interesting... well done!
could you post a bit more details to replicate your mod? what have you used, a dremel or something like that?
Apart from measurments, have you noticed any difference also in perceived SQ?
Good!
The end plates are made from FR4 which has 100% metal on one side. Except for the logo, this metal is covered by white solder stop lacquer. It is not connected to the box.
For the next model I'll open up the metal around the connectors. And I'll connect the metal to the screws. This is easily done with a scalpel and the right screwdrivers (PH2 and TX10).
Now I realize the panels may act as patch antennas!
Børge
The end plates are made from FR4 which has 100% metal on one side. Except for the logo, this metal is covered by white solder stop lacquer. It is not connected to the box.
For the next model I'll open up the metal around the connectors. And I'll connect the metal to the screws. This is easily done with a scalpel and the right screwdrivers (PH2 and TX10).
Now I realize the panels may act as patch antennas!
Børge
I Added 19 Khz -9 dBFS and 20 Khz -9 dBFS (32 bit).
After the sum normalized to 0 dBFS and 24 bit.
These are photos of detail: I hope you see.
After the sum normalized to 0 dBFS and 24 bit.
These are photos of detail: I hope you see.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Front and rear panels make contact with the rest of the box without having to do anything.
The important thing is the miniUSB conector not touch the box. And how to attach the box to the circuit ground so the noise not go up over time. I guess with a Varistor or resistor and a capacitor in series or in parallel, but do not know how, and typical values.
1audio from your proposal I am simulated three circuit to compare, your supply, one variation from yours, and Salas supply .
I have change the model components of your circuit and the salas supply. I hope not bother you. You can put yours in the comparisons.
I have searched for low-cost components that allow power to 3.3 V 100 mA The idea is to design a mini circuit board.
Here I leave the scheme and the models used so they can compare. I would appreciate comments.
An externally hosted image should be here but it was not working when we last tested it.
oneoclock:
Take an look at TeddyPardo.
The allow positive and negative versions using the same PCB and can delivery
5 and 3.3V as an alternative.
Regards Juan
Take an look at TeddyPardo.
The allow positive and negative versions using the same PCB and can delivery
5 and 3.3V as an alternative.
Regards Juan
If the "open-source" spirit of the project is to be continued, personally copywrited designs like the TeddyPardo must be avoided.
Besides, I this line from the website:
"The SuperTeddyReg is extremely sensitive to component choice, I therefore strongly recommend buying the built-and-tested version."
does not inspire confidence in repeatability..
I still have a lot to learn about PSU design. Just one question: why do the LEDs set up a voltage below filtered Vin, not above GND?
Børge
Børge
The simulation model looks like a good effort. However this is also an object lesson in understanding the importance of understanding the application's requirements.
The series regulator I came up with was intended for ultra low phase noise oscillators. The load changes essentially never with a stable oscillator. The key issue is low noise and high line isolation. A driver for a headphone has much different issues. the peak current needs to be accomodated, the load can change as much as 100%, transient response is important and noise is less of an issue. providing power for a digital circuit is still somewhat different with very high peak currents usually at the clock edges.
Add an ac source at the input to simulate line isolation. And then you could add resistors to simulate the effects of drops across the grounds.
Shunt regulators make things easier since you know where the regulator is referencing to. But they are not right for a headphone load since they need to dissipate a lot of power when idle. The thermals will also cause modulation in the voltages. Remote sensing for a series regulator is more complex and prone to instability.
Lots to think about.
The series regulator I came up with was intended for ultra low phase noise oscillators. The load changes essentially never with a stable oscillator. The key issue is low noise and high line isolation. A driver for a headphone has much different issues. the peak current needs to be accomodated, the load can change as much as 100%, transient response is important and noise is less of an issue. providing power for a digital circuit is still somewhat different with very high peak currents usually at the clock edges.
Add an ac source at the input to simulate line isolation. And then you could add resistors to simulate the effects of drops across the grounds.
Shunt regulators make things easier since you know where the regulator is referencing to. But they are not right for a headphone load since they need to dissipate a lot of power when idle. The thermals will also cause modulation in the voltages. Remote sensing for a series regulator is more complex and prone to instability.
Lots to think about.
Because you want a constant bias voltage for M1, which has the source at input voltage level
The more I studied Demian's PSU I more I like it
. Especially after reading his "object lesson" Alex
The series regulator is on the lower right of the simulation schematic.
I have serious mixed feelings about using LED's as references. They will have the following issues; temperature sensitivity with a significant tempco, noise and optical sensitivity. I got burned in the past with LED's being the source of hum in a circuit. Others disagree. I need to get some representative LED's and test them to see if they have changed. However getting a stabilized bandgap with a suitable amp in a TL431 for less than $1.00 it a great deal.
I have serious mixed feelings about using LED's as references. They will have the following issues; temperature sensitivity with a significant tempco, noise and optical sensitivity. I got burned in the past with LED's being the source of hum in a circuit. Others disagree. I need to get some representative LED's and test them to see if they have changed. However getting a stabilized bandgap with a suitable amp in a TL431 for less than $1.00 it a great deal.