Congrats for your build. Using a crock lead adds inductance and resistance to your capacitors test. Thin cheap leads that come in many colors especially. I have found 1 to 20 Ohm randoms in such bundles. Seriously. Even if good wire, still adds inductance. Got to repeat with very short thick pin extentions. All soldered. Else the ESR and ESL of capacitors get swamped by the test lead's own.
Try R6 vs J2 option and four wire vs two wire option too. So to find best synergy in your configuration.
Try R6 vs J2 option and four wire vs two wire option too. So to find best synergy in your configuration.
Hello,
I completed my first R-D with success, it took place of an Bib in powering 8v\160mA. supply for a DAC. It is one only ic Philips TDA1543.
Sound is diffenent, how to say... R-D is prefered, easy to listen, but all components brand new simply can't deliver as much 'detail' as the fully broken in Bib unit.
No need to say, I appreciate so much advice here, and wants share also my own findings. -despite my limited English and electronic knowleadge.-
Concerning CM cap, as stated before, I agree to be very influential to final sound. But not only in brand ALSO in value. After arranging an array of caps. I have on hand:
-10.000uf. Nichicon FW
-4.700uf. Nichicon KG
-2.200uf. Elna Silmic II
-1.000uf. Nichicon KZ
The negative leg is tied both toghether and soldered to neg on pcb.
Then the + leg is free, keeping to easy swap with a cocodrile wired to Pcb still when music is playing.
In first was a play for the bigger 10.000 as specs in diagram, but soon enters the play, as much as I decreases the capacitance it is increase the detail in sound. Till finally reaching the small 1.000uf. which removes completelly the loudness related of excess capacitance. But nothing thin, otherwise great strenght in overall frecuencies including clean punch bass notes. I have experienced this in many circuits, as in Bib one for instance, where its recommended main cap. 4.400uf., decreases for me till 1.000uf - 1.500uf . I use nothing but Nichikon Muse KZ, however by using non audio grade capacitors, must prefer higher filtering value as in schematic diagram. I barely remember but reading somewhere or heard that audiophile electrolitic capacitors not direct replacement by same value.
Once found the value by hearing, it's a nice, very nice addition upgrade then to CLC.
Although bulky and expensive, this solution is worth the effort, as it transforms the sound to another league. Is needed another capacitor, same value, and a choke in between wired in series with + rail. I use 10H amorfe big choke, other cheaper choke, less quality iron or filtering amount, not same results.
Try also Mundorf 4 pin Ag+ caps. They are excellent. Especially on C2 (1000uf) position
. Among many caps I tried here (elna silmic II, black gate Std, Elna Silmic ARS, Nichicon FPCAP, Nichicon muse BP non polar ...) Mundorf 4 pin are of the widest audio bandwidth (especially high frequencies are very clear and deep bass), best resolution and clarity. You connect your LEDs between Out(-) and In(+) pins then In(-) to GND in PCB and Out(+) to PCB + trace. You would have to just modify/destroy your board in good way to fit this cap or you could wait until Salas releases Reflector-D Mini Flexy version, some day, maybe . Anyways it is really worth mod.And as you already mentioned give some time to burn in.
So, for digital part of chips (DVDDs) I am using Nichicon FPCAP on C2 position (best 1/f filtering, but terrible tone). However for AVDDs, analog Vrefs or powering resistor ladders dac directly (Soekris R2R), generally places from where audio signal is produced from, Mundorf 4 pin I found to excel any other cap in this position.
And as you already mentioned give some time to burn in.
So, for digital part of chips (DVDDs) I am using Nichicon FPCAP on C2 position (best 1/f filtering, but terrible tone). However for AVDDs, analog Vrefs or powering resistor ladders dac directly (Soekris R2R), generally places from where audio signal is produced from, Mundorf 4 pin I found to excel any other cap in this position.
Mini & Flexy were released during the last GB
Mini & Flexy were released during the last GB
Well yes for rectification board you can mount 4 pin cap. However what I am referring to is place for 4 pin capacitor in regulation board for LEDs noise filtering (C2 1000uf cap).
