PDA

View Full Version : Upgrading my power supply



gandt
2005-06-07, 03:22
I have been surprised by the performance of the SB2 on my hifi system. I'd say that after a lot of careful listening it gives me 95% of what my jolida jd100s (used as a transport) provides with all of that extra convenience. I wasn't sure to what extent a unit like the SB2 would benefit from tweaking until I changed over my all purpose cheapo digital connect for a much better one - the sound was substantially improved. Armed with that I decided to see what else could be done - the obvious answer being the power supply which, though decent enough, can certinaly be improved. To cut a long story short I have asked Russ Andrews to make me a version of their PowerPak II (http://www.russandrews.com/static/PowerPak_II.html) its due thursday so i'll try to post some thoughts then if anybody else is interested in following this experiment.

Graham

(kit is Tact RCS plus 2175 into Dali Helicon speakers)

seanadams
2005-06-07, 14:33
To cut a long story short I have asked Russ Andrews to make me a version of their PowerPak II (http://www.russandrews.com/static/PowerPak_II.html) its due thursday so i'll try to post some thoughts then if anybody else is interested in following this experiment.

That page says 12V. Make sure he configures it for 5VDC and the right connector.

SB2 has over-voltage and polarity protection so it should be pretty hard to damage it by connecting the wrong supply. Caution is in order nonetheless.

gandt
2005-06-07, 14:57
yes not only that but to check the actual voltages I've sent them the actual unit to measure on the testbench. The conversion to 5 volts is apparently quite simple and available on the earlier version (powerpak I) as standard.

Graham

John Stimson
2005-06-07, 16:11
Sean, is the SB2 internally regulated? I assume that it's designed for a 5V input voltage. What range of voltages can the regulators tolerate? Are there components that are powered directly from the power input, unregulated?

seanadams
2005-06-07, 16:39
Sean, is the SB2 internally regulated?

Yes, but there are a variety of internal things expecting 5V, including the regulators, both linears and switchers, both up and down - I don't want to get into ALL the details.

Operating range is about +/- 400mV. Various things stop (safely) below about 4.6, and above 5.4 (or below 0) a protection circuit on the input will trip a resettable fuse, shutting everything off safely.

gandt
2005-06-08, 02:35
Sean,
Do you think (theoretically even) that the sound on the SB2 could be improved by changing the power supply? I was surprised at how much the digital cable change over made a difference but I'm no engineer and know that such tweaks are sometimes dismissed by those that are engineers. I can only say that in the past my experience has been that such things do make a difference but that the law of diminishing returns kicks in pretty quickly i.e. the first change from the mundane interconnect costing £10 to the decent £100 one is pretty convincing but the move from the £100 to the £300 one is much less apparent. In any event I was happy to risk £80 on the Russ Andrews power supply to find out.


(And congratulations Sean on the Squeezebox 2 after my experiences with the Netgear I was about to abandom the whole music server thing - now I'm listening to long lost tracks and discs from my collection in wholly new and unexpected combinations.)

seanadams
2005-06-08, 07:57
Do you think (theoretically even) that the sound on the SB2 could be improved by changing the power supply?

Sure, the thing is, there are NINE power supplies in the system, depending on how you count. My comment in the other thread was just that changing out the wall-wart isn't a big win. There are internal linear regulators for the output stage that one could swap out that might have a bigger impact.

If you want to improve something, but don't want to hot-rod the internals, then changing the external PS is the thing to do. I don't have measurements to say what you might gain though.

gandt
2005-06-08, 09:58
I hear you. But hey i have to know! Even if the external power supply only represents one element among nine I figure its a good starting point. By the way the Russ Andrews unit has arrived a day early - no time for serious listening yet so I wont say anything prematurely. Its nicely made anyway with an seriously thick mains cable (their yello unit) and a multistrand "woven" conector between the unit and the SB2. I plan to play a bunch of albums simultaneously on the SB2 (Apple lossless) and the Jolida Transport and compare that using both the original and the new power supply. At £80 I hope ill be immune enough to the old "i've spent loads of money it MUST be better" syndrome and be able to say "hmmm no real change" without feeling like a dork if thats how it seems.

