SB3 Jitter reduction mods

art · « **on:** 6 Mar 2008, 08:30 pm »

Those of you who are following this thread:

http://www.audiocircle.com/index.php?topic=51898.0

Might want to watch this space. Something may be in the works for that. 80% certain, will know more in a day or two when the proto PCBs arrive.

Pat

Geardaddy · « **Reply #1 on:** 6 Mar 2008, 09:41 pm »

Kool Pat. Keep us posted....

sts9fan · « **Reply #2 on:** 6 Mar 2008, 10:01 pm »

I am excited to see what you got. I2S output?

ted_b · « **Reply #3 on:** 6 Mar 2008, 11:36 pm »

Does it use your 16' U-byte SPDIF cable? Interesting reflections/jitter reducing ideas.

art · « **Reply #4 on:** 6 Mar 2008, 11:51 pm »

Anyone with BNCs can use the cable.

Nope, part of it is running the clock from a low-noise regulator, powered by a really cheap wall-wart.

Speaking of which........

Would you guys rather that whomever adopts this idea sell you a $10 wall wart for $30, or that you guys buy your own? (I bet we could get one with some stupid logo on it, and then sell it for $50!)

Probably the former, right?

The rest of the scheme has a reclocker for the SPDIF output, and an improved drive circuit. Attention being paid to really keeping the output Z 75 ohms, and as resistive as possible. Not hard to do, unless you add a transformer. Enough hints.

The trick is too shove all of that onto a PCB that is 0.75" x 1.5", and clears all the other stuff in there already.

Pat

rajacat · « **Reply #5 on:** 7 Mar 2008, 12:02 am »

Is there anything unique or innovative about your U-byte cable other than it being unusually long and employing BNCs? For instance, would a 16' Blue Jeans digital cable with BNCs be equivalent?

-Roy

art · « **Reply #6 on:** 7 Mar 2008, 12:48 am »

Maybe close. I am not fond of the cable that they use. Nothing wrong with it, I just feel that there are other choices. We didn't grab the first coax we could find in stock, and go with it. Actually, it is a pain to acquire. Plus, it needs a custom-made BNC.

As for "unique"......no, don't think so. Wire is wire, when you get down to it.

Innovative..........back in '92 when we first built it, it was. Depends on your perspective as to why it is so long.

Pat

Bill@LakeGeorge · « **Reply #7 on:** 9 Mar 2008, 07:38 pm »

Quote from: art on 6 Mar 2008, 11:51 pm

Anyone with BNCs can use the cable.

Nope, part of it is running the clock from a low-noise regulator, powered by a really cheap wall-wart.

Speaking of which........

Would you guys rather that whomever adopts this idea sell you a $10 wall wart for $30, or that you guys buy your own? (I bet we could get one with some stupid logo on it, and then sell it for $50!)

Probably the former, right?

The rest of the scheme has a reclocker for the SPDIF output, and an improved drive circuit. Attention being paid to really keeping the output Z 75 ohms, and as resistive as possible. Not hard to do, unless you add a transformer. Enough hints.

The trick is too shove all of that onto a PCB that is 0.75" x 1.5", and clears all the other stuff in there already.

Pat

So does this mean I need to send my SB3 back to you?

art · « **Reply #8 on:** 9 Mar 2008, 08:23 pm »

No. If things work out, I'll let you know, and we can think about it then.

Pat

daz_bike · « **Reply #9 on:** 10 Mar 2008, 04:19 am »

Can it be installed in a SB2?

I removed the inductor (L5 on SB2) in the path tracing back from the SPDIF last night.

Nice improvement.

Daz

art · « **Reply #10 on:** 15 Mar 2008, 05:32 pm »

Getting closer to being a reality. Tweaking the output for minimum reflections. Goal is <-30 dB.

FedEx seems to have lost my last shipment from Mouser.......how hard is it to take $5 of parts from one end of DFW to another.

So, maybe it will be done Monday? Tuesday?

Pat

art · « **Reply #11 on:** 19 Mar 2008, 06:26 am »

OK, getting closer to reality.

Output return loss. Goal is <-30 dB, with a coupling transformer. Guess I made it. Increase at lower level is due to amount of iron in the core. Not much I can do about that.

Rho on TDR is around 0.04.

Most units on the market can't reach -20 dB. Even without a transformer.

So, anyway..........the plan is to offer a mod that has an improved drive circuit (as shown above), that has been reclocked. Low-noise reg that powers mod PCB as well as the clock chip on the SB3. Preliminary tests shown noise sidebands down almost 10 dB over the stock configuration.

Still deciding on who will actually offer this mod. Not the business that we usually engage in.

Pat

art · « **Reply #12 on:** 19 Mar 2008, 06:43 am »

Anyone remember these pics of the transformer all by itself?

http://www.audiocircle.com/index.php?topic=41593.msg381942#msg381942

Gee, Mr. RF actually flattened out the rho of the transformer by sticking one into his gizmo PCB.

