M³ Project Announcement - Page 23

At the very least, the positive leads for the caps are a square shape around the hole.

-Ed

hrmm shame about the extra output pads going. i woulda quite liked that. no big deal tho really...

I think the parts labels are a matter of taste. One could argue that it's better to have the label unambiguosly within the component outline rather than not be certain of which part is referred to by a label, especially when space is tight.

I'm looking forward to this project. So many amps to build, so little time.

as i said nbd, it was just for dual outs, i will just run one set of wires to a couple of pads on stripboard near the output jacks then wire the jacks to the stripboard.

The second output pads could include pads for output resistors, normally shorted but enabled by cutting traces. Then one could wire up two headphone jacks, one of which offered Etymotic 4p to 4s 75 ohm conversion with no additional switching contacts in the signal path.

This way, the two outputs can have different impedances, not as easy an option when daisy-chaining.

Ety 4p's are my best headphones other than my electrostatic Stax. I have no idea how many other people are in my boat. I realize that the MMM is not intended for portable use, but I wish that shorted output resistor pads were a standard idea for anyone designing new amp boards. Wouldn't this help squish the "fried buffers" PPA debate, if this were also an option in PPA v2?

Some people like to add impedance to other phones, but this makes a crucial difference in the sound for Etymotics; the 4p is a compromise for portable devices. Their conversion cable is $65 and introduces a new contact.

Here are a couple of quick snapshots of the blank prototype pcbs. As Morsel stated, it doesn't have any silkscreen or soldermask to save cost (the production boards will have these).

Top side:


Bottom side:


Syzygies, as for output resistors for a second jack, you can still use the daisy-chaining idea and just bridge the resistors between the jacks. I think that's better than cutting board traces.

Nice!


/U.

six of one really..... my jacks are board mount not solder lugged so i generally find it easier to only solder one wire to them.

If you will have them board mounted there should not be much problem to mount impedance resistors and connector wires on that small board, so you would still only need to have one set of wires to the M³ board.

PSU questions
If you build your own PSU (I guess quite a few of us will do that), what would be the prefered way to build it?
Opt1: Simplest possible: just transformer+rectifier bridge+caps
Opt2: Use 7824 regulator (would 1.5A rated do fine?)
Opt3: Something a bit more advanced. Preferably with link to schematics.

What voltage would be best? 24V? 28V? or would anything 20-28V work fine?
What VA is suggested? A continuous current draw of 300mA-500mA has been mentined in this project, so I guess around 12VA would be the absolute minimum. Usually a slight bit overkill is nice to have some margin, but any suggestion both on minimum and maximum that will make a practical difference would be nice.
What capacitance is recommended in the PSU?

Bias
So far all bias is regulated with trimpots, and that is fine at this point.
For the finished board I guess that the need to adjust bias would be very little. I would guess that at least 95% of users will never want to adjust the OP-amp bias. Adjusting the output stage bias could be usefull depending on what kind of phones will be used vs if you like to reduce heat output a bit.
So, I guess it would be convinient for most people to just use a plain resistor instead of pot. This can be done allready if mounted as standing resistor.
In a layout this big and nice I would prefere if resistors did not have to stand.
Would it be possible with small means to make room and holes to mount plain resistors flat (at least for opamp-bias)?
Ofcourse this would only be usefull if you can figure out that appropriate value for fixed resistors during your testing with prototype PCB.

This is strongly discouraged as the M³ pcb has no onboard voltage regulation. There is only a capacitance multiplier serving to stablize and isolate the opamp power rails.

This is ok, but the LM317 or LT1085 adjustable regulators have better performance.

I recommend something like the STEPS. You could of course use fancier designs if you wish. You could probably also get by with an Elpac WM080 regulated wallwart (available from Newark) but you'd be operating near its maximum rated current.

With the AD8610 opamps you should have no more than 27V. With OPA627s you could go up to 36V. These are absolute maximum ratings for the opamps but since the capacitance multiplier will drop about 1.6V, and if you use the D+ input on the M³ pcb, there will be an additional 0.6V drop, that will give us enough margin. I would not recommend anything less than 24V to maintain adequete voltage swing.

If you're going to design your own PSU, I suggest the Amveco 700xx series PC-mount toroidal transformer, either 25VA or 35VA. These are available from Digikey.

It's up to you. I like to have ample amount of reservoir capacitance. Just be mindful that your power switch and rectifier diodes could handle the inrush current at turn-on.

You need to adjust this with a trimpot because there is no reliable way of knowing what resistor value will give you the desired CCS current. The Idss characteristics of the cascoded JFETs will influence the current. The only way that you'll be able to eliminate the pot is to use a CRD in place of the cascoded JFETs, with a slight loss in CCS performance.

Here too a trimpot adjustment is necessary because MOSFET device variations will influence the outcome. If we put a fixed resistor there, on one amp it may cause too little quiescent current through the MOSFETs while on another amp, the same value resistor could cause too much current.

I know that adjusting trimpots is inconvenient but there is just no way around this. I will set up a detailed web site for the M³, including an initial powerup adjustment procedure to help guide builders through all this. It won't be that bad, really; and you only need to do this once after building the amp.

Just a little more eye-candy for all to see...
The prototype pcb being assembled.

Lookin good

I'm anxious to hear how it sounds...