Me being me, I just couldn't wait to try some tweaks before parts arrive.
I have some good reasonably low ESR 3000uF / 35V caps, decided to try them out to see if that path was worth investigating.
It is not easy to assess the sonic improvements without any doubts. Once the are caps ran-in, it might be easier to find out sonic differences by removing the caps (instead of fitting them) because:
Whatever, I did a quick and (very) dirty test rig. Mini on big feet to enable quick fit and removal of underneath additional bypass caps of various sizes, no I didn't short the cap's legs too much (not good at all re ESR and other parasitic constants, I know, but then it can only improve at least), listening through my dummy loads (very old closed box), defo not the last say regarding resolution or sound.
The point is simple: IF I hear an improvement with that very basic set up, then it is worth investaigating further on a proper system with more definitive caps & proper fitting. The only thing I need to be carefull about is that some improvements (say for example an unbalanced mid-bass boost) could be wrongly perceived as an imrpovement whereas it is just mitigating a majour LS problem I have here. Oh, and if I hear no improvement at all despite all my care and my positive bias as some might say (LOL) then it means that on my proper hifi system(s) the improvement is either nile or small enough... to be disregarded vs hassle.
IF the experiment turns out for whatever dark reason to be positive, say as my addition of the filtering coil for example, then I will consider smaller and possibly better caps for that tweak. It is very difficult to to find real low ESR caps as most manufacturers aren't documenting the ESR across frequencies (I am more interested in audio band), they rather give one value at 100kHz and sometimes enable possible extrapolations through TANd. But anyway, I am now looking around the following promising caps: low advertised ESR with some luck it might still not be too bad at lower frequencies while also being quite small form factors caps to fit neatly under the board ... not the way it is now.
IF my experiment works well, then I shall compare my current quick-test bypass caps vs
https://www.mouser.fr/ProductDetail/Nichicon/UHW1E152MHD?qs=Ajmft%2BTTukE7/UDuqMyoRQ==
or
https://www.mouser.fr/ProductDetail/Rubycon/25ZLH1500MEFC12.5X20?qs=T3oQrply3y/K3q0O5G8L7w==
To be followed...
Claude
PS: and yes, as an engineer, I see (nearly) no reason why this cap addition should make ANY change AT ALL given the Mini has defo enough filtering capacity and power reserve with its 15000uF. The addition of my caps is within the tolerance band and the cut off frequency of the RC filter definitively low whatever - but then I have seen so many odd things in other well proven eng. branches. "Nearly"... because other than possibly tuning the sound to my liking (or making me believing it) by lowering the PS impedance at given frequency bands (ESR in parallel & Co...), I see nothing else, but exactly that could be the answer ;-)
PS2: the second pix is just to show you the caps I am trying, of course they aren't fitted that way, they belong under the board...
I have some good reasonably low ESR 3000uF / 35V caps, decided to try them out to see if that path was worth investigating.
It is not easy to assess the sonic improvements without any doubts. Once the are caps ran-in, it might be easier to find out sonic differences by removing the caps (instead of fitting them) because:
- It takes less time than installing them, hence not mixing things up in my brain /memory and also due to the Mini warm-up time / its associated change in sonic signature
- In terms of variations, we are usualy more sensitive to a loss than to to a progress LOL
Whatever, I did a quick and (very) dirty test rig. Mini on big feet to enable quick fit and removal of underneath additional bypass caps of various sizes, no I didn't short the cap's legs too much (not good at all re ESR and other parasitic constants, I know, but then it can only improve at least), listening through my dummy loads (very old closed box), defo not the last say regarding resolution or sound.
The point is simple: IF I hear an improvement with that very basic set up, then it is worth investaigating further on a proper system with more definitive caps & proper fitting. The only thing I need to be carefull about is that some improvements (say for example an unbalanced mid-bass boost) could be wrongly perceived as an imrpovement whereas it is just mitigating a majour LS problem I have here. Oh, and if I hear no improvement at all despite all my care and my positive bias as some might say (LOL) then it means that on my proper hifi system(s) the improvement is either nile or small enough... to be disregarded vs hassle.