Graham

RACliffe
2005-06-08, 12:17
Keep us posted with what you think, Graham.

Andrew L. Weekes
2005-06-08, 15:08
Do you think (theoretically even) that the sound on the SB2 could be improved by changing the power supply?

In my experience (having done it) there's no doubt at all. I do have some reservations over the RA Powerpak's though - they are SMPS, which is part of the 'problem' in my view, with the supplied PSU. I've not measured the RA unit, but would not expect it to come close to a good linear PSU.

I've had a bit of a prod around inside and outside my SB2 (wireless version) and discovered some interesting stuff (like Sean and the guys did a good job on the design ;) ).

For starters the supplied SMPS supply is pretty noisy, compared to what is possible from a well-designed linear PSU: -

http://www.alw.audio.dsl.pipex.com/Downloads/SB2/Noise%20Comparison.gif

BUT, as Sean has pointed out, this external supply has a long way to go before it reaches the circuits inside. Even so, the linear supply has a quite marked effect on some of the internals, for example: -

http://www.alw.audio.dsl.pipex.com/Downloads/SB2/14V_internal_supply.gif

The above shows the effect of the noise out of the 14V internal supply (derived from the display) when fed from the SMPS and the linear supply (red trace is the noise floor of the measurement system). The 14V supply is a lot noisier, especialy at HF, when the SMPS is used.

Interestingly it also changes with the brightness level of the display: -

http://www.alw.audio.dsl.pipex.com/Downloads/SB2/14V_vs_display_brightness.gif

Both of these plots are worse again with the SMPS supply (not shown).

Having said all that, the most critical supplies, the 5V and 9V linear internally regualted supplies are commendably quiet and show good engineering, design and layout around the audio sections: -

http://www.alw.audio.dsl.pipex.com/Downloads/SB2/5V_internal_supply.gif

and

http://www.alw.audio.dsl.pipex.com/Downloads/SB2/9V_internal_supply.gif

These don't tell all in my view or experience though - whilst there's little evidence in most cases of the 14V supply (that feeds the 5V / 9V regs) breaking through to the audio circuits, the reg's will be affected by this noise, and in my view (and actual experience) the effect of a better external supply is clearly audible. Measurement in this relatively simplistic manner doesn't tell the whole story.

To support Sean's previous view though, this is useful: -

http://www.alw.audio.dsl.pipex.com/Downloads/SB2/Internal_5V_from_linear_PSU.gif

This shows how much the quality of the linear PSU degrades once it gets inside the SB2 - due to the large current drain (750mA - 1A) this produces significant noise voltages, developed through simple Ohms law - the current throught the impedances of the wiring, connectors, traces and even the self-resetting fuse internally give rise to an increase in supply modulation (the upper two traces, the higher one being after the fuse)

Again though, it looks worse for the SMPS input: -

http://www.alw.audio.dsl.pipex.com/Downloads/SB2/Internal_5V_from_SMPS_PSU.gif

Note: the 5V @ D1 cathode trace isn't directly comparable on the last plot - i forgot to enable peak hold when taking the measurement - the supply current varies with SB2 activity so peak hold shows the worst case noise effects.

Andy.

P.S. One key, with a good external supply, to getting the best benefits is to get the external PSU as close as possible to the SB2, electrically. This means a very short output cable, since the cable represents a significant impedance otherwise, over which there will be self-noise generated by the current drain of the SB2, the high current bits of which are non-audio related and potentially more harmful, sonically. If you are up to it, you could make a far better performing supply for the squeezebox specifically, in my view than the RA unit, just using basic off-the-shelf 3 terminal adjustable regulators.

seanadams
2005-06-08, 18:42
[QUOTE=Andrew L. Weekes]
http://www.alw.audio.dsl.pipex.com/Downloads/SB2/Internal_5V_from_linear_PSU.gif

Please clarify - in this plot, are all three measurements powered by the same linear supply? I'm guessing light blue was measured right at the regulator?

