[randytsuch]

Hmm, pretty interesting that there is only a small difference between the SC and Newava transformers.  But, could be that the SPDIF receiver side (DAC) is the limiting factor, and is keeping you from hearing more of a difference.

I am currently working on my SB2 project again, putting it in a bigger case, with a built in Felicia, and a huge, way overkill transformer, with at least a few added regulators.

There is an interesting thread over here
http://www.diyhifi.org/forums/viewtopic.php?f=2&t=699
started by John Swensen, about jitter problems he was having.  He uses a SB2/3 (don't remember which), fed by a tent clock as a master clock, and brings I2S out of the SB, to feed into a FPGA, then DAC.  The FPGA changes it from right to left justified (or maybe vice versa), so the DAC can read the data correctly.
John was saying he was getting a lot of jitter, which surprised me, seems that this approach would eliminate most jitter.  The FPGA appeared to be the source of most of the jitter.

I plan to feed to  I2S to my beta version of the twisted pair DAC, and see if that sounds better than the SB2 DAC.  I am going to use either a tent clock powered by a tent shunt regulator, or a Hagclock, have not decided which way to go yet.

I eventually want to go balanced, which is why I am playing with the TP dac.

Slightly off topic, but I was looking at how they do things on the SB2, and I found some things I have not seen on any forums.

I think most people have seen this picture somewhere


What I did not know, is that on a SB2, the switcher and the multiplier that generate teh 11VAC and 55VDC are on the display card.  They feed 5VDC through the ribbon cable to the display card, then sned back about 15 VDC through the ribbon cable.  From there, it goes to around a 100 ohm (I think) resistor, and the "Big Ass" cap.  This powers the two regs that generate 9VDC and 5VDC.

Since I am not using the 9VDC, I will just generate a 5VDC for the DAC, then generate a seperate 5VDC to use to power the display card.  I want to try to keep noise from the display card from getting back into the SB.

Randy

[jhm731]

Hmm, pretty interesting that there is only a small difference between the SC and Newava transformers.  But, could be that the SPDIF receiver side (DAC) is the limiting factor, and is keeping you from hearing more of a difference.

I don't use a DAC.

The digital inputs on my RCS have been optimized for SCs.

Both the SB3 outputs sound the same when I run them through an upgraded Genesis Digital Lens.

I think they sound the same, because Pat optimized the circuit for each transformer and I'm using his cable which reduces reflections.

PS- If you want to eliminate noise from the display card, you should put your money & effort into the new Duet.

[tanchiro58]

PS- If you want to eliminate noise from the display card, you should put your money & effort into the new Duet.

I am sure every SB2 and SB3 owner knows this already since you have an option to shut down the display card when listening your music. I myself noticed the reduction of noise to my SB2 with no visible display card.

The Duet has more simple circuit that would help to have less noise too.

[crooner]

As far as I know, you can dim out the VFD completely on the SB2/SB3, but the actual display is still active. You can see the faint red glow of the filament in a dark room. It also stays warm to the touch.

I managed to isolate the display to a certain degree on my modded SB2, by shielding it with an aluminum panel and ERS sheets. This after using an extended flat cable for the display PCB. Seems to work just fine.

I agree the Duet looks pretty interesting. I'm getting the new "controller" for sure!

[Nuuk]

Apart from trying to buy a few metres of RG-179 (I don't want to pay 90 pounds for a reel) I am still scratching my head over what I actually have to do.

After reading and re-reading the posts, here are two versions of what I think is shown in the picture on page one of this thread. I would guess the top version is correct, except for the statement

You can not see the 0.1 uF film cap, used for the DC block. It is under the transformer. One end is soldered into a ground via, the other to the transformer.

which would suggest the lower version is? 



I am currently using a traffo soldered to the pads of the inductors that were removed, and a BNC socket. The traffo is Murata 76601/3c with the specs:

    * TRANSFORMER, PULSE 1:1
    * Capacitance:12pF
    * Diameter, pin:0.56mm
    * ET constant:5.5Vµs
    * Flammability rating:UL94V-0
    * Inductance, primary:219µH
    * Length / Height, external:6.35mm
    * Ratio, turns:1:1
    * Resistance, winding max:0.5R
    * Voltage, proof pk:500V
    * Inductance, leakage:0.25µH

I am fairly sure this can be bettered (as Pat has suggested of course) as the Transporter sounds norticably better when feeding the same DAC!

