GCSS boards? Brian?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Since Nelson Pass has given his blessing to produce a GCSS board for the diy community, I was wondering if Brian (or anyone else?) is working on one? I could really use 6 channels of them for an active dipole speaker I'm building (currently using a collection of amps, original GC board from Brian for the tweeter, NAD C272 for the mids and a Rotel RB850 on the woofers). Would love to put it all in one box, including a 6 channel attenuator, sort of a monster integrated amp ;)

Regards,
Ron
 
Yes.

I have one in the works one that uses a pair of LM4780's per channel (the two amps per chip paralleled). It's in the GCSS thread. In fact, the answer to your question is over there. :p This would probably be great for your bass-amp but overkill elsewhere.

Ironically, I'll be using mine in a dipole setup (as yet not built).

I believe I may be (finally) assembling a parts order this weekend for the rest of what I need but we'll see. I still don't have any idea if it's worth farming out the PCB or not - depends how much interest there is, whether there are a few folks wanting proto boards to test-fill. If BrianGT is interested I have already expressed a willingness to work on getting a usable design out and allowing it to be used in a GB - I can order boards at any time.

C
 
cjd said:
Yes.

I have one in the works one that uses a pair of LM4780's per channel (the two amps per chip paralleled). It's in the GCSS thread. In fact, the answer to your question is over there. :p This would probably be great for your bass-amp but overkill elsewhere.
C

I saw your posts in that thread. I was really wanting to know if Brian will proceed with producing boards, as Nelson has given his okay.

I'm mainly interested in the distortion reduction from this circuit for the mids and tweeters as opposed to the overall power (I agree with you that I wouldn't need that much power on the tweeter, the mids might need it, there's a lot of eq needed for dipole).

cjd said:


I believe I may be (finally) assembling a parts order this weekend for the rest of what I need but we'll see. I still don't have any idea if it's worth farming out the PCB or not - depends how much interest there is, whether there are a few folks wanting proto boards to test-fill. If BrianGT is interested I have already expressed a willingness to work on getting a usable design out and allowing it to be used in a GB - I can order boards at any time.

C

I'm interested in boards, to simplify construction. I'd be tempted to go with the 3886 for heat dissipation though, although I have several of each chip on hand.

Ron
 
I have been giving serious consideration to doing a 3886 variant (single chip, not paralleled) so perhaps I will dig in and do that. Heck, if my work-day doesn't change I just may have something later. :p Never know.

Part of my wait has been realizing that to really test it the way I want, I need more than just the amp. :) So, I have a smattering of other small projects that sidelined the main effort. Of course, I also have so many projects that I tend to take forever anyhow. Fine when it's just me, but not so fine where others are waiting. This will probably give me a little motivation in this direction again though. :smash: :smash:

I know the interest is there at some level - my work was not specifically intended for broader use but I do not mind contributing. But since it seems otherwise quiet, I figured I would pipe up at least.

C
 
Retired diyAudio Moderator
Joined 2002
I have a dual LM3886 layout that I made a while back from the initial circuit from Nelson. I was thinking about redoing the layout to Terry's (metalman) current schematic using a single LM4780 chip and getting a couple of prototype boards. (might also try sticking to the dual LM3886 chips, due to the higher heat dissipation of the dual packages)

It would be nice to see a few different layouts prototyped for this project created and tested. The dual LM4780 design sounds good in theory, but might take some more time to get working well (best to prototype without pcb first, and then migrate to prototype pcbs).

As for my current project status, the new miniAleph boards are in the shipping phase now from the board vendor, and I will receive them in the middle next week. I am going to start working this weekend to get all of the packages ready to ship out when the boards arrive. Once I get the boards shipped out, I will put more time into this SSGC project.

I would like to recommend that this project proceeds in the same route as the mini Aleph project, getting prototype boards made, and then migrating to production boards. I would urge others to follow the same path. This project is nowhere near as mature as the proven Aleph designs, so multiple prototype board runs might have to be done. I have been paying about $80-90 for each prototype pcb run (1 day leadtime, 2 day shipping from AC), which has been accounted for in the production board prices. It would be wise to follow the same model for a GB of SSGC boards.

I would like to consider this project in prototyping mode, and no formal GB organized until working prototypes are completed and tested.

--
Brian
 
Retired diyAudio Moderator
Joined 2002
cjd said:
Yes.

I have one in the works one that uses a pair of LM4780's per channel (the two amps per chip paralleled). It's in the GCSS thread. In fact, the answer to your question is over there. :p This would probably be great for your bass-amp but overkill elsewhere.

Ironically, I'll be using mine in a dipole setup (as yet not built).

I believe I may be (finally) assembling a parts order this weekend for the rest of what I need but we'll see. I still don't have any idea if it's worth farming out the PCB or not - depends how much interest there is, whether there are a few folks wanting proto boards to test-fill. If BrianGT is interested I have already expressed a willingness to work on getting a usable design out and allowing it to be used in a GB - I can order boards at any time.

