SSLV1.1 builds & fairy tales

diyAudio Chief Moderator
Joined 2002
Paid Member
salas i have used silchar cores even on larger transformers what I felt it had little bit of lack of low frequency grip compared to a solid EI or Toroid. I tried from a transformer manufacturer with primary and secondary winding done as whole. The designer said that the regulation is quite better in toroid than in R-core.

In my experience the R-Core is smoother for HF in preamps. For regulation it does not lose voltage easier by loading than Toroid. For heat is the coolest. For stealthy magnetic field the king. But Maybe you had best Toroid in big amp.
 
sorry to bother you again, salas. the progress of my project is pretty slow cause i can only do it in week end. also need to reorder that shunt mosfet for ref-d first.

my positive rail on sslv 1.1 doesn't have any problem at all but my negative rail only able to reach 9 volts from target voltage of 12 volts. no matter where i adjust my trimmer it stays 9 v max.

for this negative rail i use 2sk117gr with 4 mA of Idss on Q203, 1.8 kohm resistor, 5 k trimmer and 2 red leds on VRef section.

1 strange thing that i found is the VGS on both shunt mosfets on positive and negative rail is zero. but one reach target voltage, the other doesn't.

fyi, the top part of both mosfets on negative rail seems to touch the heatsink, unlike the positive rail which remain isolated. does this matter ?
 

Attachments

  • bib1.1.jpg
    bib1.1.jpg
    813.3 KB · Views: 281
diyAudio Chief Moderator
Joined 2002
Paid Member
In both polarities Qxx1 Mosfet's drain pin connects to Qxx6 Mosfet's source pin. The tabs are drain so they should not show continuity to each other when properly isolated from the sink because they are not connected through the circuit either.

Your Vref components choices look potent enough to reach your target Vout.
You are maybe using too low a dummy load resistor value for your CCS setting that sucks almost all available current and/or the isolation to sink isn't good and the Mosfets are shorted. Another scenario is the isolation is good but a Mosfet is internally shorted (failed). In any case on a correctly working build you should have voltage drop across each Rxx1 and VGS voltage between the first and last pin of each Mosfet.
 
i have changed both mosfets on negative rail, make sure there's no continuity between drain tab and heatsink also measure resistance on trimpot to confirm it's working but still the same problem exist.

the VGS on shunt mosfet on negative rail (Q206) is still zero while on positive rail turns out to be 0.2. does this mean the problem is in Vref section or there's still real possibility that Q206 is faulty?

i'm also curious about the arrows on trimpot symbols in schematic. it shouldn't be a problem even if trimpot is reversed, right ?
 
diyAudio Chief Moderator
Joined 2002
Paid Member
The voltage would go up when turning the screw the other way but apart from that there's no other change when reversing a trimmer that is connected in rheostat mode.

Maybe there is a discontinuity between parts in the Vref section? Use the DMM's continuity function between them. Or some dead LED? Try to light them up one by one as they are installed with a DMM's diode mode that has enough potential or with a 9V battery. Power should be off in this test.

Faulty MOSFETS from static by handling or soldering can occur. But is there any possibility of an accidental mix of MOSFET types between the pos/neg regulator sections?
 
all LEDs are shinning like in that picture so i think there's no discontinuity at all. there's no mixed part between positive and negative rail. i have change that BC 560C (Q204) and shunt mosfet for the 2nd time but still the same problem. maybe the problem is in those 2sk117 but i only have 1 right now which i intended to use it as spare part.

for now i will reduce the volatge on positive rail to 9 volts, put temporary jumper on LM329 pins in Ref-D to reduce voltage from 12 v to 5 v and see if this works.
 
0.537 volt between R201 legs and i use 39 ohms dummy load. R201 is 2.2 ohms but i remember measured it as 2 ohms. almost forget : i remember before changing the last 2 components that i measured output voltage (F+ and F0) without dummy load at all and its value is 9 volts, the same value with dummy load attached. so as if the voltage reference doesn't exist at all.
 
finally got the time to finish my project. it turns out this DAC can be powered with 9 volts on analog and 5 volts on digital. reducing the voltage output especially on Ref-D made CCS mosfets get really hot like my amp. but it has played for 4-5 hours without problem so i'll leave it as it is.



i know they seem to be not ideally positioned and messy but i still need room to build I/V converter in the future. although i still use cheap op-amps as I/V converter right now, it does sound very good. never been a fan of foobar, but this NOS DAC made foobar very enjoyable. i guess the naturalness of treble and more transparency balance things up. except for small mechanical hum on my transformers, it's very quite. if you see on the photos i don't use any noise reduction tantalum caps as the AD1862 datasheet suggested. i do use reservoir caps : 330 uF elna silmic 2 on analog and nichicon polymer (RNU series) on digital in the bottom side of the DAC.


overall this project is worth every penny of it. thank you Salas and Tea-Bag !
 

Attachments

  • inside.jpg
    inside.jpg
    493.8 KB · Views: 402
  • dac.jpg
    dac.jpg
    717.3 KB · Views: 392