I hope I don't open up another can of worms here. Be careful with the CLC filter. Modern components have pretty high Q (low parasitics) so the transient response of the filter needs to be controlled to avoid overshoot on load transients. This will likely mean that you need to insert some resistance in series with your filter inductor, which increases the output impedance. I'm not saying it'll blow up on you. I just suggest that you pay attention in the design phase.
With the Modulus-86, there is no need for fancy supply filtering due to the high PSRR of the amplifier. Personally, I'd fit the largest supply capacitor you can get away with and be done.
In my opinion, soft-start aka inrush limiting should be done on the transformer primary. That way any impedance imposed by the soft-start circuit is reduced by the square of the transformer turns ratio and will have negligible impact on the amplifier performance.
~Tom
With the Modulus-86, there is no need for fancy supply filtering due to the high PSRR of the amplifier. Personally, I'd fit the largest supply capacitor you can get away with and be done.
In my opinion, soft-start aka inrush limiting should be done on the transformer primary. That way any impedance imposed by the soft-start circuit is reduced by the square of the transformer turns ratio and will have negligible impact on the amplifier performance.
~Tom
I think he is a fan. He and I just disagree on how to quantify what "good" or "high performance" means. He's obviously free to use whatever method for determining whether an amp meets his requirements. A listening test is a pretty reasonable way to do this. If he'd rather use a dousing rod, that's fine by me as well, though I will reserve the right to shake my head in disbelief in that case...
I happen to prefer measurements because they are objective and correlate with the perceived sound quality. Measurements taken under comparable conditions can also easily be compared.
Others are free to disagree. However, I would prefer that those who wish to continue the discussion of measurements vs listening experience take it to the Lounge section.
~Tom
Craig, can you post some details, preferably a schematic, of this 'high PSRR equipment' which is audibly improved by using SLA batteries?
Apologies Tom, but I'm really intrigued.
______________________
On CLC PSUs, here is Prof. Cherry on the subject.
A New Distortion Mechanism in Class B Amplifiers
.. many important insights & useful pointers though you don't have to go the whole hog to get the advantages.
You don't have to use series resistance in the inductances to damp the PSU for transients etc. The ESR of the electrolytic capacitors AT the output devices does just fine. It turns out that bigger electrolytics need less ESR to do this properly .. a very serendipituous situation.
______________________
Tom, I'm lurking here in the hope you will publish the schematic for your excellent amp. As an itinerant beach bum for nearly 2 decades, I won't be able to buy your PCBs ... but would very much like to see how you have done it ... at least on a schematic level
Merry X'mas
Last edited:
"Newness" is certainly in the imagination of the beholder. It would be fair to say mutual inductance and the resultant coupling of signal was known before 1981.
I see your point on the high Q; just my luck to end up with a resonant tank. Unfortunately, I'm not a designer and certainly would not be able to develop and run Spice to model. Thank you also for your advice in the inrush limiter.
Correct. When I read that AES article, I was worshipping at the feet of (corresponding with) Great Guru Baxandall on another subject and he pointed out his own 1970's take on it (in Wireless World 'Letters' IIRC) There are probably others even earlier.
However, Cherry's treatment is comprehensive, accurate & useful and I no longer have GG Baxandall's letters.
Halauhula, Cherry shows how to do this properly in his original ETI NDFL amp articles. Note the layout and physical positioning of the bits is critical.
Both AES & ETI articles by Cherry are at
http://www.linearaudio.nl/linearaudio.nl/index.php/my-library/on-feedback-feedforward-and-error-correction Thanks Mr. Didden
Last edited:
Ooh. Wonky! Thanks for digging those up. Interesting design. It looks like a hybrid between a cable connector and a panel connector.
No worries. Let's just keep the discussion scientific and factual.
I'm guessing the inductor Halauhula is talking about is the Triad Magnetics C-94U. It's specified at 32 mH, 190 mΩ.
I dug up the largest 25 V electrolytic cap that Mouser has in stock: CDE P/N: 380LX563M025A062. 56 mF, ESR @ 120 Hz: 15 mΩ.
The resonant frequency is 3.8 Hz. The Q of the filter works out to 3.7. That's pretty peaky. You'd need to add 0.8 Ω to get the Q below 0.7.
I think we've covered this earlier in the thread...