Your design look great, and definately worth prototyping. I wouldn't consider the amp overkill, and probably much better suited for lower impedence loads, rather than a single pair of LM3886 chips.

--
Brian
 
Hi Brian,

This all sounds good. I would really encourage you to go with the 3886 for heat dissipation reasons.

I agree with you re low impedances. Some of the tweeters I'm using are rated as 4ohm, but more like 3ohm. Also, paralleling a couple of 8ohm mids gets you down there too.

Do you foresee any problems in using Metalmans 3875 design with the 3886?

Any idea on when you might be in a prototype stage? I'm available to help test if you need, I'm running balanced out from a DCX2496, with balanced attenuator (building that up right now).

Cheers,
Ron
 
Retired diyAudio Moderator
Joined 2002
transducer said:
This all sounds good. I would really encourage you to go with the 3886 for heat dissipation reasons.

Do you foresee any problems in using Metalmans 3875 design with the 3886?

Any idea on when you might be in a prototype stage? I'm available to help test if you need, I'm running balanced out from a DCX2496, with balanced attenuator (building that up right now).

As for the dual LM3886 design, I have this already partially done, so I am considering just modifying this layout and using it for the first prototype. The LM3886 is almost exactly the same as the LM3875, and is not much harder to implement. The only additional requirement is an extra resistor and capacitor for the mute pin input.

I will see what i can work up.

Also, the LM4780 is simply 2 - LM3886 dies in a single package. The only difference is the lower heat dissipation figure, due to the smaller service area vs. 2 - LM3886 ICs.

metalman said:
The 3886 should be a straight drop in substitution, with the exception of the pin-out.

Brian, when do you think you'll have some time. Got a few ideas we should discuss.

Terry

You are correct about the LM3886. I will probably some free time at the end of next week. I will drop you an e-mail.

--
Brian
 
A "fun" day at work means I have adapted the layout I used for the LM4780 to the LM3886. Meaning, I tore out most of it and re-did it. :clown:

3.050"x2.475" at the moment though I am sure there is room to trim - I did not look particularly at capacitor options for the input cap. In fact, picking components and how much room to leave is probably the one part I'm least sure about. :)

I'll try to remember to post a quick snapshot of this tonight when I get home. I can also post the current LM4780 version.

And, the ability to drive a more difficult load is one of my primary concerns. :) Many of the speakers I design end up being compromised just to manage impedance.

C
 
Well... as I said, it's a first shot just for the heck of it.

This will fit a big film/foil premium cap (you know, the one the crossover guys like to argue about) and perhaps that is not what should be used here. :). There's room for tweaking and improvement on the LM3886 version I think.

LM3886
susyLm3886.gif


Paralleled LM4780
susyLm4780Pv2.gif


I have some other thoughts that may find their way into these, but not sure at this point.

C
 
I already started working on a prototype, using my new LM4780 board. This board allows conection of both amps (in a chip) separately, so the actual GCSS needs only front end to be attached. This can be done pretty easy and I will be able to do it tomorrow. If this works fine, I will compare Terry's version to front end using 2SK389. I will not go LM3886 route, as for me, LM4780 is much more practical: if you want to push the limits, just parallel two of them. That's another possibility: why not creating separate front end board and scale the output according to power output demands. One can use as many as four LM4780 if needed.
 

Attachments

  • 2.jpg
    2.jpg
    85 KB · Views: 367
Originally posted by Peter Daniel That's another possibility: why not creating separate front end board and scale the output according to power output demands. One can use as many as four LM4780 if needed. [/B]

There is sense in this - however, there are still a couple elements that exist as it currently is being used by folks that do not exist on a basic LM4780 layout.

I actually started with a simple LM4780 layout I did and considered this. It's small - 2"x1.5" - and decided that it made as much sense to put it on a single board.

My other concern is that it compromises the circuit some by the simple fact that you have to connect across boards. Grounding and all.

Of course, that doesn't eliminate the potential or the value in persuing this approach!

C
 
cjd said:

My other concern is that it compromises the circuit some by the simple fact that you have to connect across boards. Grounding and all.

Building a circuit in a 3-dimensional plane may not be everybody's cap of tea, but it is my view that this approach can actualy lead to less compromised layout, as the connections can be simplified, especially the grounding. However, for mainstream approach, it is probably not practical. It is however very suitable for prototyping.

I used that approach previously, and it worked just fine: http://www.diyaudio.com/forums/showthread.php?s=&threadid=39582
 
Peter Daniel said:


Building a circuit in a 3-dimensional plane may not be everybody's cap of tea, but it is my view that this approach can actualy lead to less compromised layout, as the connections can be simplified, especially the grounding. However, for mainstream approach, it is probably not practical. It is however very suitable for prototyping

You summed up my thoughts perfectly. :) I would go so far as to suggest that most people don't deal with three dimensions particularly well. I suppose that goes for grounding too (and in this last, I must include myself I think.)

Now, when you finally get to 4 dimensions. . . heh.

C
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.