Alex, great job!
Almost everything looks good
My last 2 pieces of concern are these:
1.) I'm not sure if I completely like the grounding scheme yet. For example, where will the negative (ground) terminal of the speaker out connect to? It can connect anywhere on the ground plane, but it should be right next to the speaker + (out). Therefore it should have a hole and a thermal pad on the ground plane. Same goes for the protected speaker output.
This maybe Bigpanda can help... For example, how can one solder the radial capacitors to the ground plane? Once the capacitor is sitting on the board, it will cover the solder pad. Will the hole in the PCB be through plated?
2.) Is everyone happy with the choice of the relay? Omron will work nice here. But if you want something a little higher quality like Panasonic, they'll have slightly different leg/pin layout. So the board might have to be modified a little to make different kinds of relays work, unless we all agree on a specific relay.
Other than these 2 concerns, the board is flawless and is ready to be made.
Almost everything looks good
My last 2 pieces of concern are these:
1.) I'm not sure if I completely like the grounding scheme yet. For example, where will the negative (ground) terminal of the speaker out connect to? It can connect anywhere on the ground plane, but it should be right next to the speaker + (out). Therefore it should have a hole and a thermal pad on the ground plane. Same goes for the protected speaker output.
This maybe Bigpanda can help... For example, how can one solder the radial capacitors to the ground plane? Once the capacitor is sitting on the board, it will cover the solder pad. Will the hole in the PCB be through plated?
2.) Is everyone happy with the choice of the relay? Omron will work nice here. But if you want something a little higher quality like Panasonic, they'll have slightly different leg/pin layout. So the board might have to be modified a little to make different kinds of relays work, unless we all agree on a specific relay.
Other than these 2 concerns, the board is flawless and is ready to be made.
Hi Nagys,
The board will be through plated of course. However, there is another problem about the resist. Do you really want the upper copper to be un-resist? I had been talking with the fab. house for this subject and they will confirm with me very soon if the silkscrren can be printed on the tin plated surface. BTW if they can, what color will you think it is best for the silkscreen? Remember, it has to have a nice contrast to be read.
I have also give some thought to the AC traces that has been discussed by other members. I have make a small change so there is no AC trace running around at all. Sounds childish but it may be a solution. See attach. Just a rough sketch. Needs refinement.
The power trace can be broaden (if it is necessary) if we expand the board in the top direction say by 3mm?
Alex, maybe you have missed this : the left and right led at the lower corners are arranged differently. Do not know why.
Hope I am making some sense. Not trying to be picky, just want to make things better (if that can do it).
The board will be through plated of course. However, there is another problem about the resist. Do you really want the upper copper to be un-resist? I had been talking with the fab. house for this subject and they will confirm with me very soon if the silkscrren can be printed on the tin plated surface. BTW if they can, what color will you think it is best for the silkscreen? Remember, it has to have a nice contrast to be read.
I have also give some thought to the AC traces that has been discussed by other members. I have make a small change so there is no AC trace running around at all. Sounds childish but it may be a solution. See attach. Just a rough sketch. Needs refinement.
The power trace can be broaden (if it is necessary) if we expand the board in the top direction say by 3mm?
Alex, maybe you have missed this : the left and right led at the lower corners are arranged differently. Do not know why.
Hope I am making some sense. Not trying to be picky, just want to make things better (if that can do it).
Attachments
Hi Alex, please see the drawing below. Is it possible for you to make these changes?
1.) Circled in RED. Ground connections for speaker outputs (protected and not protected) and ground connections for 60VAC. This will make it a lot easier to solder.
2.) Circled in BLUE. Can you please add thermal heat pads for these?
3.) Circled in YELLOW. There's a slight aesthetic mistake on the traces.
4.) Traced in ORANGE. Can you make these traces thicker?
5.) Traced in GREEN. These traces are not needed.
6.) Traced in PURPLE. Can you please make these as traces on the bottom of the PCB and not on the top ground plane? It's better not to interrupt and divide the ground plane. Especially in that area.
7.) Capacitors C33 (1uF) and C36 (1uF) are missing from the PCB layout. Can you please include these? They can definitely affect the sonic quality of the amplifier.
Also Alex, after these changes, would it be possible for you to post a picture here of the ground plane as well? I would like to see what it looks like and make sure there's no isolated areas.
Bigpanda, I cannot clearly understand what you did in the picture. Would it be possible to make another drawing? If I understand correctly, there will be 2 places where the 60VAC is connected to. I think it's better if it connects at just one place. Less wires and less chances of noise.
