Squeezebox 3 mods thread

[firestorm]

Ron,
you're a mine of information - thanks

not sure exactly what I want to do with it, I've had a sb3 for a while now, still ripping my cds, but the thing is not as satisfying to listen to as my old Arcam Alpha cdp (reclocked). So instead of shredding my sb3, I've decided to leave it as standard and mod the sb2.

My thoughts are :

Power supply
decent local regulation, a la ALWSR or similar

Clocks
Reclock with fleas?

Replace DAC with KE version

Change output opamp, may be change to valve buffer

 

[ron]

Well if your Arcam sounds anything like mine its going to take some beating
The modded SB sounds very good tbh, much better than the stock unit, I actually tried a Flea XO set to 3.3v, not a big improvement really compared to say when I fitted it in my Arcam but I suppose the SB's are already quite good regarding low jitter.
The better caps, psu and output stage made the biggest difference so far, as stock I found the sound to be small, closed in, it reminded me of a mid priced cdp but once you start modding that does change.
Let us know how you get on with your SB2, it will be nice to keep this thread going.
Make sure you have a good iron and take your time, the track is delicate, mine is ROHS so it was a little more tricky to work on, the only good thing was that my SB3 already had the KE grade dac fitted so can't comment on what difference it makes
Its hard going back to messing about with cd's once you've lived with these units a couple of weeks

 

[hacker]

Could you point me to the data sheet please ?

Sure: http://www.tranzistoare.ro/datasheets/400/37286_DS.pdf

Hey Carl I have fitted the THS4032, it seems to be ok, no problems yet.
I won't comment on the sound I'll see what you think when you try it
The feedback caps is still the 22pf inplace of the original 330pf

I can't be bothered reading back to find out what kind of cap you're using for this... what are they? And come on.... let us know what you reckon to the THS4032!! Don't keep us waiting

 

[London Lad]

Thanks Hacker

 

[ron]

Sure: http://www.tranzistoare.ro/datasheets/400/37286_DS.pdf



I can't be bothered reading back to find out what kind of cap you're using for this... what are they? And come on.... let us know what you reckon to the THS4032!! Don't keep us waiting

Sorry, just wanted to try and get you off your arse to fit it Well actually I wanted a second opinion but to me I'd say this THS4032 wipes the floor with the AD8066 IMO, it seems extremely clean and clear, the bass is still slightly lacking right at the bottom but theres so much more details now.
I still think the 10uf BC128 coupling caps may be holding it back.
The caps in the feedback for the op-amp was originally 330pf, I changed them to 22pf, I can't remember the pcb number without taking it to bits but you can easily follow the track to see where they are

I want to try the LM4562 next

 

[hacker]

Sorry, just wanted to try and get you off your arse to fit it

I'm working! Well, I'm supposed to be writing a report but actually reading PFM... dammit!

Well actually I wanted a second opinion but to me I'd say this THS4032 wipes the floor with the AD8066 IMO, it seems extremely clean and clear, the bass is still slightly lacking right at the bottom but theres so much more details now.

Great result I take it both amps are better than the stock JRC..

I still think the 10uf BC128 coupling caps may be holding it back.

Lol, is there any room for SMRs...?

The caps in the feedback for the op-amp was originally 330pf, I changed them to 22pf, I can't remember the pcb number without taking it to bits but you can easily follow the track to see where they are

I'm guessing they were SMD? Replacing those with some nice polystyrenes should be worth a crack. Out of interest, why reduce the feedback cap...? Don't you get any bass suckout?

 

[ron]

I'm working! Well, I'm supposed to be writing a report but actually reading PFM... dammit!



Great result I take it both amps are better than the stock JRC..


Lol, is there any room for SMRs...?



I'm guessing they were SMD? Replacing those with some nice polystyrenes should be worth a crack. Out of interest, why reduce the feedback cap...? Don't you get any bass suckout?

Sorry, shouldn't be disturbing your work

I personally love SMRs but I don't think they would fit in the existing case tbh, I also only have a couple of 2.2uf and 1uf so their values may be too low.
With the op-amps I thought the AD8066 was an improvement over the standard one, things open out, you get more details, its not so boxy sounding. change that 1uf lytic biasing the op-amp too, I fitted a 1uf mks
Regarding the fb caps, well I never liked higher values going across them but that could be just me, you can try just the standard value or at least replace with polystyrenes

 

[hacker]

Meh, reports suck - that's what the day is for. Evenings are for music

I've stuck my meter in the path of the SB's 5V supply and taken some current readings (mine is SB3 using wireless):

At switch-on, it peaks at 980mA
It settles at 920-930mA during use and music playback
When the dimmed screen kicks in, it settles at 820-830mA (playing music)
When it's 'off', it consumes 790-800mA. Thats 4W just idling!
Even with the 'off' screensave set to none, it uses 780mA while off.