[randytsuch]

Nuuk
The top picture shows a DC blocking cap, in the bottom picture, the cap will filter the signal, don't think you want to do it.

Pat never actually shared what he did, for the final circuit.  I have learned enough to know that to do it correctly, it takes more that just a couple resistors.  For a SPDIF transmitter, you need to have an output impedance of 75 ohms, and also generate the correct voltage (I don't remember the voltage off the top of my head).  There was a long thread at diyaudio, where a guy was figuring out how to do this, but I could not see how he came up with his final R values for the circuit.

Others,
A few days ago, I posted a link to a pic of the internals of the new Duet box.  Not sure if I like that design better or not, although it's simplicity does have an advantage.  And, I know lots of people who are very happy with modded SB2's or 3's, so for now that is how I am going to proceed. 

JHM,
I am not trying to knock your setup, I am sure it sounds good. 
But, what do you mean your RCS was optimized for the SC's?
I am not sure how you optimize an RCS for a transformer.  It should be optimized to receive data over a 75 ohm transmission line.  Sounds basic, but it is not done very often, AFAIK.  Art's pictures show what happens if there is any impedance other than 75 ohms in the SPDIF path.

Randy

[GBB]

As Randy said, the correct schematic is the top one, which I've reproduced here.



I've taken a photograph of the spdif circuit in the Squeezebox so that you can see how this maps to the existing circuitry.  Here's a photograph after I've removed one of the resistors in the circuit and cut the trace going to the output inductors and the SPDIF jack.



C86 is a 0.1uf capacitor and is the 0.1uf capacitor that you need to block DC.  We can just reuse this capacitor.
R66 is a 249 ohm resistor and is in series with the 0.1uf capacitor and the chip outputting the SPDIF signal.
R67, which I've removed, was a 107 ohm resistor and went from R66 to ground.  The SPDIF output was taken from the junction of R66 and R67 and ground.  You can see the trace going to the output inductors and then the SPDIF output going to the right and then up.  I cut this trace so the stray wiring was no longer connected.

The primary of the transformer should be wired between the right hand side of R66 and ground.  I use the round hole just to the right of the R67 bottom pad for ground and hooked a small wire to the correct side of R66.
The secondary of the transformer should be loaded with a resistance to give an effective 75ohm impedance.  This would be 107 ohms if you reuse the 249 ohm resistor.  Of course you can play with different values of resistors, but this is a simple starting point that should work well.

---Gary

[tanchiro58]

Gary,

If I accidentely removed 249 resistor at R66 in my other SB3 could I solder another resistor (not the SMR) from the C86 (0.1 cap) and use the ground of the right hole at R67 to connect to the pulse transformer secondary would that be okay? Thanks in advance.

Regards,
Tan

[art]

The top picture is the correct one, except that the cap is on the "cold" side of the primary winding. It helps to make it larger than 0.1 uF, if you carefully read my posts.

I did not show a diagram, as I thought it was fairly obvious. Ok, I was wrong!

On paper, that Murata part has a chance.

Pat

[GBB]

. .  could I solder another resistor (not the SMR) from the C86 (0.1 cap) a

Tan - of course, you can use any resistor you want.  There's nothing special about the surface mount resistor that was originally in the SB3.

The top picture is the correct one, except that the cap is on the "cold" side of the primary winding.

Pat,
Are you suggesting that it makes a difference whether the cap is on the hot side or the cold side of the primary?  I can't think of a reason why that would be true.

---Gary

[art]

It was just easier to connect one primary pin to the chip, via a resistor, and the other primary pin to ground via a cap. Simple as that.

Pat

[tonyptony]

You can see the trace going to the output inductors and then the SPDIF output going to the right and then up.  I cut this trace so the stray wiring was no longer connected.

Gary, if you have a shot of the part of the board where the inductors are located I'd appreciate it very much if you could post it. Thanks.

