Hi Ed,
Read your work.
It seems you are heavily immersed in simulations and not the real world at all. A simulator driver. You're arguing against empirical evidence, which happens to be true regardless of what simulators with incomplete variables may "prove".
Some people here have spent decades plus at the bench with real world equipment. At this point I'm going to politely suggest you drive a soldering iron with real test equipment at hand. This isn't cheap like free simulators and time. This requires skill, knowledge, an investment in time, and investment in test equipment and parts. Until you have done this, arguing with folks who live this life is just noise.
Read your work.
It seems you are heavily immersed in simulations and not the real world at all. A simulator driver. You're arguing against empirical evidence, which happens to be true regardless of what simulators with incomplete variables may "prove".
Some people here have spent decades plus at the bench with real world equipment. At this point I'm going to politely suggest you drive a soldering iron with real test equipment at hand. This isn't cheap like free simulators and time. This requires skill, knowledge, an investment in time, and investment in test equipment and parts. Until you have done this, arguing with folks who live this life is just noise.
Hi Chris,
My profession is computer architecture. Many people have owned computer hardware that I had designed.
Audio is an old hobby. I designed and built a series of amplifiers from the 1970s through 2001. The last one was so good that I haven't felt a need to build another.
My recent hobby efforts have been in writing simulators to better understand why circuits behave the way they do. The models are based on fundamentals. They are accurate for the problems that they solve. They may or may not be accurate for real hardware, depending on the design.
Measurements can be a great confirmation, but are often conducted in a way that has too many variables to draw conclusions about why things happen.
I don't dismiss models or measurements, but interpret them in light of what they are.
Ed
My profession is computer architecture. Many people have owned computer hardware that I had designed.
Audio is an old hobby. I designed and built a series of amplifiers from the 1970s through 2001. The last one was so good that I haven't felt a need to build another.
My recent hobby efforts have been in writing simulators to better understand why circuits behave the way they do. The models are based on fundamentals. They are accurate for the problems that they solve. They may or may not be accurate for real hardware, depending on the design.
Measurements can be a great confirmation, but are often conducted in a way that has too many variables to draw conclusions about why things happen.
I don't dismiss models or measurements, but interpret them in light of what they are.
Ed
Hi Ed,
Actually, simulations don't come close to the real world in what they take into account. The variables are far too great. Everything from component parameter spread to the layout to PCB effects.
When you lay out a computer PCB, you do model the PCB along with (hopefully) power supply. The per seat license for such a program to do audio accurately is beyond what the industry will support. I know that from seminars I attend from Keysight and others. As I said, you're talking to a range of people who have little to no knowledge, to folks who designed, built and / or repaired audio for decades. Those people know.
You did say you can't measure real amplifiers. If you desire an understanding, you need to get the required equipment and work with your hands, change things and measure again. Only then can you work to make a simulator mimic real life. Otherwise you get the very rosiest of pictures that may not apply to actual constructed circuits. Some simulated circuits don't even work in real life.
I'm not trying to minimize the use of a simulator, but like test equipment you have to understand it's limitations or you live in a world not really connected to the real one. Ever see someone drive for the first time if they have only ever done it in a simulator? How about fly?
Actually, simulations don't come close to the real world in what they take into account. The variables are far too great. Everything from component parameter spread to the layout to PCB effects.
When you lay out a computer PCB, you do model the PCB along with (hopefully) power supply. The per seat license for such a program to do audio accurately is beyond what the industry will support. I know that from seminars I attend from Keysight and others. As I said, you're talking to a range of people who have little to no knowledge, to folks who designed, built and / or repaired audio for decades. Those people know.
You did say you can't measure real amplifiers. If you desire an understanding, you need to get the required equipment and work with your hands, change things and measure again. Only then can you work to make a simulator mimic real life. Otherwise you get the very rosiest of pictures that may not apply to actual constructed circuits. Some simulated circuits don't even work in real life.
I'm not trying to minimize the use of a simulator, but like test equipment you have to understand it's limitations or you live in a world not really connected to the real one. Ever see someone drive for the first time if they have only ever done it in a simulator? How about fly?
I have measured my amplifiers, but only with a DMM, signal generator, and oscilloscope. I do not plan to buy an expensive analyzer.anatech said:
I have seen enough circuits on this board that "work" only in simulation. I am decades past that stage. I feel pretty confident that my designs perform as expected. If I were doing audio for a living, I would buy the expensive analyzer.
