Modulus-86 build thread

Neurochrome.com
Joined 2009
Paid Member
I have not measured (or heard) any difference in performance between the Connex SMPS300REh and a traditional "linear" power supply with the Modulus-86.

That said, I stop short of offering a full-throated recommendation of Connex for a number of reasons. Connex trims the component pins so they're flush with the PCB before the board goes through wave soldering. This means that the spade terminals used for the output don't get hot enough to form a good bond. I've had a couple of SMPS300RE and SMPS300REh where this resulted in no connection to the terminal. I'd pull the terminal straight out of the solder joint when I tried to unplug the mating terminal. This is easily fixed by reflowing those solder joints and adding a bit of solder. You can do this from the top of the board pretty easily. I just don't think I should have to.
A client of mine had a failure of an SMPS300RE. It'd been in service for about three years at that point. Maybe that's a single premature failure (which happens). But it's a bit disappointing to have it fail after only three-ish years of service.
The customer service at Connex can be a bit frustrating at times. It helps if you've spent money there, but email often seem to go unanswered.
That said, Connex is making improvements. The original SMPS300RE could benefit from a heat sink, so they created the SMPS300REh. I also at one point had an issue with the packaging of their boards. Some boards I ordered were damaged in shipping. Next time I ordered Connex had changed the packaging to prevent that. So they're trying to do the right things.
I would say that Connex is good enough for DIY, especially if you're willing to reflow those spade terminals, but for a commercial build I'd be a bit leery about them.

There aren't really any good alternatives to Connex. At least not at the same price point. I've seen Mean Well RPS-200 used in Modulus-86 builds with good results. Two Mean Well RPS-200-27 would be suitable but that's roughly 2x Connex in solution cost.

I think there's a company out of Jordan (or in that part of the world anyway) that offers some competition to Connex. The owner wanted me to promote his products for him but also wanted me to pay for them. At the same time he couldn't provide a data sheet with enough information for me to use the products (no mechanical dimensions, for example) so I let that run out in the sand. I forget the name, but some have had better luck with them than I did. That may be an alternative to Connex.

Tom
 
  • Like
Reactions: 1 user
Neurochrome.com
Joined 2009
Paid Member
The rectifier in the SMPS300 emits some switching hash, so keep the amp input circuitry and wiring away from that corner of the SMPS.

Too many geek out about twisting wires. I doubt it does much, especially for the speaker wiring. I keep the mains wiring tightly bundled with wire ties. The key is to reduce loop area. Keep the wiring away from the amp inputs. The amp outputs are low impedance so it'll be hard to induce any disturbance onto them.

Tom
 
  • Like
Reactions: 1 users
Member
Joined 2008
Paid Member
Because someone asked in a PM, I’ve had my Mod86 playing in my main system for a few days now.

In short, I love it! :yes:

There is absolutely no detriment to the use of this amplifier. It provides clarity and detail right up there with the best that I have heard (there have been moments of “I haven’t heard that before in this song”). I am convinced it meets the “wire with gain” philosophy. I confess I think I enjoy my favorite class A amps slightly better, overall, but not by much.

As others may have gathered in their audio experiences, it is very difficult to differentiate between amplifiers of such high quality. In my diyAudio journey, I‘m finding that I prefer just a bit of 2nd harmonic distortion, but not very much. However, the clean effortlessness of a vanishingly low distortion class AB amp like the Mod86 is damn near enough to keep me happy.

I try to not dwell on what the differences are and why. That’s a recipe for neurosis and I have enough of that already with this hobby. It’s more fun to build, switch out, enjoy, and futz. It’s ultimately for my enjoyment, so I’d like to keep it that way.

I will continue to be a stalwart champion for the Mod86 (and its higher-power brethren) because it is an impeccably engineered amplifier that also sounds exceptional. If you haven’t built and listened to an ultra-low THD amp and are curious about that sound, the Mod86 is a no-brainer. I highly recommend it.
 

