial, but it can be complicated to get right. Basically, a star ground can yield better performance than a plane, but it's so much easier to screw up and yields worse performance when it's bad. Do you have some specific concerns here? The design could be done as a one layer board, especially if I were to use jumpers. The pricing for a two layer is cheap enough to not worry about trying to make it a one layer.

I am not calling for a blanket ban on ground planes. Simple analog circuits don't usually need them since they just add stray capacitance if used for the sake of using them. The serious audio designers I know like tin plated copper and no solder mask with fairly thin signal traces.
Star grounding is a good idea and makes you think about you grounding and keeping the sensitive signal returns from sharing ground impedances with dirty power supply and decoupling return currents. Mixed signal grounding and decoupling is almost an art and there have been countless numbers of articles and books written on the subject. I would be surprised if I have read less than fifty articles of the subject.

I worked on a telecom line card that had four different grounding subgroups for digital, analog, lightning protection, and electrostatic discharge. Capacitive and electromagnetic coupling between all these grounds had to be considered. Keeping the ground bounce from resisting the micro or ASICs was a task I would not wish on anyone. I had the best layout guy
I will ever meet using Mentor Graphics software on a Sun workstation. Both of us nearly lost our minds and it took over a year out of my life. The board was only 2 layers in smaller in area than a shoe box. When someone comes out and says that competently designed audio products should sound the same, the word competently covers a lot of ground (no pun intended.

[fdieck]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

I've seen a virtual ground driver circuit drawn both ways (with the feedback before and after the series resistor). I think the reason for having the feedback after the series resistor is that the op-amp will compensate for any voltage droop caused by current draw through the series resistor. Which do you think is better? I will yield to your expertise.




Ok - now I understand you. I thought I had designed the circuit to avoid any interactions between the pot and the filter. On the master volume pot (R1), U1A and U1B form stereo buffers to prevent any filter interaction (U2/U3). On the subwoofer volume pot (R2), U4B is a buffer designed to prevent any filter interaction. In case anyone is wondering, the series resistors (R13/R14) are there prevent crosstalk between the left and right channels when the pot is at maximum volume.



I chose to keep the solution as a dual pot because that increases purchasing power and helps reduce the cost. For example, the person who builds the design can easily source two of the same pot form eBay, etc. to help save money. A single pot + a dual pot usually doesn't save any money as versus two dual pots. A lesser reason is that it prevented me from having to create two separate symbols in the DesignSpark software....

I do have to ask, though - why "no ground plane"??? I've been doing audio layouts in PC's and consumer electronics for 15 years, and I'll admit that careful attention must be paid to the grounding, but I certainly wouldn't call out a blanket "no ground plane" rule. My experience has shown that planes are preferred to shield grounding. Star grounding can be beneficial, but it can be complicated to get right. Basically, a star ground can yield better performance than a plane, but it's so much easier to screw up and yields worse performance when it's bad. Do you have some specific concerns here? The design could be done as a one layer board, especially if I were to use jumpers. The pricing for a two layer is cheap enough to not worry about trying to make it a one layer.

I am not calling for a blanket ban on ground planes. Simple analog circuits don't usually need them since they just add stray capacitance if used for the sake of using them. The serious audio designers I know like tin plated copper and no solder mask with fairly thin signal traces.
Star grounding is a great idea and makes you think about you grounding and keeping the sensitive signal returns from sharing ground impedances with dirty power supply and decoupling return currents. Mixed signal grounding and decoupling is almost an art and there have been countless numbers of articles and books written on the subject. I would be surprised if I have read less than fifty articles of the subject.

I worked on a telecom line card that had four different grounding subgroups for digital, analog, lightning protection, and electrostatic discharge. Capacitive and electromagnetic coupling between all these grounds had to be considered. Keeping the ground bounce from resisting the micro or ASICs was a task I would not wish on anyone. I had the best layout guy I will ever know using Mentor Graphics software on a Sun workstation. Both of us nearly lost our minds and it took over a year out of my life. The board was only 2 layers in smaller in area than a shoe box. When someone comes out and says that competently designed audio products should sound the same, the word competently covers a lot of ground (no pun intended.

[Melby Audio]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

So, the completed board without amplifiers or power supply or case comes to around $55? Also, could it be used with any amplifiers? or is there something that an amp needs for compatibility?

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

So, the completed board without amplifiers or power supply or case comes to around $55? Also, could it be used with any amplifiers? or is there something that an amp needs for compatibility?

That's probably about right. $40-$45 if you use NE5532 (an excellent choice for this type application). Add about $10 for using the LM4562. Just about any dual op-amp should work.

The base board could be used with any amplifier. The only compatibility issue with any generic amp is that the base board does not include any DC blocking capacitors on the outputs of the preamp (and it DOES have a DC offset of 1/2VCC). The target amplifier needs to have DC blocking caps, or you can simply add some 1uF film caps between the outputs of the base board and the inputs of your amplifier.

The way the board is designed, you can choose to use LR2 slopes or LR4 slopes (or any other type filter you want). It's designed with sallen-key architecture, so it's extremely flexible. You don't have to choose the values I listed, those were just for reference for an 80Hz LR4 crossover. Changing the values in the sallen-key filters will change the slope, Q, and topology.

[Navy Guy]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Definitely interested.

This might be a wierd question, but would it be possible to use teh base board to send a mono out to a sub that already has an amplifier as well?

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Definitely interested.

This might be a wierd question, but would it be possible to use teh base board to send a mono out to a sub that already has an amplifier as well?

