The Clock Reviews (part 4)
What’s on the menu today? A short meal this time, but hopefully as interesting as short 🙂
- “Jürgens” Clock – with thanks to Mamal from Berlin
- Listening to the possible impact of power supply to the Andrea Clock
Mamal from Berlin wrote me an email after reading the earlier posts and offered to send a DIL14 miniaturized version of Andreas first versions of the Driscoll oscillator. The whole thing is on a DIL14 size. It looks plug and play, but you need 12 Volt input, so you cannot just drop into a FiFoPi. Nevertheless this is very close to plug and play…
It is an 11.289 MHz version. He also has a 5,6 MHz and prefers that one. Which is in line with the feedback on the Andrea oscillators that the lower the frequency the better the impact of the clock on perceived sound quality
I have been listening also to this clock extensively and must say the implementation worked out just fine. I could recognize the same sound print as the V1 Andrea clock Stefan sent me. This Jürgen clock is ahead of the Crystek and Accusilicons. No big surprise, it is still behind the new Driscoll 5,6 MHz I have as a reference now. Still big thumbs up for this miniaturization project !
Power supply for the Andrea Clock
Some claims were made already on the DIY Forum and they were not all aligned. Andrea claims the impact is relatively small to nil. Still Andrea recommends batteries to avoid ground loops (Just in case I guess – not every power supply creates a ground loop of course) Some others claim benefit from good and or low noise supplies.
Time for me to check a few and see for myself what happens… from top to bottom and left to right:
- 16,5 Volts on an Ultracap Bank of ~ 60F – This will play for at least 6-8 Minutes before voltage gets too low (under 16,0 volt) I keep the bank charged and disconnect the charger when listening starts
- LM317 standard solution at 16,5 Volt
- DDDAC Power supply at 16,5 Volt
- Worst case as reference: a 2 Euro buck converter at 16,5 Volt – terrible quality (50kHz 100mV ripple)
- A LM3045 regulator on 15 Volt (did not have a 16,5 version…)
- No A123 batteries at hand, this will be later this week and I will update this post
This came in a bit later… I made a battery pack with 5 LiFePo4 batteries with the option to have the Ultracap module in parallel. The A123//UC combination was slowly charged to 16.5 Volt so everything stays nicely in balance and no rush currents are flowing back or forward. See the below listening impressions for this update…
To jump ahead… The impact is indeed very low. If at all, the differences are definitely not the same as between the clocks in the earlier post.
Lets put in in this way, the LM317, DDDAC and LDO regulator all were similar and I could not find any preference.
Switching to the buck converter, OK, now you know not to make life too simple and or cheap. Soundstage gets a tad flatter and pixelized. It is just less engaging. Again small, but hey, why would you do this? Of course not, but it shows how good the Driscoll Board from Andrea is. The influence from the power supply is very low, even when pure crap is used 😉
Connecting the Ultracap Module does something positive. Transients are getting cleaner and more precise. Not much but going back and forward you pick it out and the little voice comes: “yep, that one is a keeper” – By the way, the impression did not change to my ears when the dddac power supply was connected to keep the charge on. So the ultra low impedance of the Ultracap is supposedly doing its work…
Update for the LiFePo4 // UltraCap bank
I listened to both variations, the A123 pack with and without Ultracap bank. And again back to only Ultracap. To be very honest, the differences were relatively small. So small, I would not like to state something like a final ranking here. To avoid misunderstanding, I do prefer the of combination of Ultracap and A123 over the simple LM317 power supply and a little bit over the A123 without the Ultracap. But also here the difference is not large. So the whole thing of battery pack plus yes or no Ultracap might be a personal choice to not do this when there is no space or budget. Even with a “normal” power supply the results of the Andrea Clock are excellent. And the latter ranks undoubtedly Nr. 1 of all clocks I had here.
Having said so… If you have space and budget, it is actually a nice choice. It does give that extra little dot on the “i” so to say… My mantra in listening tests is always: “[=listening….] ….. well, do I like what I hear? yes… so it have to stay”. Hence eventually it do has to stay. Eventually meaning, when I am ready to build an external clock with selectable 44.1 and 48 families and insert up to (selectable or automatic) two doublers in the circuit when the FS of the track requires this. Next “small” upgrade project. Still need to build extra clock and doublers 😉
I also realized a kind of benefit with the setup of battery pack with UC in parallel. First of all, this combination will easily power one clock without new charge for at least a day (2500mA / 70mA ) – This would allow for an automatic charge at night – Secondly, there is no need to take any extra care of charging the UltraCap, as it will always be in balance with A123 pack and slowly charged, still thanks to the A123 always available ….
