PA1002A workaround / replacement

The board will need to match the existing PA1002A 16 pins on the bottom, and accommodate the op-amp and rail splitters on top
thats exactly what i got from your sketch, so that much is clear
The board will need to be compact to fit into the existing space
If the new board kidna hangs off the side too much, we could use those long 0.1mm headers to it would sit above the other parts. like these
HTB1OpUFCFuWBuNjSszbq6AS7FXa5.jpg_q50.jpg


i have a rough idea of a PCB layout. could maybe use SMD for bypass caps and resistors to keep space down. but once i get a full schematic ill start working. with what i could understand from your sketch i did get this far. clearly not finished. might throw together a mouser BOM as well. im excited for this, and i don't even have a need for somithing like this.
upload_2019-1-16_19-31-27.png
 
Yes, that schematic above is correct. What is missing are the connections to the 2 rail splitter parts. Here's the pdf with i/o pins
https://www.ti.com/lit/ds/symlink/tle2426.pdf

Those TLE2426CLP connections are fairly easy as follows
In - VCC (for both)
Common - ground (for both)
Out - one output to pin 6, the other output to pin 7.

The rail splitter is a precision part that divides the supply voltage exactly in half, with better isolation than resistors. This provides the function of the original PA1002A pins 6 ,7

The cap is a low ESR bypass cap, 0.1uf, 50V. I usually use monolithic ceramic caps for this. You could also add a 100uF 25V panasonic FM if it will fit, in parallel with the 0.1uF.
These should be located as close to the ground pin of the op-amp as possible, especially the 0.1uF, which should be small in size.
 
IPa1002A_proto5.jpg here's a photo of my proto adapter
components are, starting from the bottom - machined pin 16 pin socket to TX9500II circuit board, 16 pin DIP solderable proto socket, wires to op-amp and TLE2426 sockets, 8 pin DIP socket and 4 pin sockets for TO-92 devices, op-amp ("testable" JRC 4565 w/0.1uF bypass cap), 2 X TLE2426
 
Last edited:
Working on an incomplete schematic, this is what i got together.
upload_2019-1-17_21-55-39.png
for C1 power supply filter, maybe we could use an SMD 1206 MLCC cap in the 10uF to 22uF range. could maybe even make C2 0.1uF an SMD part as well. 0805 maybe. or stick to the larger, easier to hand solder 1206. the very first incomplete PCB i drew up has the op amp sitting on top of one side of pins. it may be best for have the board hang off to the side a little as i dont see how to fit all those parts on a PCB the size of a DIP package. i didn't want to do the PCB untill i had the complete schematic, but i was bored so i started on it
upload_2019-1-17_22-8-22.png

upload_2019-1-17_22-9-3.png
 
That looks great
A few comments -
I would make the board as narrow as possible to better fit the space. It can be longer on both ends instead.
That means move the 0.1uF cap, which should not use any surface mount parts, between U1 and U3. Forget the 100uF cap, there is one on the Pioneer board, and the only one that really needs to be close to the op-amp is the 0.1uF
And chop the top and bottom to make it as slim horizontally as possible.

Also it needs a wire connection pad from pin 8 to go to the 2nd mute board. That can be on the right end of the board. That should also have a pad for a ground wire, so it can be a twisted pair wire.

I assume this is a 2 layer board. Do you plan to have them made, or just doing the PCB for fun?
 
That looks great
A few comments -
I would make the board as narrow as possible to better fit the space. It can be longer on both ends instead.
That means move the 0.1uF cap, which should not use any surface mount parts, between U1 and U3. Forget the 100uF cap, there is one on the Pioneer board, and the only one that really needs to be close to the op-amp is the 0.1uF
And chop the top and bottom to make it as slim horizontally as possible.

Also it needs a wire connection pad from pin 8 to go to the 2nd mute board. That can be on the right end of the board. That should also have a pad for a ground wire, so it can be a twisted pair wire.

I assume this is a 2 layer board. Do you plan to have them made, or just doing the PCB for fun?
Its actually just a 1 layer PCB. i was able to hand route it and the ground plane on 1 layer. could maybe make the bottom layer all ground for an even lower impedance path. i could get them made from JCLpcb for probably less than 20 bucks for an entire panel. but ill post the gerbers on here when its finished. but here's the latest version i got made. i really want to make that 0.1 C1 an SMD part since its going to be right against the DIP8 socket BUT the electrons dont care. only will look packed

upload_2019-1-18_17-29-47.png

upload_2019-1-18_17-31-10.png
 
Looks good guys, U2,3 pin spacing is pretty tight, might want to open it up a bit, you have a sliver or no solder mask between the pads, I like to use 100 mil spacing, like you get on the paper tape. I call it footprint TO-92-B1, use a 50mil pad and a 30 mil drill, it can be oblong.
 