Yep.

You wouldn't even need to populate all the components if you didn't want the high-pass crossover. You'd have to populate U1, U4, and the associated passives to get a sub output only, depending on just how little functionality you need...

Let me see if I can break this down a little more:

• R1 is the volume pot. If not using volume control, then you can just input the signal directly into the wiper of R1, or you can add two jumper wires to bypass the pot's location. This pot controls the master volume (left, right, and sub).
• U1 is a unity gain buffer amp. It helps keep the volume pot from affecting the performance of the sallen-key filters. It also helps maintain DC biasing through the design, since we are running this off of a single voltage source (rather than +/- rails).
• U2 is the crossover for the left channel. When using both op-amps, you'll obtain a -24dB/octave slope. It's possible to populate the passives for only one of the op-amps and bypass the second. This would allow for -12dB/octave slopes.
• U3 is the crossover for the right channel. When using both op-amps, you'll obtain a -24dB/octave slope. It's possible to populate the passives for only one of the op-amps and bypass the second. This would allow for -12dB/octave slopes.
• U4 is a dual function chip. The first op amp serves to produce the mid-rail biasing voltage. This is needed to operate the board off of a single DC voltage supply. The second op-amp serves as a buffer to protect the subsequent crossover from being affected by the subwoofer trim pot (R12).
• U5 is the crossover for the sub channel. When using both op-amps, you'll obtain a -24dB/octave slope. It's possible to populate the passives for only one of the op-amps and bypass the second. This would allow for -12dB/octave slopes.
• R12 is the trim pot for the subwoofer volume. It's basically used to just set the gain relative to the left/right channels. The left and right channels are unity-gain through this preamp. The subwoofer trim pot will allow for +6dB of gain down to -∞ of attenuation. If you don't want to populate a pot here, you can use fixed value resistors.
• R13 and R14 are the summing resistors. After this point, the left and right signals are summed together. These need to be a relatively high value to maintain crosstalk performance through the stereo left and right signals.
• C24 and C25 are extra bulk capacitors for the amp boards. These aren't required to be populated, but they'll likely help out by adding additional bulk capacitance to the power rail, and they'll help bypass the boards which can prevent intermodulation distortion and beat tones. On these amp boards, there's usually a SYNC signal that can be connected between two different amp boards that synchronizes the clocks of the two amplifiers. If they are not SYNC'd, there is a possibility for IMD and beat tones. However, if the amp boards are well bypassed, the chances of needing the SYNC signals is greatly reduced and virtually not necessary.

I hope this helps!

[fdieck]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

I think I would put some coupling caps on the board with 100k resistors to ground at the out put. If somebody plugs this powered up unit into an amplifier that is powered up they are going to put a big transient to the speakers even with a coupling cap on the amp input. Or someone will not understand that there is big DC offset and plug in a DC coupled amp and ....... Well you get the picture. Yes people should plug cables into powered up amps but we have all done it in a moment of carelessness.

[ani_101]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

am interested in two boards.... though i don't at this point understand everything that's going on, but hopefully it will be easy enough to learn. Hope you make it beginner friendly!!!

[Melby Audio]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

I am also interested. Not sure how many I would buy though, I'd need to think about what projects could use them. Would this be a one time purchase or are you thinking of selling them on a continuous basis?

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Melby, I'll probably pony up the money to build 50 and then sell them until they run out. After that, we'd have to re-visit.

Fred, you're totally right about the series caps at the outputs, but I'm opting to leave them off because I'm out of room in the layout. In the user's guide, I'll highlight the issue and provide some recommendations. There's nothing stopping an end user from adding one in series if the need arises. However, I just can't fit it and all of the amp boards I'm targeting have series caps built in.

Here's the initial layout image. Any feedback from anyone is welcomed:


At this point, I'm hoping some crowd sourcing would be available from the kind folks out on PETT. I'm really hoping to obtain the following before sending any boards off to the FAB. This will help me add amp board mounting holes in the base board in correct locations, and it will help me make sure everything actually fits. It will also help me identify any reworks required to get the amp boards configured into a 2.1 mode. I'm looking for the following:

One amp case (PE 320-326).
Two TDA7492 amp boards. They could come from PE (320-606) or from eBay.
Two TPA3116 amp boards.

There's also some TP2050 boards and a pre-configured 2.1 TPA3116 board that might fit well also, but those are secondary to me.

If anyone is willing to help out, send me a PM. I'm willing to provide a bare PCB for each $10 spent on equipment. For example, if you're willing to ship me an amp case (about $20), I'll ship you 2 PCB's when I get them. I hope people find this interesting and worthwhile... Enjoy!

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Just FYI - One member has purchased the TDA7492 boards for me, but I'm still looking for help on the amp case or TPA3116 boards.

Thanks to all those who are interested in helping out! I think this is going to be a useful board for the whole community, even if you don't put it into a case.

[scottq]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

What kind of timeline do you expect for the boards being done/shipped?

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

That depends on a few factors, but I expect about 30-45 days.

[tyger23]

Re: The DIY 2.1 amplifier. Should I build it? Would you buy it?

Just a quick status update. We have an extraordinary community here at PETT, and members have stepped up to provide all the needed items. The project is proceeding, and I hope to gerber out the board within a week of receiving the ordered items.

There are a few changes I want to make to reduce costs or increase flexibility, and I'm hoping to get some of my audio industry friends to review the schematics and layout prior to sending the board to FAB. I'll keep you guys updated! Thanks to all who donated to the cause!