• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

Modulus-86: Composite amplifier achieving <0.0004 % THD+N.

I appreciate the effort and engineering that went into this interesting design. Buy for a diy effort it is expensive. $220 for boards and parts, by the time you add the PSU, case, and incidentals you're at around 500 bucks. I'd love to give it a try, but limited budget.
 
Member
Joined 2011
Paid Member
On the topic of price, $299 + $25 shipping in the US, buys the LM3886 based GT-101 kit. In my opinion, its quality is excellent: solid chassis, three quite large and heavy extruded aluminum heatsinks, beefy power switch, nice binding posts, Avel Lindberg 240VA toroidal transformer, good fit & finish. (Why three heatsinks? Because it also includes a regulated DC power supply (!) and the third heatsink is for the big whomper power MOSFETs on the regulator PCB). The schematics and build guide are online, Google will find them easily, if you are curious.

To me, a zip-a-dee-doo-dah circuit improvement on the basic 3886, feels like it is worth maybe an extra hundred bucks or so. To me.
 
Attached is what I would consider a reasonable build budget. I went a bit overboard on the chassis with 6061-T6 aluminum rather than the less expensive 2024. It would probably be less expensive to use one of the chassis available from the outfits in Hong Kong.

Sources:
Electronic parts and connectors: Mouser
Heatsinks: Heatsink USA
Power transformer: Antek
Aluminum: Online Metals

The less expensive amplifier options I've seen so far have all been basic LM3886 amps. I am offering a composite amplifier that beats the performance of a typical LM3886-based amp by an order of magnitude. In addition, it offers dramatic reduction of hum, differential inputs, easy bridging and paralleling, etc. I have pointed to this many times now. If the features and performance of my design have no value to you, then I can't help you.

~Tom
 

Attachments

  • MOD86_Budget.png
    MOD86_Budget.png
    17.8 KB · Views: 1,986
Last edited:
I missed reading the Electronics Now article (referenced above) then, but a quick perusal of the AD711-LM1875 schematic posted at these forums from that article shows that it is significantly sub-optimal - as jcx points out, it throws away a lot of loop gain in a resistive divider in an effort to gain stability

Geez I just noticed that this article was resurrected here, not the finest hour. We got a per page bonus for using any of our parts and a couple of techs wanted to take advantage, one of whom insisted that the public should not be given a circuit that had any chance of oscillating no matter how sloppily built. The circuits worked and the measurements are real, no claims of any sonic nirvana. I expect some slack for 23 years of intervening op-amp evolution.
 
Scott, I enjoyed reading the article. In particular, I liked the outline of the tradeoffs between stability, transient response, and bandwidth. To me, it's worthwhile to trade off a few dB of loop gain in exchange for a stable circuit. With modern parts, you can get 250+ dB of loop gain (at DC), hence, stellar performance of the overall amp even if a few dB are sacrificed to ensure stability of the circuit.

Lots has happened in the past 20+ years. It was time for an update to the circuit. :)

~Tom
 
Member
Joined 2011
Paid Member
... there are three different gain versions ...

Yes I suppose that modifying the DC gain of the front end stage, while keeping the global feedback ratio constant (around the entire composite amplifier), does allow fine-grained tailoring of the Bode plot at the far left hand side. All three variants of the IC appear to have the same input impedance.

_
 

Attachments

  • that1200.png
    that1200.png
    42.6 KB · Views: 1,901
Member
Joined 2011
Paid Member
It just occurred to me that an excellent answer to the question, "Why socket the THAT chip?" might be,
Because I suspect that some of my customers will socket it. I want to test it in this configuration, just in case the extra parasitics from the socket affect the measurements and/or listening results.
 
A few nooby Questions.

Tom,

I have been recently been dipping ny toe into the basics of composite amplifiers.

Can I ask why the output power is only 25W?

