Salas hotrodded blue DCB1 build

Finally started soldering my DCB1. Two questions came up:

- Are there any issues with reversing the FETs and diodes so that they can be mounted to the chassis, underneath the board? I've seen many pictures of this, but wanted to check if there was anything that needed to be altered to facilitate it.

- Is a plate of 7" x 11" x 1/8" aluminum (approx 180x280x3mm) enough of a heatsink for the DCB1?
 
Hi everyone

First of all, I would like to really congratulate Salas, Tea-Bag and all for the wonderful work you guies are doing with the DCB1 project. You are very knowledgeable people. Showing a lot of of patience and respect in answering all those various questions and/or inquiries. Lot of passion and dedication.

Very much appreciated. A big thanks on behalf of everybody !!!

Well now, I am a newbie in the DCB1 building project. Was made aware of this wonderful "preamp" when I listened to it at my fellow audiophile's place a few weeks ago. Very very impressive stuff. Since then, I ordered boards from Tea-Bag and am part of the GB.

I read close to a thousand posts on DCB1 with great interest, but still have some remaining questions (got to admit I might have missed some answers as there are so many replies...so, sorry if I am repeating questions already answered somewhere).

Anyway, here they go:

1) I noticed there is a wide range of capacitance possible in the shunt regulator (with a significant gap ranging from 0,22-10uf if using film caps or from 10-100uf if using electrolytics). I know that these caps are not in the audio signal but still, since there is such a wide difference in values (almost 50X between 0,22 and 10uf), I was wondering if someone has compared different values of same quality caps and noticed a difference in sound quality ? I plan to use Audyn true copper film caps.

2) Could I replace the 100uf @ 35v electrolytics for 220uf @25v as I do have some extra Black Gates left that I would gladly install (yes size does fit) or is 100uf the max capacitance that could be use ? If 220uf was to be too much capacitance, wich other changes would then need to be done to be able to install them?

3) Wich relay would be best between Omron G6H-2-100-12DC or the Japanese made Takamisawa A-12W-K (Mouser# 817-A-12W-K) as both have very similar specs (including gold plated silver alloy contacts) ?

4) I am still contempling the idea of building a balanced DCB1 with one power supply per board, of course Is it possible to use RCA ouputs connectors when building balanced ?

5) Talking about power supply, if I was to use say 60 000uf (4 X 15000uf) with 100VA E-core transfos per board, would I need to change (upgrade) some components'values to be able to deal with bigger inrush current ?

6) If was NOT to go the balanced route, I would then build it with two separate power supplies, one for B+ and one for B-. Each channel with their own 100VA E-core transfos and 20000-40000uf caps. In this case, wich changes would need to be done ?

7) I found some ultra high-precision Vishay thin film resistor, Series UXB 0207 with 0,01% tolerance and 2ppm...Assuming I would be willing to spend the $$$ for them, would their be any benefits?

Oouuff !!! Lots of questions for a 1st post.
Sorry !!! ;-) ...but there is just no other place on earth where I could find such valuable anwers !!!

Thanks in advance for your kind help.

Sincere regards

Marc-André Rodrigue aka "Scorpion"
Montréal, Canada
 
diyAudio Chief Moderator
Joined 2002
Paid Member
The proposed component types and values in the typical Hypnotize DCB1 are both fixed for technical reasons, like the 100uF should stay 100uF, and for subjective reasons like a 0.22uF Wima MKP10 was a good compromise for few $. Yes, they have tested bigger values or alternative types and they settled there back then, it was a small group of volunteers. That does not mean that any excellent capacitor that fits should be exactly 0.22uF.
I see that you are willing to spend enough for a "deluxe" build. The resistors you mention are good for the audio section ones, main filter capacitance beyond four 10000uF -good quality- pieces tends to "slow it down" IMHO, for relay there is NEC EA2-12NJ with excellent quality and has been tested OK, and 600mA CCS is a good stage of operation. On the other hand you could just clone the one build you liked at your friend's for parts and CCS setting and avoid experimentation. Was it a standard kit level or a luxurious attempt?
 
Thanks for the prompt answer Salas.

Ok, I will max capacitance at 100uf and forget about those extra Black Gate caps (I have 21 extra caps of different values, FK and NX series that I wish I could use. Will have to save them for another project. Might even sell them. Anyway).

Ok si if I understood correctly, a few volunters made some experiments with different capacitance values and finally settled on 0,22uf because of performance to price ratio but also assuming compromise ? Ok, will probably go with Audyn true copper 0,47uf as they have the same size as the 0,22uf...still wondering how would, say a 5uf, compare soudwise.

