Arcam Alpha 5 CDP upgrades

I think that looks correct.

Just for reference, I think Felix got the name confused for the anode and cathode - cathode is the line on the symbol and on the diode itself when you get them.

 

Oops good catch! But Mike's diagram is correct

 

Hi Tenson, Felix,

Well I finished the planned modifications to the audio board power supplies as discussed in the previous posts.

I'm pleased to say that the CD player is still alive and I've even managed to convince myself of an audible benefit too so all together it's been a success!

Here's the proof:

windows 7 screen shot

Before connecting the CD player up for a listen I checked all of the voltages. Here are the results:

+15v regulator output (Z201) now reads +14.23v
-15v regulator output (Z202) now read -14.31v

+5v regulator output (Z203) now reads +5.25v
-5v regulator output (Z204) now reads -5.27v

I used a std 1N4004 diode in series with a 13v 1W zener (1N4743A) on the +/- 15v regs and 2 small green LEDs on the +/- 5v regs.

Like I said the CD players sounds great but I may try to get the outputs of the +/- 15v regs a little closer to specs at some point. I'll probably swap out the 13v zeners to something more like 13.6v.

I've also been busy with some other mods to the casework too. The standard casework is a bit flimsy so I decided to mount the casework onto a MDF plinth. The plinth is 19mm MDF sprayed in satin black paint to match the front panel.

I used the standard rubber feet off the player and put 1" wood screws through these to firmly fix the player to the plinth. For good measure I also added some more bitumen sheet to the underside of the casework.

The case is now really solid which must be an improvement. I use some squash balls cut in half as vibration isolators under the plinth.


upload foto


windows 7 screen shot

 

Haha.. I used to have half squash balls under an Alpha 9

You can just about see them in this old picture



I think if I were you I'd leave the 15V regs at that voltage. If anything, the 5V regs are a little hot. You might try a 5K resistor in parallel across the two LEDs. That won't add much noise, right Felix?

 

Not in the least, since it's 5K// ~max 100R (LEDs at 5ma). Not significant, but a move in the right direction.*

Nice job Mike


I have squash balls under the slate supporting the Impulse H2s! (Only put there to help levelling/roll to position. These days however they're so squashed, well, I can't usefully take a photo...)


* Lower impedance at the Vadj pin is better, always. Ultimately the best you can do with the adjustable 3pin regs is approach the datasheet performance for what is (effectively) a 1.25v regulator.

 

I'm amazed they can take an H2 on top of them! You didn't fancy trying golf balls?

 

So it's not just me with the squash balls then!

I assume you are suggesting a 5K resistor in parallel to the LED pair and not on each LED?

I reckon I could easily put a resistor in parallel across the LEDs by soldering it in place on the underside of the PCB.

Any suggestions for further tweaks would be welcome.

 

Stop tweaking and enjoy what you have

 

OK Tenson you might have a point. I promise to put the soldering iron away for a bit.

In the meantime here are some details of further mods which I have already made to the player but never got around to posting:

I have also fitted a dedicated power supply for the SNOS board.

The NET audio SNOS board contains a better quality (low jitter) clock. This replaces the standard clock crystal which is removed when fitting the SNOS conversion.

The benefits of fitting a dedicated clock power supply separate from the noisy circuitry inside a player seems to be well documented so I thought it was worth a try.

I debated building my own clock PSU for a while and even considered one of the cheap ready made regulated PSU modules which can be found advertised on ebay (they usually come from Hong Kong or China).

In the end I decided to splash the cash on one of the ready made regulated clock PSU modules sold by NET Audio. They had given me great service before so it felt right to put some more business their way.

I understood that the wires leading from the PSU to the clock should be kept as short as possible so the best location for the PSU module seemed to be on the casework behind the LCD display. I removed this panel and drilled holes for some small plastic 'standoff' feet which I sourced from Maplin.

The AC power was tapped from near the IEC plug inlet where conveniently the board has two empty hole in the PCB next to C401 and C402 (the blue discs). I soldered in some PCB pins into these holes and connected everything up from there.

For the wiring I used some modestly priced silver plated copper wire with Teflon insulation and I connected up the optional ground for the PSU to the case using a nut, bolt and split washer. The AC carrying wires were fixed to the case with some self adhesive cable clips.

The -ive lead of the PSU power out goes to what would have been pin 13 of the SAA7220P/B chip and the +ive goes to the power pin of the SNOS board.


upload foto


windows 7 screen shot

Fitting the clock PSU made an immediately noticeable improvement. It's hard to describe the differences but essentially the player just sounded clearer.

This was certainly not one of those upgrades where careful hours of listening is needed to establish whether any benefit has been gained. It just sounded obviously better straight away.

 

Next up I decided to play closer attention to the DC blocking capacitors (C13 and C113).

I've every faith that the Elna Stargets provided in the SNOS kit were much better than the standard electrolytics but even so I much preferred the sound of the Elna Silmic 2 capacitors which I later fitted.

It's clear that these two DC blocking caps have a big influence on sound so I wanted to try to optimise this area further.

