Mods to Naim CDS3
- Thread starter rontoolsie
- Start date
rontoolsie
pfm Member
The inductors are capable of 10A continuous, about 50 times the current capacity of any other choke you will find anywhere of that physical size and Henry value. They are hand machined and are very proprietary in their design. Their very high inductance coupled with very high current rating and very low DC resistance makes them a perfect match with Naim power supplies. Up to this point they are not available at any price to third parties, and have only been used (so far) in Naim power supplies (XPS2, 555PS, 552PS and 500PS). And one Epona SPS supply (hi Fox!).
Well yes it was clear to even get 1H at a decent current level was difficult. Are you going to produce them commercially?
Also on the PSU front do you feel no need at all to replace the 'standard' reg by something more super reg or Teddy like, given that they are much better regulators?
I guess I'm asking, wouldn't it be better to start from scratch?
Also on the PSU front do you feel no need at all to replace the 'standard' reg by something more super reg or Teddy like, given that they are much better regulators?
I guess I'm asking, wouldn't it be better to start from scratch?
rontoolsie
pfm Member
Firstly, I have no commercial or financial interest in these modifications or their custom built components...I can only steer those interested in the right direction.
As far as building a custom PSU or using 'better regs'- you have a valid question. I haven't tried either as the final result I have gotten is so good, I don't feel tempted by the next step. Which was not the case in my DBL crossover experiments, where finally I sold the DBLs and comissioned a no-holds-barred speaker to be built.
Technically I now have a PSU with great current capacity, vanishingly low AC ripple and a very low output impedence with a very acceptable noise spectrum. Subjectivally I have a player that is the best I have ever owned or heard. I am happy enough with the end result to not want to mess with trying out (and paying for) a dozen different PSU topologies. Dont get me wrong, I would love to hear a Teddy XPS next to mine, but cannot see that happening as nobody I know has one.
As far as building a custom PSU or using 'better regs'- you have a valid question. I haven't tried either as the final result I have gotten is so good, I don't feel tempted by the next step. Which was not the case in my DBL crossover experiments, where finally I sold the DBLs and comissioned a no-holds-barred speaker to be built.
Technically I now have a PSU with great current capacity, vanishingly low AC ripple and a very low output impedence with a very acceptable noise spectrum. Subjectivally I have a player that is the best I have ever owned or heard. I am happy enough with the end result to not want to mess with trying out (and paying for) a dozen different PSU topologies. Dont get me wrong, I would love to hear a Teddy XPS next to mine, but cannot see that happening as nobody I know has one.
Would that be the same design that Marc Newman is getting?
rontoolsie
pfm Member
Genau so
Hi Ron, sorry if it seemed like I had been challenging, that wasn't my intention.
The intention was that I was genuinely interested in the things you have done and wonder if you will take them further as you have certainly put the work in, rather like Les did way back and Teddy has done of late.
I'm not at all surprised that the PSU is very good, the fact that you are feeding the 317s with a very good quality signal must give them a very easy life indeed.
I wonder if just a couple of comments might do the job due to the very smooth input to the Regs. Just a simple gyrator? (perhaps just the last part of a Teddy for example).
Look forward to the extreme speakers, quite a system you will have then.
The intention was that I was genuinely interested in the things you have done and wonder if you will take them further as you have certainly put the work in, rather like Les did way back and Teddy has done of late.
I'm not at all surprised that the PSU is very good, the fact that you are feeding the 317s with a very good quality signal must give them a very easy life indeed.
I wonder if just a couple of comments might do the job due to the very smooth input to the Regs. Just a simple gyrator? (perhaps just the last part of a Teddy for example).
Look forward to the extreme speakers, quite a system you will have then.
rontoolsie
pfm Member
Hi Brian/trancera,
No I did not think your reponses were unduly challenging. After all, one of the main goals in the DIY forum is to get less for more. And superficially taking a very expensive power supply and doing lots of very expensive mods to it sound counterintuitive as compared to going the full DIY route. However, I chose this route because of the following.
1. I already had paid for the CDS3/XPS2, and their second hand values are dismally low right now, so selling them to fund a project that has no guarantee is risky.
