Thanks. It hasn't been that many reviews on the 1.1 vs 1,3 so it's
interessting to finally see one, accept for the few ones in the
begining of the thread. I have five 1.1's and three Ref D's as
pre-regulators in a dac project that I haven't even started up yet.
I'm a bit hesitant to replace them before I even have had a chanse
to listen to them. If someone else have had a possibillity to compare
the two (or three with the Ref D), please don't hesitate to describe what's there.
The analogue supply in my CD player. The perfect candidate for a couple of UltraBiB's?
Analogue PS
Regards,
Dan
Analogue PS
Regards,
Dan
Last edited:
@ dahlberg
I tried all the Salas volt regs ; 1.1, 1.2r, reflektor and 1.3. The later is absolutely the better
I tried all the Salas volt regs ; 1.1, 1.2r, reflektor and 1.3. The later is absolutely the better
A successful 3.3V version for my DAC Crystek oscillators. I used 2SK880-GR with gate and source shorted and IRL530N for the shunt transistor.
It confirmed my previous listening tests — sounds truly outstanding! Thank you Salas!
PS: Now my AD1865 DAC has seven Ultra BiB power supplies 😀
It confirmed my previous listening tests — sounds truly outstanding! Thank you Salas!
PS: Now my AD1865 DAC has seven Ultra BiB power supplies 😀
Attachments
Last edited:
Yes, the 2SK880's are on the bottom side of the PCB. I use a slightly modified schematic with series RC branch (1nF & 51R) between Q4 base and ground to increase stability.
A green led gives approximately 3.2V which is well within the CCHD-957 specs 😎
As with v1.1 the CSS resistor has significant impact on the sound quality. I'm waiting for some Mills for R1...
A green led gives approximately 3.2V which is well within the CCHD-957 specs 😎
As with v1.1 the CSS resistor has significant impact on the sound quality. I'm waiting for some Mills for R1...
The low threshold M2 type ate away some margin and you had to compensate? Possibly because it lowers Q4's Vce. Used a diode instead of R6 also?
Last edited:
I think I remember those mods. Ring of two BJTs instead of a single JFET tail, alright more parts, but does the job more or less alike. So where we see an LSK170 there is a 2SK880? Modeling with LSK instead of PF5102 includes more capacitance and transconductance so the comp Zobel maybe looks even more necessary. In any case it works steadily as a final assembly with final parts and board so no worries. The standard Ubib can be configured for 3.3V too, see post #631. Don't allow less than 100mA spare when you can, for a clock especially. Try 150mA even. Fast edges load draw appreciate good reserve.
There is some chance I already posted this, say for 5V output 🙄
LSK170 is used only for simulations. I’ve also tried to use BF861 model, but with resistor in the source.
LSK170 is used only for simulations. I’ve also tried to use BF861 model, but with resistor in the source.
You could even use bit of Rx to hit 3.3V output exactly by the way. An upright small value resistor in the one empty Led space.
Hello
So I need a +/- 24v @250mA out would it be best to go with a 20Vac transformer that would give me around 26Vdc after the rectifier. Or would it be better to go with a 22Vac transformer which will give me around 29Vdc to have a little more headroom?
Thanks
Dave
So I need a +/- 24v @250mA out would it be best to go with a 20Vac transformer that would give me around 26Vdc after the rectifier. Or would it be better to go with a 22Vac transformer which will give me around 29Vdc to have a little more headroom?
Thanks
Dave
What a good idea! Tanks! On my PCB there is a place provided for trimmpot (R9). That’s where I’ll put the resistor to reach 3.3V. I happen to have some small value Takman REY’s lying around 😎