|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
68.223.10.145
In Reply to: RE: cMP - the open source high-end Memory Player posted by cics on December 30, 2007 at 05:42:01
One of those things that seems like a good idea.
Did not have room to mention the plus and minus 12 volts rails, too.
Seems like it would offer that much more isolation from the rest of the computer. Even with strictly linear supplies there are still the onboard regulators. I have no idea if they are switchers or not.
Any reason not to?
Follow Ups:
Sorry to call you by your last name on my last note.
Don't know what I was thinking.
Would you mind connecting a voltmeter to your JULI@'s power supply leads and measuring the voltage when you turn your machine off?
I have checked everything possible and am 99% sure I have no leakage from JULI@'s 5 volts tabs into the MB, yet when I turn off the machine my voltage will drop as if it is being sinked into the MB. I cannot imagine what would cause this.
Not having any problems with the machine coming up tonight. Running fine though I am still needing to install another (higher voltage)transformer to get the voltage right. I have no idea of how it sounds but it runs fine with 4.5 volts. Using my traditional choke input supply the voltage goes all over the place at start up but settles down to a solid reading.
With traditional supplied one must do some experimenting to get the voltages right. Next thing to try is a nine volts secondary. Hope that is the one.
Not sure if your regulated supply will eliminate what I observe at turn off but when the mood strikes I would appreciate you seeing what happens.
If there is not voltage change, maybe you could do a current measurement?
(I thought while I was asking ...)
Thanks,
Rick McInnis
Rick,
I had a memory (they are coming less often as I gain seniority) of doing this measurement about a year ago.
I found it.
Here's what I recorded:
Juli@ 5v PCI
73.7 mA computer off
147.8 mA on computer on in cMP mode
189.2 playing music
Again, I'll confirm these measurements this weekend, but this is what I got a year ago.
Greg in Mississippi
Everything matters!
LOL again.
Rick, calling me 'Stewart' is not a problem and works too... I'm not too picky about things like that.
I can make that measurement, likely this weekend, but might get to it Friday night. In my setup, since I have the 5v fed from a regulated supply, measuring the current will give you the best information, since the voltage will stay at 5v no matter what. But if the current rises with the motherboard off, that may shed some light on what you're seeing. It's easy to do on my setup, just unsolder two wires and insert a meter set to measure current.
Will let you know what I see.
Everything matters!
I am impressed with your note keeping and your files.
Your measurements do point to something going wrong.
I looked for problems extensively last night and could not find anywhere the "voltage" could be leaking out of the board into the MB. (as I said)
Very confusing.
Of course, I need to get it connected and hear it. If it sounds OK I guess it is fine (?).
No need to set all of that up again for measuring. I have confidence in your numbers. You are very kind to offer and this is appreciated.
I am of the opinion that when one takes copious notes this does lessen the importance of memory and the brain is at the ready to work a little less hard. So your good habits are what is to blame for what you think is a function of age. Of course, if we did not take notes we would not have got where we are, so as with everything with every gain there is a loss. One has to choose their compromises and consequences of those decisions . Just like designing an amplifier or speaker. One is either frustrated or satisfied.
I am fifty five though I prefer to think I have celebrated my thirty-eighth eighteenth birthday and when the memory seems to fail I figure it is the same as an HDD getting full and in need of de-fragmentation. Please humor me.
As close as you are I would like to take a drive to hear what you are up to. (how's that for inviting oneself?)
Thanks again for the help and your experiments!
Rick McInnis
Ah, Grasshopper... look very closely! The 3.3v PCI connections come already pre-cut on the Juli@!
Seriously, they aren't connected to the board, so there's nothing to cut. The digital side only uses the 5v and a single regulator to drop that down to 3.3v. Some chips on the digital side use the 5v, most use the regulated 3.3v. That regulator is a not-to-bad linear device, but SQ improves with a better regulator. I've used the Dexa one in the past and have a Paul Hynes one here for the one I'm working on slowly now. Others have had good luck with a battery for the 3.3v and I have one of them here to try also.
