Heh, yes, balanced world firstly allows to decouple grounds between signal source and amplifier, between signal ground and output power ground, between left and right channels. Single ended have it's own benefits, but really it must be forgotten as bad history of our marketologists to sell as much high-priced interconnecting wires as possible.
I suppose and this is my hope, your next design will be much more sophisticated if you really care about grounds. Now community have well-known 4.4 mm TRRS jack connection and 4-pin XLR. Have no doubt, amplifier is not alone in our system, it's better to have common-mode choke and input filter at the input, output filter with per-channel or triple-wire common mode choke at the output, mains low-pass and common mode filter with 20-30 pF parasitic interwinding capacitance.
IIRC one reason may be that the usual measurement technique for those DS numbers is to tie drain and source together then measure reverse breakdown of the PN junction.
Those are nice-looking boards, Jam. I don't know how much symmetry helps sound quality, but a clean layout is satisfying to look at. It means the designer cares about details, which likely carries over to the design as a whole.
FWIW, the low-res board layouts appear to show things as being misaligned, but everything is fine when you zoom in.
About balanced circuits and extreme grounding and shielding techniques, I've read about these things but haven't tried many of them yet. Like anything else, the deeper you look at it, the more complicated it gets. If you want to go for extreme performance, you need to really know what you're doing. It's probably not good enough, for instance, just to throw a filtered AC socket on the amplifier. You need to start thinking about balance in your power supply transformer windings, and all kinds of parameters that don't show up in most spec sheets and cost a lot of money if they do. And you need specialized measurements. I'm also not keen on doubling the amount of active circuitry. Nothing comes for free. In the end, I try to be realistic about my abilities and how far I want to go with this.
FWIW, the low-res board layouts appear to show things as being misaligned, but everything is fine when you zoom in.
About balanced circuits and extreme grounding and shielding techniques, I've read about these things but haven't tried many of them yet. Like anything else, the deeper you look at it, the more complicated it gets. If you want to go for extreme performance, you need to really know what you're doing. It's probably not good enough, for instance, just to throw a filtered AC socket on the amplifier. You need to start thinking about balance in your power supply transformer windings, and all kinds of parameters that don't show up in most spec sheets and cost a lot of money if they do. And you need specialized measurements. I'm also not keen on doubling the amount of active circuitry. Nothing comes for free. In the end, I try to be realistic about my abilities and how far I want to go with this.
Oh.. I've a suggestion that you might, or might not, like - simply this:
Mirror-symmetric circuit layout is the pointlessly hard way - it's only cultivated for interior shots for magazine reviews we all despise at some level, and leads to compromises in one channel, or the other, but usually both. And it takes twice the time.
My view, for a long time: set up the layout to best-serve the circuit and performance goals. Refine that one layout.
Then just repeat it for the other channel.
That's the only way to drive out common gremlins, or esp. differences between channels - when all we really wanted was identical performance in each.
Mirror-symmetric circuit layout is the pointlessly hard way - it's only cultivated for interior shots for magazine reviews we all despise at some level, and leads to compromises in one channel, or the other, but usually both. And it takes twice the time.
My view, for a long time: set up the layout to best-serve the circuit and performance goals. Refine that one layout.
Then just repeat it for the other channel.
That's the only way to drive out common gremlins, or esp. differences between channels - when all we really wanted was identical performance in each.
LOL, I'm not gonna lie... You're right. I did the mirror image thing for the fun of it, and I'm not worried about the minor differences between the layouts. If I have problems later, I will come back and admit in front of everyone that I should have listened to you.
Hi Martin,
If you think about there are numerous benefits to a symertical layout and not pointless as you have indicated, well worth thinking about the reasons......🙂
I leave you to your thought process.
Regards,
Jam
If you think about there are numerous benefits to a symertical layout and not pointless as you have indicated, well worth thinking about the reasons......🙂
I leave you to your thought process.
Regards,
Jam
Curious, Jam, can you elaborate? Seems like the argument for identical performance makes sense.
Edit: I admit, I slightly prefer my left channel layout over the left. I would like to have the input/output connectors centered, really need the outputs together because they share the relay, and this was the main motivation for the symmetrical layout at first. But I can easily move the output trace to the other side if I were to replicate the layouts. The input on one side poses a problem; I would have to go under rather than over with the connector traces. TBH, you could argue these accommodations themselves could lead to potential performance differences. Some people would say build it both ways and choose the best-sounding one by listening. I doubt I could hear the difference myself.
Bottom line is I like the way it looks, I'm pretty sure it will work fine, and I'm already negotiating with the Ferrari dealership 'cuz I just know this is gonna make me a rock star in the audio world. (Because of the magazine cover, I mean.)
Edit: I admit, I slightly prefer my left channel layout over the left. I would like to have the input/output connectors centered, really need the outputs together because they share the relay, and this was the main motivation for the symmetrical layout at first. But I can easily move the output trace to the other side if I were to replicate the layouts. The input on one side poses a problem; I would have to go under rather than over with the connector traces. TBH, you could argue these accommodations themselves could lead to potential performance differences. Some people would say build it both ways and choose the best-sounding one by listening. I doubt I could hear the difference myself.
Bottom line is I like the way it looks, I'm pretty sure it will work fine, and I'm already negotiating with the Ferrari dealership 'cuz I just know this is gonna make me a rock star in the audio world. (Because of the magazine cover, I mean.)
