Asynchronous I2S FIFO project, an ultimate weapon to fight the jitter

Jump to Latest
A

A123

Disabled Account
Joined 2020
kunlun121 said:
I emailed sales at Skeleton’s tech asking them by when the SCI3400 would become available at the major wholesalers. This morning I received a reply that currently they are manufacturing these caps solely for large direct orders from industrial customers. He wrote that they “can't specify yet when we make it available for smaller quantities through wholesalers.” So nothing new there unfortunately.

I couldn’t help myself from asking what the MOQ would be. The answer probably won’t be useful to aspiring UcPure users, but who knows.
Click to expand...

kunlun121 said:
Latest news from Skeleton is that they can't do small quantities at this point. So that's a no to a direct order.
Click to expand...

Cheers for making a attempt directly. Hopefully the threaded versions start populating on some distribution channels (primarily Mouser). But since they have demanding corporations and governments, the chances may be quite low.
 
A

A123

Disabled Account
Joined 2020
kunlun121 said:
I'm not sure what "too cutting edge" means. The SCI3400 has slightly better specs, but nothing shocking it seems. Ian's UcPure manual says: "Other 3000F/2.7V or higher capacitance/voltage ultracapacitors are also good for UcPure." That would imply that more "cutting edge" caps should work, but Ian would know better. Perhaps he'd like to comment.

Of course there are the Tecate caps that you mentioned, and the Eaton ones. Both look like reputable companies selling quality products. Personally, I am a sucker for that "made in Germany" thing but would not feel bad about using either of these.
Click to expand...

I just meant the latest and greatest of their model, the 3400. The 3000 may have been out on the market longer. The 3400 may not even be released yet.

It's too bad, it would be my #1 choice too as German and European engineering is really appealing. I can't imagine how fast the Graphene will charge with the UcPure. If it takes an hour to fully charge a non-Graphene Ultra Capacitor, maybe a few minutes charge with Graphene? That would make a pretty robust solution.

I'm a good year behind in these Ian projects, so I may skip out on the UcPure as I just realised you need a third-party balancing board after skimming the manual. I just want to keep things simple for now and slowly learn. So I will try to hold out for Skeleton Technologies in late 2022. I have enough on my plate ATM to keep me occupied.
 
Last edited:
A

A123

Disabled Account
Joined 2020
iancanada said:
Conclusions

5. I could be wrong, but my own experiences have shown that the sound quality of a OCXO with a multiplier is not as good as OCXO at the same finial frequency without a multiplier. I would suggest using the OCXO without multipliers as possible.

Ian
Click to expand...

Thanks. Since I run exclusively Redbook on the RPi, the 5 MHz is all I will ever need along with the SinePi.
 
A

Arnolddew

Member
Joined 2018
PinoNL said:
I was thinking about that as well. Hopefully it’s possible to use that kind of XO’s
Click to expand...

It's possible afaik. But i also wanted to know if the 3.3v regulators to the clocks can get bypased so only the clocks get 5v. Tentlabs seems to preform better on 5v. But maybe there are other sugestions. But not a lot people on this topic seems te be interested in this option.
 
A

A123

Disabled Account
Joined 2020
Hi Ian,

I noticed LapTech has an OCXO section. Because I moved on from MagicXtal, looking for alternatives for future use.

The OXO214 - 1000 may have worked out, but unfortunately they only have one freq and is 12V. They have under-performing 3.3V or 5V DIL, so will have to pass until they release an "optimised" version.

index%20pic%20OXO214-1000%20500x329x72.jpg


https://www.laptech.com/pdfs/OXO214_1000_007.pdf

Anyways, I stumbled upon another product that may look interesting for SinePi pairing:

ocxo4.jpg


https://www.laptech.com/pdfs/OCXO_series_OS364-13_generic_data_sheet.pdf
 

Attachments

  • rfx.jpg
    rfx.jpg
    91.3 KB · Views: 492
  • rfx1.png
    rfx1.png
    77.2 KB · Views: 499
  • rfx2.png
    rfx2.png
    127.4 KB · Views: 485
A

Arnolddew

Member
Joined 2018
Arnolddew said:
Already have seen that manual. Thought it was for Q1 alone. R58 and R59 needs to be bypased. But there is a max. voltage warning for 3.3 volt. I can ignore that for Q2?
Click to expand...


"You can solder a 0R 0805 jumper resistor or just a short piece of wire at positions R58, R59 which are located at bottom side of the FiFoPi. CRITICALLY IMPORTANT NOTE: After you short R58 and R59, the ABSOLUTE MAXIMUM VOLTAGE you can connect to J5 is 3.6V!!!! Any higher voltage means you will likely damage your FifoPi and clocks."

Ian, when you are back, can you for sure confirm that i can connect 5v after bypassing R58 and R59 with the Q2 J5 (assiming the clocks can handle 5v)?
 
A

A123

Disabled Account
Joined 2020
dddac said:
Hi A123, I looked at this as well. Point is that the phasenoise round 10Hz is significantly worse than the Andrea clocks. 115 vs 150.
Click to expand...

Ouch. Okay I'll keep that in mind when checking the clock specs. I was just focused on the SMA connection now that the SinePi is soon to be released, but for that the Neutron Star or Andrea's is clearly the better way to go.

I'll stop getting sidetracked with clocks and just be happy with Andrea's clock.
 
S

SimonJ

Member
Joined 2013
A123 said:
The SCI3400 is too cutting edge. Can you try the SCL3000? That's more equivalent to the Maxwells.

Another alternative? Tecate Group TPLH-2R7/3000SL60X138




TPLH-2R7/3000SL60X138:

ESR DC (mΩ): 0.26
ESR AC (mΩ): 0.17

Capacitance: 3000F
Surge voltage: 2.85V
Rated voltage: 2.7V

Tecate Group - TPLH-2R7/3000SL60X138 Data Sheet

Digikey has stock @$57 each

Tecate Group - About Tecate Group
Click to expand...


Thanks for this data and I now have some of these. All I need now is the group buy to start to get the UcPure ;)
 
You must log in or register to reply here.
Top Bottom