I will try to go all out with your design and go maximum quality on parts. Where it will bring me, I'll experiment and see.

Think I will take the B+ B- road and use the extra board for experiments on the sound of different quality parts at specific locations. Will put extra attention to components in the audio signal. This way, will be able to fine tuned sound to max musicality and performance.

Will start with Vishay naked Z-foil in the signal for max transparency. Will then try different quality parts at position#2 in the signal. Make the optimal choice, and then, move to position#3 and so on. Don't care if I have to try 10 different resistors per position as long as I end up with max musicality.

Btw, are you insinuating that NEC EA2-12NJ relay would be a better choice then Omron G6H -2-100-12DC or Takamisawa A-12W-K....or just another good option ? Would there be an even better option available ?

Also, assuming I got what you wrote correctly, even if I go B+ B- with 100VA and 20000-40000uf per side, I wouldn't need to change anything else on board ?

Finally, my friend's board was not a fully Deluxe version but kind of a "middle-of-the-road" one. He has Vishay Z-foil, IXYS Hex Freds, Takman carbons, Vishay s102 resistors, Nichicon KZ/FG caps, WIMA MKP 10 etc.

Main difference is that he uses Mills 1,2 ohm resistor instead of 3,3 ohms (or 10 ohms as per basic)...so his board runs VERY hot (as something like 1,2A) !!! Friend says the sound is even more liquid. Who knows !!!

One thing is fore sure, sounds outstanding.

Regards
Marc-André Rodrigue
aka Scorpion
 
diyAudio Chief Moderator
Joined 2002
Paid Member
-Nichicon FG and KZ are good 100uF alternatives, Silmic II also. The BGs ain't fresh anymore, that could prove an issue even if you had the right values at hand.
-I don't think you will need more than 1uF for the film cap LEDS bypass. More foil brings more self inductance. Capacitor intrinsic spec matters most there.
0.22uF was good regarding MKP10. Which proved a solid VFM choice in the DCB1 synergy context. There are better of course, but $$, see about Clarity MR also.
-Good luck with your numerous parts subjective tests plan, I could not disagree to investigations when you are willing and able. See about Rhopoint Miniohm also.
-The NEC brand has a rep of being the top dog in relays quality though I could not fault the OMRON or the NAIS, about the Fujitsu Takamisawa I don't know, I haven't used it.
-20000uF per side max IMHO. The TO-220 MUR 820 or IXYS equivalent will take the inrush when with 100VA. Use the proper primary fuse as recommended in the transformer's spec.
-Your friend's build isn't a slouch for parts though.
-Your friend's CCS level helps it further, you could be content near but below 1A too.
-About your subjective evaluation compliments, thanks.
 
Thanks Salas, that clear things up a lot.

Btw, I was thinking about using big Siemens/Epcos Sikorel 125°C or SIC-SAFCO Felsic Capax in the PSU for their very high quality, reliability, excellent sound and very low ESR... but since both models are screw-terminal types, I guess I would end up with increased resistance (from routing wires from caps to board). Is there any way to counteract or to lower this effect ?

Of course, I would keep wire lenghts to a minimum and keep your advice to max capacitance of 10 000uf X2 per side B+ B-

I was thinking about using very high voltage caps (350v) because the higher the voltage the lower the ESR...Also had the idea to link big 200uf @350V SCR aluminum metallized polypropylene caps as "current drawing caps" to the Sikorel or SIC-SAFCO in order to lower ESR even further (big SCR have ESR OF 1,3 !!!).

In fact, would like to imitate the way PSU are built by Audiomat (known for the excellence of their PSU and their exquisite sound quality). My source is composed of Audiomat D1 drive and Maestro 2 DAC....love them both. Had Audiomat Opera Reference before and loved it too.

Would it be feasible/logical in the DCB1 context ? I am willing to build the aluminum enclosure to allow required space to fit the hig caps.

Please see the following links for Audiomat:
6moons audio reviews: Audiomat Opéra Référence
AUDIOMAT MAESTRO REFERENCE | Convertisseurs | Audiomat
Audiomat Récital MKIV

The following link for SCR (look for PC20000AL capacitor when from pdf file at bottom of page:
AC capacitors with aluminium enclosure | SCR

The following for SIC-SAFCO Felsic Capax (pdf file...look for 10000uf @350-450V):
Screw terminals

And for Epcos Sikorel (pdf 125°C):
EPCOS - Capacitors with Screw Terminals - Aluminum Electrolytic Capacitors - Capacitors - Product Catalog - Technical Parameter - Product Search - Products - Home


Sorry if there are some mistakes here and there as english is not my first language (french is).