I'm amazed by the 40+ dB delta between the light blue and dark blue. Is this attributable solely to cable length and connectors?

Fabrice Rossi
2005-06-09, 01:18
Andrew L.Weekes a écrit :
> These don't tell all in my view or experience though - whilst there's
> little evidence in most cases of the 14V supply (that feeds the 5V / 9V
> regs) breaking through to the audio circuits, the reg's will be affected
> by this noise, and in my view (and actual experience) the effect of a
> better external supply is clearly audible in my experience. Measurement
> in this relatively simplistic manner doesn't tell the whole story.

Andrew, do you hear differences with an external DAC or with the
internal DAC? I understand that noise from the PSU can introduce
distortion and other problems in the internal DAC, but I thought the
digital path was quite immune to this kind of problem. Isn't jitter the
main problem for the transport? Could noise from the PSU affect the clock?

Fabrice

gandt
2005-06-09, 01:49
Before this becomes way to complex (for me) let me just tell you what i am hearing with the change to the Russ Andrew switch mode power supply. I'm assuming no sane person would like to read my individual notes on each album (besides it would reveal my peculiar musical tastes) so the summary is this so far:

I said before that the SB2 seemed to be 95% of what i got from my Jolida jd100s used as a transport (mid-end hifi i guess but pretty decent and very solid) The missing 5% roughly amounted to a certain missing energy and a lack of dynamic punch - particularly on tracks that seemed to demand it (try the new Kraftwerk live double)The detail was all there and there was even grace and delicacy but - a certain lack of oomph.

Ooomph is what Ive got back - more energy, more slam - just a sense that there is more current in the system (metaphorically). What i would say though is that a price has been paid - I'm finding the treble a bit overbearing now and will need to tame it by recorrecting on the Tact RCS though ill probably wait a while and let it settle down.

So: early days (and i missed a couple of hours play with last nights nightly having a firmware issue) but I've got my £80 worth in power alone, perhaps a cost of extra treble brashness that I will need to tame. Bottom line - I'm going to start ripping the rest of my record collection. But if there were to be a "no holds barred" linear mode, full on audiophile SB2 (hey the SB2 signature edition maybe...) for (lets say) £499 I suspect (on uncompressed or lossless files) it would lay to waste transports costing thousands of pounds.

Andrew L. Weekes
2005-06-09, 02:39
I'm amazed by the 40+ dB delta between the light blue and dark blue. Is this attributable solely to cable length and connectors?

It's not so surprising when you work out the maths, they are all of the same supply, a linear regulated 5V feed to the SB2: -

Those plots are calibrated, so if you take the worst case scenario of a change from around -145dBV to -80dBV, from light blue (as you say, measured at the regulator) to the top magenta trace, that represents only around 1mV of noise, give or take a few nV.

For ease of maths, make the current consumption 1A (it's actually lower than this) and using good old Ohms law, we can see than the impedance that represents is: -

R=V/I = 1e-3 / 1 = 1mOhm, not a lot, really!

Remember these are sensitive measurements, that reveal tiny voltages and as alluded to before, they don't necessarily correlate directly with sound, that depends more upon the nature of the noise and what it's correlated with and what the noisly supply feeds.

They also exclude all the stuff above 90kHz or so, that I can't measure ;)

Andy.

Andrew L. Weekes
2005-06-09, 09:46
Andrew, do you hear differences with an external DAC or with the internal DAC?

Internal DAC - I don't have an external one at present.

What i would say though is that a price has been paid - I'm finding the treble a bit overbearing now and will need to tame it by recorrecting on the Tact RCS though ill probably wait a while and let it settle down.

