advertisement


Naim XPS and CDPS voltage

teddy_pardo

Trade: Teddy Pardo
I’m trying to understand how Naim XPS and CDPS works, and something escapes me.

There are three outputs:

+/- 22 Volt
+/- 10 Volt
2x 15 Volt

Looking inside the CDX it seems that the first two are used for the analogue and the 2x15V is used for the digital and mechanics. So far so good, however if I look at the CDPS or the XPS pictures at acoustica.org it seems like the 2x15V is getting the large capacitors and a large transformer (on the CDPS. the XPS uses one common transformer), while the analogue part gets smaller capacitors and a share a smaller transformer, doesn’t make much sense to me as I would expect that the analogue part would benefit from better supply. What am I missing?

I followed the thread burndy pin layout for XPS2, but got confused since there are contradicting opinions WRT what goes where.
 
IIRC:

+/- 23V analogue
+/- 10V servo/motor
+/+ 15V digital filter / DAC

You've got the voltages for the XPS right, but the CDPS uses a +/- supply for the older Philips chipset, which required +/-.

Voltages for the servo / motor sections of the CDPS are higher than those used in the XPS.

The XPS mk I transformer is a SUPERCAP transformer; the CDPS uses a HICAP transformer + a smaller transformer for the digital supply.

XPS mk II has a dedicated transformer wound for it, which actually generates lower secondary voltages than its predecessor.

Location within the case is everything.
 
Thank you for the info.

There is still something that bugs me. Even in the CDX, when it is auto powered, the 2x15V get the larger capacitors, why do they care so much about the digital, I would assume that there's more value in providing good supply for the analogue?

Do you know which capacitor size is used for each voltage in the XPS?
 
I tracked the supply inside the CDX to understand where the +/-10V is used. My motivation was to know is it is worthwhile to keep the Super Regulators in my home made PSU for the +/-10V.

The answer is that it is used to power the PCM1702 DAC. Looking at the data-sheet it says:

"No real advantage is gained by using separate
analog and digital supplies. It is more important that both
these supplies be as “clean” as possible to reduce coupling
of supply noise to the output. Power supply decoupling
capacitors should be used at each supply pin to maximize
power supply rejection, as shown in CONNECTION DIAGRAM
regardless of how good the supplies are."


There are internal 317/337 regulators that further reduce the supply to +/-5V. I wonder if they don't nullify/reduce the qualities of the supper regulators?

Teddy
 
Teddy,

An interesting bit of research! Have you spotted what supply the clock is fed from. This gets its own regulator in the 555PS. If you havn't seen an internal picture, look here http://home.comcast.net/~rtoolsie2/Naim555PS.JPG

Those huge caps are going to give better LF dynamics. Martin made the comment on the LM1086 thread today that clocks also like a low LF noise PSU, so maybe thats why these digital supplies get the biggest and best PSU.

David
 
bivalve said:
Teddy,

An interesting bit of research! Have you spotted what supply the clock is fed from. This gets its own regulator in the 555PS. If you havn't seen an internal picture, look here http://home.comcast.net/~rtoolsie2/Naim555PS.JPG

Those huge caps are going to give better LF dynamics. Martin made the comment on the LM1086 thread today that clocks also like a low LF noise PSU, so maybe thats why these digital supplies get the biggest and best PSU.

David

It looks like an XPS with larger capacitors and a double price...:mad:

They are lucky that there are people who are willing to pay for it. We all here know that it wouldn't cost more than on tenth of their price to build something much better with super regulators and CAP6s.

I think that the clock is fed from one of the +15V (which is than further regulated to 5V using a local simple LM317). If that's the case it might be worthwhile to add another super reg for one of the +15V (the other one is for the mechanics). I'll check it next time I open the CD.

Teddy
 
We all here know that it wouldn't cost more than on tenth of their price to build something much better with super regulators and CAP6s.

Please please make this PSU and let us know how to build it at that kind of low cost! "Something much better" you said... If this is true, I wonder why we don't sell our replacement PSUs for Naim owners!
 
I have two for sale (based on slightly different concepts) will publish details soon including build instructions for those who want to DIY.
 
