M³ Project Announcement - Page 2

Thanks morsel for your answers. Sorry, my bad as I barely looked at the schematic.....BTW remember that sometimes what is obvious for you, is not that obvious for others, and there are a lot of members that would like to venture themselves into a first good DIY project, but they are not DIYers in any way.
Even though I will be really interested in trying this out, definitelly, I own a PPA, and I love it, and I was always curious about this Sheldon D. Stokes project, just that nobody seems to care too much about it, and offer it as a DIYFSE option (other that himself, a very expensive labor BTW)......

nevertheless you should consider 2sj76 / 2sk213 series. A mosfet designed
for amplification isn't the same as one designed for switching. Rds means much less in an amp than linearity. also I don't recall them costing more than a few bucks and easy to get from mcm or wherever.

The spec'd transistors look reasonably linear to me, at least from the datasheet, unless Vgs is very small.

I would be a big fan of a well considered voltage regulation system on the board to give us a greater choice of power sources. If we can take control of the voltage regulation, then we can make it as robust as necessary. Since the circuit board is targeted for such a large size, it seems reasonable to aim for independent regulation for each channel, with, perhaps, some sort of jumper option to use a single regulator.

Maybe a Jung regulator? Is it called that anymore?

-Drew

Very nice work on this long anticipated project!

The LaRocco and Glassman discrete buffers have shown the merit of this approach, having an integrated design - opamp voltage amplifier + discrete buffer - seems like the logical next step.

I think it would be nice if there would be a configuration that could be built in a case smaller than the Par 20 series, especially if there is an outboard PSU. The 1455N series case used in the PPA seems to me an ideal size for a headphone amplifier. I do like the idea of the outboard PSU, and also the idea of more options for board-mounted components.

Whatever you decide, I look forward to building it.

sweet. what an exciting news for the diy audio community. you mentioned it's goign to have a better price and performance ratio than the ppa, but any idea what the price range is goign to be? (200-250, 300-350, etc...)

cant wait to hear it/hear impressions of it.

Hi all,

I am glad that hardly a half day has passed and this announcement already generated so much interest. This is going to be a very exciting project because it combines the essential qualities of the PPA with that of a very robust class-A MOSFET output stage (I love the term "studly" that morsel used in the announcement ), and makes several significant enhancements over the SDS Labs MOSFET amplifier as well.

I built and tested several iterations of prototypes on breadboards, exploring some different possibilities. As Morsel said, we tried a diamond buffer output stage, except with MOSFET output devices. It worked well enough, but ultimately didn't measure as well as a simpler, directly-driven MOSFET topology. Just for kicks, here is a photo of the diamond buffer-based amp prototype:



As for the choice of MOSFETs, we had to consider several factors. Since this amp is going to operate in class A, we would like to choose high current devices for reliability. The Ciss should not be too high, and yet high current devices always have higher Ciss due to their larger internal junctions. Stable and assured availability is also an important factor in our consideration. Linearlity, performance and suitability for the application, and last but certainly not least, the cost is also considered because we want this amp to deliver a lot for the money.

The Hitachi 2SK213/2SJ76 series were considered. These are nice devices with low Ciss, but IMO just a bit "weak" with a Ids rating of 0.5A. The 2SK1058/2SJ162 are stronger, but the IRF devices we chose are even more robust yet with comparable linearlity, Ciss. Plus they are stocked in quantity at Digikey and Newark Electronics. The 2SK/2SJ devices are scant by comparison (at least in the US) and more expensive. The choice then became clear. The Renesas/Hitachi parts are not pin-compatible with the IRFs, so we had to choose one or the other for our PCB layout.

As for the op amps, both Morsel and I like the AD8610. However builders/users can substitute others like the OPA637/627 just like you could with the PPA.

We would definitely welcome more comments about the other open choices as listed in the announcement. In particular the power supply, casing, heat management, etc.

-Ti

I think it would be nice to have one larger heatsink in the center of the case, dividing it in two halves, one half for the amp itself, the other for the PSU w/ transformer.. I think having everything in one box is a better approach for such project, you're choosed rather big casing where there is enough space to prevent input section being interfered by PSU.. as for the PSU, the output section is a Class A beast and as such I'd choose a capacitance multiplier for it and keep the voltages rather high! for the opamp power I'd simply go with 78xx regulator for each channel with 317 as a preregulator for them all.. or some TL431 + pass transistor job..

I don't think those Hitachi MOSFETs are hard to get either.. the question is if they're sonicaly better or not.. considering one bigger heatsing for all MOSFETs, you can add pads for both types of MOSFETs.. it won't take much space..

yeah I know.. maybe you can arrange it like Kevin did with dynahi - output devices mounted horizontaly on L shaped extrusions.. that way it's possible to use Aavids as well as one big heatsink.. I'd just like to have the ability to use big heatsink profile at one side of the board.. keep that in mind when laying out the board, I think it's doable

oh and jung regs are most likely overkill here, opamps have very low current requirements as well as very good supply rejection, it's more important to focus on good bypassing close to the supply pins.. I'd go with 100n polypropylenes.. and also I'd like to have SOIC pads available..

MMM, Mosfets
Have you ever tried a M² version?

ask IpsilonSound and you'll realise there's more then two people wanting SOIC pads regarding my location at the time of discussing PPA v1.1, well I might have been busy with designing the buffers

what about M4 and balanced operation ?

I agree with Glassman that having all the hot parts lined on the side for a big heatsink would be a big plus in my view.