Replace caps of Arcam Alpha 9 CDP?

Okay Eldam. I’ll measure the caps C47, C48, C44, C45, C36 before eventual replacement. Can this really be done (safely) without desoldering? Btw, I forgot to list C36 earlier (3300 uF 25V). Unfortunately my multimeter only measures capacitance up to 20 uF, so I first have to get another multimeter.

Cerafine caps C4 and C104 are not on the main PCB of the Alpha 9 (page 9). The output circuitry has been moved to the DAC board (page 16..). And the DAC board of the Alpha 9 is different from the DAC board of the CD23T. So my previous plan to swap the DAC boards of the players won’t work..

The Alpha 9 CDP sounds very good, but the CD23T appears to be in a different class. I suppose it is possible that the audible differences between the Alpha 9 and CD23T are entirely the result of the improvements/upgrades of the CD23T relative to the Alpha 9:

• A separate toroidal transformer to supply the Ring DAC board
• New dual layer motherboard, with improved power supply grounding and buffering of clocks and audio data
• New Ring DAC board which reduces distortion in the audio band
• Improved master clock and clock output distribution circuitry on the Ring DAC board

However, I’ve always understood that the audible differences between these two players are subtle. The differences I’m hearing (e.g., in naturalness, ease, detail, sound-stage) are quite astonishing..
 
Last edited:
Frankly the first caps of the Pi are not a problem if a lower capacitance and a worst esr than when brand new ! But of course if it brokes, but the CD player has a off switch as you said and is not too old so.... !

Before changing your caps measurement device : try it on a bare 2 000 to 4700 uf cap 😉 !

In your choose (shoes) I will try the c47/48 tip as the come back is easy : it's just // caps on the main ones you keep !

Try to identify what the DC blocking caps are (in serie after an opa and before the RCAs plugs !

I believe you are here :

1) Not sure swaping C36 changes your life : but here are you looking for security because you believe all those caps are too old : C36/C44/C45 ?

If you can not sleep about security and collapsed device from this brand : just change these 3 by their capacitance/voltage clone in the Panasonic FC range (105° and gives often an excellent result in power supply smoothing/reservoir position).

2) If you want to try to tailor the sound : try what I wrote in the post above as it's reversible ; finding if there is a DC blocking cap and howit sounds with an other can be an important part of the process as well !

You see about security 2) is the opposite of 1) : tweaking a device with a lot of roms chips is dangerous ! In the post above I tryed according my own experience and at the reading of the shematic to give a good enough improvement just based of the sound targett (so not 1) point) and a maximum of security (not playing around the chips and regs : shorts guaranted !)

You could do 1) + 2) as well !

Choose your poison and level of risk !

regards
 
Last edited:
What does your cap meter actually read ? Is it just capacitance because if so then its a poor guide as to the condition of a cap. Measure a known good cap out of circuit and then add say 1 ohm, then 10 ohm and then 100 ohm in series with it because that is the way caps fail, hey develop high internal resitance (E.S.R. or equivalent series resistance). What does the meter do in that case ?
 
Frankly the first caps of the Pi are not a problem if a lower capacitance and a worst esr than when brand new ! But of course if it brokes, but the CD player has a off switch as you said and is not too old so.... !
Well, actually it appears it doesn’t have a true on/off switch, but I turn of the grid power to my audio stuff with one central switch when I’m not using it, so that would significantly extend the life of the caps 🙂 I bought the Alpha 9 in 1998 and the CD23T is from 2004 (not used much according to the previous owner)..

Before changing your caps measurement device : try it on a bare 2 000 to 4700 uf cap 😉 !
It just doesn’t measure it. It says 1 in the display..

In your choose (shoes) I will try the c47/48 tip as the come back is easy : it's just // caps on the main ones you keep !
Okay, I might try that..

Try to identify what the DC blocking caps are (in serie after an opa and before the RCAs plugs !
There are no blocking caps after the OPA (see page 19 of the service manual)..

