Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › M³ Project Announcement
New Posts  All Forums:Forum Nav:

M³ Project Announcement - Page 13

post #181 of 565
9/19/04 at 9:42pm
Quote:
Originally Posted by amb
My feeling is that the ALPS blue will probably be the more popular choice amongst builders simply because the DACT is several times its cost (~$16 rather than ~$65+). As such, we should lay out the PCB for direct-mounting of the ALPS and provide a second set of pin headers not far from where the rear of the DACT would be (for hand wiring to a DACT or other models of pots or stepped attenuators). A short run of wires is easier to deal with shaft extenders, and should not adversely impact performance. I'd like to hear what your opinions are about this.
I think this would be an excellent choice.
Reply
post #182 of 565
9/20/04 at 2:03am
Quote:
Originally Posted by z2trillion
This might have been mentioned already, but would the smaller CP601 work in the place of the CP2500? It appears to have leads that would enable it be mounted to the pcb more easily. Is this a possibility, or is the CP601 inferior to the CP2500?
The 2CP-601 would work, but indeed it is quite inferior to the 2CP2500. The 601's channel matching is rated at +-2.0dB to -45dB, compared to the 2500 which is +-0.5dB to -50dB.
Reply
post #183 of 565
9/20/04 at 2:26am
If this is a no-compromise design wouldn't it make more sense to tailor it towards higher end components? Especially considering the phenominal price we are getting on the DACT attenuators.
Reply
post #184 of 565
9/20/04 at 5:22am
Thread Starter 
It is simple enough to hand wire the DACT CT2. The board should be friendliest to the most popular parts, and the Alps Blue will probably remain the most popular volume control, don't you agree?
Reply
post #185 of 565
9/20/04 at 11:09am
Quote:
Originally Posted by morsel
It is simple enough to hand wire the DACT CT2. The board should be friendliest to the most popular parts, and the Alps Blue will probably remain the most popular volume control, don't you agree?
FWIW I agree with you. Even if this is a "high-end" amp, making the DACT the default volume control would probably scare many people away (and draw others in no doubt )


/U.
Reply
post #186 of 565
9/22/04 at 10:18pm
Thread Starter 
Team M³ is working on opamp power issues. In the mean time, here are a couple of draft layouts to look at. There are no opamp power components yet. The main focus is choice of heat sinks and channel component arrangement. The parallel vertical lines are 2 inches apart.

What is your opinion on ground planes on the bottom layer .vs. the top layer of the pcb? Putting the traces on the top layer and the ground plane on the bottom layer results in fewer traces through solder pads on the bottom layer, and thus easier soldering, but perhaps there are some compelling reasons to have the ground plane on the top layer. Kevin points out that a top layer ground plane sandwiches the traces between the ground plane and metal case bottom.
Reply
post #187 of 565
9/26/04 at 5:21pm
Thread Starter 
M³ v008a


M³ v008b
Reply
post #188 of 565
9/26/04 at 6:06pm
That looks very nice Morsel. I think I prefer the heatsinks on v008a. I think they look better and I like the fact that they're fixed firmly to the board



/U.
Reply
post #189 of 565
9/26/04 at 6:17pm
Thread Starter 
Aavid 531202 heat sinks in v008a are firmly attached to the board with screws.

Aavid 504222 heat sinks in v008b only connect to the TO-220 cases, probably should be drilled since the existing holes are meant for horizontal mounting and result in the heat sink floating in mid air, and probably should be used with insulators since they are so close together.
Reply
post #190 of 565
9/26/04 at 6:25pm
Quote:
Originally Posted by morsel
Aavid 531202 heat sinks in v008a are firmly attached to the board with screws.
That's what I meant, v008a (mistake corrected in the above post)


/U.
Reply
post #191 of 565
9/26/04 at 6:46pm
Quote:
Originally Posted by morsel
Aavid 504222 heat sinks in v008b only connect to the TO-220 cases, probably should be drilled since the existing holes are meant for horizontal mounting and result in the heat sink floating in mid air, and probably should be used with insulators since they are so close together.
I should elaborate this with a picture. As you can see in the picture below, the 504222 heat sink comes pre-drilled with two holes. Using the "top" hole leaves not enough of the TO-220 device's leads to go through the board. Using the "bottom" hole causes the heat sink to sit unsupported about 0.25" above the board (assuming you insert the TO-220's leads into the PCB holes up to the point where it changes thickness). Since the heat sinks "face" each other in the layout, the direction of instability will likely cause the heat sinks to touch each other. If the TO-220 MOSFETs are mounted without insulators we'd get a short circuit from positive power rail to the negative power rail.