You can get many hints from the references listed in Post #1.~Tom
For the real advantages that Cherry shows, he uses 15uH & 470u.
I did say "Cherry shows how to do this properly"
It's also important to get the caps close to the output devices & VAS and its difficult to do this with your 56mF
I'm not sure very large inductors have a benefit for other than a true Class A design. For that, choke input ie LC, might be better than CLC
Yes. And I pretend to be versed in the art at times.I think we've covered this earlier in the thread...
But its details eg the position and value of your local decoupling, your precise compensation, etc .. that will be critical for your 4ppm THD
The really huge factor is, of course, your PCB layout.
_______________
I also think you don't pay enough homage to our Golden Pinnae sensibilities.
When you add the note about tacking optional snubbers on the back of your PSU PCB, you should specifiy the work must be carried out by Virgins and the snubbers must be of Unobtainium. Otherwise, the mod is likely to introduce a definite roughening of midrange clarity
But its details eg the position and value of your local decoupling, your precise compensation, etc .. that will be critical for your 4ppm THD
The really huge factor is, of course, your PCB layout.
All true. That's the secret sauce. That's my value proposition.
When you add the note about tacking optional snubbers on the back of your PSU PCB, you should specifiy the work must be carried out by Virgins and the snubbers must be of Unobtainium. Otherwise, the mod is likely to introduce a definite roughening of midrange clarity
I'll make sure to mention that in the design documentation...
~Tom
Last edited:
Power-86 boards open for pre-order
Speaking of... I've opened up for pre-orders on the Power-86 boards. Please order through the Power-86 page on my website.
Fedex says the Power-86 boards will be here by Monday.
~Tom
Speaking of... I've opened up for pre-orders on the Power-86 boards. Please order through the Power-86 page on my website.
Fedex says the Power-86 boards will be here by Monday.
~Tom
Indeed I am - while in the throes of working out my own gainclone I did much research and thinking about doing a composite amp of my own design, but decided that it was adding another level of complexity to all the areas I was trying to optimise, and it didn't seem at the time that it was going to add enough to the table to make the extra effort of sorting the stability, etc, issues worthwhile. So, I'm impressed that Tom has gone the extra distance, and created a very solid design that shapes up beautifully - I would be happy to use one of his units to create good music, if the opportunity arose,
.As regards the rest, yes, that pretty well nails it, terrence - I don't want to be aware of any character in the sound, apart from what the recording has encoded ... part of how I how assess systems is how much the playback system succeeds in not intruding when the pressure is on. I have mentioned hearing a Bryston system at an audio show several times, which did a remarkable job of being completely unflappable, even at "extreme" volume levels - Tom's unit should be similar, but at more modest SPLs,
.
Yes, and if I know Frank, that's exactly what he's sayin', and me also.
I have tons of album cds of every genre you could poke a stick at.
I have tons of VA cds of every genre you could poke a stick at.
Between all of these recordings it is possible to discern system signature.
Once the system signature is learned and recognised, any small change to the system signature becomes quite apparent, not always immediately apparent, but any change will become apparent.
In the longer term, ultimately any new signature may be subjectively pleasant or unpleasant.
I find the best long term signature is close to no real system signature.
When the system has relatively little self signature, all recordings reproduce as they were mastered.
Of course that brings up a whole new subject, that said, I reckon my system is a whole lot better than than a whole lot of mixdown monitor systems of the past, and modern mixdown monitors too.
The result I get is that my system is dead clean, clear, dynamic and detailed, BUT without false embellishments, and wonderfully musical, as should be.
This means done right recording/masterings really do sound right.
Faults in other recordings/masterings are apparent but not necessarily objectionable per se.
Many recordings would benefit from on the fly effecting....I take note of Andrew's (Bonsai) tone controls, and the Quad tilt/eq controls also.
Tom I congratulate you on your whole concept..wonderful, even class leading schematic, implementation and documentation informations.
So, your amp should have close to no character of it's own according to your measurements data.
What is your now longer term subjective appraisal of this amplifier technique, and your physical implementation ?.
Respect, Dan.
Last edited:
Look at Digikey
LNR1V105MSE Nichicon | 493-7811-ND | DigiKey
Aluminum Capacitors | Capacitors | DigiKey
Those should do the trick!
Do
Thank you.
I find the Modulus-86 open and revealing. The transient response is precise and smooth. There's good punch in the bass, natural voicing in the midrange, and precise detail in the highs. I like it...
