A Diy Xps...

[Richard Nichola]

...is a scary undertaking. Andrew said in the other forum that he doubted that many people could do as well. I want to examine this for a moment.

The XPS uses a multi secondary transformers and thus multi earths (they are linked in the CDS/CDX). Since a similar transformer will be next to impossible to source in small quantities, a different approach may be better for DIY.

How about using a simple (one or two secondary) transformer driving giving us +/- and earth via a couple of caps. This feeds multiple super regs. These regs remote sense the earth in the CDX/CDS using the XPS earth lines and the extra earth (is there one?) back from the CDX/CDS via the XPS is used to carry the main earth current.

Could it work? Would it sound any good?

 

[Andrew L Weekes]

Andrew said in the other forum that he doubted that many people could do as well.

I said that simply because most people would try to copy what is superficially there, and fail to achieve the same performance.

It would in all likelihood be better than no PSU.

I'd look to the pre-XPS supply (WTF was it called?) - this had a Hicap dual-secondary transformer + a 110 (SNAPS) transformer.

Given information about what the supplies power, it would be realtively simple to analyse the currents and their nature and determine the best course of action.

The improved line rejection of the SR's will allow greater isolation between supplies running from the same rails, so your idea is sound (no pun intended).

Anyone care, in the interests of reasearch to give us the pinouts and functions?

Andy.

 

[Patrick Dixon]

Andy, you mean the CD-PS!

You may be interested to know that at the Bristol show, Doug Graham said that the XPS2 was based around a new transformer. They evaluated transformers from a number of suppliers and 1 was much better than all the rest for the XPS, but didn't make as much (any) difference to the SC.

The new case & transformer may well be the only difference between XPS & XPS2, despite the major sonic improvements reported - shows how much devil is in the details.

 

[AshleyD]

I'm up for a joint venture on this project, I already posted the pinouts of the link plug on the previous PFM forum, but of course that is only half the story.

From memory the link plug distributes three supplies to two pins each, which would suggest that the PSU could be broken down into 6 separate supplies (including +ve and -ve).

More investigation is required within the CDX innards to find out what these separate supplies do, Richard would you like to lend me your CDX so that I don't run the risk of blowing mine up?!!!
Only joking, I have some very good insulated probes that I've learnt to love from previous mistakes. Only one problem here, a leak in my garage roof filled my newish meter with water - but the replacement should arrive soon.

Obviously , Richard if we can collerate results that would raise confidence factors immensly.

Ashley

 

[Andrew L Weekes]

They evaluated transformers from a number of suppliers and 1 was much better than all the rest for the XPS

Well, I guess since they seem to stick to those TO3 can regulators, they'll struggle to do better than this - there's only transformer, rec's and caps left to play with

I reckon Les could do wonders for us on the transformer front, his are much quieter and of lower impedance (for a given VA rating) than conventional toroids.

Of course better regulation is likely to have a big impact too, the requirements for digital and analogue supplies may be slightly different too in some areas, based on my experience.

A good starting point would be simply to find out what circuit areas / chips those supplies feed, that will give the base to build on.

Andy.