In my experience swapping electrolytics capacitors in speaker crossovers to quality polypropylene types (e.g. Clarity Caps) can make a staggering improvement. I therefore wanted to try to fit some polyprops in place of the Silmic 2's but the problem is that most decent polypropylene capacitors with a sufficiently high capacitance value are massively larger than the standard electrolytics and won't fit without major headaches.

After a bit of research I settled on some 1.5uF 250v Mundorf M Caps (MKP)
from Hi Fi collective. These were inexpensive and small enough to fit in place of the electrolytics when mounted vertically on the board.

I hoped that 1.5uF was enough in place of the 100uF electrolytics.

At the same time as fitting the M Caps I decided to swap resistors R14/R114 and R13/R113.

I went with genuine Shinkoh tantalum resistors for R13/R113 (22R 0.5W) and Audionote tantalum (47K 0.5W) for R14/R114 as I couldn't get Shinkoh resistors with the correct value.

I've heard it said many times that 'oversized' (higher wattage) resistors sound better than smaller ones hence the choice of 0.5W.

Initially after fitting these parts the player sounded a bit muddy however slowly but surely over the next few hours of listening the muddiness completely cleared.

After some more listening I think that the bass had also tightened up slightly too.


upload foto

 

I have also replaced the some of the resistors in an around the output stages.

R4 and R104 (1K2)
R5 and R105 (2K2)
R9 and R109 (2K4)
R10 and R110 (4K7)
R11 and R111 (1K)

I used the PRP PR372 resistors from Precision Resistive Products Inc (sourced from Hi Fi Collective). These are the red coloured resistors in the picture.

http://www.hificollective.co.uk/comp...resistors.html

These are non magnetic Ni-Chrome metal film resistors with 1% tolerance. I used the 0.25W rated ones.

I'd read some very positive comments about these on other forums and they are also very reasonably priced too.

Here is a picture of the resistors in their new home:


upload foto

 

Whilst I had the player in pieces again I also replaced C416 with a Panasonic FC 470uF 35V. I'd originally put an Elna Cerafine in this location as I had one lying around in my spares pile but the larger Elna didn't fit too well and it also made sense to save the expensive Elna for something in the audio path.
C146 is reputed to be a bit of a liability in the Alpha 5 and sometimes fails on older players. If you are doing some work to your Alpha 5 it may be worth replacing this capacitor at the same time.


upload foto

 

I also finally got around to removing the transport mechanism in order to gain access to the very last electrolytic on the main board which hadn't already been replaced. It's a nice feeling to know that EVERY SINGLE electrolytic has now been replaced, no more dried out leaky caps to worry about.



I also finally noticed that C219 was missing from the audio board. You'll notice that in some of my previous pictures this isn't fitted. When I re-read the fitting instructions for the NET Audio SNOS kit it's clear that this should have been replaced with a 220uF 16V Rubycon ZLH but for some inexplicable reason I'd obviously forgotten to fit it....doh!

According to the manual;

'The +5V supply (to the DAC chip) is filtered by RC network R216, C219 and C220.'

For now I fitted an Elna Cerafine 220uF 16V as I had one handy. Interestingly I can't honestly say that the player sounds any different with the missing C219 now reinstated!


upload foto

 

The 1.5uF cap should be fine - it looks like there is a 47K resistor in parallel after the cap just before the output plug, right? The low res image is tricky to read. This will set the cut-off frequency to 2.3Hz. The trick with DC blocking caps is to set the turnover frequency significantly lower than the audio signal you want to pass undistorted. Once the cap/resistor combo starts to cut the output level it will also introduce distortion. For this reason Arcam chose the large value of 100uF so that despite being an electrolytic cap it will have little impact on the audio range distortion.

P.s. You should be able to change that 47K resistor to something bigger like 200K without problem and lower the turnover freq.

 

This is what the output looks like:


windows 7 screen shot

As you have seen I have used a 1.5uF Mundorf MKP at C13/C113.

Resistor values at R13/R113 and R14/R114 are as standard but swapped to 'audiophile' equivalents.

Some people recommend the use of a smaller value 'bypass' capacitor used parallel across the large value DC blocking capacitor. Is this something which you would advocate?

 

Not with a film cap there to begin with. Parelleling caps is rarely a satisfactory answer to anything - just use one good cap, and you have.

Nice job

 

Thanks Felix, Tenson,

Next time I'm ordering some parts I'll get a couple of 5K resistors to put across the LEDs on the 5v regs.

Just to clarify, is it definitely the value of R14 rather than R13 I would need to change to lower the turnover frequency? (see schematic above).

 

Yes R13 is just a small value 22 Ohms to buffer the opamp output from cable capacitances. R14 47K sets the frequency with C13. I doubt it will really do anything to lower the turnover freq. more but it can be done easily for the cost of a couple resistors.

 

I wouldn't worry about it; R14 will be swamped by whatever amplifier you connect this cd player to. If your (pre)amp has an input loading of 47K also, then the Alpha 'sees' the two in parallel, roughly 22k load.

With the 1.5uF coupling cap that puts the -3dB point at about 5Hz. Fine for music!

 

Oh yes good point!