IF the experiment turns out for whatever dark reason to be positive, say as my addition of the filtering coil for example, then I will consider smaller and possibly better caps for that tweak. It is very difficult to to find real low ESR caps as most manufacturers aren't documenting the ESR across frequencies (I am more interested in audio band), they rather give one value at 100kHz and sometimes enable possible extrapolations through TANd. But anyway, I am now looking around the following promising caps: low advertised ESR with some luck it might still not be too bad at lower frequencies while also being quite small form factors caps to fit neatly under the board ... not the way it is now.
IF my experiment works well, then I shall compare my current quick-test bypass caps vs
https://www.mouser.fr/ProductDetail/Nichicon/UHW1E152MHD?qs=Ajmft%2BTTukE7/UDuqMyoRQ==
or
https://www.mouser.fr/ProductDetail/Rubycon/25ZLH1500MEFC12.5X20?qs=T3oQrply3y/K3q0O5G8L7w==
To be followed...
Claude
PS: and yes, as an engineer, I see (nearly) no reason why this cap addition should make ANY change AT ALL given the Mini has defo enough filtering capacity and power reserve with its 15000uF. The addition of my caps is within the tolerance band and the cut off frequency of the RC filter definitively low whatever - but then I have seen so many odd things in other well proven eng. branches. "Nearly"... because other than possibly tuning the sound to my liking (or making me believing it) by lowering the PS impedance at given frequency bands (ESR in parallel & Co...), I see nothing else, but exactly that could be the answer ;-)
PS2: the second pix is just to show you the caps I am trying, of course they aren't fitted that way, they belong under the board...
Attachments
Claude, I bought (this fancy LCR meter, model ET4510) on AliExpress . . . . for a much lower price than Amazon quotes. It will measure at any frequency you like, between 10 Hertz and 100 kiloHertz. The meter ships with 4-point Kelvin probes so you get reasonably accurate measurements even for loooooow impedances, like the ESR of an Aluminum Polymer electrolytic capacitor. Also handy for finding the self resonant frequency (where |Z| is minimum) of big caps. Once you know the S.R.F. you apply the good old EE resonance equations to calculate ESL.
Mark, many thanks! (again...)
I didn't even know that such thing existed! Amazing... even if understandably limited to 100kHz (not such an issue as for small or "HF tech caps" SRF is often given)
I will look if I can get one at a reasonable price somewhere, they seem sadly to charge a lot (same as Amazon) even on Ali once delivered in Europe.
It looks like a great tool to measure values for both analogue... and digital projects.
And, indeed good point re ESL!
At the "knee point" /SFR, things get indeed even easier... A first degree equation with one unknown and with a square root being "the most complicated sign in it" should hopefully still be within my (very rusty) reach LOL...
Good for me that the filtering coil ("yours"...) sits between my SMPS and any Mini PS cap (including my additional ones) re possible resonnance. Also at least the ESL effects given the SFR of "similar caps" (when manufacturers dare sharing it) seem to be well beyond audio range for my lowish lytics values here, so I guess that if I hear any difference it is more down to reduced PS impedance "overall" in the audio range, but that is just a guess. Indeed, lowering further the already very low Fc of the RC filter is IMHO not the point here.
All very interesting and, as ever, trully great added value - again thanks for all this!
Enjoy your day very much
Claude
I didn't even know that such thing existed! Amazing... even if understandably limited to 100kHz (not such an issue as for small or "HF tech caps" SRF is often given)
I will look if I can get one at a reasonable price somewhere, they seem sadly to charge a lot (same as Amazon) even on Ali once delivered in Europe.
It looks like a great tool to measure values for both analogue... and digital projects.
And, indeed good point re ESL!
At the "knee point" /SFR, things get indeed even easier... A first degree equation with one unknown and with a square root being "the most complicated sign in it" should hopefully still be within my (very rusty) reach LOL...
