That's what I wrote on post #215. 😉there must be a reason.
I answered in post #216. I don't see the point in telling people what they hear and don't hear, especially if you weren't there to listen for yourself.
To everyone that has a question, when I conducted this experiment, levels were matched to 0.1 dB (same components the had been turned on 24hrs prior) and everyone in the listening panel noticed the difference.
That is all I have to say on the subject.
Evenharmonics,
It has been your MO to disrupt other threads with your personal opinions on what people are supposed to hear and not hear and frankly we don't need that noise in this thread. Please feel free to migrate to another thread to carry on your antics.
Best, regards,
Jam
That is all I have to say on the subject.
Evenharmonics,
It has been your MO to disrupt other threads with your personal opinions on what people are supposed to hear and not hear and frankly we don't need that noise in this thread. Please feel free to migrate to another thread to carry on your antics.
Best, regards,
Jam
The layout looks great, care to share how the bottom layer looks?
Sure. The main update for today was to the relay circuit. I hadn't changed the signal layout at that point, which is why I just posted the one image. Just now, I did some minor work around the bias spreaders. And I've made many, many small adjustments, aligning things perfectly and so on.
I like the OSH Park Gerber viewer for design verification. I'm trying to decide if it's worth spending $106 for three boards to get ENIG finish and support the home team. I'm not very fond of purple.
Perhaps you misread what I wrote. I posted why they hear.I answered in post #216. I don't see the point in telling people what they hear and don't hear, especially if you weren't there to listen for yourself.
As I suspected, no bias control implemented.To everyone that has a question, when I conducted this experiment, levels were matched to 0.1 dB (same components the had been turned on 24hrs prior) and everyone in the listening panel noticed the difference.
Don't shoot the messenger. https://www.aes.org/e-lib/browse.cfm?elib=3839Evenharmonics,
It has been your MO to disrupt other threads with your personal opinions on what people are supposed to hear and not hear and frankly we don't need that noise in this thread.
"I have heard listened to the same circuit laid out by different people sound totally different ........exactly same components, board material and pcb house." is an extraordinary claim without extraordinary evidence. Perhaps you can follow up with supporting evidence for such claim. 🤔
I like the OSH Park Gerber viewer for design verification. I'm trying to decide if it's worth spending $106 for three boards to get ENIG finish and support the home team. I'm not very fond of purple.
Maybe try their "after dark" service instead? I used to order those PCBs all the time before I had to pay import stuff for every Oshpark order. Click the spoiler for an example. It costs just as much, you just have to tick the right box after adding the design to your cart. Oshpark is really great for tiny PCBs because they exclusively charge by PCB area, for larger ones it gets expensive quite quickly.
Last edited:
Good suggestion. The After Dark option is cool, but for some reason I can't decide if I like it or not. Probably the best thing to do is to build the prototype on a cheap board so I don't waste a lot of money if I need to make changes. If I really have the urge to build one of these on a gold-plated board, I can always make another one later.
BTW, reviewing the purple images above, I spotted two errors from my changes tonight. They're on the inboard traces connected to the bias pots. So no need to call those out if you see them, but if anyone finds anything else, by all means let me know.
BTW, reviewing the purple images above, I spotted two errors from my changes tonight. They're on the inboard traces connected to the bias pots. So no need to call those out if you see them, but if anyone finds anything else, by all means let me know.
Henry,
That is one awesome looking board..... you got me beat.
One small comment....one of my mentors insisted that traces should enter pads at 90 degrees whenever possible.........so this is a no no according to him.😉
Congratulations are in order.
Jam
That is one awesome looking board..... you got me beat.

One small comment....one of my mentors insisted that traces should enter pads at 90 degrees whenever possible.........so this is a no no according to him.😉
Congratulations are in order.
Jam
Attachments
Thanks so much, Jam. LOL, that is one of the mistakes I mentioned above. I moved two resistors up and forgot to fix the traces on the bottom pads.
Edit: I'm impressed you caught it! I appreciate the review.
Edit: I'm impressed you caught it! I appreciate the review.
Henry,
One suggestion with your layout - and only that - is that in my experience a long line of resistors can confuse with stuffing and should be avoided.
Along the bottom of each channel, you have eleven resistors in a row. If you break it up with larger gaps to say 4, then a gap, then 3, another gap, and then another 4, stuffing them is easier to select correct resistors. This breaks the memory issue to only 4 max resistors, not to 11. It is not better in terms of appearance or function, but it can reduce stuffing errors.
EvenHarmonics,
You are opening up a can of worms of subj v. obj. This is a perennial dilemma and does nothing to assist a designer laying out his excellent circuit.
Effectively it is only ever a urinary match. Asking for difficult, complex and measurement differences is beyond most designers anyway unless they have a laboratory and deep understanding of measurement and setting up a double blind test with a large sample. Besides, there are thousands of products slanging it out there in the marketplace and in the finish people select what they like to hear anyway. Those who buy only on specs and measurements are more than catered for. In short, your post is superfluous.......