Wonder how he did that? Sorry, trade secret. Bottom line is: this mod works. Period. Just took some time, hard work, and a little bit of engineering know-how.

Mind you, this is the actual output. Takes into account the BNC connector, cable to the transformer, the active circuit that drives it, etc.

Pat

F-100 · « **Reply #13 on:** 19 Mar 2008, 01:48 pm »

Quote from: art on 19 Mar 2008, 06:26 am

Still deciding on who will actually offer this mod. Not the business that we usually engage in.

Pat

You should rethink about this. I would rather pay extra $$$ for this mod and any other digital mod.

art · « **Reply #14 on:** 19 Mar 2008, 01:59 pm »

Yes, but we are not normally in that business. We, well, mostly me, would rather sell the idea to someone who does this stuff on a routine basis.

We do have some interested parties. My gut feeling is that this will end up being done by someone outside of the normal current crop of modders. Some of them have expressed interest, but if I were a betting man, it will end up being someone relatively, if not totally, unknown to y'all.

Pat

mgalusha · « **Reply #15 on:** 19 Mar 2008, 02:21 pm »

Pat, any plans to offer your board/technology to intrepid DIY types? Granted there are not that many who should be hacking away at SMT parts but for those of us who can...

mike

tonyptony · « **Reply #16 on:** 19 Mar 2008, 03:00 pm »

Quote from: art on 19 Mar 2008, 01:59 pm

Yes, but we are not normally in that business. We, well, mostly me, would rather sell the idea to someone who does this stuff on a routine basis.

We do have some interested parties. My gut feeling is that this will end up being done by someone outside of the normal current crop of modders. Some of them have expressed interest, but if I were a betting man, it will end up being someone relatively, if not totally, unknown to y'all.

Pat

Hmm, I'm interested, Pat. Some questions and comment...

1) Will this mod necessitate the removal of the 12.28MHz crystal? If so that may affect users of some internet radio content.

2) Can you characterize the quantitative improvement of this new design over simply bypassing the L8 inductor in the SB3?

3) While I'm a big fan of the American way, sometimes things like mods tend to be a bit overpriced. In this case, especially since there would be no need for ROI on the intellectual property if this work was sourced out to someone else (it's not their IP!), would you consider making public the cost structure of the mod and a labor estimate for implementation? Frankly I suspect making that information public may help "optimize" the price if offered by a second source.

art · « **Reply #17 on:** 19 Mar 2008, 03:43 pm »

Why take out the crystal???

I had to sand the PCB to fit as it is. I need all the space I can get. Too much stuff for too small of a space.

How to quantify it............well, since I have neither the h/w or s/w to listen to one of these, my opinion could only be based on measurements.

The output Z is flat, and and on the money. The output is reclocked. The oscillator has a much quieter supply. I dunno......how would you quantify it?

Cost.........looking at what some of these guys charge for mods.......i don't think that what they pay us will break their bank. They seem to make a good living on selling modest improvements at not-so-modest prices.

The plan, as I think it will evolve, is such:

In order to sell a new product, that needs a good transport for it to work, you first have to come up with a good transport. So, we are talking to someone with that goal in mind. We may then help to design that product, as well. Since that is what we mainly for a living around here.

Sell to DIYers.........not sure I have the patience for that. Make it public domain, so some weasel in China can get rich making an inferior version: no way. We make our living selling designs, not giving them away.

(Manufacturing amps and the like is usually an outgrowth of a design process, intended for use elsewhere.)

As far as how much the parts cost...........definitely under $50. Time....good question. 2-3 hours to shove one of these into an SB3. If there was more space, you could shave some time off. Connecting the PCB to the existing clock and SPDIF is really tricky. Routing the wires, so that they don't short out, come loose in transit, and are not 6" long is a challenge. So, whoever does this will probably earn their pay.

My gut feeling is that $200 would be a starting point, but the labour time would probably up it to $300. I see mods that have a lot less for a lot more.

Trying to keep my personal biases aside, if other mods cost in that range, something that has good engineering behind it should be worth as much.

More would seem folly, as the unit only costs $300 to start with.

Well, I'm fixin' to play plumber for the rest of the day. Gotta shut the water off to the house. Will be back.......????????

Pat

tonyptony · « **Reply #18 on:** 19 Mar 2008, 04:09 pm »

Quote from: art on 19 Mar 2008, 03:43 pm

Why take out the crystal???

Thanks Pat. Appreciate the pragmatic view. I asked about the crystal only because I recall some of the pulse transformer DIY mods required the removal of that crystal in order to get it to fit, IIRC.

mcullinan · « **Reply #19 on:** 19 Mar 2008, 04:30 pm »

So if I am using an outboard DAC this will improve the sound greatly? Count me in if thats so.
Mike