Modulus-86 build thread

He includes a multitone test, which no manufacturer dares to do. Also if you go here https://www.neurochrome.com/modulus-86-rev-2-1/ and scroll down you will see a graph of THD+N vs frequency.
What is missing that you need and what are you comparing to? I am not sure exactly sure.
Ah, yes. Didn't see that; 'twas exactly what I was looking for:
40 watts into 8 Ω at <0.001% (-100 db), 20-20KHz
And the multi-tone test is impressive.
Quite a match for an LX Mini speaker system, methinks.
 
The LX Minis I listened to were pretty nice, well balanced and good sound for small group performances. A bit less liveliness compared to a live performance, some subtle note swaying and vocal skills presentation are missing. I think this is due to the highly damped material in the driver cones.


Sent from my iPhone using Tapatalk
 
Ah, yes. Didn't see that; 'twas exactly what I was looking for:
40 watts into 8 Ω at <0.001% (-100 db), 20-20KHz

Yep. Most of that is the +N of the measurement setup.

And the multi-tone test is impressive.

It's a neat test. According to Belcher (1978) it also correlates well with the perceived sound quality and it's a much more representative test of actual music signal than a pure tone test. I'd like to see it used more.

Quite a match for an LX Mini speaker system, methinks.

Indeed it is. That's in fact how I use mine. I like the minis. They're not LX521.4 level, but they're also a rather small fraction of the LK521.4 build budget. For residential settings, they're really quite nice and my Modulus-86/286 drives them well.

Have a look at Baxandall and Cordell publications regarding distortion measurements. There is nothing new, but...

Will do. You're right. Nothing new. I just need to push my test equipment to the max (and beyond).

Tom
 
push my test equipment to the max (and beyond).

"Work smarter, not harder"(c)

True. It'd have to be a rather steep filter to provide any meaningful attenuation at H2, though.


Steep doesn't mean a high order, can be a high-ish Q band pass or it can be something that looks like a Chebyshev type II low pass, but with nulls adjusted to harmonics etc.
Using internal AP sig gen can have some benefits. It may allow to do meaningful vector averaging (if necessary), synchronization FFT etc
 
I like the minis. They're not LX521.4 level, but they're also a rather small fraction of the LK521.4 build budget. For residential settings, they're really quite nice and my Modulus-86/286 drives them well.
Tom
OK, so what's best to drive the LX Minis---four Modulus-86s or two Modulus-286s? For some reason, the 286 is MORE than 2x the cost of two 86s ($140 vs $130)
And, why is your recommended supply at +/- 28 volts instead of the datasheet suggested +/- 35 volts? It seems that the higher voltage would provide more headroom, which you cannot have too much of--eh?
And, what about regulated supplies? Wouldn't they improve performance?
 
Last edited:
So I'm finally getting some listening in given that I have finished another project, a pair of Frugal Horn MK IIIs with MarkAudio Alpair 7P drivers.

What this has immediatley done is revealed the deficencies in the rest of the system. Namley my MP3 collection that was ripped for maxium device compatibility. So my first port of call is to re-rip the CDs to FLAC and find an improved media player over the current Squeezebox.

I'm also considering pre-amp options. Current thinking is something based around Tom's THAT driver so that I can use single ended sources.
 

Attachments

  • Frugal Horn 1.jpg
    Frugal Horn 1.jpg
    860.6 KB · Views: 423
  • Frugal Horn 2.jpg
    Frugal Horn 2.jpg
    763.5 KB · Views: 403
Last edited:
OK, so what's best to drive the LX Minis---four Modulus-86s or two Modulus-286s? For some reason, the 286 is MORE than 2x the cost of two 86s ($140 vs $130)

Since we're apparently speaking in all-caps now: It also provides BETTER PERFORMANCE (lower noise). It is also a FOUR LAYER BOARD (more expensive than two MOD86 boards). And it allows for PARALLEL OPERATION of the two channels (richer feature set).

And, why is your recommended supply at +/- 28 volts instead of the datasheet suggested +/- 35 volts?

That's one of the frequently asked questions on the LM3886. I've answered that enough times both in this thread and in others that I really should write a page on that and put it on my website. Here's the summary:

There are two things at play:
  1. Power dissipation in the LM3886 (thermal issue)
  2. Output current limit of LM3886 (output drive issue)
In the isolated package (LM3886TF), I recommend that you stay below about 40 W of dissipated power to avoid tripping the SPiKe protection circuit in the LM3886. In practical terms, this means staying at ±28 - ±30 V if you're driving a 4 Ω load. At ±28 V you can get the full rated power out of the LM3886. At ±35 V, this drops to about 55-60 W (4 Ω) due to the power dissipated in the LM3886.
Secondly, the LM3886 is "only" guaranteed to be able to provide 7 A (peak) of output current. This means the highest supply voltage with a 4 Ω load should be ±7 * 4 = ±28 V.

It seems that the higher voltage would provide more headroom, which you cannot have too much of--eh?

Depends on how you define headroom. I can market the MOD86 as a 40 mW amp and get 60 dB of headroom or I can market it as a 40 W amp and have 0.2 dB of headroom. Headroom is mostly a marketing term. It basically means, "we designed a 100 W amplifier, sized the heat sink for 30 W and the power supply for 200 W. We then claim it's a 100 W amplifier (10 dB crest factor) with 3 dB of headroom (200 W supply)." In reality the amp will only deliver a 30 W sine wave continuously before overheating.
I don't operate that way. I prefer a honest approach. You can run my amps continuously at the specified output power under the conditions I give in the spec table and design documentation. To me that's the honest approach. Others will probably chastise me for crappy marketing, but honesty makes me sleep better at night.

And, what about regulated supplies? Wouldn't they improve performance?

With the regular LM3886 by itself, sure. You'll see an improvement from using regulated supplies. With the MOD86/286 due to their composite topology, you won't see a lick of difference below clipping.

Both the MOD86 and MOD286 are suitable for driving the LXmini. I use 4xMOD86 to drive my LXmini. If I was to build the amp today, I'd use 2xMOD286 for a lower build cost, slightly better performance, and easier assembly (I prefer SMD over leaded for assembly in many cases).

Your question about pricing does make me wonder if it's time to adjust the MOD86 pricing for inflation. It's been at $65 for coming up on three years now. 😉

Tom
 
Last edited:
As a contented user of an LMS-based server, very interested to see what you decide on this one. I see that IQAudio now do a version of their pi-DAC with balanced outputs. That was going to be my first port of call.

Colin,
Yes I have to say that over the years LMS & Squeezebox has got to be my most long lived and enjoyed bit of tech! I wake up every day to the Boom. I was wondering about modding it but given the tech has moved on so much I was thinking of a Pi and DAC combo. Volumio comes to mind with the Kali & Piano. I have also done a coulple of PiCore builds

How did you route the wires from the binding posts to the speaker?

I got my router out an created a slot on the underside and then filled it.
 

Attachments

  • IMG_2591.jpg
    IMG_2591.jpg
    811.6 KB · Views: 555
  • IMG_2595.jpg
    IMG_2595.jpg
    714.1 KB · Views: 528
  • IMG_2609.jpg
    IMG_2609.jpg
    660 KB · Views: 529
Hi Gary,
I used Frugalhorn with Alpair 7p drivers for quite some time on the end of my Modulus-86 amp too. Then I tried the XL's with 10p drivers, both were very good indeed so you must be having fun. Just recently sold both pairs as they weren't getting used much and the new owner is delighted.

Bill, glad you're making progress, it was you who set me on this neurochrome path so it's about time you got yours finished!