Attachments

  • IMG_0292.jpeg
    IMG_0292.jpeg
    364.1 KB · Views: 119
Last edited:
  • Like
  • Thank You
Reactions: 2 users
Neurochrome.com
Joined 2009
Paid Member
The company out of Jordan you mentioned is Micro Audio.

Ah, yes. It looks like they haven't made much progress on their "data sheets" since I last spoke with them. For example, the SMPS630G's output voltage is specified to "40 - 84 V DC" never mind that there's a ±28 V selection on the product page and no ±20 V so where does 40 V come from? Also, does "40 - 84 V DC" mean that a ±28 V supply can put out ±84 V under some conditions? Mechanical dimensions are (now) given but the hole pattern is not. Sorry. I can't take that seriously. I don't have enough information to be able to see if the product would work in my target application.

Tom
 
Member
Joined 2008
Paid Member
As a layperson, I would imagine they have a selection of transformer windings available for different output voltages and can get custom voltages made upon request. I have found them easy and pleasant to work with for hobbyist projects, but that use case has different requirements than higher-volume purchases, for sure.
 
Neurochrome.com
Joined 2009
Paid Member
It seems to be a "you get what you get" type of operation. Not a "here is what you will get" operation. I'm not willing to send money into the ether even for QTY = 1 without knowing what I'll get. But that's me.

All I'm asking for is a spec that shows the tolerance on the output voltage and the mechanical specs. I'm also assuming that lower voltage options would have higher output current, but the spec says 6 A (with tolerance given).

Have a look at the data sheets from Mean Well and you'll see what I'm looking for.

Tom
 
@tomchr
is there a post you can share(sorry to ask) where you explained using a power 686 with 2x modulus 86 boards and why/benefits? i see that in the quad 405 build (https://www.diyaudio.com/community/threads/modulus-86-build-thread.267802/page-293#post-7059344).

i didnt want to expedite my build in and wanted the experience the other facets of building this amp, so i think i am still going to go for a transformer + power86+ ISS, but i dont know where im going to spend more or less in my build yet.

sorry for not finding your recommended parts page on the modulus 86 showing the entire formula for a $550 build :( a SMPS definitely saves a bit of money!
 
Last edited:
Member
Joined 2008
Paid Member
Valid points, @tomchr.

Have you had a conversation with Sami from MicroAudio about what you’re looking for in terms of specs and documentation? The discrepancies could merely be a language barrier.

It’s also possible in the time between when I had some email conversations with them for my purchases and now (this is a few years now) they grew very fast and have an entirely different business.

I am not diminishing your concerns, nor making any excuses for gaps in their documentation. I’ve just been happy with their products and would hope to see a merging of two audio vendors with which I have interacted successfully in this forum. I expect there wouldn’t be much difference between your just-shy-of-endorsement of Connex for their products and those of MicroAudio. I‘ve used and like Connex SMPS products, too (my Mod86 uses a SMPS300RE).

Unfortunately, it seems that for the hobbyist there’s MeanWell, and then there’s “the rest” in the SMPS business for audio, at least for documentation and trust.
 
Last edited:
  • Like
Reactions: 1 users
Neurochrome.com
Joined 2009
Paid Member
is there a post you can share(sorry to ask) where you explained using a power 686 with 2x modulus 86 boards and why/benefits?
The Power-686 has twice the capacitance of the Power-86. Some prefer the added capacitance as it results in less voltage droop on heavy current transients, which some associate with "firmer bass" and such. The Power-86 has about twice the capacitance you typically find in a commercial amp at the same power level, so I don't personally see a need to add more capacitance, but if you have to geek out or splurge the power supply is a good place for it.