Good for me that the filtering coil ("yours"...) sits between my SMPS and any Mini PS cap (including my additional ones) re possible resonnance. Also at least the ESL effects given the SFR of "similar caps" (when manufacturers dare sharing it) seem to be well beyond audio range for my lowish lytics values here, so I guess that if I hear any difference it is more down to reduced PS impedance "overall" in the audio range, but that is just a guess. Indeed, lowering further the already very low Fc of the RC filter is IMHO not the point here.
All very interesting and, as ever, trully great added value - again thanks for all this!
Enjoy your day very much
Claude
Sooo.... over 100h run-in, over 20 on/off cycles, my additional caps are latest now at their operating best.
Listening performed over several days, over several hours, and final check done A/B quickly with just desoldering 4 tiny points.
At mine, these additional bypass caps definitively brought a plus, as did the filtering coil previously. Removing them the sound became immediately muddier... although it was of course already VERY enjoyable.
The first word that comes to mind to describe the addition of the caps in parallel to the C3s is "chiseled".
My notes on the listening impressions are:
These are small improvements, as the coil on other registers, but IMHO by no mean negligeable ones.
It is easy to try yourself: look in your drawer, take 2 low ESR caps say between 1000 and 3000uF/25V (or more V) and just give it a try.
As of me, my task is now to keep the same positive results while using smaller sized caps, as posted already.
The last experiment I could then try is daisy chaining the coil-filters, for the sake of it. It easy to do so will order another coil.
And then that's it unless someone pops up with a great idea. I won't bypass the big output lytics with some fancy PPP caps as doing so on my VFET wasn't really worth it, but who knows. So I guess I am nearly done - well bi-amping tries are also on the agenda.
It is trully an amazing little amp as designed. IMHO it becomes fabulous with the non expensive addition of (just) a filtering coil and 2 bypasses caps.
I hope this triggers some interest despite me being very late to joining the party
Claude
Listening performed over several days, over several hours, and final check done A/B quickly with just desoldering 4 tiny points.
At mine, these additional bypass caps definitively brought a plus, as did the filtering coil previously. Removing them the sound became immediately muddier... although it was of course already VERY enjoyable.
The first word that comes to mind to describe the addition of the caps in parallel to the C3s is "chiseled".
My notes on the listening impressions are:
- all notes are quicker on the rising front on all registers, hence the perceived dynamic is enhenced,
- the bass punches indeed quicker, but remains at its place, it gains in fact in precision and agility, bass lines are easier to follow
- everything is more precise while not being intrusive, in fact details stand out while at the same time fitting better together (integration!)
- complex instrument combos are easier to follow, the message is more intelligible
- the sonic stage is both more precise while also more airy / it gained in dimension
- No draw back whatsoever, just allow some run-in time for the caps, say a few hours at least
These are small improvements, as the coil on other registers, but IMHO by no mean negligeable ones.
It is easy to try yourself: look in your drawer, take 2 low ESR caps say between 1000 and 3000uF/25V (or more V) and just give it a try.
As of me, my task is now to keep the same positive results while using smaller sized caps, as posted already.
The last experiment I could then try is daisy chaining the coil-filters, for the sake of it. It easy to do so will order another coil.
And then that's it unless someone pops up with a great idea. I won't bypass the big output lytics with some fancy PPP caps as doing so on my VFET wasn't really worth it, but who knows. So I guess I am nearly done - well bi-amping tries are also on the agenda.
It is trully an amazing little amp as designed. IMHO it becomes fabulous with the non expensive addition of (just) a filtering coil and 2 bypasses caps.
I hope this triggers some interest despite me being very late to joining the party
Claude
Thanks for your interest 
If you zoom in on the first picture post 2802 you will just see them under the board. I admit... barely.
As said, it was just a quick and dirty try to see if there were any tendencies, I sadly didn't make any underboard pix. The real fit will anyway come afterwards (hopefully with the smaller caps), as will the frame I intend to build around the Mini (with proper LS cable connectors). I will try not to forget to post pix then.