Thanks for the report - I found that the most obvious change from SMPS to linear supply was a removal of a treble harshness and lack of clarity that actually bothers me when using the supplied PSU.

The linear one really helped here, I wonder if the RA unit is suffering a similar problem?

Andy.

seanadams
2005-06-09, 11:24
I'm amazed by the 40+ dB delta between the light blue and dark blue. Is this attributable solely to cable length and connectors?

It's not so surprising when you work out the maths, they are all of the same supply, a linear regulated 5V feed to the SB2: -




How about running a separate sense line back to the regulator? I suppose if the supply were really close, like right under the SB2 with a 3" cable, it wouldn't matter so much...

Anyway I'm still more interested in the internal supplies. :)

Triode
2005-06-09, 11:37
Andy,

Interesting stuff..

Any chance of a measurement of the 3.3V digital line at the 74hc04. [Next to the spdif output]

Adrian

seanadams
2005-06-09, 11:43
Andy,

Interesting stuff..

Any chance of a measurement of the 3.3V digital line at the 74hc04. [Next to the spdif output]

Adrian

Good question - this is, by the way, a pretty easy place to put a dedicated regulator (lift pin, add decoupling cap) if you wanted to.

NealG
2005-06-09, 14:27
Andy, that is a really insightful post, very interesting.


If you are up to it, you could make a far better performing supply for the squeezebox specifically, in my view than the RA unit, just using basic off-the-shelf 3 terminal adjustable regulators.

Which is exactly what I did if you remember for my SB1, re-housed the SB and placed a simple 3 terminal regulator close to the SB.

I powered an SB2 from the same supply and noted a sonic improvement as well, however, not as significant as the SB1 but noticeable all the same. I think it would sound even better with one of your more sophisticated designs

I look forward to hearing more about your experiments, keep it up!

topster
2005-06-09, 14:33
In my experience (having done it) there's no doubt at all. I do have some reservations over the RA Powerpak's though - they are SMPS, which is part of the 'problem' in my view, with the supplied PSU. I've not measured the RA unit, but would not expect it to come close to a good linear PSU.

Hi Andrew, that's all fine and dandy, but where did you get or how did you build the linear PSU?

Inquiring minds who want to improve the sound of their SB2's want to know.

Thanks!

Robin Bowes
2005-06-09, 14:56
topster wrote:
> Andrew L. Weekes Wrote:
>
>>In my experience (having done it) there's no doubt at all. I do have
>>some reservations over the RA Powerpak's though - they are SMPS, which
>>is part of the 'problem' in my view, with the supplied PSU. I've not
>>measured the RA unit, but would not expect it to come close to a good
>>linear PSU.
>>
>
> Hi Andrew, that's all fine and dandy, but where did you get or how did
> you build the linear PSU?
> http://www.alw.audio.dsl.pipex.com/potentially_positive.htm
> Inquiring minds who want to improve the sound of their SB2's want to
> know.
>

http://www.alw.audio.dsl.pipex.com/potentially_positive.htm

R.

Andrew L. Weekes
2005-06-09, 15:21
Any chance of a measurement of the 3.3V digital line at the 74hc04. [Next to the spdif output]

I'll try and do this at the weekend - I kinda missed this one when poking around.

Andy.

topster
2005-06-09, 19:38
http://www.alw.audio.dsl.pipex.com/potentially_positive.htm

R.Thank you!

Stuart Mason
2005-06-09, 20:38
Andy, that is a really insightful post, very interesting.



Which is exactly what I did if you remember for my SB1, re-housed the SB and placed a simple 3 terminal regulator close to the SB.

I powered an SB2 from the same supply and noted a sonic improvement as well, however, not as significant as the SB1 but noticeable all the same. I think it would sound even better with one of your more sophisticated designs

I look forward to hearing more about your experiments, keep it up!