What does your cap meter actually read ? Is it just capacitance because if so then its a poor guide as to the condition of a cap. Measure a known good cap out of circuit and then add say 1 ohm, then 10 ohm and then 100 ohm in series with it because that is the way caps fail, hey develop high internal resitance (E.S.R. or equivalent series resistance). What does the meter do in that case ?
My cap meter only measures Farads up to 20 uF. I might need a dedicated cap meter that also measures E.S.R. Here’s a video of a guy that simply feeds an audio signal through capacitors in order to test them. 🙂 If the volume of the treble is down, then the cap is bad: https://www.youtube.com/watch?v=J9-k0HgGzbU
 
By the way, when I disassembled the Alpha 9 to get out the main PCB, I noticed that most of the screws that fasten the main PCB where loose.. 😕 Could that possibly have been a problem in terms of internal rigidity and fibrations altering the sound?

Maybe I should just reassemble the player and listen again..
 
Caps are so cheap its just not worth deliberating. You could change them all for what, 15£$ or so.

The method in the video... lol, don't temp me 😀 Clueless comes to mind. If a cap has even 200 millohms series resistance at high frequency it could be considered bad. That guys method is hopeless.

Do it properly, use a wide bandwidth oscilloscope (with correct grounding) and look at the noise/ripple across the caps or spend the money you have for the ESR meter on new caps.
 
Caps are so cheap its just not worth deliberating. You could change them all for what, 15£$ or so.
Yes. Cost obviously is not an issue. But some of Eldam’s remarks have me a bit worried, such as: "change the reservoir caps only and that's all: you risk to collapse a chip" and "do not play around the chips and regs: shorts guaranteed ! "

Anyway, would you agree with Eldam to use Panasonic FC (not Rubycons) for all caps? And what about the indications ZA, RA, nonpolar, in the specs? And isn’t the Oscon cap (C53) a better quality cap?
 
I tend not to worry to much over individual brands as long as they are good quality parts from recognised manufacturers.

In other words I'm not a believer in boutique parts 😉 If it meets the requirements of size, voltage rating, temperature rating, ripple current rating at the frequency involved and has respectable ESR (all modern parts do) then I tend to buy on price tbh.

Non polar are specific... they are non polarised parts and you need to see whats available from the suppliers.

Oscons have a good reputation (in the right place) but are of a different construction to 'wet' type electrolytics. They have been around for decades.
 
Okay, thanks Mooly. Personally I’m inclined to use Rubycons because that’s the brand Arcam used originally. I’ll just see what’s available. And shall I leave the Oscon then or replace it with a new Oscon? And apart from discharging the caps (over a resistance) before desoldering and soldering the new caps in, is there anything in particular I should or shouldn’t do to reduce the risk of damaging a chip?

Sorry for the many questions. Apart from knowing how to solder components in and out, I’m a complete novice to this.. 😱
 
Probably not much to be gained changing Oscons for the same. They don't deteriorate like wet electrolytics.

Provided the player is turned off and unplugged then the caps will have discharged into the players circuitry. It does no harm to check the residual voltage across the large PSU caps before working on it but I think you'll find them all down at the few hundred millivolt level.

Use solder braid and a large hot iron to remove parts. If its double sided print then its more difficult.
 
Thanks again. I planned to use a so called solder sucker instead of desoldering braid. Would you advice against that? The main PCB of the Alpha 9 is a pretty densely laid-out single layer board. But I don’t think there’s a big risk of unintendedly shorting anything..
 
I always prefer braid tbh. I've just today removed and refitted some parts on an old player using braid and a large (probably 5mm) tip.
 

Attachments

  • Capture.PNG
    Capture.PNG
    730 KB · Views: 176
I contacted a guy here in the Netherlands who would replace all elcaps (including the DAC PCB) with Panasonic types and possibly some low capacity Vishays for bypasses, for about €130. He also suggested to use ultracaps (aka supercaps), which would open up the sound and give more detail and better dynamics, for an extra €80.

I’m seriously considering his offer, but I’m really not sure about the ultracaps option, exactly because it may (?) impact the sound. And seemingly more detail and dynamics doesn’t necessarily improve a CD player.. Anyone familiar with the use of ultracaps?
 
I've always known supercaps as the battery backup variety i.e. a mega capacitance value such as 0.2F that can be used for keeping volatile memory powered. Surely he can't mean those.