All this is good news for me and my LM1086 pre-regged SR linear supply... looks like I won't need that LM0184 after all. I guess I can stick a VBE board before the SR, too After that, I can think about adding more rails for 3.3V, 9V etc... I'll have to get the lid off this thing soon!

 

[ron]

Nice one Carl
Well it should be interesting once we all start reporting our results, first I'm just going to see what I can get out of it whilst its in the standard case then probably stick it in a new case and go mad
Next thing is to fit in a 1:1 Pulse transformer and a 75R BNC for the spdif

 

[hacker]

Nice one Carl
Well it should be interesting once we all start reporting our results, first I'm just going to see what I can get out of it whilst its in the standard case then probably stick it in a new case and go mad

Hehe, looks like we're thinking along the same lines! Although I am kinda tempted to see if I can get an 8-pin preh socket on the back and run power into it like that... of course, SRs and fleas should be right up against the circuits they're powering, so a seperate case is the way to go.

Next thing is to fit in a 1:1 Pulse transformer and a 75R BNC for the spdif

You use an external DAC?

 

[ron]

Hehe, looks like we're thinking along the same lines! Although I am kinda tempted to see if I can get an 8-pin preh socket on the back and run power into it like that... of course, SRs and fleas should be right up against the circuits they're powering, so a seperate case is the way to go.



You use an external DAC?

I don't have much luck with cases so have been putting it off but yes a separate case with all the extra's etc in there is the way to go.
Having the Flea's closer to the circuits is a good idea, some of the tracks are quite long on the pcb.
Could do with four 10uf SMR's, I'd like to try those in the SB3 if it means mounting them outside the case

Yes I have a few dacs, actually trying to get the few bits needed to finish a PCM1794 based one, my current favourite is Pedja's TDA1541 based

 

[S-Man]

Well chaps, I seem to be going in the oposite direction to you all....

After several days of listening to my SB3 with the Jung reg doubts were starting to creep into my mind, basically the sound was boring. The most obvious failing was dynamics, it just had no grunt, slam or excitement, although there was bags of detail and no nasties.

I plugged my CD2 back in and was instantly struck by the dynamics, clarity, tunefulness and sheer music making ability compared to the SB.

Back to the SB, firstly I tried a big transformer and big caps feeding the SR, then my totally OTT Aleph C-L-C supply run off a variac. Bad result.

I then went back to LM340T5. Tunes and rythym returned. I then included my preamp and the whole lot tightened up and gained some drive and dynamics.

I have no idea why adding an extra amp stage and using a technically inferior regulator should sound better, but I know which setup I prefer!

Given that several times I have preferred 3 terminal regs to SRs, I don't see the point in replacing the internal regs with anything else. For what it is, I thing the SB3 is very well designed and it would take a major redesign to really improve on it (SB+?).

 

[ron]

Thats interesting you got the details for your LM340T5 supply like type of caps etc? wouldn't mind knocking one up to compare it against the T-net/ALW SR

Although the type of supply makes a difference with the stock unit I thought doing the few mods on the pcbs makes as much a difference tbh

 

[hacker]

I think we'll find the real differences in PSU upgrades will come when we start giving dedicated psus/regulators to sensitve circuitry such as the DAC's analog section and the output opamp. At the moment we're trying get improvements out of a device that's making all of it's 3.3/5/9V rails from a single 5V rail...

 

[London Lad]

At the risk of repeating myself, you guys really should talk to ALW. The external supply he made for me is really something special and isn't just a super-reg.

 

[hacker]

What's in it?

 

[London Lad]

Cant tell you, I was sworn to secrecy. PM ALW.

 

[ron]

At the risk of repeating myself, you guys really should talk to ALW. The external supply he made for me is really something special and isn't just a super-reg.

I already did weeks ago but not had anything back yet I think he must be busy

 

[hacker]

Andy is a busy chap. Perhaps you'd just like to post your PSU to me so I can, erm... evaluate its sound character... ?

 

[S-Man]

Thats interesting you got the details for your LM340T5 supply like type of caps etc? wouldn't mind knocking one up to compare it against the T-net/ALW SR

It's nothing special. It uses a 30VA split bobbin transformer with 9VAC secondaries in parallel. Schottky diodes (50DQ060 I think), mainly for low Vf. Then (IIRC) 0,47R - 2,200uF - 1R - 2,200uF - LM340T5 - 100uF 63V Elna Cerafine. The SR was connected in parallel iwth the LM340 so that I could swop from one to the other quickly, also to provide a level playing field for comparing the regs.

I must try some of those 4 terminal caps at some point!