2. The XPS2 already has the right cosmetics.
3. The XPS2 already has a transformer that is massively overspecified and with the correct secondaries.
4. As you aleady pointed out, the LM317/337 are very capable, provided most of their work is already done for them.
5. There is massive amounts of local regulation with the LM317/337s internal to the CDS3 head unit (a total of 42 of them, as compared to just 6 within the XPS2), so putting in other regulators upstream within the XPS2 may not really be a better choice. Far better instead to give them much less work to do.
6. A gyrator as I understand is just a synthetic inductor, but one with a higher output impedance and lower current carrying abilities. Why not use a mo-fo sized choke instead?
The engineer who built the chokes and did all of the mods did mention to me that for the price of an XPS he could build a far superior PS with 6 orders of filtration prior to the voltage regulators, instead of simply one (in the XPS case, a bank of filter caps, in the modified XPS case, two- the chokes and the filter caps ). This would be point to point wired without circuit boards and have wiring looms composed of much better Cardas Litz cable. You should hear the difference-as I have- that is to be had by replacing less than two inches of the generic hook up wire from the CDS3 PCB to the DIN sockets with this Cardas stuff.....or the effects of using this same wire from the rectifiers to the circuit boards.
I have also considered having the wiring loom replaced with a litz constructed equivalent that would be hard-wired into the CDS3 head unit, getting rid of one of the Burndy sockets/plugs and an additional 30 mOhms or so of unwanted impedence.
But the next mod, due to be done in the next week or two is to replace the 0.1uF cardas capacitor that bypasses the 10uF output coupling cap with a 1.0uF value one. These suckers are HUGE and will require some careful installation, but will then give the ideal ten-to-one main-to-bypass value (instead of the current one hundred to one).
I have listened to the CDS3 head unit with no bypass caps on the output, with some 0.0063 uF ones and some 0.1uF ones, and there are huge steps in each of these, with the larger value giving great subjective improvements lower down in the frequency spectrum as compared to the 0.0063 ones which seemed to tame the ringing top end but did not get rid of the buzzy, cardboardy sounding mid-range, which the 0.1uF did so well.
If you lived in the US, I would be happy to ship you my XPS2 for you to try out yourself. But you are probably still in Scotland.
No I did not think your reponses were unduly challenging. After all, one of the main goals in the DIY forum is to get less for more. And superficially taking a very expensive power supply and doing lots of very expensive mods to it sound counterintuitive as compared to going the full DIY route. However, I chose this route because of the following.
1. I already had paid for the CDS3/XPS2, and their second hand values are dismally low right now, so selling them to fund a project that has no guarantee is risky.
2. The XPS2 already has the right cosmetics.
3. The XPS2 already has a transformer that is massively overspecified and with the correct secondaries.
4. As you aleady pointed out, the LM317/337 are very capable, provided most of their work is already done for them.
5. There is massive amounts of local regulation with the LM317/337s internal to the CDS3 head unit (a total of 42 of them, as compared to just 6 within the XPS2), so putting in other regulators upstream within the XPS2 may not really be a better choice. Far better instead to give them much less work to do.
6. A gyrator as I understand is just a synthetic inductor, but one with a higher output impedance and lower current carrying abilities. Why not use a mo-fo sized choke instead?
The engineer who built the chokes and did all of the mods did mention to me that for the price of an XPS he could build a far superior PS with 6 orders of filtration prior to the voltage regulators, instead of simply one (in the XPS case, a bank of filter caps, in the modified XPS case, two- the chokes and the filter caps ). This would be point to point wired without circuit boards and have wiring looms composed of much better Cardas Litz cable. You should hear the difference-as I have- that is to be had by replacing less than two inches of the generic hook up wire from the CDS3 PCB to the DIN sockets with this Cardas stuff.....or the effects of using this same wire from the rectifiers to the circuit boards.
I have also considered having the wiring loom replaced with a litz constructed equivalent that would be hard-wired into the CDS3 head unit, getting rid of one of the Burndy sockets/plugs and an additional 30 mOhms or so of unwanted impedence.
But the next mod, due to be done in the next week or two is to replace the 0.1uF cardas capacitor that bypasses the 10uF output coupling cap with a 1.0uF value one. These suckers are HUGE and will require some careful installation, but will then give the ideal ten-to-one main-to-bypass value (instead of the current one hundred to one).