Trying the caps mod is dead easy. If you are unsure, just put some adhesive film under the board to avoid any potential shorts were you put your caps. Of course take care of the polarity (same orientation as C3 obviously). And keep legs of the additional caps short, directly soldered under the C3 soldering points.
I hope this helps!
Claude
If you zoom in on the first picture post 2802 you will just see them under the board. I admit... barely.
As said, it was just a quick and dirty try to see if there were any tendencies, I sadly didn't make any underboard pix. The real fit will anyway come afterwards (hopefully with the smaller caps), as will the frame I intend to build around the Mini (with proper LS cable connectors). I will try not to forget to post pix then.
Trying the caps mod is dead easy. If you are unsure, just put some adhesive film under the board to avoid any potential shorts were you put your caps. Of course take care of the polarity (same orientation as C3 obviously). And keep legs of the additional caps short, directly soldered under the C3 soldering points.
I hope this helps!
Claude
I've been out of touch with the Pass forum for a few years now, so imagine my surprise to discover the ACA mini! I immediately ordered the essentials kit (a bargain, thank you, Papa) and filled out the BOM with parts I had, plus a few I needed to order. During biasing, I made some adjustments with my meter on the wrong terminals and let out a bit of the magic smoke from R8. Fortunately, although Mini smoked, she didn't inhale, so all was still good.
It turns out I'm a "jumpers in" kinda guy and this amp is an interesting and satisfying change from the Pete Millett NuClassD amp it replaced in my desktop system. That consists of a Win 10 PC feeding a Topping D10s DAC (with OPA1612A replacing the stock LME49720}, a BA2018 preamp (with linear supply modeled on the Whammy) and ADS L400 loudspeakers. I'm using a 3 amp supply on the ACA mini because in the past I have found with my headphone amp designs that oversizing an SMPS a bit over the theoretical need can improve the sound quality. Also, the 3 amp was only $15 on Amazon. I may just buy the 2 amp $9 supply and do a comparison.
Needless to say, I was disturbed like many others by the thump when turning on my BA2018 preamp. Like many Greedy Boyz, I couldn't wait for a fix, so I immediately laid out a PCB for Mark Johnson's SMPS noise filter (Thank you, Mark!) that included a SPDT switch with the "off" position connected to ground through either a 68 ohm 3 watt resistor or a 60 ohm inrush current limiter. I also included a status LED. This limits the ACA mini cap drain current to less than 500 mA, saving the switch contacts and any damage to the caps and preventing the boredom of watching the status LED drain down for minutes on end. It now takes about 5 to 6 seconds to extinguish with the resistor or about 1 to 2 seconds with the inrush limiter.
For anyone interested, I've attached the gerbers, schematic and BOM. It's as compact and slim as I could make it, requiring only four holes drilled in the Hammond plastic box, although their location is critical. You can always oversize them slightly if you end up off a bit. It has both input and output jacks. The cable required for the output connection is from Amazon. You can get two 18ga/5A 5.5mm/2.5mm plug to 5.5mm/2.5mm plug cables for about $4 each at https://www.amazon.com/dp/B09NBKRF8H
You can build it for other projects without the switch and/or LED, just jumper the switch pads. 5 pcs of the PCB from JLCPCB are less than $4 TOTAL (that's 80 cents each!), which includes delivery to continental USA addresses if you request standard delivery and just wait an extra week. That's about $20 total parts cost for the switched version. I included pads for both the original open air resistors and the Ohmite solid body current sense resistors, since the OARs have been unobtainium for what seems like years now.