Last edited:
Hi Henry,
Some points to ponder
In production
1) Think of checking for mistakes(inspection process)
2) Reduction of errors in assembly
3)Repatebility of the assembly proccess especially in complex products.
From the engineerineg side of things
1) Generally symetrical layout leads to shorter trace lengths with the attendant reduction of capacitive and inductinve coupling which leads to
reduction of noise and crosstalk
2) Power supply lines ( for split rail supplies) would be similar ........something that you need to think about and the ramifications with regard to star and bus grounding
And ther are many others of course there are always some excertions mainly in non-complementary single ended circuits but some basic rules still apply.
Not trying for a symetrical layout in my humble opinion a lazy way out (I can't emphasize this more) and can lead to problems sometimes, someone I respect highly that works in RF design, says the same thing is true in the RF field.
I am not saying your design won't work but I think striving for the best layout is a noble goal, I have heard listened to the same circuit laid out by different people sound totally different ........exactly same components, board material and pcb house. Part of the reason some Chinese products do not sound too good.I I could tell you horror stories of bad layout issues but maybe another time.
That is why I admire the way you are going about your board layout, don't let the naysayers get you down....short cuts do not lead to a better design or sound for that matter.
-Jam
Today's funny.
Some points to ponder
In production
1) Think of checking for mistakes(inspection process)
2) Reduction of errors in assembly
3)Repatebility of the assembly proccess especially in complex products.
From the engineerineg side of things
1) Generally symetrical layout leads to shorter trace lengths with the attendant reduction of capacitive and inductinve coupling which leads to
reduction of noise and crosstalk
2) Power supply lines ( for split rail supplies) would be similar ........something that you need to think about and the ramifications with regard to star and bus grounding
And ther are many others of course there are always some excertions mainly in non-complementary single ended circuits but some basic rules still apply.
Not trying for a symetrical layout in my humble opinion a lazy way out (I can't emphasize this more) and can lead to problems sometimes, someone I respect highly that works in RF design, says the same thing is true in the RF field.
I am not saying your design won't work but I think striving for the best layout is a noble goal, I have heard listened to the same circuit laid out by different people sound totally different ........exactly same components, board material and pcb house. Part of the reason some Chinese products do not sound too good.I I could tell you horror stories of bad layout issues but maybe another time.
That is why I admire the way you are going about your board layout, don't let the naysayers get you down....short cuts do not lead to a better design or sound for that matter.
-Jam
Today's funny.
Attachments
Last edited:
Thanks for the information. I joked about it, but I've been thinking seriously about the suggestion. I hadn't considered design for production at all, but it's good to think like a pro even if you're just doing it as a hobby. There's always more to learn.
I set out just now to try a layout where both sides of the board are the same. It turns out to be harder than just copying and pasting because Kicad wants every component to have a unique annotation. If there's a way in Kicad to maintain a base layout and tile it on the board, I don't know the technique yet.
I mostly avoided doing anything ugly to achieve the left-right symmetry. I put some real thought into the ground bus routing and it's the same, side-to-side. There are slight differences in a few trace lengths and routing but I don't think enough to make a major difference. I don't take any comments as trying to get me down. It's all good stuff to think about.
I totally believe you about different layouts sounding different. I've seen some good articles explaining some of the reasons for that.
My goal here is to mostly to have some fun, and if I'm lucky, I might get a decent amplifier out of it.
Thanks, Hugh.
Martin, the thread belongs to everyone, so please feel free to post anything that seems relevant.
Hi Martin,
Point taken.
I mentioned exceptions....for example noisy traces in a symertical design should be moved away from small signal traces but on the whole these rules are still valid. You have to weigh several factors in the final layout. Trying to keep ground and power rail lengths the same has some value..........when considering the capacitance between each rail and a ground plane for example. It migt be worth an experiment on one of your own designs, to see this works. Phono equilizer preamps are a great way to try this type of layout.
Regards,
Jam
Point taken.
I mentioned exceptions....for example noisy traces in a symertical design should be moved away from small signal traces but on the whole these rules are still valid. You have to weigh several factors in the final layout. Trying to keep ground and power rail lengths the same has some value..........when considering the capacitance between each rail and a ground plane for example. It migt be worth an experiment on one of your own designs, to see this works. Phono equilizer preamps are a great way to try this type of layout.
Regards,
Jam
Last edited:
https://www.audiosciencereview.com/forum/index.php
Ok, the thing is, very objective stuff like ground layout can have a significant measurable effect on noise and distortion. Poor routing of the output stage can couple harmonic currents into low-level signal traces. Differences in parasitic capacitance and inductance can affect high frequency response and loop stability. Leakage current into high-impedance nodes can be different. EMI susceptibility may be different. If you have an output inductor, you can get distortion due to magnetic saturation in nearby components.
I'm not saying listener bias isn't a factor some or most of the time. But changing the layout can definitely affect performance in non-trivial ways.
Last edited:
I have a release candidate. Since I think it's absolutely perfect right now, I will keep checking for a few more days.
What about same circuit laid out by different people making them sound totally different?
Well it is getting a bit off topic but if four different people layout the same RF power amplifier circuit you can get back any of the following:
1. A linear amplifier.
2. A not very linear amplifier.
3. An oscillator.
4. A firecracker.
Do a bad job of the grounding layout on a power amplifier and you can definitely hear it.
1. A linear amplifier.
2. A not very linear amplifier.
3. An oscillator.
4. A firecracker.
Do a bad job of the grounding layout on a power amplifier and you can definitely hear it.