SMT-on-copper-clad

Using "surfboards" to mount about a half-dozen SMT cmos logic to a 
copper clad, how many by-pass/de-glitcher caps do you put on and where 
do you put them when you use this "style"? frequencies in the 10s of 
MHz. Do you put one on each chip (seems like overkill) or do you bypass 
at the power input and rely on distributed capacitance of the board?

On Tue, 7 Apr 2020 13:15:09 -0400, bitrex <user@example.net> wrote:

>Using "surfboards" to mount about a half-dozen SMT cmos logic to a 
>copper clad, how many by-pass/de-glitcher caps do you put on and where 
>do you put them when you use this "style"? frequencies in the 10s of 
>MHz. Do you put one on each chip (seems like overkill) or do you bypass 
>at the power input and rely on distributed capacitance of the board?

All these worked:

https://www.dropbox.com/s/v5m4i3v8gvlnikb/Z356_Top.JPG?raw=1

https://www.dropbox.com/s/opnxfnk79o5lk1s/Z466_2.JPG?raw=1

https://www.dropbox.com/s/dvmothm0oawnqhb/BB_Fast_6.jpg?raw=1

For fast stuff, I like to run two wires from chip ground to the PCB
plane, and add a cap on the proto board if possible, or another two
wires to a bypassed power island. Surfmount caps can go sideways
sometimes, between the main board and the adapter. Or use a short fat
axial ceramic cap from the surfboard to the copper plane.

I have made some of my own adapter boards with really close bypass
caps, for oddballs like US8.

-- 

John Larkin         Highland Technology, Inc
picosecond timing   precision measurement 

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

On 4/7/2020 1:54 PM, John Larkin wrote:
> On Tue, 7 Apr 2020 13:15:09 -0400, bitrex <user@example.net> wrote:
> 
>> Using "surfboards" to mount about a half-dozen SMT cmos logic to a
>> copper clad, how many by-pass/de-glitcher caps do you put on and where
>> do you put them when you use this "style"? frequencies in the 10s of
>> MHz. Do you put one on each chip (seems like overkill) or do you bypass
>> at the power input and rely on distributed capacitance of the board?
> 
> All these worked:
> 
> https://www.dropbox.com/s/v5m4i3v8gvlnikb/Z356_Top.JPG?raw=1
> 
> https://www.dropbox.com/s/opnxfnk79o5lk1s/Z466_2.JPG?raw=1
> 
> https://www.dropbox.com/s/dvmothm0oawnqhb/BB_Fast_6.jpg?raw=1
> 
> For fast stuff, I like to run two wires from chip ground to the PCB
> plane, and add a cap on the proto board if possible, or another two
> wires to a bypassed power island. Surfmount caps can go sideways
> sometimes, between the main board and the adapter. Or use a short fat
> axial ceramic cap from the surfboard to the copper plane.
> 
> I have made some of my own adapter boards with really close bypass
> caps, for oddballs like US8.
> 

Is that an ECL xor for fast phase detector, there? Very fast phase detector

74VHC seems pretty happy with a bypass cap across every 2nd or 3rd chip 
and the grounds tacked to the copper-clad with short jumper 
interconnects up to about 40-50MHz, don't have to be super-precise other 
than the scope probe connections to catch the edges. a 3" ground clip 
and the edges look like slop past about 10

On Tue, 7 Apr 2020 19:03:57 -0400, bitrex <user@example.net> wrote:

>On 4/7/2020 1:54 PM, John Larkin wrote:
>> On Tue, 7 Apr 2020 13:15:09 -0400, bitrex <user@example.net> wrote:
>> 
>>> Using "surfboards" to mount about a half-dozen SMT cmos logic to a
>>> copper clad, how many by-pass/de-glitcher caps do you put on and where
>>> do you put them when you use this "style"? frequencies in the 10s of
>>> MHz. Do you put one on each chip (seems like overkill) or do you bypass
>>> at the power input and rely on distributed capacitance of the board?
>> 
>> All these worked:
>> 
>> https://www.dropbox.com/s/v5m4i3v8gvlnikb/Z356_Top.JPG?raw=1
>> 
>> https://www.dropbox.com/s/opnxfnk79o5lk1s/Z466_2.JPG?raw=1
>> 
>> https://www.dropbox.com/s/dvmothm0oawnqhb/BB_Fast_6.jpg?raw=1
>> 
>> For fast stuff, I like to run two wires from chip ground to the PCB
>> plane, and add a cap on the proto board if possible, or another two
>> wires to a bypassed power island. Surfmount caps can go sideways
>> sometimes, between the main board and the adapter. Or use a short fat
>> axial ceramic cap from the surfboard to the copper plane.
>> 
>> I have made some of my own adapter boards with really close bypass
>> caps, for oddballs like US8.
>> 
>
>Is that an ECL xor for fast phase detector, there? Very fast phase detector

The EP08? It just has one signal input. I don't remember why I did
that.

https://www.dropbox.com/s/48g4z9tg035p7g0/Z466_WB.JPG?raw=1


>
>74VHC seems pretty happy with a bypass cap across every 2nd or 3rd chip 
>and the grounds tacked to the copper-clad with short jumper 
>interconnects up to about 40-50MHz, don't have to be super-precise other 
>than the scope probe connections to catch the edges. a 3" ground clip 
>and the edges look like slop past about 10
>

A solid copper plane with bypassed Manhattan power strips works well,
like my first pic.

 

-- 

John Larkin         Highland Technology, Inc
picosecond timing   precision measurement 

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

On 08/04/2020 03:15, bitrex wrote:
> Using "surfboards" to mount about a half-dozen SMT cmos logic to a 
> copper clad, how many by-pass/de-glitcher caps do you put on and where 
> do you put them when you use this "style"? frequencies in the 10s of 
> MHz. Do you put one on each chip (seems like overkill) or do you bypass 
> at the power input and rely on distributed capacitance of the board?

You can mount the SOIC chips right on the copper clad, the right way up, 
without surf boards. Just bend up horizontal all of the pins that are 
not ground, solder the ground pins straight to the copper clad, and lean 
an 0402 100nF or 1uF diagonally between the VDD pin and the copper clad. 
Then wire up the signals with 0.125mm or 0.2mm solderable enamelled 
magnet wire.