Modulus-86 build thread

Neurochrome.com
Joined 2009
Paid Member
Is it possible to set the gain even lower than 20dB?

20 dB is the lowest gain for the MOD86. Higher gains are possible.

In the simulator, I was able to reduce the gain of the MOD86 to lower values, but these circuits did not work well in the lab. The LM3886 has a few mood swings as the output voltage approaches the supply rails. I suspect the saturation clamps (baker clamps) are to blame. Regardless of the cause, these mood swings aren't modeled by the LM3886 model deck published by TI and result in a discrepancy between modeled performance and actual performance. In case of the lower gain options, it resulted in instability near the supply rails.
This would apply to all LM3886-based circuits and likely the LM3886 derivatives as well (LM3875, LM3876, LM4780, etc).

Tom
 
20 dB is the lowest gain for the MOD86.
Really? 'ere, let me quote the product page (emphasis added):

Neurochrome Audio said:
Should a lower gain be desired for further optimization of the system gain structure, the Modulus-86 circuit supports the use of the THAT1203 and THAT1206 for a total amplifier gain of +17 dB (7x) and +14 dB (5x), respectively.

Perhaps there is some confusion over end to end total gain versus gain of the power stage versus local gain of the chipamp. The latter is necessarily 20dB at the intercept for most of National's chipamp family due to their design. However, if the power stage (control device + chipamp) operates at unity it's not particularly difficult to guarantee a few volts between peak output and rails, thereby avoiding the gremlins around onset of clipping.

It would be interesting to see distortion measurements with a lower gain to be able to compare it with the LPUHP v2 at low power levels with a similar noise floor.
I'm not seeing an LPHUP v2 schematic on a quick search. Looks like the v1 is 4dB on the frontend and unity thereafter. With the LME49990s and resistors shown in the schematic this is about 2.4uV into the power stage versus 3.1uV from a THAT 1206. That suggests the Mod is going to be ~22dB noisier due to having 20dB power stage gain. So there may be some difficulty with the similar noise floor bit.

Expect also considerable difficulty obtaining the distortion products for any properly executed composite design (which the LPUHP is not since the 49600s aren't within the 49990 diff amp's feedback loop). More about that in the earlier parts of the Mod thread in the vendor's bazaar.
 
Last edited:
Reading the design document and the orther Modulus 86 thread on this site, the reduced gain comes from use of front-end THAT12xx chips with lower gain
Yes. If one needs lower gain it is necessary to insert a passive loss stage at the input. Fortunately such things are not difficult either to design or to implement, and "ready made" devices that even include a handy "variable" feature are readily available from both Digi-Key and Mouser.
 
Neurochrome.com
Joined 2009
Paid Member
Reading the design document and the orther Modulus 86 thread on this site, the reduced gain comes from use of front-end THAT12xx chips with lower gain, not by reducing the gain of the LME49710/LM3886 circuit.

Yep, hence the distinction between end to end (total) gain of the amplifier and the power stage gain. For those not familiar with THAT's part number conventions the 1200 has 0dB gain, the 1203 -3dB gain, and the 1206 -6dB. Thus the figures quoted in post 843.

Sorry. I tend to get rather focused on one specific context and forget to take other contexts into account. It is correct that by swapping out the THAT1200 for a 1203 or 1206, the end-to-end gain can be reduced. My bad. I was thinking specifically about the stability of the LM3886+LME49710 Composite Amp, hence, my previous answer.

The gain of the LM3886+LME49710 Composite Amp is 20 dB minimum. Higher is possible. Lower is not.

The default gain of the Modulus-86 and Parallel-86 is 20 dB. Lower (17, 14 dB) is possible by swapping out the THAT1200 for a 1203 or 1206. Higher is possible by changing a resistor.

Tom
 
Last edited:
Neurochrome.com
Joined 2009
Paid Member
I was thinking about trying MiniDSP products.

There aren't a lot of stand-alone DSP solutions out there, so MiniDSP has certainly found a good niche. That said, I'm not so impressed with the their products or their business model - at least judging from my experience with the MiniDSP 2x4.

I went with the MiniDSP because I'm building the LXmini speakers. I'm building one pair for a friend and one pair for me. He'll get the amp with the MiniDSP built-in. I'll use an older laptop running Reaper as the crossover, driving a quad of Modulus-86 through a Focusrite Saffire PRO24 sound card.

The PC crossover is nice if you don't mind the extra mess of yet another computer floating around, external sound card, etc. For a plug-n-play solution, the MiniDSP provides a half-decent solution.

Tom
 
Neurochrome.com
Joined 2009
Paid Member
Tom, have you actually tried out the minidsp before you reached that conclusion..just curious..

Yes. See my previous post (emphasis added).

There aren't a lot of stand-alone DSP solutions out there, so MiniDSP has certainly found a good niche. That said, I'm not so impressed with the their products or their business model - at least judging from my experience with the MiniDSP 2x4.

The MiniDSP works well enough to satisfy most people's needs, but I see lots of room for improvement both in technical performance, software UI, and documentation. I also disagree with the business model of having to spend $10 for the control software after spending $115 on the hardware. Compare that with, say, Audio Precision's model of giving their absolutely awesome, award-winning software away for free so prospective customers can play with it. After all, the software is rather useless without the hardware. And vice versa.

Tom
 
Last edited:
Tom, you ask 90 usd for an empty board. MiniDSP is a stuck board of far greater complexity for just 80 usd. I don't think you are best placed to judge other peoples business model. It is what the market carries.

In addition, if I may, they know how to get there product to their customers in 2-3 days, in stead of 2-3 weeks.
 
Last edited:
Neurochrome.com
Joined 2009
Paid Member
I can get my boards to you overnight if you wish. As long as you pay the shipping charges...

I deliver state of the art performance. My circuits are well designed, well engineered, well tested, well characterized, and well documented. I charge accordingly. Judging by the feedback I have received from paying customers, many people find my products quite reasonably priced.

Tom
 
Last edited:
Member
Joined 2014
Paid Member
And everyone is free to comment on value. I don't think the software should be extra.

FWIW I have a 2x8 for prototyping I bought second hand. It has 24/96 operation so could be considered the entry level 'proper' DSP. And it's nearly 4 times the price of the miniDSP, but you still have to buy the plugins separately. I would not have paid $300 for it new tho.