Thanks Salas for kind att.,
I take your advice into account, it happens now, my VALAB dac, has stopped accepting further modifications, jajaa. It is malfunction and feel it's over, I wringed it enough, yet!. I hope receive soon new DAC, Only an small and simple board filled with as much ICs as needed, one DIR9001 and one TDA1543 (I will fed it with three 1Kohm 'anode' choke on output). For clocking a Singxer F-1 usb bridge module. All in the same case.
My idea is a power supply with three Rflktr-D modules. 8v., 5v. and 3.3v.
First one is finished, I menage to set at 8v. with 3 leds I have on spare, one big white led and then two small round red.
To find properly R1 value, I set in place a wirewound pot 20ohm. decreased to around 4 ohm by paralleling it with 4ohm. cement resistor. Then adjusting by ear while music was playing. I found very picky, only a narrow corridor at exact 3 ohm. inbetween the swamp. Higher value is everything thin sounding, lower value not vibrancy, projection neither detail.
I take your advice into account, it happens now, my VALAB dac, has stopped accepting further modifications, jajaa. It is malfunction and feel it's over, I wringed it enough, yet!. I hope receive soon new DAC, Only an small and simple board filled with as much ICs as needed, one DIR9001 and one TDA1543 (I will fed it with three 1Kohm 'anode' choke on output). For clocking a Singxer F-1 usb bridge module. All in the same case.
My idea is a power supply with three Rflktr-D modules. 8v., 5v. and 3.3v.
First one is finished, I menage to set at 8v. with 3 leds I have on spare, one big white led and then two small round red.
To find properly R1 value, I set in place a wirewound pot 20ohm. decreased to around 4 ohm by paralleling it with 4ohm. cement resistor. Then adjusting by ear while music was playing. I found very picky, only a narrow corridor at exact 3 ohm. inbetween the swamp. Higher value is everything thin sounding, lower value not vibrancy, projection neither detail.
3 Ohm R1 means 200mA approximately. Lower mA and you challenge the current availability vs your DAC's 160mA needs and healthy margin to limit. More mA and you challenge the raw supply's reserve when using small electrolytic. So its logical you home in to a best compromise. Also make sure there is no treble small deficiency in DAC or loudspeakers that you try to compensate by "drying up" the supply near to the load+little margin working limit point. Try the J2 instead of R6 option also. Its a tighter vs plusher tuning tool depending on subjective synergy system needs.
Try also Mundorf 4 pin Ag+ caps. They are excellent. Especially on C2 (1000uf) position
And as you already mentioned give some time to burn in.
So, for digital part of chips (DVDDs) I am using Nichicon FPCAP on C2 position (best 1/f filtering, but terrible tone). However for AVDDs, analog Vrefs or powering resistor ladders dac directly (Soekris R2R), generally places from where audio signal is produced from, Mundorf 4 pin I found to excel any other cap in this position.
Hi Tiofrancotirador, thanks, so you use then a second Mundorf ag+ on a C2 place? Nice to know, beacuse of great footprint in this place and more affordable too. I will give a try on this.
Also I start considering 5v. supplies separated for digital and analog. It will sum a total 4 boards Rflktr-D.
I already fit the J2 straight, after reading many recommendations.
I understand this kind of balance is possible, yes yes.
My amplifier is a modified low power class D, with great synergy to drive a fullrange Rullit with fronthorn Azurahorn.
It's very balanced with other analog source. Although very dinamic and fast, is obviously a need for deep bass, but an smooth rollof prevents noticing at all. Preamp is Lightspeed LDR.
Also same unusual low settings on powering the Tripath TA2024 IC.
Drawing only 80mA.\12,8v. if I remember. I set your Bib board with input filter cap. decrease to 1500uf. and resistor current setting 14,5 ohm.
The main weakness on my system is bad main home supply two live fase 115v. and a ground 0v. No neutral coming to my home. I can't hear music during day time the labor days.
I wonder being able to power this high 12,8v. with Rflktr-D and compare also to Bib..
