Joachim, the offset is fluctuating within a range of more than a 100mV every second - it is not steady and is too fast for the servo to correct. Finally took a listen using a DL103: there is a very prominent hiss, completely against expectations. Bad batch of electrolytics at the input stage? Oscillations?
I had a short discussion with MiiB. He is in Hong-Kong at the moment preparing to sleep.
That is what we came up :
Preforming the Elcaps, especially the ones in the input. Load them up to the rated voltage with a battery or a lab PSU. Let them soke up the voltage for some hours so that they form. Then put them in the circuit.
Thermo couple the input transistors with a piece of metal. The circuit has a lot of gain in the lower frequency region so thermal drift can challenge the servo.
Decouple the mirrors locally with ceramic caps, that can take away oscillation.
That is what we came up :
Preforming the Elcaps, especially the ones in the input. Load them up to the rated voltage with a battery or a lab PSU. Let them soke up the voltage for some hours so that they form. Then put them in the circuit.
Thermo couple the input transistors with a piece of metal. The circuit has a lot of gain in the lower frequency region so thermal drift can challenge the servo.
Decouple the mirrors locally with ceramic caps, that can take away oscillation.
What brand of Elcaps do you use ? I use Panasonic, a long life version of the FM. I think they are part of the group buy and should be good. I also tried Rubicon ZL and ZY without problem. Elna should be good too.
Another option is to make the servo faster by reducing the 1uF cap. It will still reach down low enough.
Another option is to make the servo faster by reducing the 1uF cap. It will still reach down low enough.
I had similar issues...How did you choose the new top fet ? Different idss ?
In my build the output offset was inittially around 100mV but after some hours it came down to around 10mV, so I agree it might have something to do with the input emitter caps.
We will attack that DC offset problem and we will find a solution. I mean, it can not be,
My Paradise R3 worked right away after eroding two stuffing mistakes.
Ether you have made a stuffing mistake, or a part is faulty. When it is something else we have to find out. There must be a reason.
Hesener mentioned some time ago that the hunt for ultra high Hfe COULD create some of the oscialltion problem we had on the Paradise 1.
I repeat is now again : an Hfe of 400 is more then abundant, in ALL positions, also the input.
My Paradise R3 worked right away after eroding two stuffing mistakes.
Ether you have made a stuffing mistake, or a part is faulty. When it is something else we have to find out. There must be a reason.
Hesener mentioned some time ago that the hunt for ultra high Hfe COULD create some of the oscialltion problem we had on the Paradise 1.
I repeat is now again : an Hfe of 400 is more then abundant, in ALL positions, also the input.
The supply current should be 110mA ...130mA, not more. This is irrespective of whether the amplifier is connected or not (the big power jumpers in the middle), because the shunt will reduce its current accordingly. But I read the problem is solved already.....
