Arcam Alpha mods

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hin again Simon,

Thanks for the input! I do have the service manual and it helped me of course though I do not have the test cd. These super regs seem to be a must fo good performance I will def look into this. I did not change the psu caps yet, any suggestions for what to replece them with? The motherboard caps I replaced now was with Sanyo OSCONs. Changeing the regulators should be a straight forward soldering action I guess?
Any other must do actions to make this into a killer drive?

All input greatly appreciated!

Cheers/Mikael
 
Well another thing would be to add voltage regulators on the servo section but I'd do the other things first as this mod's a bit trickier and doesn't bring massive gains in my experience.

With PSU caps you want a nice low ESR and ESL (series resistance and inductance) but really the key thing is finding ones of the right capacitance (more uF than standard is usually a good thing), voltage rating (same or higher than standard) that will fit into the space allowed. Panasonic FC and FM are good quality industrial capacitors that can fit into small spaces. Great smoothing caps (F&T, Epcos, BHC, Mundorf, Kendeil etc.) will not fit the spaces without some "bodging".

Another tweak that I don't want to do for fear of breaking a working player (applies to any player) is to replace key signal line traces with coaxial wire. Some people, including myself, did this to one or more lines (HF signal, for example) in the Marantz CD63 and got a smoother, cleaner sound.

As for adding regulators to the two chips I mentioned, you'll need to locate where standard 5V supply goes in (it passes through a low-value resistor before each chip) and cut the resistor at the chip side. At this point you will feed your new regulator's output (usually pin 3) into the SAA7220 and SAA7310. The ground (usually middle pin) goes to ground anywhere on the player. The easiest thing is to use a fibreglass pencil to scratch into the ground plane on the mono board or to find a ground jumper link and use that. The input pin (usually pin 1 on a positive voltage reg) must be fed via a wire link to a voltage source of 8VDC or higher. This can be taken from before the standard on-board 5V regulator. Checking around the player turned on using a multi-meter will quickly reveal where to tap this voltage from.

Simon
 
Further clock upgrades

I recently had the opportunity to acquire a new Fidelity Audio C2 master clock, now called "Premium Clock". This I used to replace my C1 clock, which is a simpler and now discontinued product.

Initially I tried running the C1 on the SAA7310 (decoder) and the C2 on the SAA7220 (digital filter) at the same time. This played music but, presumably due to sync issues, there was a ticking noise superimposed over the music. I quickly abandoned that idea, I have somewhere else I may try the C1 clock later.

The C2/Premium clock has 7 separate outputs: unbuffered F1 (11.2Mhz on mine), 2 x buffered F1, 2x /2 and 2x /4. So we have 11.2Mhz, 5.6Mhz etc. The clock has onboard regulation (very good voltage regulation at that) so I could leave the super reg attached to the old C1 clock for later use with that old clock.

The first thing I wanted to test was to see if just replacing the C1 running the decoder and digital filter would provide audible gains. I connected the two buffered /1 outputs to each of the SAA7220 and SAA7310 clock inputs. I listened for days and came to the conclusion that I like the clock a lot. The sound became more solid, lifelike, more open and three-dimensional and with less fatigue - particularly on busy passages, which are prone to hardening up on lesser CD players.
 
After enjoying this for a while I went back in to finish what I'd started. Now it was time to re-clock the DAC, namely pins 2 (BitClock) and 4 (SystemClock). Pin 2 receives a /2 feed, pin 4 receives /1. For pin 4 I took the clock feed from the spot labelled T1 on the Premium clock, as this is the lowest jitter point (the unbuffered output).

The PCB link between pins 2 & 4 was cut and R214 was lifted (original clock feed).

Some TDA1541 DAC versions *may* not accept this clocking arrangement. My chip is a Taiwan-manufactured TDA1541A and seems happy.

In fact, the sound is quite excellent. Gains were made in every conceivable area. The sound became more dynamic and less superficial once again. The bass, in particular, has taken on a new level of hardness and agility. The pitch of individual bass notes is much easier to discern. Overall, the sound is again easier on the ear and it's more possible to turn the volume up to very high levels due to the non-fatiguing source quality. This is, of course, somewhat material dependent but then again such upgrades seem to benefit ALL music, even the bad stuff.

The clock runs off the same power supply that ran my C1. It's a big frame transformer off-board, feeding a standard bridge (onboard), then 10,000uF smoother, a home-made coil, then a further 25,000uF with various bypass caps.
 

