There are 4 boards left, price is 25euro each including VAT ex shipping, leave me a PM if you are interested. An initial english version of the manual is expected coming week. This is the board's schematic :
An externally hosted image should be here but it was not working when we last tested it.
Non Disaster Shipment
Folks, the shipping fees for 1 to 25 pcb's with our national PostNL are eu: 18 euro and worldwide: 24 euro. These shipments are insured for a max of 500 euro.
Folks, the shipping fees for 1 to 25 pcb's with our national PostNL are eu: 18 euro and worldwide: 24 euro. These shipments are insured for a max of 500 euro.
In this case post #473 about
http://www.diyaudio.com/forums/chip-amps/226437-optimizing-tda7294-output-48.html
provides also a good description.
I still want to know commercial amplifiers, where this kind of building was realized.
Check out this thread concerning parallel mode operation:
http://www.diyaudio.com/forums/chip-amps/265290-about-jeff-rowland-lm3886-amplifiers.html
To answer your original question: The abundance of poor quality PCBs is probably because few people have the experience to lay out a PCB that delivers good performance. Also, many don't bother with measurements, so they're blissfully unaware of how their layout affects the overall performance of the amp.
Point-to-point is an easy way to muck up the layout as well. I measured the performance of various P2P layouts here: http://www.diyaudio.com/forums/chip-amps/252436-lm3886-pcb-vs-point-point-data-3.html#post3846783
I highly doubt you'll find P2P construction in any commercial design. It's simply too expensive. It's also rather difficult to get consistent performance, so in addition to the high assembly cost, the manufacturer would have to perform extensive testing to ensure that the product was assembled correctly.
Rowland's circuit is interesting. I think some improvements could be had in the PCB, but the circuit looks well executed. Do note the adjustment pot for each channel, however. Paralleling LM3886es is tricky business. Often better results can be had by paralleling the two channels in an LM4780 as the two halves tend to be closer in performance than two LM3886es.
Tom
Point-to-point is an easy way to muck up the layout as well. I measured the performance of various P2P layouts here: http://www.diyaudio.com/forums/chip-amps/252436-lm3886-pcb-vs-point-point-data-3.html#post3846783
I highly doubt you'll find P2P construction in any commercial design. It's simply too expensive. It's also rather difficult to get consistent performance, so in addition to the high assembly cost, the manufacturer would have to perform extensive testing to ensure that the product was assembled correctly.
Rowland's circuit is interesting. I think some improvements could be had in the PCB, but the circuit looks well executed. Do note the adjustment pot for each channel, however. Paralleling LM3886es is tricky business. Often better results can be had by paralleling the two channels in an LM4780 as the two halves tend to be closer in performance than two LM3886es.
Tom
Seems to me there are many advantages of simple-to-understand layouts. Many of these PCB will be first builds. Easy to follow traces are appreciated by new builders. If you can optimize the board and hear a difference then +1.)
A good layout can still have easy-to-follow tracks and good assembly instructions. My point is that if you're bothering to design or buy a PCB, why not make it a good one?
Tom
Tom
A good layout will not look busy and thus the tracks will be easy to follow, it will flow (as the schematic flows), will be neat and will take consideration of electrical requirements for layout including EMC requirements and current flow/loops🙂
I guess I'm really trying to express that there's a 'good enough' point. Refinement? Absolutely! I'm all for it.
'Good enough' like so many things will vary from person to person -almost- as much as it varies from one component's integration to the next. With one op-amp a certain sub-optimal layout might be a non-starter while in others it could be inaudible.
In the case of the LM series and pre-made PCBs my bigger concern would be using a good, solid layout as a teaching tool. "This is why the layout looks as it does". It would also be a good selling point, IMO.
Sonics are important, but more of us knowing what we're about is even more so.
'Good enough' like so many things will vary from person to person -almost- as much as it varies from one component's integration to the next. With one op-amp a certain sub-optimal layout might be a non-starter while in others it could be inaudible.
In the case of the LM series and pre-made PCBs my bigger concern would be using a good, solid layout as a teaching tool. "This is why the layout looks as it does". It would also be a good selling point, IMO.
Sonics are important, but more of us knowing what we're about is even more so.
PCBs for Audio are either going to be analogue or digital (or both on the same board) so the rules for analogue or digital layout apply....
Basics for analogue... keep all current loops as small as possible especially local feedback and decoupling, avoid running low level signals where they can be influenced by high di/dt signals....
And for theses chip amps I'd recommend reading Tomchrs' threads....
Your starters for PCB design related links...
Basics for analogue... keep all current loops as small as possible especially local feedback and decoupling, avoid running low level signals where they can be influenced by high di/dt signals....
And for theses chip amps I'd recommend reading Tomchrs' threads....
Your starters for PCB design related links...
This is quite possibly the best LM3886 PCB thread I've participated in: http://www.diyaudio.com/forums/chip-amps/270206-my-take-lm3886-layout.html
Main selling point: A good layout is arrived at in Post #7. The entire thread is two pages at this time.
Tom
Main selling point: A good layout is arrived at in Post #7. The entire thread is two pages at this time.
Tom
Modulus-86, Parallel-86... 😉
See my website: Neurochrome Audio: Precision high performance audio circuits for the DIY market.
Tom
See my website: Neurochrome Audio: Precision high performance audio circuits for the DIY market.
Tom
It takes quite a bit of experience to get pcb layouts right.
Separating signal and power paths is vitally important.
I remember my first project with a USB mixer.
With the inputs shorted I was getting 1 volt ac on the output !!
I had mixed in the small signals with the power supply charging pulses.
This modulated the signals nicely.
So I sat down and worked out what was high power and what was low power and separated them the best I could. The next pcb revision was almost completely hum free.
I have found the same with chip amps. Great care has to be taken on they arent hum free.
Its too easy to just take the pcb as it comes and hope it works.
It can take a lot of effort to get it right.
Separating signal and power paths is vitally important.
I remember my first project with a USB mixer.
With the inputs shorted I was getting 1 volt ac on the output !!
I had mixed in the small signals with the power supply charging pulses.
This modulated the signals nicely.
So I sat down and worked out what was high power and what was low power and separated them the best I could. The next pcb revision was almost completely hum free.
I have found the same with chip amps. Great care has to be taken on they arent hum free.
Its too easy to just take the pcb as it comes and hope it works.
It can take a lot of effort to get it right.
Yep. Anybody can connect the dots. Actually designing a good layout takes experience and thought.
Tom
Tom
People dont always take care when designing the pcb.
Its vital the feedback resistor path is as short as possible.
1mm of pcb track = 1nH of inductance.
Care should be taken using star grounding where possible.
Decoupling should be close to chip.
Its vital the feedback resistor path is as short as possible.
1mm of pcb track = 1nH of inductance.
Care should be taken using star grounding where possible.
Decoupling should be close to chip.
I agree with those.
Actually, star grounding is about the worst you can do. A well-designed ground plane will give you better performance. I describe why here:
Taming the LM3886 Chip Amplifier: Grounding – Neurochrome
Tom
I have had good results from star grounding.
It depends where you star ground and how.
Some people do it at the pcb connections which is almost right.
The main star ground should be at the connection between the two smoothing caps.
So the best way to do it is to star ground pcb then connect that to star ground of smoothing caps. The speaker return should only go to smoothing caps and never the pcb.
Of all my many designs the only one that hummed was a LM3886 pcb. All the rest were dead silent.