As for the +-12v lines, that's one where I think it won't make any difference... those lines are not used in the digital side and just get passed through to the analog card. But it wouldn't hurt to try... run it without cutting them at first, then do that and let us know if it makes a diff.
BTW, GA is not far from MS. You wanna try my linear-hybrid PSU out sometime? Shipping wouldn't be too bad. And as I report, it is actually surprisingly good compared to the fully linear supply, switchers or not!
Later!
Greg in Mississippi
Everything matters!
Stewart,
Thanks for the information.
That would have been much delicate work for nothing. I was looking at the list in Mihaylov's LINEAR COMPUTER SUPPLY article and not the board.
Thanks for the offer. I had used the PICO before, though I was using it with a giant battery instead of a AC based supply. I do remember that it sounded, to me, more different, than truly better. Of course, I was feeding a bit too much power to P4 with an unfettered battery.
I have not got the thing attached to the DAC or the line stage yet, anyway.
Powering JULI@ outside of the computer has been a rather interesting experience. I wonder if I have not completely cut the traces for the 5 volts. (?) Last night at turn on (after making some changes) I had BIOS failure. The system would not accept the recommended BIOS changes. I had to implement about half of them, and left CPU voltage at 0.75 volts to get it to turn on. I will make the rest of the changes in stages and see what happens.
Does any of that sound like a problem due to voltage leakage from the linear supply back into the MB to you?
Thanks again for the offer and all of the good advice.
Rick McInnis
Rick,
It does sound like you are feeding 5v back onto the MoBo... especially if the 5v that you are putting into the Juli@ is higher than that being fed to the MoBo by the SMPS (like 5.1v vs 4.95v).
What I do when I see an anomoly like you saw with the BIOS is to pull out the change (the modified Juli@), see if that causes it to go back to the original condition, and then if it does, troubleshoot the changed item (that darned JulI@).
Hopefully the MoBo is ok.
On cutting the fingers, I always cut them (funny that I've I've done this a number of times!) just where they get full width. I use a new hobby knife blade and put several layers of masking tape on either side of the finger I'm cutting to help keep from cutting the adjacent fingers. I do a single slice across the finger, but go over it several times to make sure it's cut all the way through. Then I test for lack of continuity with my meter.
Doing it as a single slice, I can (and have) done a solder bridge across the slice to restore the finger.
I will try to get pix later today, my regular camera doesn't do the tiny stuff well, but I've found that my cell phone does a pretty good job, so I may be able to capture them.
I'm sure you know that there are 8 +5v fingers to cut, four towards the front of the card and four towards the back.
You probably also have this diagram, but just in case, I use one from here:
http://pinouts.ru/Slots/PCI_pinout.shtml
And let's see if it comes through ok: (EDIT - It didn't, so I took it out. See it on the web link.
On the Pico, let me know if you'd like to try it. I hope to be done with it (except for use as a spare) after this weekend). I think the secret to it sounding good are:
1. Replacing the PicoPSU on-board caps with Black Gate N's (suggestion from Peter Daniel comments on DIYAudio).
2. Adding Black Gate N additional filtering across the 3.3v & 5v (this isn't shown in the pictures, I did it after that shot was snapped).
3. Using great components in the linear supply (oversized Hammond split-bobbin transformers, Jensen 4-pole main filtering caps, Black Gate Std output filters).
4. Using a separate raw DC & regulator for the 12v P4.
I need to try a battery setup too. JackWong suggested that LiPo sound better than Lead-Acid and I have some serious LiPo for my model airplane hobby. But I'd still regulate it to precisely 12v... those digital circuits on the MoBo (and on the Juli@ too, for that matter) don't tolerate large (> 1v) over or under-voltage situations.
Later!
Greg in Mississippi
Everything matters!
wish I had thought of that!
I feel sure your PICO set-up is exceedingly well thought out. I have never been one to have to try everything. I just get confused.
I feel confident the pure linear approach has to be better so I will slwoly creep along until I have it implemented. Nonetheless I do appreciate your generous offer.
Those exploding batteries seem to be intriguing. My philosophy has always been the more exotic and dangerous the device the better it has to sound. I am not trying to be funny. I am thinking of those mercury vapor rectifiers, the KEITH MONKS tonearm with the little mercury baths to eliminate the twisting of the lead wires, running gear without fuses (it does make a BIG difference), you get the idea.