As for the solder mask/resist. I personally would strongly 100% prefer if the board did not have any. The original Goldmund amplifier does not use solder mask on the top ground plane. I also prefer that the board is only tin plated, I no longer care for the electroless nickel finish. And for the color of the silk screen, how about black?
Thank you everyone for putting in so much effort in making this project available for everyone to build
1.) Circled in RED. Ground connections for speaker outputs (protected and not protected) and ground connections for 60VAC. This will make it a lot easier to solder.
2.) Circled in BLUE. Can you please add thermal heat pads for these?
3.) Circled in YELLOW. There's a slight aesthetic mistake on the traces.
4.) Traced in ORANGE. Can you make these traces thicker?
5.) Traced in GREEN. These traces are not needed.
6.) Traced in PURPLE. Can you please make these as traces on the bottom of the PCB and not on the top ground plane? It's better not to interrupt and divide the ground plane. Especially in that area.
7.) Capacitors C33 (1uF) and C36 (1uF) are missing from the PCB layout. Can you please include these? They can definitely affect the sonic quality of the amplifier.
Also Alex, after these changes, would it be possible for you to post a picture here of the ground plane as well? I would like to see what it looks like and make sure there's no isolated areas.
Bigpanda, I cannot clearly understand what you did in the picture. Would it be possible to make another drawing? If I understand correctly, there will be 2 places where the 60VAC is connected to. I think it's better if it connects at just one place. Less wires and less chances of noise.
As for the solder mask/resist. I personally would strongly 100% prefer if the board did not have any. The original Goldmund amplifier does not use solder mask on the top ground plane. I also prefer that the board is only tin plated, I no longer care for the electroless nickel finish. And for the color of the silk screen, how about black?
Thank you everyone for putting in so much effort in making this project available for everyone to build
Attachments
Last edited:
The speaker returns shouldn't be next to the speaker terminal! Ever heard of a star ground? Although there's a "ground plane" (not really) the ground for some nets needs to be arranged like a star ground. The upper connections Nagys proposed are definitely wrong! The speaker return should be close to the main ground connector (the other red encircled ones).
Input ground. Catastrophic. This is definitely worse than before. How long is the distance the input ground has to flow over the "plane" before it reaches the main ground connector? Almost the whole pcb! On it's long way it shares what's left from the plane fill (bottlenecks) with the currents of the huge capacitors and the speaker return. I'm sure Alex intended to connect the stub (the GREEN one Nagys proposed to delete ) to the main ground connector and have no other connection to plane for the sensitive ground?
Nagys ORANGE. Never run a dead end to a capacitor.
Added thermals are good. Still some missing.
Please can someone explain me why you don't want solder mask??? There can't be a technical explanation that makes sense. Nagys, please?
Input ground. Catastrophic. This is definitely worse than before. How long is the distance the input ground has to flow over the "plane" before it reaches the main ground connector? Almost the whole pcb! On it's long way it shares what's left from the plane fill (bottlenecks) with the currents of the huge capacitors and the speaker return. I'm sure Alex intended to connect the stub (the GREEN one Nagys proposed to delete
) to the main ground connector and have no other connection to plane for the sensitive ground? Nagys ORANGE. Never run a dead end to a capacitor.
Added thermals are good. Still some missing.
Please can someone explain me why you don't want solder mask???
There can't be a technical explanation that makes sense. Nagys, please?
Lee Knatta - I don't like solder mask because it makes it harder to build an amplifier. What if I decide to change something, or use different size parts and I need to solder to the ground plane? The solder mask is tough and hard to remove/scrape off. After it's removed in spots, the board looks ugly. Without the solder mask, I can solder anywhere on the ground plane. Plus, the original Goldmund Mimesis amplifiers did not use solder mask.
Why is the ground point on the bottom of the PCB the main ground connection? The main ground connection is the entire ground plane! I don't understand why I would ever want my speaker wires to be separated by an entire board? At the very least it will look ugly and unorganized.
Here are some pictures of the original Goldmund Mimesis 6 (similar to Mimesis 3 and Mimesis 9.2) amplifier. As you can see on the bottom left, the speaker output (white and black wires) are right next to each other. The negative is connected directly to the ground plane. Also, the input is right between the two large capacitors and directly above the A1 module. The ground is again connected directly to the ground plane. If you look carefully, you can see that every ground point in the entire schematic is connected to the ground plane. This ground plane/circuit board is NOT connected to the amplifier's chassis. The chassis is separately connected to ground on the IEC.