If you want the 504222 heat sink to sit against the board in a stable manner, you will need to drill a third hole on the heat sink somewhere between the two existing holes, and not exactly in the center. This has to be measured fairly precisely to achieve the desired result. The bottom edge of the heat sink will probably also need to be insulated to prevent shorting to the ground plane or traces (in case the soldermask wears through).

Another point that I'd also like to bring up is that the two 504222 heat sinks as oriented on the board form a semi-enclosed area. Even though there are slits on the heat sink's wings I think it may trap some heat.

The benefit of the 504222 heat sinks and the the v008b layout is that it allows for very direct PCB traces to the MOSFETs, and saves some board space.

My personal preference is v008a, but your opinions are welcome.
Reply
post #192 of 565
9/27/04 at 12:33am
Haven't been here for a while. Its lookin great! I prefer version a heatsinks as well.

Can't wait to build one of these suckers.

JV.
Reply
post #193 of 565
9/27/04 at 5:27am
Morsel, your layout looks alright but I would make it a bit more tight and get shorter traces for the feedback. I think you should strive to get it as compact as possible. Can't you use only one heatsink for both mosfets? Check the most left heatsink below. I have two transistors plus a temp sensing transistor.

Your holes in the groundplane are hardly necessary if you have fairly low values of your feedback parts. On the other hand it's making no harm either.

Reply
post #194 of 565
9/27/04 at 11:25am
Quote:
Originally Posted by peranders
Can't you use only one heatsink for both mosfets? Check the most left heatsink below. I have two transistors plus a temp sensing transistor.
We considered putting both MOSFETs back-to-back on a single (bigger) heat sink like in your pic, but abandoned the idea because not only must insulators be used between the MOSFETs and the heat sink, the screw and hardware for fixing the MOSFETs to the heat sink will also have to be insulated or non-conductive or else you short out the power rails (The drain pins on the MOSFETs are connected internally to the TO-220 body). Also, it makes for a messy PCB layout, everything has to go around the heatsink.

We don't need a temp sensing transistor in this amp. It is thermally stable without it.
Reply
post #195 of 565
9/27/04 at 1:33pm
Thread Starter 
Quote:
Morsel, your layout looks alright but I would make it a bit more tight and get shorter traces for the feedback. I think you should strive to get it as compact as possible.
Hi Peranders. We could shorten the feedback loop by swinging C6+ around to C6- and sliding R5 away from the bias circuit at the expense of not bypassing the opamp right at the supply pin. Alternatively, we could rotate the entire opamp section 90° clockwise and slide the bass boost section towards the MOSFETs at the expense of bringing the input resistors closer to high current areas.
Quote:
Can't you use only one heatsink for both mosfets? Check the most left heatsink below. I have two transistors plus a temp sensing transistor.
AMB covered the bases, but I'll expound a bit on the messy layout issues. Since it is undesirable to thermally couple MOSFETs from different channels, complementary pairs would be mounted on the same heat sink. The similar pinout of N and P channel MOSFETs results in input and output resistors being positionally reversed since they face outward, so they must cross paths on the N channel side, also crossing the power trace for that MOSFET. The pairs of input and output resistors must lie outside the domain of the dumbbell profile of the heat sink, and then come together again on the far side. The Vbe bypass cap C2 no longer has a niche to sit in, which results in greater distance from the MOSFETs and a longer feedback loop.

While we are on the subject, another advantage of v008b (besides the shortest possible signal paths with or without the suggested modifications) is the option for a longer C2.
Quote:
We don't need a temp sensing transistor in this amp. It is thermally stable without it.
Not only is it thermally stable, AMB noted that switching to the smaller heat sinks actually lowered the bias current.
Reply
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › M³ Project Announcement