HD
One suggestion with your layout - and only that - is that in my experience a long line of resistors can confuse with stuffing and should be avoided.
Along the bottom of each channel, you have eleven resistors in a row. If you break it up with larger gaps to say 4, then a gap, then 3, another gap, and then another 4, stuffing them is easier to select correct resistors. This breaks the memory issue to only 4 max resistors, not to 11. It is not better in terms of appearance or function, but it can reduce stuffing errors.
EvenHarmonics,
You are opening up a can of worms of subj v. obj. This is a perennial dilemma and does nothing to assist a designer laying out his excellent circuit.
Effectively it is only ever a urinary match. Asking for difficult, complex and measurement differences is beyond most designers anyway unless they have a laboratory and deep understanding of measurement and setting up a double blind test with a large sample. Besides, there are thousands of products slanging it out there in the marketplace and in the finish people select what they like to hear anyway. Those who buy only on specs and measurements are more than catered for. In short, your post is superfluous.......
HD
Cosmetically maybe, but otherwise you gotta be joking.One small comment....one of my mentors insisted that traces should enter pads at 90 degrees whenever possible.........so this is a no no according to him.😉
Electrically you have a point but there are other issues to be considered. My friend was a profesional board designer and insisted on using a set of best practice rules for consistancy. In real world applications when tiny traces are used to attach to a pad at the wrong angle it can cause etching problems during the board fabrication process, which could lead to a weak connection. They are known as acid traps.
Most designers are aware of the problems caused by acute angles in a circuit board and are therefore trained to avoid them. However, mistakes do happen. Often, PCB acid traps are the result of simple human error, although some design software programs may also set circuits to acute angles if the settings are not properly adjusted.
Now on the other hand some audio designers insist on not laying out right angle traces (lose electrons when they spill over the edge
) Seriously this has large ramificaions in RF design. Having a set of guidelines and rules can't hurt, surely, it leads to better and more consistant designs and products.
Some light reading
https://resources.altium.com/p/top-5-pcb-design-guidelines-every-pcb-designer-needs-knowhttps://resources.altium.com/p/pcb-impedance-control
Regards,
Jam
ACID TRAPS
“Acid trap” is the common term for acute angles in a circuit. They are referred to this way because these acute angles trap acid during the PCB etching process, allowing the acid to build up in the nook of the angle. The angle functionally keeps the acid in the corner for a longer period than the design calls for, causing the acid to eat away more than intended. As a result, the acid can compromise a connection, making the circuit defective and causing more serious problems later on.Most designers are aware of the problems caused by acute angles in a circuit board and are therefore trained to avoid them. However, mistakes do happen. Often, PCB acid traps are the result of simple human error, although some design software programs may also set circuits to acute angles if the settings are not properly adjusted.
Now on the other hand some audio designers insist on not laying out right angle traces (lose electrons when they spill over the edge

Some light reading
https://resources.altium.com/p/top-5-pcb-design-guidelines-every-pcb-designer-needs-knowhttps://resources.altium.com/p/pcb-impedance-control
Regards,
Jam
Last edited:
They probably have read e.g. TI's high speed layout guidelines:Now on the other hand some audio designers insist on not laying out right angle traces (lose electrons when they spill over the edge)
https://www.ti.com/lit/an/scaa082a/scaa082a.pdf(see chapter 2.5)
Thanks for the PCB suggestions. I've read that acid traps are less of a problem nowadays because of improved production techniques, but I try to avoid them anyway. Traces entering pads at off angles look bad, though, IMHO. WRT breaking up long rows of resistors, that's a good idea and I will incorporate it in my next design. I used to avoid "tee" junctions but apparently they're not a problem for low-speed circuits so I use them sometimes. For instance, the output ground traces are tees. I played with changing that and couldn't decide which way I preferred. The electrons probably don't care.
I have this acute angle here, and it annoys me. I may change it, but for now I've decided it's the best compromise for mainly arbitrary aesthetic reasons.
It would be nice for the vertical trace going to that pad to be the same distance from the board midline as its twin on the other channel, but then that forces me into another compromise.
For the power supply, I plan to start with a pair of prototype super regulators I have left over from my first DCG3 project. Eventually I will build the "official" regulator and then I'll be able to compare.
I have this acute angle here, and it annoys me. I may change it, but for now I've decided it's the best compromise for mainly arbitrary aesthetic reasons.
It would be nice for the vertical trace going to that pad to be the same distance from the board midline as its twin on the other channel, but then that forces me into another compromise.
For the power supply, I plan to start with a pair of prototype super regulators I have left over from my first DCG3 project. Eventually I will build the "official" regulator and then I'll be able to compare.
- Home
- Amplifiers
- Headphone Systems
- New Headphone Amplifier Design