Its actually just a 1 layer PCB. i was able to hand route it and the ground plane on 1 layer. could maybe make the bottom layer all ground for an even lower impedance path. i could get them made from JCLpcb for probably less than 20 bucks for an entire panel. but ill post the gerbers on here when its finished. but here's the latest version i got made. i really want to make that 0.1 C1 an SMD part since its going to be right against the DIP8 socket BUT the electrons dont care. only will look packed

One layer is fine, I just wasn't sure if I was looking down through 2 layers, or just one.
I think maybe I was not clear that my board view was from the top, on the component side. Normally the copper side with traces and ground plane would be opposite the component side, on the bottom. In the case of this board, it appears the copper traces are on the top along with the components.
This would make it harder to solder. If not too much trouble, move the trace/ground plane to the bottom, keeping the layout of components on the top, as it is now. This is usually a view setting on the PCB software to select the layer.

I always do routing by hand, for me it's impossible to do it any other way. That price sounds good, how big is a panel? Would it make say 10 boards? The guy I use charges about 115 bucks for 155 sq. inches, however many boards that makes. That would likely make 80 boards or more.
Their minimum construction is 2 sided. Almost no one makes production qty. of single sided boards anymore, at least from the companies I look at.
 
One last input - I would move the mute signal pad right next to pin 8, make the board a little longer if needed, then put the ground pad for the mute twisted pair next to pin 9. I'm concerned there is an top/bottom ground plane isolation, so filling in the right side edge with ground plane would help.
 
This is what i got now.
upload_2019-1-18_22-45-46.png
One layer is fine, I just wasn't sure if I was looking down through 2 layers, or just one.
I think maybe I was not clear that my board view was from the top, on the component side. Normally the copper side with traces and ground plane would be opposite the component side, on the bottom. In the case of this board, it appears the copper traces are on the top along with the components.
This would make it harder to solder. If not too much trouble, move the trace/ground plane to the bottom, keeping the layout of components on the top, as it is now. This is usually a view setting on the PCB software to select the layer.

I always do routing by hand, for me it's impossible to do it any other way. That price sounds good, how big is a panel? Would it make say 10 boards? The guy I use charges about 115 bucks for 155 sq. inches, however many boards that makes. That would likely make 80 boards or more.
Their minimum construction is 2 sided. Almost no one makes production qty. of single sided boards anymore, at least from the companies I look at.
How many boards do we want? a 4x5 panel for 20 1.6mm boards would cost $12.85 through JLCPCB. i believe there's an extra charge for V cuts. have to look into that. Only issue im not too sure about it ive never ordered a panel and i can only do a minimum of 5 quantity, so i THINK we would get 5 panels of 20. ive ordered from them 5 or 6 other times and all my boards have came out great.
 
That last board looks great.
I'll let others chime on buying boards. I think there is a hobby group that does small board quantities, maybe someone remembers it?

It would be wise to get a small qty and try them out before going gung ho on a quantity buy. I can maybe make a couple single sided units here. But it would be easier to buy them somewhere. I can pay half the cost to get it going.
 
It;s looks fine to me. These parts do not get even lukewarm, they run cold.
It was a comment on DFM, possible shorting during soldering. You will find this out when you solder one up. Go for it.
A pcb production panel from my experience is usually 18" x 24", that is why they offer these services to combine orders with batch runs.
 
It;s looks fine to me. These parts do not get even lukewarm, they run cold.
It was a comment on DFM, possible shorting during soldering. You will find this out when you solder one up. Go for it.
A pcb production panel from my experience is usually 18" x 24", that is why they offer these services to combine orders with batch runs.
im not a fan of the TO-92 footprint with the pads that close. i made a board before with that footprint and you just have to be very careful
 
im not a fan of the TO-92 footprint with the pads that close. i made a board before with that footprint and you just have to be very careful
OK, now understand the comments. Use a wider spaced TO-92 component then, rotate the right one 90 degrees to place both the pads horizontal and move the right one to the right, move the ground pad up.
 
One important detail to work out is the exact hole diameter for the 16 pins. The Brown Dog adapters are a good example of what I had in mind using machined pins from the board. I also have the the fatter square wire type shown at the top of this page, they don't fit into sockets very easily. Using a 16 pin socket leaves options for comparing to the stock chip.

https://www.cimarrontechnology.com/browndog-adapters

Mill max makes these pins, and they suggest a hole dia. for each of their parts. Not sure if this allows for production tin plating or not.

I also have a mute board design coming, not done proto test yet. That board is also small, uses 3 active parts - a to-92 mosfet and two 6 pin dip form A switches. So boards will be needed for that one too. The plan is to put it at the board edge next to the wire wrap i/o pins.
 
ill look in the parts library for a different TO-92 footprint.
okay after looking i found one with the pins 2.54mm apart but now its really crammed in there. but the PCB is the same size its unchanged.
 

Attachments

  • upload_2019-1-19_20-41-58.png
    upload_2019-1-19_20-41-58.png
    183.6 KB · Views: 52
Nice and tight, I can look at the gerber/drill data if you want, PM me if you do not want everyone to have it. I can look at the padstacks better
There are the other 2 staggered pin orientation versions of the TO-92 footprint to choose from.
Chinese new year is coming fast, get your pcb orders in ASAP.
 
Back
Top Bottom