Hi Neal,

Being a lazy sod, I still haven't built the PS you sent me the design for some time ago in PFM land. I don't recall if you were using an external DAC with your SB1 or not. If you were, did the improved PS have an impact on the SB1's performance via an external DAC?

Thanks,

Stuart.

NealG
2005-06-10, 05:10
Hi Stuart,

Yes, I started with an old MF X-ACT which was dire TBH. I had to extensively re-work the internal PSU and opamp cct, I managed to get a very good performance IMHO but then I went and bagged a MF A3.24 off Ebay and have stuck with it ever since.

With the SB1 the PSU is very influential it's affects can heard with the DAC, in fact it's more noticeable the better the DAC IMHO.

I have not tried the same as yet with SB2, only compared my simple LT based PSU against the supplied OEM unit.

Also my SB1 is re-clocked so I don't have a relevant base to compare SB1 vs SB2 overall performance. All I can say is the SB1 with the extensive mods betters my previous setup of Naim CD5 and Flatcap II.

The SB2 with simple linear PSU and using internal DAC does not IMHO but it is ever so close.

Fifer
2005-06-10, 10:54
Being a lazy sod, I still haven't built the PS you sent me the design for some time ago in PFM land.

Also my SB1 is re-clocked so I don't have a relevant base to compare SB1 vs SB2 overall performance. All I can say is the SB1 with the extensive mods betters my previous setup of Naim CD5 and Flatcap II.
Neal, are the details of the PSU and (especially) your extensive SB1 mods detailed anywhere on-line? I'm very curious.

NealG
2005-06-11, 01:37
Neal, are the details of the PSU and (especially) your extensive SB1 mods detailed anywhere on-line? I'm very curious.

The PSU is just a data sheet clone of a 3 terminal LT regulator. I'll see if I can dig it out and send it to you, the only deviation is to use some specific caps instead of the datasheet recommended.

Some pics of my crude SB re-housing are here:

http://www.pinkfishmedia.net/forum/showthread.php?s=&threadid=10685&perpage=15&highlight=&pagenumber=3

Personally I blame Mike Hanson for all of this. ;)

Mike Hanson
2005-06-11, 02:48
Personally I blame Mike Hanson for all of this. ;)
Hey, sorry Neal! :) The ironic thing that I'm still using the original wall warts with mine. I'm debating whether I need Andy's Super Regulators, considering I'm using them only as transports. The alternative is my own pre-regulator design, which isn't "great", but is certainly much better than the SMPS wallwart.

-=> Mike Hanson <=-

Fifer
2005-06-15, 13:14
Thanks for the link Neal. Looks interesting. I'm just back from a long weekend away so I'll have a good read through it tomorrow.

WSLam
2005-06-18, 12:51
Are there actually any readily available PSU that one can buy to upgrade?

Since it's DC, maybe make it battery run?

Mike Hanson
2005-06-19, 03:44
Are there actually any readily available PSU that one can buy to upgrade?
I'm not aware of one, although that doesn't mean that one doesn't exist. The problem is that making a great power supply for audio reaches into the realm of the audiophile, wherein most components cost $$$'s. For example, Naim's Super-Cap power supply sells new at close to $5K, IIRC. The standard explanation for this is that it takes lots of tinkering and research to get things right. In the DIY world, however, the parts are cheap, and the research time is your own (hobby). This is why you'll find more cheap DIY supplies than pre-made ones.


Since it's DC, maybe make it battery run?
That would work, if you can find a 5 volt battery than can provide peaks of 1.5 amps, and is rechargeable. Car batteries can provide lots of current and are rechargeable, but provide 12 volts. I haven't done any research, but I'm not familiar with a 5 volt variant that does the same.

-=> Mike Hanson <=-

sugarmonster
2005-06-20, 16:23
Probably a stupid question, but I seem to remember that the Naim I-supply is their "budget" PSU and has the option of different cables for different devices - they do one for an iPod IIRC(!)

Any chance of using this on an SB? Would it work, or would it just be no better than the standard PSU?

Marc.