Ed
A few weeks ago when i was working with this:
I tapped on all resistors with no result, looked at the PCB layout without finding any directly wrong.
Finally i found it was the input RCA connector.
I did find out that the distortion i measured in the amplifier was not in the amplifier at all!
That is one of the faults that can happen if you only look in the real world.
Other usual faults is influence of the power cords when working with loose pcb:s on the desktop. When you are working in the ppm range wiring is critical.
Something that is impossible to correct in a simulator.
A small desktop current dumping amplifier
i had mysterious distortion. It was 0.001 to 0,003 and varied. It was almost pure 3rd order so i guessed it was a bad resistor.I tapped on all resistors with no result, looked at the PCB layout without finding any directly wrong.
Finally i found it was the input RCA connector.
I did find out that the distortion i measured in the amplifier was not in the amplifier at all!
That is one of the faults that can happen if you only look in the real world.
Other usual faults is influence of the power cords when working with loose pcb:s on the desktop. When you are working in the ppm range wiring is critical.
Something that is impossible to correct in a simulator.
Hi Ed,
Okay. From vast experience I can only say you don't know much yet.
To look at this stuff in any detail, you need that expensive test equipment. I have it, so do many others. That means we can see the truth.
An oscilloscope will only ever show to most gross signal problems. It's better than nothing for checking for oscillation. You would be very hard pressed top detect 1% distortion with an oscilloscope. Sorry, that's the truth.
Even an ancient HP 331A is light years better than what you have now, and that is not expensive. So if you have the slightest interest in this subject, you coud at invest in one. Then you could actually see crossover distortion with your scope. At the moment you are completely in the dark.
Okay. From vast experience I can only say you don't know much yet.
To look at this stuff in any detail, you need that expensive test equipment. I have it, so do many others. That means we can see the truth.
An oscilloscope will only ever show to most gross signal problems. It's better than nothing for checking for oscillation. You would be very hard pressed top detect 1% distortion with an oscilloscope. Sorry, that's the truth.
Even an ancient HP 331A is light years better than what you have now, and that is not expensive. So if you have the slightest interest in this subject, you coud at invest in one. Then you could actually see crossover distortion with your scope. At the moment you are completely in the dark.
Anatech your comments disappoint me with you as a person. What you wrote is pure judgment without knowledge!
When I finished college, I was trained at Studer Revox - Sweden to repair their products. At the time, I worked at the brand's official importer. I repaired many amplifiers and worked a lot with open reel-to-reel tape recorders. After that I went to the telecommunications industry which paid much better.
Attached the picture of one old audio amplifier project I did about 20 years ago. NO USE OF SIMMULATORS IN THAT TIME. I understand very well the relations between all components and currents in circuit without use of simmulator - for sure. It was tested with a small trafo (old picture) and I finnaly use a big toroidal over sized to final amplifier in a wood box.
The result was good, but Marantz PM-11S3 is better. What I am working is to make a better amplifier than PM11S3.
I also tested more bias in my PM-11S3 as all of you advocate. It uses only one 2SC2922-2SA1216 pair in output to 120W output power with a big heatsink. No audible advantage, only heating and more ac power consuption as result. I use B&W 803D4 as speaker. Sorry but I don't change my opinion related to bias current. Lot of work without result.
I also used HP 8903A as distortion test in that time.
Regards to all and thank you for your comments.
Ronaldo
When I finished college, I was trained at Studer Revox - Sweden to repair their products. At the time, I worked at the brand's official importer. I repaired many amplifiers and worked a lot with open reel-to-reel tape recorders. After that I went to the telecommunications industry which paid much better.
Attached the picture of one old audio amplifier project I did about 20 years ago. NO USE OF SIMMULATORS IN THAT TIME. I understand very well the relations between all components and currents in circuit without use of simmulator - for sure. It was tested with a small trafo (old picture) and I finnaly use a big toroidal over sized to final amplifier in a wood box.
The result was good, but Marantz PM-11S3 is better. What I am working is to make a better amplifier than PM11S3.
I also tested more bias in my PM-11S3 as all of you advocate. It uses only one 2SC2922-2SA1216 pair in output to 120W output power with a big heatsink. No audible advantage, only heating and more ac power consuption as result. I use B&W 803D4 as speaker. Sorry but I don't change my opinion related to bias current. Lot of work without result.
