Salas DCG3 preamp (line & headphone)

[jimk04]

Hello again guys

I'm struggling a bit here. Tried again yesterday with different power amp...tpa3255evm.. same noises. Also tried through input selector and made it worse. Haven't tried volume pot yet but I'm pretty sure now there is an issue with something.

I will struggle to diagnose this...no scope....and wandered if anyone in the UK I could post to and for small fee help me out with this please?

Admittedly not 100pc tested C1 and C2 yet. I cant read the print and can't measure in circuit so will remove and check but I do remember double checking them before soldering in.

[IDM]

Originally Posted by IDM

Hi,

After all the issues with a fried J1. I finally got the DCG3 finished and whilst not completely boxed up yet, I have been able to drop it into the system. I am using to feed my Pass F5 and I have to say it is truly superb. It just seems so agile, detailed but musical. I am totally delighted.

So again thanks for the advise in getting it sorted.

I am using it with dual mono PSU with 4X ultraBIB regs and I see a few posts back that I should change the R1 values for the negative and positive regs.

Originally Posted by Salas

Congratulations!

Using BC327 for Q1 Q2 already I suppose. What R1s you use in the UBiBs, and is the dissipation manageable?

Yes, I am using BC327-40's.

The R1 resistors for the regs are all 2.7R. I had the preamp on for about 3-4 hours yesterday evening and the regs were running quite hot. I will measure and monitor the temperature tonight.

I have been using valve pre-amps for years either 2p29L wired as DHT or your 6V6 linestage, and what really surprised me with the DCG3 was the vocal presentation. With the DCG3 so much more detail, without any hint of harshness.

[Salas]

Hi

2.7R R1 in the UBiB gives it 210-220mA CCS current. If your DCG3 output stage bias is chosen circa 100mA then for the negative rail you got OK extra shunt current margin. For the positive rail when you use it as a power amp driving preamp only, as much R1 is still OK.

But if you use it for lowish Ohm headphones also and listening loud, it will be pushing enough power, and you should up the PSU's current margin on the positive. Because the positive side is the one that swings signal current. Its not push-pull, its a single ended Class A CCS biased output stage. Can go twice its bias current peak to peak before reaching one sided clipping level.

Rule of thumb for knowing you always got 100mA reserve, massaging the particular shunt regs to good effect in all DCG3 uses, is: Negative rail R1 setting = DCG3 bias level +100mA. Positive rail R1 setting = 2xDCG3 bias level + 100mA. Sinks size must be adequate. The UBiB's MOSFETS are tough but for robust reliability better not have more than 60C on their sinks in the long run. With so much high idling circuitry in a box, good ventilation is a must too.

I am happy to know that you like it so much. What you describe was my goal i.e. not to only sound hi-rez but musically right also. To stay so when driving long cables and low impedance amps strongly too. Actually, the headphones function was a side effect of that. I did not start the development with headphones in mind.

Knowing the F5 I understand it has a particularly good synergy with the DCG3. This preamp fills it up because F5 maybe is pure and agile but it can also come across a bit stark at times if on somewhat lean speakers or ancillaries.

From where you got the J1 replacement part in the end? UK?

[Salas]

Originally Posted by jimk04

Hello again guys

I'm struggling a bit here. Tried again yesterday with different power amp...tpa3255evm.. same noises. Also tried through input selector and made it worse. Haven't tried volume pot yet but I'm pretty sure now there is an issue with something.

I will struggle to diagnose this...no scope....and wandered if anyone in the UK I could post to and for small fee help me out with this please?

Admittedly not 100pc tested C1 and C2 yet. I cant read the print and can't measure in circuit so will remove and check but I do remember double checking them before soldering in.

Could it be some problem with bad uPA for you too? Is there any DC offset on the signal inputs? Still try with the pot also, its important. The pot's metal body should be grounded. What happens if you mock up the whole preamp as now but having it also grounded on a metallic resting plate? Maybe IDM can practically help you debug it in the end if there's nothing more you can do?

[IDM]

From where you got the J1 replacement part in the end? UK?

I have ordered some from Teabag, which are probably somewhere over the Atlantic. So due to my impatience I also bought one from here: nikkoe.com

Maybe IDM can practically help you debug it in the end if there's nothing more you can do?

I would be delighted to try and help. jimk04, if you do get totally stuck you are welcome to send it to me to try and debug. Obviously I cannot guarantee anything but will try. It is certainly worth persisting with as its a great preamp.

Cheers
Ian

[jimk04]

Will remove C1and C2 to double check they are correct, and try a better mock up with grounding.

I really want to succeed with this as I have quite a lot of money in it now (for me!) what with the transformers etc so I dont really want to have to sell it all on at a loss.

My fingers are crossed. I dont have lots of spare time so I won't be back on here tonight with any findings.

Edit.....sorry Ian you must have posted whilst I was writing my reply. That is fantastic comfort to know I have your backup....thankyou .

[Salas]

I wish you will solve it alone, but if no luck still cheer up, because fellow diyer IDM located also in the UK just volunteered to the rescue. Hey, he even has a bunch of BC327s to select from and maybe fit in your build as well.

He already deserves a crate of Newcastle Brown Ale only for extending a helping hand.

[jimk04]

Yes Salas I am lucky with the support of DIYA members. Thankyou for you help also. My knowledge and experience is very limited, sometimes I dont understand what you guys are saying....but I try and learn! It would be nice to solve this myself with the help of others.