Nice FR speaker but I wonder about HF beaming angle as with any FR. White LED can be noisy. Try LM329 precision reference as Zener instead of LEDS. That will give you little above 8V without anything else in the Vref. Can try use smaller value C2 also when having only LM. Because low noise source with better voltage stability. Orient its cathode opposite to if it was LED. LM+LEDs can take you to 12.8 also. You will need external sinks. When you want more CCS than recommended with M1 IRF9610 try Fairchild FQP3P20 for M1 which has better junction to case theta and its only little more gate capacitive.
Hi Tiofrancotirador, thanks, so you use then a second Mundorf ag+ on a C2 place? Nice to know, beacuse of great footprint in this place and more affordable too. I will give a try on this.
Also I start considering 5v. supplies separated for digital and analog. It will sum a total 4 boards Rflktr-D.
Yes, changing resistor R6 to JFet is a must for me, then from my experiences quality of C2 is of second/same importance in terms of final sound signature/quality.
Not sure what you are using for I/V but if this is op-amp or some descrete stage try powering it with Salas BiB with Mundorf EVO Oil Impregnet Caps on output position (zobel network) and for LEDs filtering. This gave me spectacular results in terms of tone and final smooth/esasy sound signature. This was when I was playing with PCM1704. For chip powering I used reflectors. For I/V stage I tried reflectors with different cpas, BiB with different caps, but BiB with those Oil impregned caps excels in I/V stage everything else in this case.
TioFrancotirador,for C2 and CM,you have try with different value also,joimonf relate improved result with lower values.
Yes, C2 is prone to capacitor rolling also but just I don't do it in first instance, and now is no fun without my DAC.
My TDA1543 DAC set a resistor 1Kohm for I/V. But instead, I use a kind of choke, so called tube 'grid' choke in place. Actually I use two choke in parallel 200Henry/2K75ohm + unknown henry/1K67ohm. to set the desired 1Kohm value.
I also use PIO on the BIB boards and suppose to do so after in fine tuning R-D. I will use russian k75-10. value 4,7uF to bypass electrolitic. When removed the magnetic iron case. But doing that is scary, not easy, is need tools and time, better use latests Mundorf Oil, don't know TioFrancotirador but I will look at 4,7uF.
My settings:
On the Tripath amp BiB , 12,8 volts:
input: 1000uF+470uF Nichicon KZ
Led filter: 470uF Nich KZ + 4,7uF Paper In Oil
zobel output: 220uF + 4,7uF PIO + 0,01 uF Russian Teflon.
On the Dac BiB , 8,2v :
input: 1000uF Nichicon KZ
Led filter:220uF Nich KZ + 4,7uF Paper In Oil
zobel output: 4,7uF PIO + 0,01 uF Russian Teflon.
My proceeding to find C2 in Reflektor-D will be as follows, in first same value as in schematic, 1000uF, then lowering step by step in the whole range Nichicon KZ has, while listening to bass quality. Too much capacitance do loudness without detail, if too less then thin bass. When got, bass is pleasant with extension and detail. Then focus the listening to middle frequencies and try if a quality 4,7uf Paper In Oil is need and finally try if a teflon 0,01uf can enhance the high freq.
My TDA1543 DAC set a resistor 1Kohm for I/V. But instead, I use a kind of choke, so called tube 'grid' choke in place. Actually I use two choke in parallel 200Henry/2K75ohm + unknown henry/1K67ohm. to set the desired 1Kohm value.
I also use PIO on the BIB boards and suppose to do so after in fine tuning R-D. I will use russian k75-10. value 4,7uF to bypass electrolitic. When removed the magnetic iron case. But doing that is scary, not easy, is need tools and time, better use latests Mundorf Oil, don't know TioFrancotirador but I will look at 4,7uF.
My settings:
On the Tripath amp BiB , 12,8 volts:
input: 1000uF+470uF Nichicon KZ
Led filter: 470uF Nich KZ + 4,7uF Paper In Oil
zobel output: 220uF + 4,7uF PIO + 0,01 uF Russian Teflon.