Here's a picture of the other side of the circuit board. There are no ground traces of any kind. Every single ground point is soldered to the top ground plane.
Why is the ground point on the bottom of the PCB the main ground connection? The main ground connection is the entire ground plane! I don't understand why I would ever want my speaker wires to be separated by an entire board? At the very least it will look ugly and unorganized.
Here are some pictures of the original Goldmund Mimesis 6 (similar to Mimesis 3 and Mimesis 9.2) amplifier. As you can see on the bottom left, the speaker output (white and black wires) are right next to each other. The negative is connected directly to the ground plane. Also, the input is right between the two large capacitors and directly above the A1 module. The ground is again connected directly to the ground plane. If you look carefully, you can see that every ground point in the entire schematic is connected to the ground plane. This ground plane/circuit board is NOT connected to the amplifier's chassis. The chassis is separately connected to ground on the IEC.
An externally hosted image should be here but it was not working when we last tested it.
Here's a picture of the other side of the circuit board. There are no ground traces of any kind. Every single ground point is soldered to the top ground plane.
An externally hosted image should be here but it was not working when we last tested it.
Great how the perfect schematic that he kept talking about has had lots of issues
I'm still interested in this design though, especially now that you guys are working through some real issues and trying to improve the functionality of the original schematic. We'll have to see how it turns out but you guys have made move in the right direction I feel.
I'm still interested in this design though, especially now that you guys are working through some real issues and trying to improve the functionality of the original schematic. We'll have to see how it turns out but you guys have made move in the right direction I feel.
So that chassis is like 170 bucks shipped to the states? Not a bad price? Those heatsinks look big enough for just about any A/B design or low power A.
Thanks for the "closeups", nagsy. This confirms the current direction of the "other" thread. I would say the grounding schemes on the "mongrels" or on what the "goodmund" will end up with would result in a 10db or more reduction of PS hum. I would also say that an amp like AAK's symasym ported to run 3 pairs of the original SK134/J49's would be a much quieter amp. I will have 2 pair of the to-3 hitachi's (not fakes) by november to further the research.
The real goldmund is VERY well constructed. Genesis potted the trafo's similarly , along with the steel box to make the PS a "mystery" to the consumer. the goldmund's PCB's are way better than the genesis (even as I totally disagree with the grounding scheme). I still would not spend thousands as I can exceed what I see for hundreds.
PS.. you might want to recap that Mimesis , I see "bulging" caps !!!!
OS
The real goldmund is VERY well constructed.
Genesis potted the trafo's similarly , along with the steel box to make the PS a "mystery" to the consumer. the goldmund's PCB's are way better than the genesis (even as I totally disagree with the grounding scheme). I still would not spend thousands as I can exceed what I see for hundreds. PS.. you might want to recap that Mimesis , I see "bulging" caps !!!!
OS
Last edited:
I see. You think it is very likely that you will want to solder some parts to the ground plane. I thought this is the best amplifier you ever heard? Why build it breadboard style?
The reason Goldmund didn't coat his boards is totally unclear, must be some weird vodoo beliefs. The disadvantage not to coat a board is too obvious so it can't be a rational thought behind, maybe except making more money with saving some Euros for soldermask.
The magic Goldmund ground plane needs to be connected somewhere to the PSU, right? That's what I call the "main" ground connection.
A groundplane is not like a black hole absorbing everything and turn it into nothing. It is not the magic cure for everything. For a groundplane to make sense at all it needs to be continuous, without any obstacles inside.
A real groundplane actually is a conductor that has ohm resistance as well as inductance. A current flowing over the plane is like a current flowing over a wide trace. Now if you arrange your circuit in a way that currents share that trace they interact with each other. Since you are an engineer you should understand this.
I agree it looks nice to have the speaker return right next to the speaker output but this makes the return current flow all over the plane to the "main ground" connection to the PSU ground. Not a good idea at all.
I'm sure you know about the star ground concept that is used for amplifiers for good reason. The idea behind is that currents don't interfere with each other. I think I don't need to further explain this since it is considered basic knowledge, especially for an engineer, like you.
Thanks. The picture of the pcb is a very good example how not to do it.
This is VERY bad design practice showing that the one who designed the pcb has not the slightest clue what he was doing. I'm so sorry for the ones who bought such an amplifier and paid so much money for it. This is poor DIY quality, nothing to sell.
