This is easier said than done. I would think I would/could extend beyond the diode limits. Anyone have more experience with this?
SoundEasy speaker measure, model, design, optimize...
Once you set up your CAD network, SE will allow playback through your designed filters, 1 ea ch, woofer, mid, twe...
You would need to use larger diodes, maybe big 4 terminal bridges, but you could sell your completed boards here and cover the cost; people would want them.
Diodes have a very nonlinear impedance that changes with current flow, so I think this is wrong. But lets wait for others to chime in.
rgrayton said:
Diodes have a very nonlinear impedance that changes with current flow, so I think this is wrong. But lets wait for others to chime in.
OK, let's see if I have this right. I guess the V drop across ea diode is basically inversely proportional to the temp or current. If there is basically a short between ea bridge input, then the hot one will always take from the others. Wouldn't a poor man's solution be a 1.ohm 5W at each node of ea bridge? I will take the pwr hit.
My first use will be for Sound Easy speaker design, then who knows, probably some HT use. See post #18 for schematic. Ground scheme is star ground: chips to star, ps to star. More pics here.
Gargan8 pics
Sorry, but I’m not familiar with Sound Easy design. Could you enlighten me what is it?
Do you have 8 single ground points?
If I'm correct you don't use feedback caps in parallel mode operation and connected feedback grounds (two 680Ohm resistors) together, likewise you don’t use input cap?
I also have smth similar, but with 4 ch only (paralleled 4780) and 2 PS. When I designed it according to National schematics it worked pretty well, however with low frequency distortion on low power (loud) level in some cases (when there is low frequency buzz reproduction). Than I redesigned it to carlosfm 3886 schematic, similar to yours you posted (paralleled 4780) but with other resistors values, snubbers (Zobel) and 3,3uf cap between voltage rails on the chip. The last is not working. Chips are drawing much current (because toroid humming like never in normal use) and speakers reproduce only hum (like radio signal). My first thought it’s because without feedback caps.
Do you have 8 single ground points?
If I'm correct you don't use feedback caps in parallel mode operation and connected feedback grounds (two 680Ohm resistors) together, likewise you don’t use input cap?
I also have smth similar, but with 4 ch only (paralleled 4780) and 2 PS. When I designed it according to National schematics it worked pretty well, however with low frequency distortion on low power (loud) level in some cases (when there is low frequency buzz reproduction). Than I redesigned it to carlosfm 3886 schematic, similar to yours you posted (paralleled 4780) but with other resistors values, snubbers (Zobel) and 3,3uf cap between voltage rails on the chip. The last is not working. Chips are drawing much current (because toroid humming like never in normal use) and speakers reproduce only hum (like radio signal). My first thought it’s because without feedback caps.
One ground point, star ground. All PSs and amp pcb grounds are connected to that with symetrical ground lengths. In other words, the the distance and amount of copper between any cap and any pc board to the star is the same. I am not an expert. I only know I did have a ground problem and poked around on the board.I don't get any hum, unless coupled from some other equipment (PC). When I run from just a cd player and preamp I don't get any hum.
I am being advised to add an input cap, so I probably will.
I have not really run this through its paces, so I can't really talk to a lot of stuff. You mention a lot of current. I just had it on for the first time for around an hour with no input. The heat sink seemed to hot to me, so I don't know.
Sound Easy is spkr measure, design CAD/model, optimize and digital filter playback through PC.
Sound Easy
I am being advised to add an input cap, so I probably will.
I have not really run this through its paces, so I can't really talk to a lot of stuff. You mention a lot of current. I just had it on for the first time for around an hour with no input. The heat sink seemed to hot to me, so I don't know.
Sound Easy is spkr measure, design CAD/model, optimize and digital filter playback through PC.
Sound Easy
I’m not expert either.
I thought it’s not allowed to connect grounds from two separate transformers (with multiple PSU) into one point. As I guessed earlier, it must be at list two star grounds for two transformers.
So, your 8ch sound source is PC sound card, as is mine. You could use your amp for playing Blu-ray discs with 8ch HD sound tracks. That’s must be amazing. You could even use 16ch when 4780s are not paralleled. Just for future.
You said that your voltage rails are 36V, quiescent. Than I suppose you supplying chips with 40V PSU without load? You should have 157W of power/ch.
Did you isolate back of chips from heatsink?
What electrolytic capacitors, rectifiers and kind of speakers do you use?
Your ground problem was related with unsymmetrical ground wires?
What did you do when had at first lots of current? Was there any sound from speakers under such conditions?
Thanks.
I thought it’s not allowed to connect grounds from two separate transformers (with multiple PSU) into one point. As I guessed earlier, it must be at list two star grounds for two transformers.
So, your 8ch sound source is PC sound card, as is mine. You could use your amp for playing Blu-ray discs with 8ch HD sound tracks. That’s must be amazing. You could even use 16ch when 4780s are not paralleled. Just for future.
You said that your voltage rails are 36V, quiescent. Than I suppose you supplying chips with 40V PSU without load? You should have 157W of power/ch.
Did you isolate back of chips from heatsink?
What electrolytic capacitors, rectifiers and kind of speakers do you use?
Your ground problem was related with unsymmetrical ground wires?
What did you do when had at first lots of current? Was there any sound from speakers under such conditions?
Thanks.
No, 36 w/o, 33with.
The way I figured my pwr was across 4ohm load, driven w/1kHz tone input. I used a dvm to measure vrms, and a scope to visually see the output clip. It clipped at 21.5 vrms across the load. I have no idea if this is the "correct" way to measure it. It just made sense to me. I write code for a living
I built a kit, formly availble from Brian on this board.
Speakers, I design from scratch, the motivation from 8chs.
When I went to 2 xfomers, I got a hum. I did not have a star connected at the time and the only ground they got was from the source input. This was not going to work. With my prototype using a single toroid, it did not seem to matter. So, I went with a star. The problem was really that the 3 toroids were referenced to each other only through the input source, 12 ft away. It may be that an unsymetrical combination would work just as well. Since I was doing it, it really did not take any more work, so why not?
Haven't really had lots of current, so I maybe we have a disconnect, here. I had the thing powered up for an hour and it seems to be hotter than I would expect, but what do I know? -bg
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.