New Millett Hybrid MiniMAX (what happened to this thread?)

maybe you can find some info on the millett max thread?

The thread was removed (by the big boss in management), but we will not give up!!!(DIY FOR THE PEOPLE!)

The MiniMAX prototype boards have been ordered, and building will start up soon I guess. You can view more info here.

Yes, it's too bad, but I respect the wishes of the man in charge. Thanks for that link, btw.

I am still looking for 2 more volunteers to help us build the prototypes. PM me for information if you are interested.

Sorry - all the MiniMAX prototype boards are now taken. Hopefully, someone will start a build thread when they arrive. (Someone has volunteered to do that, too.)

Nooooo, I missed it

I've seen in the link provided that the PS has been redesigned and improved.

What exactly was changed and improved?

Yes I would love to know it too.
As far as I can see, that PS circuit (just like the old MH max) is lifted straight from Tangent's STEP/TREAD PS, with an added ground plane which is connected to the star-ground via a narrow strip (very SOHA II alike ).

Well, since none of us saw the SOHA II board for months, we know it didn't come from there, did it?

As for "lifting", the linear regulated schematic design comes from National Semi's own data sheet regarding the LM317. The layout is totally different from Tangent's STEPS and is derived from the original MAX and the original MAX prototype. Tangent himself has stated that the circuit is not proprietary and is in the data sheet. You would know that if you did your research, Ferrari.

I am sorry that I sent you a PM once suggesting that it was unfair to compare the MAX to the Bijou. You were a valuable contributor to the Millett MAX at the time, often offering technical support to builders in the thread. I felt it was a bit unfair for you to do that while stating publicly that you felt the Bijou sounded better in a direct comparison. So I sent you a note. The Millett family has never intended to compete directly with high voltage, pure tube amps. You seemed to understand the comparison with a B22 was unfair, but couldn't understand that one.

Ever since, your attitude has been shameful and sometimes vulgar. Your accusatory tone is not appreciated in the least.

Colin and I literally did hours and hours of testing. We found that nothing over the existing MAX PS has as much effect as changing the grounding scheme. That's kind of hard to prove out for certain until you make another board. That's why we're prototyping - otherwise it would've been an updated layout and we would've run off production boards from the outset.

Some of this is explained in the DIYForums thread, but suffice to say that the output tantalum, common to other off-board power supplies of this nature, is redundant in the MiniMAX configuration. The ESR of that cap is swamped by in the bank of electrolytics and film caps right after the PS in the Millett circuit itself. Instead, a resistance in the positive feed from the PS to the Millett circuit controls both the PS output impedance and the level of PS noise (ripple). Our tests have indicated it could be as low as ~0.027mVAC or as high as ~0.097mVAC. (Either value is outstanding for a PS mounted on the same board as the amp and with direct AC traces.)

This is not unusual and is one of the reasons that the Millett MAX has somewhat been blazing a trail with a linear-regulated PS board-mtd on the same board as the amp circuit. In many power supply scenarios, there is some distance with corresponding resistance/inductance in the connecting wire/cabling between the amp and the PS. It's possible that the wiring impedance in offboard regulator setups helps to dampen the resonance caused by the regulator output impedance and ultra low ESR of the cap bank in the amplifier it typically serves. In the MAX, that connection is a short trace of less than 1/4". We have been able to control the PS noise in the MAX by removing the ground plane under the PS completely. For instance, the original MAX protoypes had line hum and ripple that were reduced below the level of audibility. This strategy has worked well - the regular MAX's power supply is rock-stable. Bass is impressive, as almost any MAX builder can attest.

However, Cetoole has felt that the power supply noise could still be incrementally improved by using an improved grounding scheme along with the right combination of components. Cetoole's MiniMAX layout is the result of our testing in that regard. (Also with the help of Tangent's LNMP.)

The tradeoff is in regulator output impedance vs. ripple. As the ripple is reduced by adding series resistance, the output impedance is increased. This has the effect of losing some stability under a varying load. (Note that all of our testing is under the constant load of a 50ma-biased MAX with tubes lit.) What happens is that the ripple climbs slightly when a varying music signal is applied, and it can vary up and down by a few hundred uVAC. This has the effect of slightly softening the bass. However, the effect is very negligible and can be difficult to detect without direct measurements.

Bottom line, there is an optimum series resistance, given the grounding scheme, that's somewhere between 1 ohm and perhaps milli or micro ohms (a long jumper). The optimum will result in the lowest noise, while still maintaining rock-hard stability under a varying signal.

EDIT: The first MAXes and MiniMAXes that I had modded for low-noise were demonstrated at the Atlanta Mini Meet back on July 5th. You might check that thread for some listening impressions.

So a 0R47 resistor that I have lying around from Beta22 build would work perfectly right?

It's possible.

Since some accusations have been thrown out, perhaps a bit more insight into the long history of these changes is warranted.

Note the photo below. Sans arrow, this is the same photo that has been posted on the DIYForums MiniMAX page since February 24th. (bottom right photo link on that page)



The arrow is pointing to an 1/8" cut that I dug out of the top of the board and through the ground plane. That cut, with a few zigs and zags, goes all the way in front of the PS section, and continues on to the other side of the board. The ground plane in the PS is completey separated from the ground plane in the amp circuit. Underneath, however, is a small jumper in the center that connects the two ground planes through the pins of the component parts above.

Further, the suggested ground plane cut had been drawn out by cetoole in his post made over a year ago - back on March 13, 2007: DIYForums.org :: View topic - New Millett Hybrid Layout
In that post, you may note that cetoole had left a small portion in the middle uncut. I modified it slightly on my actual build by making the cut continuous and simply adding a jumper between the two ground planes.

The board in the pic above and the layout in the link is from the very first set of Millett Hybrid Maxed prototype boards that were ordered and received back on February 22, 2007. It's just that it took some of us awhile to get around to doing the surgery on the old boards. I was the first one to actually case it up in the Lansing case and posted pics back in February of this year, but the board is one of those that arrived in Feb, 2007. I felt like I noticed a difference with the Mini, but it was only recently that I built Tangent's LNMP (posted in the Pics of Your Builds thread early last week). Whereupon, we confirmed the outstandingly low noise of this grounding scheme (and continued testing almost constantly between then and the announcement of the MiniMAX).

I trust that's sufficient proof of the heritage of the MiniMAX design by cetoole (using Pete Millett's legacy, of course!). So let's get back to DIY-ing.

I've posted the layout, schematic, and BOM on the MiniMAX home page. Hold down the -Ctrl- key while re-freshing if your browser has the old page cached.

I'll fancy up the website to match the MAX's site as time goes on.