I have listened to the CDS3 head unit with no bypass caps on the output, with some 0.0063 uF ones and some 0.1uF ones, and there are huge steps in each of these, with the larger value giving great subjective improvements lower down in the frequency spectrum as compared to the 0.0063 ones which seemed to tame the ringing top end but did not get rid of the buzzy, cardboardy sounding mid-range, which the 0.1uF did so well.
If you lived in the US, I would be happy to ship you my XPS2 for you to try out yourself. But you are probably still in Scotland.
stackowax
pfm Member
Ron,
Can you provide specifics on the Cardas wire (awg size, for example) plus any hints for soldering this stuff. I've tried using 16 awg and 18 awg Cardas hook-up wire but just found it too difficult to work with (in terms of soldering).
rontoolsie
pfm Member
I believe its the 16awg chassis wire. I agree it is a b*tch to work with requiring some decent strippers, a high temperature solder pot (the litz coating is almost indelible) and preferably some Cardas eutectic lead-free solder. But the end results are worth it.
pippo
Pippo
Dear Ron
I had read earlier posts of you on the naim forum and found them always among the most valuable and informative. I read your DYI mod posts here with excitement. Gives me something to think about. What woud an intervention entail (if feasible at all) on a supercap feeding a 252? From what I read the improvements must be significant. The supercap is very similar in principle to the XPS2. Thanks for sharing this stuff with us.
Cheers
Paolo
I had read earlier posts of you on the naim forum and found them always among the most valuable and informative. I read your DYI mod posts here with excitement. Gives me something to think about. What woud an intervention entail (if feasible at all) on a supercap feeding a 252? From what I read the improvements must be significant. The supercap is very similar in principle to the XPS2. Thanks for sharing this stuff with us.
Cheers
Paolo
There is a bunch you can do...Ron is out for a few days. I'm sure he will reply as soon as he can.
Marc
rontoolsie
pfm Member
Hi Paolo,
The supercap would actually be easier to redo than an XPS2....it does not have multiple + and -ve voltage feeds. There is only a +24V supply that feeds the audio parts (in addition to a +15V or so that would feed the front panel of a 252/52. So only a single choke regulator would be needed, in addition to a couple of caps to bypass the main filter caps.
I have not had any supercap worked on..yet, but I did have an Epona SPS power supply redone this way which produced remarkable improvements on the LP12s Prefix.
The supercap would actually be easier to redo than an XPS2....it does not have multiple + and -ve voltage feeds. There is only a +24V supply that feeds the audio parts (in addition to a +15V or so that would feed the front panel of a 252/52. So only a single choke regulator would be needed, in addition to a couple of caps to bypass the main filter caps.
I have not had any supercap worked on..yet, but I did have an Epona SPS power supply redone this way which produced remarkable improvements on the LP12s Prefix.
1Theotherone
Member
It has now been over a month since my post re: the sound of the the modded CDS3/XPS2 and as expected the 'burn in' factor has upped the ante. The listening experience is beyond all expectations. What I find incredible is that regardless what I play the sound continues to be involving, balanced and extremely pleasurable.
I have a box of cds that I was going to take to the s/h store, as I had played them on my CDX2 and CDS3, and could not enjoy the sound, usually due to too much harshness or just lack of involvement. Well, the box is slowly being emptied and the cds are going back on my storage shelves.
Without a shred of doubt the expense of the mods is the best 'bang for buck' going. Priceless.
Congratulations to Ron and his engineer Daryl. Brilliant work !
I have a box of cds that I was going to take to the s/h store, as I had played them on my CDX2 and CDS3, and could not enjoy the sound, usually due to too much harshness or just lack of involvement. Well, the box is slowly being emptied and the cds are going back on my storage shelves.
Without a shred of doubt the expense of the mods is the best 'bang for buck' going. Priceless.
Congratulations to Ron and his engineer Daryl. Brilliant work !
rontoolsie
pfm Member
While modifying Theotherones player, after all the other mods were done, Daryl then went on to doing the bypass capacitors on the CDS3s output coupling capacitors (those blue bead 10uF tants), using the only value Cardas cap he had at hand, 0.0068uF. When we listened to the result it sounded nowhere as good as mine, comparatively having a boxy, slurred and rather tuneless sound at all frequencies, except the top end, which did have the same silky smooth, detailed and extended presentation. My player by contrast had some 0.1uF bypass caps to the output coupling caps. And yes, we did have both players side by side and did not have to resort to fading audio memories.