Hi Teddy and others,

I think I know why the capacitors on the 'digital' part of the power supply are relatively big. In a digital supply you need quite a bit of current to rapidly charge or discharge the gates of the MOSFETs in the circuit. We talk amperes here, not milli-amps that the output circuit normally deals with :) . In general a quality digital supply wants, no, needs big capacitors.
I wholeheartedly encourage you to build your super power supply. It's fun, you learn some on the cost of hifi gear and it lets your CDs SING!

good luck, and enjoy the music :cool:
Mark
 
Oops, could have been more to the point :rolleyes:

In case a 'digital' supply has a high output impedance OR is noisy (especially HIGH frequency noise), the edges in the digital signal function get blurred. That gives more jitter, which is what you don't want in a digital music apparatus. So there you have it: the reason why you need a low noise, low output impedance power supply for the digital circuits just as much, or actually MORE, than for the analog ones.

greetings
Mark
 
Yes, I concur. Now that I spent time working on the CDX and it's regulators I can tell that there is much current running there, one ampere at the +15V, much more than on the analogue. It's also clear to me that improving the digital part supply improved the sound. Maybe because it isolated it better from the analogue, or maybe because it helped the digital part, don't know.

On the other hand I learned not to look for logic in Naim. Sometimes they do it the wrong way just to fix it when you buy an external PSU. E.g. on the NAC102, when powered by the HiCap, the gain stage is sharing the same rail as the control circuit, while the other rail supplies the buffer boards. This way when you add a NapSc the improvement is more significant. Smart...

Same stands for the CDX. They could have had separate rails just like the XPS on the internal PSU (not much more cost), but then who would have paid so much for an XPS?

Teddy
 
Thanks David. These threads were written long before I new what PFM is... In fact at that time I thought that Naim knows best, and someone has to be really crazy to even open a Naim box to see what's inside... :)

My experience is that having super regulators in the external PSU is just a waste of super regulators. Much better to have them inside the CDX instead of the internal LM317/337 regulators, otherwise the internal regulators just reduce (not to say nullify) the high performance of the SR.

The Flea may also be a good alternative to the SRs as internal regulator, although AFAIK there's only positive voltage version. I'm currently thinking of a way to use it as a negative voltage regulator, doing some slight changes on the PCB (Martin, any thoughts on that?)

Teddy
 
It's not difficult Teddy, but you need to cut quite a few tracks and re-link, and substitute a couple of parts. Still only a tiny current output though! Will try to post a how-to later.
 
martin clark said:
It's not difficult Teddy, but you need to cut quite a few tracks and re-link, and substitute a couple of parts.

That's what I thought :)

martin clark said:
Will try to post a how-to later.

Thanks a lot, that would be great! (I assume you still have PCBs right?)

BTW, I have not received the XO modules yet, still waiting impatiently...
 
Hi teddy i have managed to buy an old cds1 without power supply and i know you have made a couple will you make the circuit schematics available in the future? I can probably manage all of the construction bits but the 19 pin layout is still not clear to me.

regards Luke

sorry i was not around to bid on your diy xps, just missed out i was out shopping.
 
Luke,

On this thread, post #13, David points to two other threads where everything is explained. Note that the ".txt" files have to be converted to ".doc" or ".jpg" in order to read them. Also read my two CDX threads for more info.

If you build your XPS, my recommendation to you is to add a VBE before each317/337 regulator. These VBEs carry high currents and thus should be based on power darlingtons like TIP122/TIP127. My current XPS is built this way, and it sounds better (did some blind tests too). Using VBEs can save you investment in expensive transformers and capacitors too.

With the CDX, the XPS outputs are used this way:

+/- 22V the six analog opamps
+/- 10V the two DACs (both digital and analog supply of the DACs)
+15V digital circuits including clock (and thus important)
+15V Servo and front panel (don't care too much about regulation quality)

Let me know if you need further help.

Teddy
 
Thanks Teddy, if I get into trouble I might tempt you into selling your latest incarnation of the beast on ebay.

Luke
 


advertisement


Back
Top