[art]

The "inductor" (ferrite bead, actually) is marked L8 on the unit. It lies behind the big 220 uF cap for the +3.3 V supply, right under where the TOSLINK is on the right angle daughter board.

It connects to R67, through 4 cm or so of PCB trace.

Pat

Edit: Typed L8, but screwed up and posted L9. L9 is a zero ohm jumper that provides ground for the RCA jack.

[Nuuk]

A big thanks to Pat and Gary. I have now modified my SB3 and even after a few minutes can say that the output is much closer to the Transporter. I still haven't located any mini coax cable so have had to compromise slightly for now.

I removed the 107 resistor (R67) and soldered it to the bottom of the connector PCB under the output socket (BNC in my case). I removed C86 but left R66 in place as it won't be in circuit. I then soldered a 0.22uF cap to the input pad of C86, with a 300R resistor attached to the other leg. The other leg of the 300R is soldered to the pad of R67.

Here are some pictures and a shot of the transformer as it is located until I get the coax cable. It also shows the inductor locations for tonyptony.

[art]

Not sure we are on the same wavelength. I may need to send you some of that cable, but it looks to me that you are not doing it right. Why is the 100 ohm resistor connected to the coupling cap on one end, and the 249 ohm resistor on the other end?

Here is a picture of a unit that a brave soul sent my way:



The only trace that was cut is the one to pin 14. Pin 14 now goes to +3.3 V through a ferrite bead. (Looks quieter than the last one........sorry, Dan.) All of the 48 kHz parts were removed, to make way for the transformer. Pin 1 goes to ground via.

It may be hard to see, but the 105 ohm resistor sits on top of the coax. Look closely, and you will see it. (Yep, used the one that I scraped off.)

I may post pics of the output and TDR, if I get the time later on. In any case, this is not that hard of a mod to do, assuming that you have the right parts and don't destroy the PCB trying to get the 48 kHz crystal off.

Pat

[Nuuk]

Not sure we are on the same wavelength. I may need to send you some of that cable, but it looks to me that you are not doing it right. Why is the 100 ohm resistor connected to the coupling cap on one end, and the 249 ohm resistor on the other end?

The 107 resistor is the one across the output of the transformer, albeit nearer to the BNC socket (temporarily).

The MF is the 300R (coming off the DC blocking cap). The 249R isn't going anywhere so it is effectively out of circuit.

That all confirms to the circuit I posted but what you show above looks different. Is this what you have done?

[tanchiro58]

I thought the 300R should be connected serially with 0.22uF on the primary site (pin3 of Newava S22083- pin4 to ground). I have done that last night and the sound is very neutral, smooth high and mid and bottom more tighter but not dull. Thanks to Pat, Randy and Gary and Nuuk for the pictures and infos.

What do you mean by 0.22uF should be on "cold side?" So the 300R is also the same side while connecting to 0.22mF. Pictures will be posted tonight.

Edit: I have installed the first diagram in my SB2 and will do to my SB3 tonight (Components are Newava S22083s, two 0.1uF I took out from my MHDT DAC, Allen Bradley 300R (actually measured 280R) and 100R (actual measure 122R). The stock cap comes out directly from the chip is C1. I bypassed the R2, R3, L2 and L5 and connected directly to a new RCA connector (not BNC which is hard to install in the SB2).

Tan

[multibit16]

I thought the 300R should be connected serially with 0.22uF on the primary site (pin3 of Newava S22083- pin4 to ground).

I think thats what most of us thought too as in that first diagram

[art]

That all confirms to the circuit I posted but what you show above looks different. Is this what you have done?

Yep......that is all. The resistor on the secondary may need tweaking, possibly a few ohms higher.

PK, pics I promised:





In this case, I will tweak that resistor to match his cable.

Pat

[tanchiro58]

The resistor on the secondary may need tweaking, possibly a few ohms higher.

Pat,

Would that be fine if I use an Allen Bradley 122R 1W on the secondary site of Newava S22083? My digital cable is a Zu Ash with RCA connectors (SB2) and BNC connectors in SB3. Thanks.

Regards,
Tan