So a couple of days later when some 0.1 caps were back in stock, out came the .0068s and in went the 0.1s. And OMFG what a difference, with the end result that the player became a sonic clone of what mine did. Which is to say very, very VERY good, and produced the result that made Theotherone so very happy with the degree of improvements as compared to his stock CDS3.
So it occurred to me, if going from .0068uF to 0.1uF as a bypass value was SO much better, what would happen if we then went from 0.1uF to a large a value as would physically fit within the CDS3 head unit. Which turns out to be a capacitor bank of 0.54uF inserted just behind the front panel and hooked up to the tant bypass caps via a 12 inch run of Litz cable.
Words can barely describe the huge increase in palpability, texture, transparency and gestalt that this extra value added over the 0.1uF value. Indeed I convinced one other bold guy who had these mods done to send his unit immediately back from Canada (not Theotherone, who is also from the great white north) for these new tweaks. He got it back two days ago, and has slept only a few hours over the last two nights as he is in a mad frenzy of pulling all his disks out just to see how much more musical insight exists in them. Maybe I could persuade him to add his own insights at some later time, but I don't think he is a contributing member of this forum.
While taking apart the head unit for one final time, we did have a detailed look to see if ANY other thing could be upgraded while we had it in house, but short of building another analog output board from scratch (based on some tasty 6SN7 driver tubes), we had taken this about as far as we could.
Next up in line for similar treatment is my NAT01 tuner head, which already has the 0.1uF bypass caps. There is enough room in there to maybe squeeze in a pair of enormous Cardas 1.0uF ones. We shall see.
So a couple of days later when some 0.1 caps were back in stock, out came the .0068s and in went the 0.1s. And OMFG what a difference, with the end result that the player became a sonic clone of what mine did. Which is to say very, very VERY good, and produced the result that made Theotherone so very happy with the degree of improvements as compared to his stock CDS3.
So it occurred to me, if going from .0068uF to 0.1uF as a bypass value was SO much better, what would happen if we then went from 0.1uF to a large a value as would physically fit within the CDS3 head unit. Which turns out to be a capacitor bank of 0.54uF inserted just behind the front panel and hooked up to the tant bypass caps via a 12 inch run of Litz cable.
Words can barely describe the huge increase in palpability, texture, transparency and gestalt that this extra value added over the 0.1uF value. Indeed I convinced one other bold guy who had these mods done to send his unit immediately back from Canada (not Theotherone, who is also from the great white north) for these new tweaks. He got it back two days ago, and has slept only a few hours over the last two nights as he is in a mad frenzy of pulling all his disks out just to see how much more musical insight exists in them. Maybe I could persuade him to add his own insights at some later time, but I don't think he is a contributing member of this forum.
While taking apart the head unit for one final time, we did have a detailed look to see if ANY other thing could be upgraded while we had it in house, but short of building another analog output board from scratch (based on some tasty 6SN7 driver tubes), we had taken this about as far as we could.
Next up in line for similar treatment is my NAT01 tuner head, which already has the 0.1uF bypass caps. There is enough room in there to maybe squeeze in a pair of enormous Cardas 1.0uF ones. We shall see.
rontoolsie
pfm Member
I have directly compared the Cardas caps directly with Solens, Hovlands, Mundorf Silver/Oil Supreme, Theta Audiocaps etc etc...and they are the ONLY capacitor that seem to be sonically transparent on either white noise or on musical signal. There is some mumbo-jumbo about their use of multiple foils in the contruct of varying sized (and in 'golden ratio proportions) that renders the capacitor free of resonances, or the high purity Litz leads (ditto). Their size is a greater disadvantage as compared to their price. I would have loved to replace the 10uF tants with equivalent values of Cardas, but the largest value they make (4.7uF) is about the size of a hand grenade, and will not fit anywhere inside the CDS3. As compared to the performance upgrade, the price of Cardas caps is relatively trivial.
I am sure that the tant bead caps are perfectly OK when used as buffer caps in a DC power supply, but when they couple AC audio signal, they are really toxic and do not belong in any product costing multiple kilobucks.
I am sure that the tant bead caps are perfectly OK when used as buffer caps in a DC power supply, but when they couple AC audio signal, they are really toxic and do not belong in any product costing multiple kilobucks.