So €130 or €210, its a lot of money for something you could do yourself for a fraction of that.

(I'm afraid I'm not a believer in exotic parts... if they are good quality commercial types and make the grade then that is good enough for me)
 
I've always known supercaps as the battery backup variety i.e. a mega capacitance value such as 0.2F that can be used for keeping volatile memory powered. Surely he can't mean those.
Yes. I’ve asked him what exactly are the ultracaps, but he hasn’t replied yet. He might have something like this in mind:

An externally hosted image should be here but it was not working when we last tested it.

See: This link
 
Ah, that's something absolutely totally different and quite complex. The theory is good, whether it can work well in practice is another matter.

Imagine the caps are like batteries. It works by running the audio circuit off one 'battery', then as that runs down it switches the audio circuit to be fed from another 'battery' that is charged up. During that time the first is quickly charged back up again and then as the second runs flat, it switches back to first and so on, but all that happens fairly quickly.

I'm saying nothing 🙂 its your money 😉
 
Is there a Mouser/Farnell/DigiKey reference for this ultracap ? I know some of it but are huge most of the time not like this little blue one of the photograph ! Notice this amp is not benchmarked as the best of the world despite the "revolutionary" ultracaps powersupply !

low voltage but very low esr permitts to assembly its in serie for the acurate voltage maybe ! A datasheet would be usefull !

Take care about a total recaping : putting brand new caps every where is not a gurantee of good sound, it's the opposite in fact ! Re read what is said about esr and output of some regs ! Or if dig on it : ask for the same original parts at the same positions !

I believe it's a waste of money too ! Trying yourself the simple tip I gave you above will let you to learn more and come back easily if not good ! Soldering in // caps is not dangerous and you certainly do not need to replace all the caps of the pcb but a subjective need thinking yourself all the caps have poor specs because too much old, etc, etc ! For the same price of the providential guy I would buy a dac at HifimeDIY, etc : you should have a better result for less money !

Again if you are worried about security you could just need to secure the 3 caps after the bridges (C36/44/45) which see the pulsed current (smoothing caps : this one is the most stressed part of a supply !) : Just this one are a weak point.... But I have a Pioneer amp from the end of 70s in a perfect working condition with the original parts (caps) ! And it sounds good 🙂
 
Last edited:
Hi Mooly,

I believe this is not the same caps Abtr are talking about ! Those ultracaps are not Supercaps for saving ROM operation in computers....or at least not the same technology than the cheap ones ! The 'ultracaps' are very expensive, very low esr, and huge case size! The only thing in common is the huge capacitance iirc ! But I never saw so little "ultracaps" !
Sort of : http://uk.farnell.com/vishay/mal210...9172?ost=1F+capacitor&categoryId=700000005418 ; notice the esr is not M ohms... but ohms ! So not so low as some polymer of 560 uF !

This battery saving caps you link as you know it very well already has a very high ESR ! So not suitable for hifi (but maybe very low voltage smoothing caps ??? but is it solid ? My idea is a smmothing cap is better to tame the pulsed current with a bad esr (so high esr). Then a resistor for a Pi where the last cap is a real reservoir cap for the load with a very low esr for transcient ! This first cap should have only one quality : long life and huge C° temperature ratting : 105° and more)

Before with some Philips CD player, I putted it after a 78005 reg as an inerty Wheel to lower the impedance curve and have a sort of ripple isolation. on the feet of the dac chips I putted a polymer this time with low esr and big enough capacitance to allow
the dac chip not to have the need of the supercap for instant demands !
After a real battery , I used to solder its in serie 1F/5V x 3 to have 0.33 F and 15V for 2604 oap 😉 ... 10 uF MKps at the feet of the oap !

It was 15 years ago, simple tweaks, not knowing what a good supply was like the actual shunt or OEM fast transcient, oscillation stable at output, OEM reg chips !
 
Last edited:
I wouldn't put Panasonics. They have become a panacea in DIY for their affordable price and wide choice of capacitances, voltage ratings and sizes. But chances are they will begin to leak DC and lose capacity sooner than your factory-fitted components. For the same or slightly higher price, I would advise you go with Nichicons.