On the Dac BiB , 8,2v :
input: 1000uF Nichicon KZ
Led filter:220uF Nich KZ + 4,7uF Paper In Oil
zobel output: 4,7uF PIO + 0,01 uF Russian Teflon.
My proceeding to find C2 in Reflektor-D will be as follows, in first same value as in schematic, 1000uF, then lowering step by step in the whole range Nichicon KZ has, while listening to bass quality. Too much capacitance do loudness without detail, if too less then thin bass. When got, bass is pleasant with extension and detail. Then focus the listening to middle frequencies and try if a quality 4,7uf Paper In Oil is need and finally try if a teflon 0,01uf can enhance the high freq.
Hi
Because it is relative to final JFET IDSS when put in circuit see post #707 first.
Also because its already built, you can measure Vout with JFET at R6 position while running the reg on dummy load, with now RX. Then you can measure mV drop across RX to know what 62 OHM already contributes. With Ohm's law you can calculate new RX if needed, or just tell us your JFET for R6 mod Vout and RX mV drop to assist you.
Because it is relative to final JFET IDSS when put in circuit see post #707 first.
Also because its already built, you can measure Vout with JFET at R6 position while running the reg on dummy load, with now RX. Then you can measure mV drop across RX to know what 62 OHM already contributes. With Ohm's law you can calculate new RX if needed, or just tell us your JFET for R6 mod Vout and RX mV drop to assist you.
Thank you Salas
I've had JFET 2SK117 with IDSS=3.1mA so I soldered it and this gave me Vout=3.4V. I've had also spare yelow LED so I replaced green LED with yelow one and I recive 3.288 ~ 3,31Vout. So resistor 62 ohm remained untouched
. Another issue: As I mentioned my circuit consume about 35mA so I set R1= 2.5 ohm (actually 2x5ohm/1W) but M1 heatsink heats up strongly.
What could I do to reduce the heat?
TIA
Thank you Salas
I've had JFET 2SK117 with IDSS=3.1mA so I soldered it and this gave me Vout=3.4V. I've had also spare yelow LED so I replaced green LED with yelow one and I recive 3.288 ~ 3,31Vout. So resistor 62 ohm remained untouched
Another issue: As I mentioned my circuit consume about 35mA so I set R1= 2.5 ohm (actually 2x5ohm/1W) but M1 heatsink heats up strongly.
What could I do to reduce the heat?
TIA
For 3.3V and with JFET mod I use only one green led and Jfet with high idss like around 6mA. This will give something like 3.25V (no need for 62ohm resistor in such case). Since most ICs has 5% tolerances on 3.3V your are still within the range here. I am using such configuration from the beginning and did not have any problems.
Hi, I did no tried different Cm values. I did try different C2 in Reflectors values:
15uf Mundorf EVO Oil
100uf 4 pin Clarity Cap TC 4
10uf Clarity Cap ESA
20uf Clarity Cap TC 2
and bunch of different 1000uF electrolytes and polymers.
I was using this to power Soekris R2R Vref directly and for I/V stage for pcm1704 also.
My conclusion are:
-going lower with C2 on reflectors made my system thin sounding (maybe if zener was used, results could be different, since it needs less filtering).
- 15uf Mundorf EVO Oil did not impact sound nicely like it did when used in BiB
- Mundorf 4 pin 1000uf/160V (that is of lowest voltage and lowest capacity) brought best results in terms of resolution, bandwidth and made treble very clear. Sound signature very similar to Black Gate STD. If it just had the tone of Elna Silmic II it would be perfect.
-for I/V stages BiB is a winner. I used 15uf EVO Oil for filtering the LEDs and 5,6uF EVO Oil on Zobel. Maybe different values would brought better/worse/different results. I just used those values, because they fit nicely in the board.
- Mundorf EVO Oil caps mod worked very well only for I/V stage. Could not get same results when using them for powering Soekris R2R Vrefs directly.
I also tried bypassing electrolytes in BIB. However most significant change was brought by skipping electrolytes and just using some possibly high value Mundorf EVO Oils (15uf for leds and 5,6uf for zobel).