Those are my subjective experiences. I can throw more verbiage at it, but considering the ambiguities in the language used to describe subjective experiences, I'd rather sum it up: The Modulus-86 is the best amplifier I have designed to date.
If you want to extract every detail of the source material, The Modulus-86 is for you. If you prefer an amplifier with a sonic signature, I suggest building my Damn Good 300B...
~Tom
The Power-86 boards are now in stock and available for purchase on my website.
One Power-86 board is optimized to power the two Modulus-86 boards in a stereo amp. The Power-86 board can easily power one Modulus-86 board in a mono block and/or multiple Modulus-86 boards in a multi-channel configuration.
The largest electrolytic capacitor diameter supported on this board is 40 mm ±1 mm. The board is designed for use with snap-in capacitors with a pin pitch of 10 mm.
~Tom
One Power-86 board is optimized to power the two Modulus-86 boards in a stereo amp. The Power-86 board can easily power one Modulus-86 board in a mono block and/or multiple Modulus-86 boards in a multi-channel configuration.
The largest electrolytic capacitor diameter supported on this board is 40 mm ±1 mm. The board is designed for use with snap-in capacitors with a pin pitch of 10 mm.
~Tom
Attachments
Mmm, this a topic Tom and I've corresponded about some on the back channel. The difficulty with Golden Pinnae is they're typically employed in tests without scientific controls and with insufficient documentation to be replicable under control. This makes them extremely difficult to action effectively when doing audio design.
To pick a recent example, the battery discussion a few pages back on this thread is essentially about one person's subjective findings in relation to the general topic of power cord audibility. Huge amounts have been written about this from both subjectivist and objectivist viewpoints but very few people do the Spice parameter sweeps needed to assess the range of common mode error injected across various real world configurations (class I and class II appliances with various cabling options and mains faults---ground lifting via batteries is a special case of class II). The folks who have done this kind of analysis and have the data to make meaningful design decisions are primarily in pro audio. The result is that something which is quite complex but is actually fairly well understood---audio receiver CMRR requirements---is often seen regarded an unsolved problem in DIY and home audio.
As I've touched on earlier in this thread, the 40dB CMRR of a basic balanced receiver is adequate to render errors related to ground offsets inaudible in most home and DIY audio configuration. But cases infrequent, albeit routinely, occur where 40dB is insufficient. In the Mod this expresses as Tom's selection of the THAT 1200 for 90dB CMRR. I don't mean to come across as disrespectful but the pragmatic reality is most folks in the Golden Pinnae crowd lack the engineering background to look at something like the Mod's block diagram and recognize one little label on one box means a very large range of problems have been considered and addressed. Typically there's a suspicion something's been overlooked but cases where a specific mechanism and a proposal to address it is advanced are exceedingly rare.
This poses problems for audio engineers, both in the very real limits on the time available to educate to folks how stuff works and in the quite considerable difficulties around sifting through subjective testimonies trying to figure out if people are hearing something which isn't already addressed by good audio design practices. The nutshell summary is, in addition to inadequate receiver CMRR, most amps lack the PSRR and loop gain to nail the sound in = sound out requirement that's the basic expectation for pro audio. If one builds an amp, such as the Mod, which is does a good job in this regard, then inevitably someone will object the amp doesn't have some characteristic someone found desirable in a particular sound in != sound out configuration. The assumed corollary is most always the amp must fail in some Golden Pinnae kind of way rather than there being a limitation in the pinnae's characterization abilities.
Net result is the DIY community unwittingly, but collectively, creates an arrangement where no amplifier can be unambiguously successful (or most anything else, for that matter). As an amp designer the best solution I've found to this rather thankless and Sisyphean task is really just not to try. Working with live sound folks who can articulate what they want in a clean channel and know how to use effects to shape the sound to their liking is a fair bit easier. The key distinguishing characteristic is one's able to talk with musicians in a way which generates testable hypotheses. I don't mean that as a slam to DIYers---it's entirely expectable and normal an instrumentalist, vocalist, sound guy, or recording engineer would be more articulate about music than someone who focuses on playback as a hobby---but oft times the best design tradeoff one can do ends up stepping on Golden Pinnae a bit.
It's not my intent to speak for Tom here. It is my perception that in this and the other linked threads and web pages he's working incredibly hard to accommodate DIYer needs in this regard. That's more patience than I have for this for sure.