I also used HP 8903A as distortion test in that time.
Regards to all and thank you for your comments.
Ronaldo
Attachments
Last edited:
stigigemla
For a while (almost 5 years) I worked in the components industry. It was a Philips factory in Brazil.
At that time I was an industrial maintenance technician. In production, transistors were tested in a test station controlled by a PDP10 computer. The test checked all transistor parameters and selected the correct box for each device acording with the datasheet classification. After that, a small sample came to Control Quality.
One day the quality control pointed out catastrophic fail in all sample and in all tested transistor. The tester was OK and all transistors that I tested were OK in a Tectronix curve tracer. The CQ sample was really bad. So, what was the problem with the large tester that pointed out OK and the quality control pointed out catastrophic failure?
This had never happened there. It involved the entire team and specialized engineers and we could not find the problem. The test equipment was OK, the transistors were OK, production resumed and soon the problem appeared again. The transistors in production did indeed appear with problems after the CQ.
This was a serious problem that compromised all production and backlog of orders.
In the rush to find the problem, a test batch fell to groud during transport. ALL TRANSISTORS FAILED.
What happened in subsequent tests! When analyzing the transistors, a fault was discovered in the soldering on the DIE.
It took a long time and was a big trouble.
When we work with electronics all can happen.
More than 30 year in electronic. I have lots of histories to tell about my experience.
For a while (almost 5 years) I worked in the components industry. It was a Philips factory in Brazil.
At that time I was an industrial maintenance technician. In production, transistors were tested in a test station controlled by a PDP10 computer. The test checked all transistor parameters and selected the correct box for each device acording with the datasheet classification. After that, a small sample came to Control Quality.
One day the quality control pointed out catastrophic fail in all sample and in all tested transistor. The tester was OK and all transistors that I tested were OK in a Tectronix curve tracer. The CQ sample was really bad. So, what was the problem with the large tester that pointed out OK and the quality control pointed out catastrophic failure?
This had never happened there. It involved the entire team and specialized engineers and we could not find the problem. The test equipment was OK, the transistors were OK, production resumed and soon the problem appeared again. The transistors in production did indeed appear with problems after the CQ.
This was a serious problem that compromised all production and backlog of orders.
In the rush to find the problem, a test batch fell to groud during transport. ALL TRANSISTORS FAILED.
What happened in subsequent tests! When analyzing the transistors, a fault was discovered in the soldering on the DIE.
It took a long time and was a big trouble.
When we work with electronics all can happen.
More than 30 year in electronic. I have lots of histories to tell about my experience.
Last edited:
Anyone with a computer has the means (hardware) to detect crossover distortion which is invisible on an oscilloscope. And I’m not talking about the simulator. Whether you can see below the level of audibility depends on the quality of that hardware. Unfortunately, my E-mu has been collecting dust since the last motherboard I had that it was compatible with died, leaving me with either internal garbage or the little Creative Labs USB unit that I was using to measure speakers outside with a laptop. Even that got dumped for the OmniMic. So it’s been a while since I’ve done any real distortion measurements. But in that time I’ve learned how to make amplifiers with so low a distortion that I can’t hear it and don’t chase it down any further.
I wont ever be buying an expensive audio analyzer either. But the QuantAsylum is on the list of things to acquire by the time the lab is operational again. I’ll need something “decent” to align the huge amplifiers for the PA rig that are coming, but that’s still 2 years out. I have everything I need for home listening already. I’ll need to get the computer part sorted first, and I have to find a way to make the entire rig independent of a pay connection back to Microsoft.
The last 5 amplifiers I built were tubes. I don’t even CARE how the distortion measures. They sound good. The point was to build out of tubes because I could. More tube amps are coming too. Big ones.
I wont ever be buying an expensive audio analyzer either. But the QuantAsylum is on the list of things to acquire by the time the lab is operational again. I’ll need something “decent” to align the huge amplifiers for the PA rig that are coming, but that’s still 2 years out. I have everything I need for home listening already. I’ll need to get the computer part sorted first, and I have to find a way to make the entire rig independent of a pay connection back to Microsoft.
The last 5 amplifiers I built were tubes. I don’t even CARE how the distortion measures. They sound good. The point was to build out of tubes because I could. More tube amps are coming too. Big ones.