- Thanks Mamal, your friend Jürgen showed it can be done 🙂
- No REAL need in my opinion for a super duper power supply. But…. If you are building this and you have space for a good Ultracap module, why not, it gives the extra notch.
- Maybe / or even probably A123 Batteries will do the same ! YES it does (see update…)
- The A123 will be tested soon versus the Ultracap Bank – Update will be below then…. (DONE !…)
- The A123 in parallel with an Ultracap is an excellent choice if you have space and just want it top notch
- I best liked an Ultracap and a tickle better when powered by A123 pack
- Oh, there is another clocks on its way to me from the UK and possibly a SinePi…. so we are not done yet 🙂
- If any one has a Neutron 3 clock for me to test ????
to be continued !
12 thoughts on “The Clock Reviews (part 4)”
We did this Ultracap stuff since 2014. It will only sound acceptable, if you ad a regulator and normal Cap behind. Use Belleson or UWB2 and a silmic II e.g. And about 1.5-3V Overhead for the Regulator.
In the end, a realy good psu will easily outperform Ultracap. I was using up to 3.000F maxwell. Dont stop feeding the ultracaps. After compleate loading, use a small 500-800mA psu to keep current flow. If you like to listen direct out of the cap, it will sound boaring.
Thanks for your feedback Sunny,
For me the benefit of Ultracap seems to be the low output impedance over a wide frequency range, but for all at low frequency (like low 1/f noise) – so adding regulators behind it seems to defeat this.
Listening to Ultracap only sounding dull is not my experience. I will do some more experiments based on your feedback. Just give me one hint what you consider as a really good PSU?
You state a good psu wil easily outperform an ultracap but that is a wrong comparison; ultracap is only one component in a psu.
I would assume the noise level and output impedance (esr) of a post regulator to be worse than the direct output from an ultracap. The Belleson regulator seems to ba a low noise design. Can you share measurements of the noise level and output impedance of your setup?
Thank you for his interesting test! I am looking forward to your battery supply test. Please test the batteries also with the ultracapacitors added to the output. I have heard a step up in sound when using ultracapacitors behind lifepo4 batteries. There was less noise end a more open sound presentation with the ultracaps added. My assumption is that the ultracaps filter the chemical noise of the batteries and present the low esr.
This combination has fed my fifopi with v1 Andrea clocks at the time (feeding your dddac of course).
Interesting suggestion Stefan, I will do that. I just came back from a few days on the road and the batteries were at the doorstep. I will charge them today and do some listening tomorrow
Your subjective testing is in line with my listening experience although in my setup the difference with the batteries and ultracapacitors is a bit more obvious in favour of the ultracap+battery set.
How did you power the STS board for the Andrea clock? I think a clean supply on this squarer will also benefit.
I power the STS Boards directly from the FiFoPi with its own on board LT3045 (I use for the tests a Q2)
guess you are right: low impedance over a wide frequency range.
I didn´t make any mesurement, but in my experience the output caps of a powersupply are decisive. I use a big polypropylene-cap + a small polypropylene-cap + a Glimmer-cap of 10 nF. Same on powersupplies as in analog or digital circutries or crossover networks. Your 1543-Dac benefits clearly (PP + Mica parallel to the 100 µF caps at TDA 1543 and 74VHC125). The pp´s clear up the mids. The Mica is the one that gives increase in detail and air between the voices / instruments.
Thanks Jürgen for useful reply. I have been using the glimmers as well in some of my tube amps. For larger PP I use the TuveCaps from Mundorf (still affordable 😉
Nice to see, there are still active DDDAC1543 users!
DDDAC 1543 is still very impressive. Two banks of towers, one with inverted data signal, fed by FiFoPi / Accusilicon, via Pass` Son of Zen to a full-horn-system. Your old dac is easily growing with any upgrade elsewhere.
Nice to hear Jürgen! May be I should re-instate my old DAC (just for fun) and see how it goes with all the new streamer sources….
another PSU mod: snubber (Siemens 100Ω + 100nF) and RF-filter-cap (Wima FKP1 10nF) over the AC-connections of the rectifier bridge. The snubber should allegedly dampen the resonance circuit given by the inductivity of the secondary winding of the transformer in connection with the capacitance of the rectifiers. It has been clearly audible in my poweramp-PSU, increasing detail. In my dac-PSU I did it without comparing while installing the FiFoPi.
For reference please see: https://neurochrome.com/pages/rectification-snubbers