A question on Cap replacement mini DSP

Hi I am upgrading the caps in my mini DSP and wanted to confirm the positive is in the normal configuration, the centre has a + in the middle and I believe the cap removed was Polar given its markings. Also thinking of using a 35V cap, wasn't sure what value this cap is 1000Uf and looked like it might be 10V. Its fed with a 12V supply and has maximum of 15V internally, so thought adding a 35v was a safe option ?



2005-10-05 1:26 am
I recommend using panasonic FM instead of larger silmics :

these caps are likely output filters for switching supplies, so high peak ripple and stable ESR at ultrasonic frequencies is critical here.
adding silmics horizontally with extended lead wires increases the series inductance feeding those caps -- could lead to more switching noise, regulator instability.

also recommend purchasing some standard series SMT caps as well, in the event the ultra-low ESR of the FM series creates switching regulator instability and you need to swap them out to get it running again. Most switching regulators are designed for output caps within a specific ESR range, too low or too high can get you into trouble.
Pay attention to tincanear.

What are you attempting to achieve by these changes? What circuit and what function? Caps have a lot of different parameters, not just value and voltage.

To some other comments, a cap that gets hot enough to need a heat sink has something terribly wrong with the design.
Adding magic sheets to , I assume reduce some magic micro-vibration is another fad without merit.
Be wary of source of your caps. There are a lot of counterfeits.
"Squeeze in a Panasonic". Panasonic makes dozens of electrolytic caps. They are a reputable company. Pick the correct series. There is no "this one is better than that one". Application matters.
Fancy brand names mentioned in glossy page magazines do not effect the actual cap parameters and suitability for use.
Ive experimented a lot with pre amps and CD players over the years and have found changing caps in both Ps and signal can make big difference sonically, not always better. But for a few £s worth a try. The miniDSP is only used to control subs and I wanted to ensure the performance is as good as possible.

I buy caps from HiFi Collective and Farnell

The AC coupling caps in my Bluray player had a significant impact on the sound. I settled on Silmic II there (All 470uf)
AC coupling caps are often too small. 470 sounds a bit big, but not much cost difference.
What parameter does the Silmic cap have that others they make ( reputable company) that make it "audio? Is it their $1 cap sold for $2 because marketing said the "audio" label makes it more valuable? Lower inductance? Lower ESR? DF, TC? What is it in series with? A 1.2K resister and ESR is important? I am not claiming no difference. A place where there likely is, but why? Why that value. Why that series? Is a 10u electro better than a .5u film? What did it replace, a Z5U?

Power supply caps are often too small. Cost constraint. But just swapping caps may lead to other issues, like improper rectifier ratings due to inrush. Easy to calculate. You can monitor ripple. You can monitor for rail collapse. Though a CD player will have regulated supplies and a pretty steady state draw, so I doubt much improvement in actuality. Now, going to the boards, close to the components that draw power, improving bypass caps can make a measurable difference. Not only choice of cap, but isolation of parasitics in the power and ground layout.

Worried about your SUB performance eh? Often overlooked. Be sure your MiniDsp has .0001 % distortion, not .001% when feeding that amp with 2% into a speaker with 4% in a frequency range humans are particularly insensitive to. Won't hurt anything. The placebo effect may tell you they sound better, and well, that's the actual goal, if you believe. Music is what your brain believes and does it really mater how we make it believe? That's the fun part. You only have to please yourself.

I buy caps form Mouser, Digi-Key, Newark and Allied. I do not know the electronic supplies on your side of the pond. I can' afford boutique resellers.
I know Silmics sound better when comparing to stock, Nichicon FG and Silmics, they do have a different dielectric - silk and have copper leads not steel. All I know is they are physically large for the rating and sound better when I have changed the coupling caps

I recognise some measurements help and some simply don't. My subs sound better as a group using more of the higher distoirtion 12" subs rather than asking my very capable 18" that is quite happy at 5Hz to pick up more around 2O Hz. I did they and ended up only reliving some 'ask' from the 12" subs on LFE only (I was seeing too much extension at 30 Hz from the 12" subs) but it sounds better for the 12's to do more.

I dod my best with measurements and theory then simply listen and A-B
Distortion comes in many forms. Some is tolerable at 4%, some is irritating at <<0.01%. Distortion that is not correlated to the harmonic structure of the sound is not masked and is therefor more noticeable. As a profession audio engineer who worked with orchestras and opera companies for many years (which are traditionally acoustic performances) I am quite sensitive to hearing sounds in electronic recordings of acoustic sources that are completely unnatural (i.e. do not exist in the character of acoustic instruments and voices), because they possess artefacts produced by the reproduction system. Thise artifacts can be immeasurable - or at least the standard THD+N tests do not qualify or quantify those distortions.

Just because someone else doesn't identify the change in character as distortion does not invalidate my experience or that of others. I started as a graduate electronic engineer but a carreer in the field of electro-acoustics has taught me to be a lot more humble about what is audible and what isn't. For example it took decades for the mainstream to recognise that 16-bit digital with 'unmeasurable wow & flutter' still sounds crap unless the clock precession and jitter is many orders of magnitude lower than early digital implementations.
The sound impairment idea is sound (no pun intended) and works well for many applications. I've assisted John Dunlavy (Duntech, Dunlavy Audio Labs) use a dual channel FFT that cost well over $100,000 in its day to measure the performance of speaker cables, and yes they do make a difference.

The problem for state of the art equipment is you need a better than state of the art analyser, and even then a LOT of care to ensure that the measurement setup is not masking changes by introducing larger changes itself.
Again with the far too large silmics with 10x the lead inductance making an improvement only the deaf couldnt hear ... makes such an improvement; yet because measurements are so terrible today, nothing is born out objectively. You dont think the famously fallible brain and ear interface is telling you fibs, vs the equipment thousands of times more sensitive?
Im yet to fully evaluate, but do trust my ears over time, if we were able to measure sonic improvements accurately that would be amazing wed buy one spec and live happily ever after, alas this simply isn't the case. As an example is I obviously use this unit to EQ and time align the bass and carefully measure all of the speakers using REW, its excellent information but you then need to finally adjust by ear. Take speakers positioning ever moved a speaker by a few degrees toe I or toe out and found sonic repeatable differences. I cant measure these with REW, but they are very audible.

Take voice recognition its taken years of complex and clever algorithms to achieve this, we do it instantly

Everyone is fallible but that doesn't stop us endeavouring to be objective, ultimately its all about the enjoyment of the end result. Unfortunately Im a hard marker...