Regards

Scorpion
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Thanks Salas, that clear things up a lot.

Btw, I was thinking about using big Siemens/Epcos Sikorel 125°C or SIC-SAFCO Felsic Capax in the PSU for their very high quality, reliability, excellent sound and very low ESR... but since both models are screw-terminal types, I guess I would end up with increased resistance (from routing wires from caps to board). Is there any way to counteract or to lower this effect ?

Of course, I would keep wire lenghts to a minimum and keep your advice to max capacitance of 10 000uf X2 per side B+ B-

I was thinking about using very high voltage caps (350v) because the higher the voltage the lower the ESR...Also had the idea to link big 200uf @350V SCR aluminum metallized polypropylene caps as "current drawing caps" to the Sikorel or SIC-SAFCO in order to lower ESR even further (big SCR have ESR OF 1,3 !!!).

In fact, would like to imitate the way PSU are built by Audiomat (known for the excellence of their PSU and their exquisite sound quality). My source is composed of Audiomat D1 drive and Maestro 2 DAC....love them both. Had Audiomat Opera Reference before and loved it too.

Would it be feasible/logical in the DCB1 context ? I am willing to build the aluminum enclosure to allow required space to fit the hig caps.

Please see the following links for Audiomat:
6moons audio reviews: Audiomat Opéra Référence
AUDIOMAT MAESTRO REFERENCE | Convertisseurs | Audiomat
Audiomat Récital MKIV

The following link for SCR (look for PC20000AL capacitor when from pdf file at bottom of page:
AC capacitors with aluminium enclosure | SCR

The following for SIC-SAFCO Felsic Capax (pdf file...look for 10000uf @350-450V):
Screw terminals

And for Epcos Sikorel (pdf 125°C):
EPCOS - Capacitors with Screw Terminals - Aluminum Electrolytic Capacitors - Capacitors - Product Catalog - Technical Parameter - Product Search - Products - Home


Sorry if there are some mistakes here and there as english is not my first language (french is).

Regards

Scorpion

Big PSU capacitor components and parallels will null the relative PCB arrangement for the snap in ones and will necessarily introduce cabling. It will be an outside from PCB arrangement and I can't predict its impact. Naturally it will work but if its worth the trouble you will be the one to judge since I only used snap in on board as everybody here without an exemption that I can remember. Side to side square profile cables can be better for inter cap and to board connections with less inductive impedance than loose wide ones.
 
I'm powering an LED lit power switch off of the LED position on the PCB with the arrow pointed to it. I am going to not run this LED power switch off of this spot now. Do I need to put a PCB LED in this spot or can I just leave this LED spot empty without issue?

Dcb1_photo_zpsb482571d.jpg
 
I mentioned that one of my DCB1 boards had rCRC fitted ON THE BOARD.
You did not pick up on that.
It clearly shows less ripple being fed to the regulator.
But I cannot hear any difference in the sound output.

Thanks for your answer Andrew T. You are right, I did not pick up on that. With so many posts, I guess I missed/skipped that part.

Btw, excuse my ignorance but what is a rCRC ?

Since you said you didn't hear differences in sound from less ripple to regulator, I am wondering if I might not be complicating stuff for nothing i.e going trough the hassles of building an external PSU with top specs like super low ESR...but also introducing higher inductive impedance from these extra cables, as pointed out by Salas.

Again, thank you for your time.

Sincere regards.

Scorpion
 
The regulator makes a pretty good job of isolating the amplifier/buffer from the changes/interference coming in from the mains.

For that reason one should not expect a good regulator to show a big improvement by improving the quality of the DC coming from the smoothing & rectifier.

An lm317 is a poor regulator especially at HF and this generally shows a big measurable improvement when less ripple is fed to the chip. So big that it is usually audible.

I would not expect that to happen with the DCB1. But the experiment was worthwhile.

RC is a filter.
rC is a filter.
rCRC is a cascaded filter.
The r is the transformer and wiring and rectifier resistance.
The first C is the first smoothing capacitor.
The R is the added resistor between the two capacitors.
The last C is the main supply of transient current to feed changes in current demand.
 
Last edited:
Hello,
I'm planning this build starting from an old kit a friend forgot on a shelf. I admit... I'm totally noob! I tried to scan this thread, but I couldn't find an answer.
I'd use a 16Vx2 trafo, used in commond with the volume/input selector board. Is it enough for a slightly hotrodded build?