Have you had a conversation with Sami from MicroAudio about what you’re looking for in terms of specs and documentation? The discrepancies could merely be a language barrier.
I have. He contacted me a few years back and we had a conversation then. I outlined my expectations. And it looks like we may be restarting that conversation as I just responded to an email from him. I would love to be able to recommend someone in addition to Connex, even if the price is between Connex and two Mean Well RPS-200. But to do so I will need the vendor to provide a data sheet that shows:
  • Input voltage (nominal)
  • Input voltage (max and min)
  • For each output voltage option: Output voltage (nominal)
  • For each output voltage option: Output voltage (max and min)
  • For each output voltage option: Output current (nominal, max, min)
  • Output power, average
  • Output power, peak
  • Mechanical dimensions, including mounting hole locations and diameters (or thread specs)
  • Connector pinouts should be provided and mating connectors should be specified
  • Additional features (standby, thermal shutdown, etc.) should be documented
Even for a product with many voltage options this would take maybe a day to do. That's a reasonable ask. A more detailed data sheet, such as those by Mean Well, would be a bonus.

If there's anyone out there who makes SMPSes for audio who reads this and would like me to review their product with the Modulus-86, please contact me via email or PM. I'm happy to ship the product back after review as long as they pay the shipping cost. Alternatively, I can ship it along to a buyer in North America (at the vendor's expense). That's a pretty standard review process in the audio business. Some reviewers still consider the product provided to be on "long-term loan", i.e., provided for free with no expectations of return shipping.

Unfortunately, it seems that for the hobbyist there’s MeanWell, and then there’s “the rest” in the SMPS business for audio, at least for documentation and trust.
And I would love for Micro Audio to be "almost Mean Well" at a price point that's between Connex and Mean Well. It wouldn't take much to stand out.

Tom
 
  • Like
Reactions: 2 users
Neurochrome.com
Joined 2009
Paid Member
THD+N vs frequency at 0.1-1 W for the Modulus-86 would only show the noise floor of the Modulus-86 and the instrumentation. That's pretty clear from the THD+N vs output power graph. So I don't see a point of providing that measurement.

I do measure the THD (no +N) at 1 W. See below. You can also find the graph in the Performance Graphs tab on the Modulus-86 product page.

Modulus-86 Rev. 3.0_ Harmonic Spectrum (1 W, 8 ohm, 1 kHz, 0 dB = 1 W, 8 ohm).PNG


Tom
 
  • Like
Reactions: 1 user
Hi Tom :)
I built a 2.3 stereo modulus-86 a couple years ago. One of the channels was always flaky, and it just died (something strange, figured it was time to start fresh!), so I bought a stereo kit of the 3.0 modulus. I've read about the changes you made, and I'm curious as to whether I'll hear a difference with the new board(the 2.3's sounded very very good!).

I was wondering if using opa2134/1612 or something similar would make any audible difference in the differential stage? I'm guessing not as our ears aren't that good, but I figured you might have tried it in your travels.
 
Neurochrome.com
Joined 2009
Paid Member
Do you have any sense of why that channel died? I don't recall hearing about the flakiness so I'm a bit surprised. Maybe we could have found a solution for it.

Anyway. The Modulus-86 Rev. 3.0 is a complete redesign of most of the amp. It has much lower noise than Rev. 2.x. That said, the noise level of Rev. 2.x is usually inaudible, so you're not likely to notice a difference. Rev. 3.0 also did away with the DC servo. I get stellar DC performance by other means.

I don't see a point of using a 40 year-old opamp in the input stage. The OPA1612 would be a better choice. I doubt it'll make any difference in the measurements or sonic signature, though. Just make sure that U3 stays an LM4562/LME49720.

Tom
 
It was an intermittent fault, myself and another tech tried to pinpoint it but to no avail. There was a possibility that it was related to a faulty through hole for one of the LM3886 rails, but in the end the board complained about all of our tinkering and some of the through holes failed.

On a side note I'm using one of those cheap chinese "dc biased lm3886 balanced input" boards while I build the modulus replacement, you certainly hear the difference.

So, another question (sorry!) Would an OPA1612 (or similarly highly specced modern Op Amp) make a better composite amp, if used in the feedback loop instead of the 4562? Would we even be able to hear the difference?