I have both 3886 and 49710/20/40. Also the 22V 'new' LM (i forget the pn). I would've hoped for slightly more power.
Recently i have considered experimenting with multiple paralleled 'new' ADA4870 buffers, and these op amps (if i recall the pn)

Why would this 3886 be better at 25W, than a multiple paralleled buffer design, with similar noise and THD behavior?

Finally, how would this design react to additional front end, and bridging?
 
Why socket the THAT chip?

I used sockets for the ICs to be able to test the circuit in blocks rather than have a Big Bang test.

By pulling the THAT, I can test the composite part of the circuit alone. Pulling the LME49710, I can test the LM3886 alone. Pulling the OPA277, I can test the circuit without the DC servo. This is pretty handy for debugging.

Also, as you point out, I wanted to make sure the circuit worked well with the added parasitics from the socket.

I have been recently been dipping ny toe into the basics of composite amplifiers.

Can I ask why the output power is only 25W?

It's totally worth it to go the composite amp route! It's the most direct route to getting good performance from the chip amps.

The output power is dependent on the power supply voltage. The lab supply I was using is limited to 30 V and the "real supply" (toroid + rectifier + caps) was using 35 V capacitors. So I was in a pretty tight spot.

I went back and looked at the data from a previous measurement. That shows the Modulus delivering over 35 W before clipping. This at ±28 V. I think in my most recent setup (that I've shown measurements from), I didn't have a good enough connection between the LM3886 and the heat sink and the power is limited due to the SPiKe circuit protection circuit in the LM3886 engaging.

You should be able to hit the same power numbers as listed in the LM3886 data sheet, provided that the LM3886 is very well heat sunk.

Why would this 3886 be better at 25W, than a multiple paralleled buffer design, with similar noise and THD behavior?

You'll need many, many buffers to be able to beat the 11 A output current of the LM3886.

Finally, how would this design react to additional front end, and bridging?

I'm not sure what additional front end you need. The Modulus-86 will interface directly to a preamp. It is very easy to bridge and parallel the Modulus-86. Two boards can be connected in a bridge by swapping the input wires on one of the boards at the terminal block. You can parallel as many boards a you'd like as well or construct a bridged parallel as shown in the National Semiconductor BPA 200. All this can be done by swapping wires and connecting boards in parallel. No component substitutions necessary.

Any reason not to use an SMPS on this amp?

I think an SMPS would be an excellent choice. I know ConnexElectronic makes some specifically for audio. From what I understand, the lead time is pretty long, though. That's the only drawback.

~Tom
 
Thanks Tom for the info.

I couldn't open/read some of your docs so I asked about bridging a little without studying the provided info. When i get to my home computer ill have a real look.

11A from a 3886? I'm probably wrong but i thought more like 3 A was more reasonable in practice.

Yes many buffers for a opamp Amplifier along the lines of Self's. But the AD buffer is (up to) 1A so not all THAT many ;)
 
Thanks for the article on composite amplifiers.

Hello Tom,

Thanks for the article on composite amplifiers and congratulations on achieving some great results.

With reference to Fig. 5. in the article on composite amplifiers in Radio Electronics, kindly note:

In Amp- A, IC1 has its polarity reversed. Amp- B, is the same circuit, has got it right.

Surprisingly, there is an article titled 'Tuning the AF amplifiers' in Elektor 303 Circuits. There too, the mistake is repeated in the figure for inverting configuration.

Tom, it seems that the magic values for inverting and non-inverting configurations are given in the table one of the article and should be useful for any pre/post amplifier combination.

I, however, agree that there is a lot more to a PROJECT than the schematic.
PCB layout, especially the return lines, choice of components, placement , all matters.

Thanks again and regards!!
 
.................... the 11 A output current of the LM3886............

..........................11A from a 3886? I'm probably wrong but i thought more like 3 A was more reasonable in practice. ..................
The 3886 is specified by National as >=7A maximum output when Tj is 25°C

In practice the variance above the worst case 7A will be balanced by the reduction due to Tj being substantially above 25°C.