Wow! $50 for an 84mm x 92mm printed circuit board that holds 10 components (2 connectors, 2 electrolytics, 1 bridge, 1 heatsink, 2 resistors, 2 film capacitors). $50 for the bare board!
DIYers who have the confidence to lay out their own printed circuit board, may be motivated to consider alternatives. Seeed Studio will custom build ten PCBs this size, for $37. ($22 for the fab plus ~ $15 for the shipping; see attached quote). The cost is $13 lower, and you get nine extra PC boards to give away or to sell at your cost to other DIYers.
Laying out your own unregulated power supply board also provides an opportunity to modify the design, if you so desire. Perhaps to install secondary fuses (if desired), reverse-polarity rail protection diodes {Cordell ch.15}, Quick Connect I/O blades, power-is-good LEDs, ferrite beads, four individual diodes rather than a 4-in-1 bridge assembly, and many other contrivances, limited only by your ambition and imagination.
DIYers who have the confidence to lay out their own printed circuit board, may be motivated to consider alternatives. Seeed Studio will custom build ten PCBs this size, for $37. ($22 for the fab plus ~ $15 for the shipping; see attached quote). The cost is $13 lower, and you get nine extra PC boards to give away or to sell at your cost to other DIYers.
Laying out your own unregulated power supply board also provides an opportunity to modify the design, if you so desire. Perhaps to install secondary fuses (if desired), reverse-polarity rail protection diodes {Cordell ch.15}, Quick Connect I/O blades, power-is-good LEDs, ferrite beads, four individual diodes rather than a 4-in-1 bridge assembly, and many other contrivances, limited only by your ambition and imagination.
Attachments
Last edited:
As I said in the first couple of pages of this thread, you get more than the raw board material when you buy my boards.
There are many supply boards on the market. I have seen many of them. Mine is the only supply board that has been optimized for a low supply impedance (performance). That adds value.
One of the main things you get when you buy the Power-86 board is good documentation. The documentation explains how to identify the different windings on the Antek AN-2222 power transformer and hook them up correctly to the mains and the Power-86 board. For experienced builders this may be no big deal, but for builders towards the beginner end of the spectrum, this is a significant source of anxiety, which is further magnified by the choice of wire colors by Antek. My step-by-step instructions will allow the beginner builder to get the wiring right the first time. This adds value.
If you buy my Power-86 board, you get a link to a project set up with Mouser Electronics. To order the correct parts, that will fit the board and provide the advertised performance, all you have to do is click "add to cart" and check out. This adds value.
In addition, I plan to fully characterize the Modulus-86 when powered by a Power-86 board in a stereo amplifier. This way, the customer can get a full end-to-end spec and be confident in the performance of their build.
The value proposition of Neurochrome is simple: I provide well designed, well tested, well characterized, and well documented state-of-the-art circuits. I perform my testing and characterization using state-of-the-art equipment. I also support my customers and take pride in their success. I believe this quality level is worth paying for.
If you want cheap products, there's always eBay...
~Tom
There are many supply boards on the market. I have seen many of them. Mine is the only supply board that has been optimized for a low supply impedance (performance). That adds value.
One of the main things you get when you buy the Power-86 board is good documentation. The documentation explains how to identify the different windings on the Antek AN-2222 power transformer and hook them up correctly to the mains and the Power-86 board. For experienced builders this may be no big deal, but for builders towards the beginner end of the spectrum, this is a significant source of anxiety, which is further magnified by the choice of wire colors by Antek. My step-by-step instructions will allow the beginner builder to get the wiring right the first time. This adds value.
If you buy my Power-86 board, you get a link to a project set up with Mouser Electronics. To order the correct parts, that will fit the board and provide the advertised performance, all you have to do is click "add to cart" and check out. This adds value.
In addition, I plan to fully characterize the Modulus-86 when powered by a Power-86 board in a stereo amplifier. This way, the customer can get a full end-to-end spec and be confident in the performance of their build.
The value proposition of Neurochrome is simple: I provide well designed, well tested, well characterized, and well documented state-of-the-art circuits. I perform my testing and characterization using state-of-the-art equipment. I also support my customers and take pride in their success. I believe this quality level is worth paying for.
If you want cheap products, there's always eBay...
~Tom
Last edited: