Have a look earlier in the thread with the measurements of regulated lab supply vs transformer that was lying around. The PSRR is supreme on this amp so it will work well with almost anything. An electrostatic shield is considered by all to be a good thing, so as long as you pick your V and VA to meet your requirements (such as the recommended Antek unit) you will be good to go.
Anything else is just keeping your inner audiophile sensibilities fed and watered. Nothing wrong with that
Bill
IMO it's all about the design philosophy of the person creating a circuit - at some point the "outside" world to a component is a very nasty place, electrically, so a barrier or cutoff point has to be introduced somewhere, to isolate the key parts from that bad stuff. Which could be done on the input, or the output side, of the transformer, or both - all are perfectly valid, so long as the job gets done.
If you're selling a PCB, as here, it's smart to make the circuit on it as impervious to outside muck, possibly passed through by the transformer, as one can ...
If you're selling a PCB, as here, it's smart to make the circuit on it as impervious to outside muck, possibly passed through by the transformer, as one can ...
Have a look earlier in the thread with the measurements of regulated lab supply vs transformer that was lying around. The PSRR is supreme on this amp so it will work well with almost anything. An electrostatic shield is considered by all to be a good thing, so as long as you pick your V and VA to meet your requirements (such as the recommended Antek unit) you will be good to go.
Anything else is just keeping your inner audiophile sensibilities fed and watered. Nothing wrong with that
Bill
I am a believer of cross comparing measurements with listening impressions. Lots of people just say what you should do, but fail to identify the difference if you don't. With an amp of such low distortion, certainly it deserves more care in addressing other issues.
Love the specs of this design, there are still so many ways you can screw it up. Seems like nobody is posting builds and results here. Is there another discussion on this?
The Power-86 board is a simple, unregulated, linear supply just as the one I used for my testing. It will have footprints for two snap-in capacitors with 10 mm pin pitch and up to 40 mm diameter (40 mm specified, so 41 mm max. actual dimension).
The board will have a terminal block for the output voltage. The intent is that you connect one wire per terminal and that wire goes to the amplifier board. In this configuration, the board will support two Modulus-86 boards. You can connect more, if you'd like. That just requires you to stuff two wires into one terminal - no biggie. As long as you can fit the supply capacitance needed for the number of boards you intend to connect, you're good.
~Tom
If you look at page 22 of this thread you can see how much improvement you could gain with an ultimate PSU. Its not a lot and down over 120dBc. Tom has designed this so that its almost impossible to screw up or improve. Build to BOM and you get the performance promised.
Yes, electrostatic shields are IMO are a must.
Most transformer manufacturers run the transformer cores very hard, i.e. high magnetisation levels, causing distortion of the AC sine wave (well not exactly) as the core rises towards the high end of the BH curve.
If you choose a larger core for a given VA rating, this will lower the transformer core magnetisation by reducing the BH curve excursion closer to saturation keeping it on the more linear part of the BH curve. Lower power line harmonics will be the end result.
Trouble is the AC power is not exactly sinewave from the power distribution companies. This results in harmonics which transformer design cannot overcome.
That is where some pretty sophisticated attempts have been made for AC regeneration systems to produce spectrally pure monotonic sinewaves to provide power in AC transformer mains powered audio equipment.
Over 10 years ago I attempted to build an AC filtering system to improve the sound quality of a test bed developmental audio system.
The system had the best medical grade Schaffner AC mains filters, balanced 1:1 mains transformer. All the transformers ran over sized cores. Electrostatic shields on everything. Multiple balanced / unbalanced line filtering.
IIRC the whole thing cost over $1,000 in parts. Not to mention many hours spent developing it.
It was progressively optimised, from a single mains filter. As each stage was added audible improvements were noted. From the finally completed unit to a raw power supply test there was a significant and what I thought amazing difference for just filtering AC power.
The next step was to power all of the audio equipment from SLA batteries.
The SLA batteries were so far ahead in performance it made all the effort used to make the AC filtering unit seam worthless. I remember I couldn't stop laughing and shaking my head in disbelief at the fruitless pursuit of trying to make good mains power. It was a learning exercise in realising how much gain could be made cleaning up the AC mains power and how superior SLA batteries were.
The AC conditioner was quickly stripped down and converted into a SLA battery charger for the battery supplies for the audio equipment. Never bothered with AC again and just used batteries after that.
For most designs, I would agree with you. Most designs don't have the 250 dB of loopgain at DC and 50ish dB loop gain at 20 kHz that my Modulus-86 amp has. Most designs don't have the CMRR or PSRR that my design does. These properties make it possible for the Modulus-86 to shine and be nearly independent of the power supply, as my measurements show.
That's why the Modulus-86 excels. It's impervious to all the crap you can throw at it. There's really no reason to douse your transformers here...
A couple of guys have given me feedback privately. I have quoted them in previous posts. I'll get their testimonials up on my website as well. It seems builders are happy with the results. Many give me glowing reviews on the documentation as well.
I encourage builders to share their thoughts here. More detailed build threads are probably better served in the Chipamp forum with a couple of links connecting this thread an the build thread.
~Tom
Bingo. That's my value proposition in a nutshell. I deliver the best documented design out there. Build it according to the BOM and you'll match the results I'm getting. You know what you get. There are no secrets or mysteries. If you get stuck, I'm only an email away.
~Tom
What did you measure?
What did you find?
__________________
I've done DBLTs on alternative transformers for a 50W commercial design with a very high powered DBL panel. The results were unexpected. This design is in the right ballpark for these results to be relevant.
A DBLT is a measurement. Your panel is your test instrument and it has an accuracy. You need to check its calibration regularly.
Well it's the PSRR that matters when it comes to power supply harmonics and EMI/RFI, not the open loop gain or CMRR.
Rubbish can be heard way down in the noise floor.
What I suggest is try it for yourself and report back.
All my tests were done with extremely high PSRR equipment. You would have trouble seeing any of it on most test equipment. You would probably need a state of the art spectrum analyser and even then you would be trying to extrapolate what you see against what you hear.
It's pretty easy really. Just hook SLA batteries up and do an A/B. You will be gob smacked.
Rubbish can be heard way down in the noise floor.
What I suggest is try it for yourself and report back.
All my tests were done with extremely high PSRR equipment. You would have trouble seeing any of it on most test equipment. You would probably need a state of the art spectrum analyser and even then you would be trying to extrapolate what you see against what you hear.
It's pretty easy really. Just hook SLA batteries up and do an A/B. You will be gob smacked.
I am using a brand new Audio Precision APx525 for my measurements. Many of the measurements shown earlier in this thread were done with an Audio Precision SYS-2712 that I had access to. For audio measurements, only the APx555 (base model starts at about $30k) beats the performance of my gear.
~Tom
~Tom
Listening test is one thing. What about living with it?
I have used battery power with Altman BYOB chip amp back before I did DIY. It did sound quiet, powerful and fast. And that chip is far inferior to 3886. The best part was listening to music during lightning storms without worry. In NC, USA that is every evening of the summer so it is a big advantage. So I am game for battery power.
You mean 4 x 12V batteries to make +/-24V PS?
Is that enough to get the best sound quality from LM3886?
Can you charge them with a standard 48V golf cart charger?
Good sealed batteries safe for interior use and a smart charger will cost >$1000 in US.
Using one battery for 12V automotive chip amp is one thing, but dual supply is not as easy.
Last edited:
In a sense that is what I had with my own version of the classic gainclone - it effectively had very large reserves of energy, by nature of the capacitor setup, and I could pull the mains cord from the wall at any time while playing music - and nothing happened. That is, the system kept playing with zero change in the quality of the sound, for up to quite a few minutes, until suddenly the music stuttered to nothingness - the time lag was totally dependent on the nature of what was being played, of course.
To me this was an excellent metric that most things were in order, and that the mains interference was having minimal impact ...
Ime standard lead/acid/SLA battery powered systems can sound ok/good, but can also sound leaden.
I will sometime get around to trying such system with terrestrial earth stake connection.....the exact terrestrial location of that earth stake I expect will have another importance.
Dan.