Of course, batteries are controversial. I know you admire Peter Daniel's work as I do and he thinks they are a dead end, along with some other folks whose opinion I take very seriously. I think it all has to do with what the rest of your system is composed of, like everything else, since, as you say, everything matters, and everything matters individually and within the system.
Back to batteries, I do suspect that PD has not used BIG batteries and if you are not using high amp-hour one's you are wasting your time. But they are a pain because of the charging and these LiPo's require being placed in a bag to contain the possibility of an explosion. I do worry I would get tires of that. At least with lead/acid you can use one of those "tender" chargers and not have to worry. My phono stage of the moment uses the battery I bought for the PICO experiment ...
I wonder how much power the regulators require and how much quicker they would deplete the charge especially for the 5 volts rail. I will stick with AC.
Thanks again for the help and the offer,
Rick McInnis
I agree that the pure linear is better... and given the experience and SQ you had when you tried the PicoPSU, I understand you not wanting to try it again. Nottaprob!You are very, very welcome. I'm trying to pay back all the support, both directly and indiretly (through your posts in the various forums) that you gave me when I was setting my cMP up at first.
LOL on the exploding batteries. Yah, IMHO, better sound comes from living-on-the-edge... I've been considering both a DAC output stage and a phono stage that each uses mercury rectifiers... and agree on the cautions with LiPo batteries and the benefits of running without fuses (although I have on good authority that good (read that EXPENSIVE!) circuit breakers can sound about as good as no-fuse.
Another set of living-on-the-edge experiences are high-biased and/or class A circuits... a different DAC output stage I'm looking at is based on a Nelson Pass design and runs the I/V FETS at 50v or more and so hot that they take amplifier-sized heatsinks. Gotta love it!
But I was never convinced on the Keith Monks arm. OTOH, I never played with one...
As for the batteries, I already have LiPos, so it's easy to experiment with them... and having used them in my model airplanes and helicopters for 6-7 years, I have the equipment to safely charge them and am very aware of their dangers and how to use them pretty safely (no fires yet, knock on wood!).
I agree that it's how you do it that makes or breaks a battery versus an AC setup. My power supplies have been similar to PD's (with a couple of diffs that I doubt he'd condone) and that's worked well for me... and that's why I've put a lot of weight on his opinions on this.
But a battery setup is intriguing to me. My biggest caution is that it's a high-maint setup, especially for a fully-linear computer... just witness JackWong's setup! My 2nd biggest caution is that I suspect that it will work best if the entire system is battery... and my super-low-efficiency Eminent Tech speakers require a LOT of power and are not good candidates for battery powered amplification. I have several other cautions, but these are the two key ones.
Finally, on your regulator comment, if you are referring to the ones in the link I sent, yah, they idle high and hot and will suck a lot of power just sitting there. That's the nature of shunt regs. But I suspect that (like class A amps) is part of why they work well.
Later!
Greg in Mississippi
P.S. Picture attached of how I do my PCI finger cuts. Note that the finger that is second from the bottom on that top lobe of the PCI connector is one of the 3.3v ones that is not connected to the Juli@ board at all. There are many other ones like that, not all are the 3.3v ones.
Everything matters!
Edits: 07/29/10
I wish my cuts were as elegant.
On some of them I was able to remove the whole strip but on others, not!
I ran PD's amps on lead acid batteries when I was using them (above 80 hz) and liked the sound through my horns (edgar/tad).
I have since got one of Mr. Pass's J2's and I am very happy with what they do.
I did not mean to imply the Keith Monks sounded all that good, but in my early twenties mind it HAD to since it was using this somewhat dangerous element! I would never use one now though I respect the elegance of the solution the problem is it wasn't a solution!
Shunt reg's and batteries will not do!
I use one of Paul Hynes BIG reg's for my DAC!
I will wait patiently to hear the result of your battery experiments. I suspect, for the computer, I will stick with AC.
THANKS, again,
Rick McInnis
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: