| soundNERD |
I know the 1875 is a very picky chip when it comes to pcb layouts.
I designed these PCBs with Eagle. How will they work?
There are 3 sizes of square pads. The largest are input, medium are power input, and the smallest (2 of them) are jumpered together. They are from one of the 100uf caps to pin 5.
I know the copper.bmp is of bad quality; thats just how eagle exported it. to see it properly, you will need to open it in eagle.
I'm new to Eagle. Can somebody tell me how to use a part from a library? When bigparsnip sent me that 3886 amp, there was also a lib file with the 3886 pinout. How to I use this file in a PCB design?
Thanks, Mike |
|
|
| GregGC |
| quote: | Originally posted by soundNERD
I know the 1875 is a very picky chip when it comes to pcb layouts.
I designed these PCBs with Eagle. How will they work?
There are 3 sizes of square pads. The largest are input, medium are power input, and the smallest (2 of them) are jumpered together. They are from one of the 100uf caps to pin 5.
I know the copper.bmp is of bad quality; thats just how eagle exported it. to see it properly, you will need to open it in eagle.
I'm new to Eagle. Can somebody tell me how to use a part from a library? When bigparsnip sent me that 3886 amp, there was also a lib file with the 3886 pinout. How to I use this file in a PCB design?
Thanks, Mike |
Can't really see it very well (the BMP file). |
|
|
| soundNERD |
| Yes, you need Eagle to see it. It didn't export properly. |
|
|
| soundNERD |
OK, heres the GIF of the schematic
and next post is the board
i've been having pc problems lately with a new vid card, so I couldn't re-export it.
On the PCB, there is a jumper from the left 100uf cap to pin 5 of the chip. |
|
|
| soundNERD |
Heres the PCB
Also, on the schematic, I forgot the gnd symbol from the speaker. |
|
|
| soundNERD |
and heres the silk layer
thanks! |
|
|
| GregGC |
| quote: | Originally posted by soundNERD
OK, heres the GIF of the schematic
and next post is the board
i've been having pc problems lately with a new vid card, so I couldn't re-export it.
On the PCB, there is a jumper from the left 100uf cap to pin 5 of the chip. |
First thing and VERY important is: you need a resistor from +in to GND. Use the same value as R1 ( I think 22k). Without it you'll have nasty DC offset at the output.
If you don't have DC coming from your source, replace C1 with 1K resistor and then you don't need a res to GND (that's what I'd do if I want to keep C7 in place). This way you have one cap less on the way of the signal. |
|
|
| soundNERD |
Greg,
Isn't that what the pot does? it is a resistor from input to gnd.
Also, how does the actual board look?
Thanks! |
|
|
| peranders |
| quote: | Originally posted by soundNERD
Greg,
Isn't that what the pot does? it is a resistor from input to gnd.
Also, how does the actual board look?
Thanks! |
You must have a DC path from each input to "somewhere" because inside you have ordinary NPN transistors and they must get base current. |
|
|
| macboy |
| You do not need C1 because C7 serves the same purpose: blocking DC at the input. As suggested by Greg, replace C1 with a 1 K resistor and just forget about adding the new 22 K to ground. |
|
|
| peranders |
Macboy, I'm sorry but you are partly wrong here. The input cap block DC but the C7 reduces the output offset mainly. 1 V DC in gives you 1 V DC out also.
There is one advantage to have an input cap and this is that you will get constant output offset due to constant DC resistance from +in down to ground. If you are lucky this offset voltage may not be so big so there may not be any trouble when you turn your volume pot. |
|
|
| GregGC |
| quote: | Originally posted by peranders
Macboy, I'm sorry but you are partly wrong here. The input cap block DC but the C7 reduces the output offset mainly. 1 V DC in gives you 1 V DC out also.
There is one advantage to have an input cap and this is that you will get constant output offset due to constant DC resistance from +in down to ground. If you are lucky this offset voltage may not be so big so there may not be any trouble when you turn your volume pot. |
Having C7 in place gives you DC gain of 1 (otherewise it will be = to the ac gain) so your output offset will depend interely on the dif of the I bias. For those small resistor values and DC gain of 1 I don't think you'll get DC offset variation more than 1mV at the output (for volume setings between min and max). As far as C1 I'd get rid of it (if no DC on the input) and replace it with 1k to provide some resistance when the pot is at 0 vlome. My GC has 10mV variation of the DC offset (-1mV to 9mV) and the DC gain is 22.
So I'd even get rid of C7 and put up with 10mV DC offset but no Caps at all on the way of the signal. Much better than no DC offset at all but I have to start looking for Wonder-caps. |
|
|
| soundNERD |
Thanks for all the suggestions about the parts.
Would somebody mind telling me what they think about the layout of the PCB?
Thank you!
-Mike |
|
|
| Aud_Mot |
soundNERD,
On this forum, you ask for a glass of water and some one starts telling you why Coke sucks.
I do not understand why you use larger pads (and I assume large pads are for larger thru holes for larger cables) for inputs and smaller pads for power and outputs.
I am also not a fan of running traces around a board when it is not needed. I beleive that the big hunk of copper in the middle of the board is your ground. You have it going up the left side of the board with no pads there. The ground traces you have drawn remind me of some of the PCB designs I did for microwave applications, where the copper trace was a capacitor. The trace that looks like it goes to pin 1 of your 1875 is surrounded by ground traces top and bottom and looks like a cap to me.
Some people hold to the theory that all ground signals should be a trace that simulates a wire. One end is the component end, the other end goes to a star ground point. The theory is that you do not want the ground signals to mix with other ground signals. Each component's electrons have their own copper highway path to go down. This is a theory first proposed to me by a famous high end audio designer, sorry - I do not have his permission to identify him.
Some people like to do what you seem to be doing. Swamp the ground with as much copper as possible. If you wish to take this approach, I would recommend that your hunk of ground pattern is roughly defined by the pads you are connecting to ground.
My 2 cents for what the are worth.
Aud_Mot |
|
|
| soundNERD |
Thanks!
I will get rid of that unneeded trace.
About the input pads, I was new to the program, and put in the first thing I could find into the schematic.
I didn't think it would be so big on the actual board.
If I remove that unneded ground, and leave only the center, would the board be good and not oscillate like I know the 1875 likes to do?
-Mike |
|
|
| theChris |
well, i'd try to get the 0.1uF as close as possible to the +V and -V pins -- they are used soley for low impedance at high frequencies, and there's no need for extra resistance and inductance by moving them further away from the pins.
the R1 feedback resistor could be a little closer to the output pin to minizime common-mode inductance (inducatance common to the feedback and load).
there needs to be some connection to ground after the cap.
i think national reccomends a series resistor on the non-inverting terminal to limit current into the chip during power down conditions. or something along that line.
i'd try more for a parts-farm and IO area -- most all the parts near the IC, and all the connectors in a nice open, unobstructed area. you could likely minimize the size this way too, though the size isn't too big from the looks of it. |
|
|
| Aud_Mot |
Will it oscillate?
That is one of the trickest questions you can ask on this forum.
I do know that the fewer extra traces you have on your board, the fewer stray capacitiances you will create on your board.
That translates into "probably" a more stable design. It could be your original design might turn out to be less prone to oscillate. Unless you are going to go with a true ground plane, keep the traces as econimical as possible is my rule of thumb.
Aud_Mot |
|
|
|