The potting is a nice idea but the transformers should be potted entirely in vacuum. This way is doesn't have too much effect on vibration unless very flexible silicon was used and there's enough of it underneath the transformers.
Potting the rectifiers or whatever is inside just has the "black box" effect and does nothing to improve anything. Repairs are impossible of course.
Looking closer I think I see another very bad thing on the pcb: The NFB takeoff point. See the uppermost "thick" copper trace connecting four screws of the output devices and goint to the relais? I bet this is speaker out. See the small trace from the second screw from the left going into the amplifier circuit? I bet this is NFB, taken off the worst point one can imagine. As D. Self pointed out taking NFB from a point somewhere between the output devices low frequency distortion rises considerably. His measurements show this effect very obviously. NFB must be taken right from the rightest screw where no current flows between the upper and lower ouput transistors.
The Goldmund technology seem to mostly rely on obfuscation, vodoo and questionable electronics design.
Lee Knatta - I disagree with you 100% on everything you wrote. This is a simple DIY amplifier kit, not a nuclear reactor. If we are going to be that anal on even something as simple as running the negative feedback, then we'll be designing this PCB for 10 years. A ground plane is a ground plane, that's what it's there for. All ground connections can connect to it without any problems. Any noise, distortion, etc. incurred will be completely insignificant. I listen to this amplifier everyday and sonically I've never heard anything superior.
BTW, the PSU is connected directly to the ground plane right behind the capacitors, you can't see it in the picture.
BTW, the PSU is connected directly to the ground plane right behind the capacitors, you can't see it in the picture.
This reply is exactly what I expected from you.
If you can't figure out trivial things like the correct NFB takeoff point and proper grounding you have no chance to get it right - never ever. Getting it right doesn't take 10 years, btw.
A lot of people gave advice to you but you didn't care.
Regards,
Lee
NagysAudio:
This is an ATTEMPTED clone of the Goldmund,
a TRUE clone would duplicate to a tee
the parts used, the circuit board precisely, the power supply and chassis and etc. etc.
This is okay, but a clone it is not.
I strongly suggest you heed the word of the wise men here, they do not have a bone to pick with you. They just want to see to it that those that don't know what is going on have a chance to choose. What they want and see to it is that somethings you say are against the grain of what is proven or normal engineering practices and conventions. Goldmund may do it, but it does not make it right or better.
Personally I have learned a great deal from the different threads here. I am not a rank amateur, but SMART men and women know that they don't KNOW everything....
Your premise that "this is the best amplifier I have ever heard" is just that YOUR OPINION.
We all have one.
Thanks
This is an ATTEMPTED clone of the Goldmund,
a TRUE clone would duplicate to a tee
the parts used, the circuit board precisely, the power supply and chassis and etc. etc.
This is okay, but a clone it is not.
I strongly suggest you heed the word of the wise men here, they do not have a bone to pick with you. They just want to see to it that those that don't know what is going on have a chance to choose. What they want and see to it is that somethings you say are against the grain of what is proven or normal engineering practices and conventions. Goldmund may do it, but it does not make it right or better.
Personally I have learned a great deal from the different threads here. I am not a rank amateur, but SMART men and women know that they don't KNOW everything....
Your premise that "this is the best amplifier I have ever heard" is just that YOUR OPINION.
We all have one.
Thanks
I made some suggestions and pointed out problems because you do your own pcb design.
For staying as close to the original as possible you consequently should clone the pcb layout as well since it contributes to the magic sound for sure.
You're right this is the wrong place for improvements. I better try to help the other guys...
For staying as close to the original as possible you consequently should clone the pcb layout as well since it contributes to the magic sound for sure.
You're right this is the wrong place for improvements. I better try to help the other guys...
This entire thread is a fascinating read! Thanks to all who have contributed.
I need to get up on a soapbox and very clearly say something, that many other people have said but the thread principles have been ignoring -
Will someone beta this board, build the amp, and see if it works as expected?
I am extremely interested in this project, but also know enough not to jump in head first in a design like this until we know, for sure, that the implementation of the design is stable and without errors.
I need to get up on a soapbox and very clearly say something, that many other people have said but the thread principles have been ignoring -
Will someone beta this board, build the amp, and see if it works as expected?
I am extremely interested in this project, but also know enough not to jump in head first in a design like this until we know, for sure, that the implementation of the design is stable and without errors.