Attachments

  • Clock_Supply.jpg
    Clock_Supply.jpg
    165.3 KB · Views: 490
  • Clock_Lines.jpg
    Clock_Lines.jpg
    153.6 KB · Views: 458
  • Cut.jpg
    Cut.jpg
    140.6 KB · Views: 465
  • TDA.jpg
    TDA.jpg
    152.6 KB · Views: 462
A couple more pictures.

Not a lot more to do to this player now. Of course there's the DEM re-clock but a close friend thinks it's nigh on pointless (ie. barely audible gains) and that puts me off a bit. I might change my CAT5 clock lines for some good mini coax when I get some, though this works just fine.
 

Attachments

  • ClockEnd.jpg
    ClockEnd.jpg
    144.8 KB · Views: 454
  • Overview.jpg
    Overview.jpg
    164.5 KB · Views: 261
Thanks for the compliment. It does sound excellent.

If I'm honest with myself I can think of more mods to do but some of them are expensive and many involve lifting the main PCB out, which I'd rather not do again.

The smoothing caps on the DAC/original output board are of low quality so that's something I'll deal with. I might also replace the bridge with HEXFRED diodes too (currently schottky).
 
Digital and analogue ground

Hi Simon
One more mod to add to your excellent reference here. I made a change that might be worth noting to the Arcam audio board. The digital ground is connected to the analogue ground by a small ceramic cap, C227. I have very good ground back through the digital source, and out through the analogue side. I cut the connection by snipping off C227. The premis is why let digital noise from the clock get into the analogue and damage the sound. To me the improvement is significant. In my case more high quality base and micro detail across the sound spectrum.

I do not know how it will work (if at all) in the real ARCAM CDP. Recall, I am driving the audio board with a PCM2706 USB to I2s converter.
 
Digital and analogue ground

I would tend to agree. I record it here in case someone surfaces with more experience on this to share. I read somewhere the recommendation to separate the ground paths, and to my ears it is good. The tda1541a spec sheet reference circuit hard wires pins 5 and 14 together. What I don't understand is why ARCAM put a cap here. Would not that block DC and allow mhz noise from the digital circuit to pass through? Seems counter intuitive somehow.

I think it was Lucas at the Lampizator site said to think of ground paths as a river where current will follow the path of least resistance to ground. The idea is to let it flow away from the sensitive areas that can most affect the signal.

On a separate note, I have yet to replace the Arcam audio transformer. I have a nice Plitron with an 18-0-18 winding ready to drop in. I hope I get the same results you have seen.
Cheers
 
Return currents follow different paths depending on their frequency, and at 1mhz and above mostly follow the path of least resistance NOT resistance. This is important when playing with digital as it is imperative to create the smallest loop area for any paticular signal.
 
woops sorry, path of least inductance.
This link gives a nice overview:
http://www.x2y.com/filters/TechDay0...log_Designs_Demand_GoodPCBLayouts _JohnWu.pdf

I would not use anything resistive to join the two grounds, the aim is to keep them as near to an equipotental plane as possible, so my preference would be a ferrite, but choosing the correct ferrite can be a pain, you nead one to cover the digital noise frequencies.
A good reference for grounding issues are articles (and books) by Henry Ott.
We have similar problems where I work, with digital and analogue, and with so many viewpoints and solutions it is almost impossible to decied which is the best solution.We have tons of fancy equipement, EMC lab etc and can afford multiple re-spins of a product as we have to get it right. What I hope to do later in the year is a analogue digital PCB to test various schemes, such as linear, smps and battery power, different grounding etc, the idea being we have a set of board where the effect of digital (RF) noise on audio can be investigated. I am persuading the powers to be to make this an open source design, and hopefuly let me have some fully populated designs to throw out for others to play with. I am in the initial stages of this, but will start a full thread later in the year.
 
Thanks for the link, it looks really helpful!

Your project sounds really good, I'll look out for the upcoming thread.

Question: is there any benefit to adding ferrite beads to ground wires "willy-nilly"? I ask as I added one on my new clock's power supply ground where it joins from its own psu to the main pcb groundplane.

Thanks,
Simon
 
A common practice is to either use a ferrite or even better a Pi filter (cap, ferrite, cap COG caps are best) to cut noise from one section to another. On a lot of digital and analogue digital it helps limit the general mush on the planes from one section to another. If its a surface mount design there are some nice 3 terminal devices where you can isolate each supply pin.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.