And I will be curious to see what Papa comes up with for the F6m. I'm assuming he might use the Triad TY-250P (still less than $6). I tried those 6 years ago and got pretty good results, although I suspect Papa can do much better. Some of ZM's ideas on that are also here: https://www.diyaudio.com/community/threads/f6-build-options.313078/
It turns out I'm a "jumpers in" kinda guy and this amp is an interesting and satisfying change from the Pete Millett NuClassD amp it replaced in my desktop system. That consists of a Win 10 PC feeding a Topping D10s DAC (with OPA1612A replacing the stock LME49720}, a BA2018 preamp (with linear supply modeled on the Whammy) and ADS L400 loudspeakers. I'm using a 3 amp supply on the ACA mini because in the past I have found with my headphone amp designs that oversizing an SMPS a bit over the theoretical need can improve the sound quality. Also, the 3 amp was only $15 on Amazon. I may just buy the 2 amp $9 supply and do a comparison.
Needless to say, I was disturbed like many others by the thump when turning on my BA2018 preamp. Like many Greedy Boyz, I couldn't wait for a fix, so I immediately laid out a PCB for Mark Johnson's SMPS noise filter (Thank you, Mark!) that included a SPDT switch with the "off" position connected to ground through either a 68 ohm 3 watt resistor or a 60 ohm inrush current limiter. I also included a status LED. This limits the ACA mini cap drain current to less than 500 mA, saving the switch contacts and any damage to the caps and preventing the boredom of watching the status LED drain down for minutes on end. It now takes about 5 to 6 seconds to extinguish with the resistor or about 1 to 2 seconds with the inrush limiter.
For anyone interested, I've attached the gerbers, schematic and BOM. It's as compact and slim as I could make it, requiring only four holes drilled in the Hammond plastic box, although their location is critical. You can always oversize them slightly if you end up off a bit. It has both input and output jacks. The cable required for the output connection is from Amazon. You can get two 18ga/5A 5.5mm/2.5mm plug to 5.5mm/2.5mm plug cables for about $4 each at https://www.amazon.com/dp/B09NBKRF8H
You can build it for other projects without the switch and/or LED, just jumper the switch pads. 5 pcs of the PCB from JLCPCB are less than $4 TOTAL (that's 80 cents each!), which includes delivery to continental USA addresses if you request standard delivery and just wait an extra week. That's about $20 total parts cost for the switched version. I included pads for both the original open air resistors and the Ohmite solid body current sense resistors, since the OARs have been unobtainium for what seems like years now.
And I will be curious to see what Papa comes up with for the F6m. I'm assuming he might use the Triad TY-250P (still less than $6). I tried those 6 years ago and got pretty good results, although I suspect Papa can do much better. Some of ZM's ideas on that are also here: https://www.diyaudio.com/community/threads/f6-build-options.313078/
Attachments
@A Jedi , you'll have to invert the input sigal and feed to one channel. Then connect the speaker to the + out of both channels.
Unlike the ACA, the ACA Mini has no eay way to get an inverted signal from the exstng circuit. Try ading a unity gain inverting op-amp buffer to get that.
Unlike the ACA, the ACA Mini has no eay way to get an inverted signal from the exstng circuit. Try ading a unity gain inverting op-amp buffer to get that.
Last edited:
@audiosteve
[sigh] so was I...
until I got the Pearl 3 pioneer kit which has the option for a balanced out...
and I remembered that the BA2018 preamp input could be configured to be balanced... or it could be doubled up to be fully balanced...
if only I had a balanced amp.
then Nelson Brock decided to make more ACA Redux (hedgehog) kits... and I bought two.
this hobby is making me unbalanced, probably in more ways than one!
[sigh] so was I...
until I got the Pearl 3 pioneer kit which has the option for a balanced out...
and I remembered that the BA2018 preamp input could be configured to be balanced... or it could be doubled up to be fully balanced...
if only I had a balanced amp.
then Nelson Brock decided to make more ACA Redux (hedgehog) kits... and I bought two.
this hobby is making me unbalanced, probably in more ways than one!
@mhenschel
The obvious answer (to me anyway) is the ZM Balanced Iron Pre!
SE in to Bal out, SE in to SE out, Bal in to SE out, Bal in to Bal out.
The obvious answer (to me anyway) is the ZM Balanced Iron Pre!
SE in to Bal out, SE in to SE out, Bal in to SE out, Bal in to Bal out.
Last edited: