Separate front panel PCB, or use right-angle switches and displays?

[Steve S]

I'm trying to decide the best approach for a new design (consumer audio
product). For a similar product in the past, I used two PCBs:

1) Main PCB, mounted horizontal in the unit, containing 90% of the
circuitry, double-sided with components mainly surface mount on one side. It
also includes right-angle coax sockets and a DC socket that go through and
attach to the rear aluminium panel.

2) Front Panel PCB, mounted vertical in the unit, attached behind the front
panel using glued standoffs glued. It contained 3 x 6x6mm tactile buttons
and 4 x 0.7" seven-seg LED displays. This is a double-sided PCB, with all
components through-hole. FWIW the front panel is quite small - 1" high, by
5" wide.

I want to simplify and reduce the manufacturing costs for a new design of
product similar to this (quantities of 2-5k per year). So I'm considering
using just one PCB, attaching right-angle 7-seg LED displays and tactile
buttons to it for use as the front panel.

As I see it, the pro's of this approach would be:
a) Only one PCB
b) Less components - ribbon cable and header
c) No need to mount to front panel

The con's would be:
a) Harder to source right-angle displays and switches, and more expensive
b) Excess force on a connector on the rear panel would push front panel
switches and displays forward into the front panel

The rear connectors are screwed to the rear panel by one screw each, but I
think that last point may still be a concern - coax connections can be very
tight and attract a lot of user force sometimes.

Does anyone have any thoughts as to the best approach in order to minimize
manufacturing costs without causing problems down the line?

Thanks,
Steve

 

[Ian Stirling]

I'm trying to decide the best approach for a new design (consumer audio
product). For a similar product in the past, I used two PCBs:

I want to simplify and reduce the manufacturing costs for a new design of
product similar to this (quantities of 2-5k per year). So I'm considering
using just one PCB, attaching right-angle 7-seg LED displays and tactile
buttons to it for use as the front panel.

As I see it, the pro's of this approach would be:

The con's would be:
a) Harder to source right-angle displays and switches, and more expensive
b) Excess force on a connector on the rear panel would push front panel

c) Excess force on the front panel switches may snap them off.

If you can, pick the switches so that any force is borne by the front
panel, not a solder joint.

 

[Spehro Pefhany]

I'm trying to decide the best approach for a new design (consumer audio
product). For a similar product in the past, I used two PCBs:

1) Main PCB, mounted horizontal in the unit, containing 90% of the
circuitry, double-sided with components mainly surface mount on one side. It
also includes right-angle coax sockets and a DC socket that go through and
attach to the rear aluminium panel.

2) Front Panel PCB, mounted vertical in the unit, attached behind the front
panel using glued standoffs glued. It contained 3 x 6x6mm tactile buttons
and 4 x 0.7" seven-seg LED displays. This is a double-sided PCB, with all
components through-hole. FWIW the front panel is quite small - 1" high, by
5" wide.

I want to simplify and reduce the manufacturing costs for a new design of
product similar to this (quantities of 2-5k per year). So I'm considering
using just one PCB, attaching right-angle 7-seg LED displays and tactile
buttons to it for use as the front panel.

As I see it, the pro's of this approach would be:
a) Only one PCB
b) Less components - ribbon cable and header
c) No need to mount to front panel

The con's would be:
a) Harder to source right-angle displays and switches, and more expensive
b) Excess force on a connector on the rear panel would push front panel
switches and displays forward into the front panel

The rear connectors are screwed to the rear panel by one screw each, but I
think that last point may still be a concern - coax connections can be very
tight and attract a lot of user force sometimes.

Does anyone have any thoughts as to the best approach in order to minimize
manufacturing costs without causing problems down the line?

Thanks,
Steve

My opinion, which is worth what you paid for it, and based on limited
information (not sure what you mean by "glued"), is not to f*ck around
with the front panel ruggedness.

As you say, you'll likely pay more for right-angle parts and find them
harder to source, *especially* in your sub-consumer quantities.

The result won't be as good. You won't save much money, if any.

I'd look at soldering the flat cable rather than plugging it and
building/testing the assembly as one piece in production.


Best regards,
Spehro Pefhany

 

[John Woodgate]

Does anyone have any thoughts as to the best approach in order to
minimize manufacturing costs without causing problems down the line?

Forget the front panel and put your controls and indicators on the 1
inch by 5 inch strip at the front of the top of the unit. You could
step this part down if that offers an advantage (use Courier font:

Upright or sloped
\ ________________
\------>| / |
| / |
Controls |/ | Side view
_\______| |
|_________________________|


That'll be USD 15000 for design consultancy. Used notes only, please.

Definitely fix the rear panel sockets firmly. I suggest 12 mm bolts.

 

[Steve S]

My opinion, which is worth what you paid for it, and based on limited
information (not sure what you mean by "glued"), is not to f*ck around
with the front panel ruggedness.

As you say, you'll likely pay more for right-angle parts and find them
harder to source, *especially* in your sub-consumer quantities.

The result won't be as good. You won't save much money, if any.

I'd look at soldering the flat cable rather than plugging it and
building/testing the assembly as one piece in production.


Best regards,
Spehro Pefhany

http://www.speff.com

Thanks Sphero. FWIW , by gluing I really mean self-adhesive standoffs (since
I can't come through the front panel due to aesthetic reasons.) That is the
part of my design I really like the least, and saw the use of right-angle
switches to be a possible way to avoid it. However, I take all your points
on-board and realise it might not be such a good idea after all! The
suggestion for soldering the ribbon cable does sound very good though. I'm
tempted to leave one end socketed to speed of repairs, although the entire
cost of the assembled front panel board is only about $6, so might not be
worth treating it as a separate item.

Steve

 

[Steve S]

c) Excess force on the front panel switches may snap them off.

If you can, pick the switches so that any force is borne by the front
panel, not a solder joint.

Thanks Ian. In my case they are those tiny 6x6 tactile switches with a tiny
throw, and mounted so they hardly protrude through the front panel at all.
Even though, I guess a frustrated customer could press them hard enough to
compromise the solder joint. It's a good point and another reason against
the use of right-angle mounted switches.

Steve

 

[Ian Stirling]

Thanks Ian. In my case they are those tiny 6x6 tactile switches with a tiny
throw, and mounted so they hardly protrude through the front panel at all.
Even though, I guess a frustrated customer could press them hard enough to
compromise the solder joint. It's a good point and another reason against
the use of right-angle mounted switches.

Feedback to the user is also important.

I've got an Iriver IFP-395. (that I bought with a button snapped off at
the PCB and fixed)
The switches are surface-mount, and mounted at a right angle.
If you're not looking at the display, there is no indication that
some of them have activated.
Some of them are overloaded, so that (for example) in FM mode, a
short press on record/mode will start recording the current program,
but a long press will let you choose another mode.
(mp3/voice dictation/...)
This leads to the tendancy to press the button harder than you might
normally, just to ensure you don't get a bounce that puts the equipment
in the wrong state.

Even knowing that the button is fragile (it's held on by two tiny
solder pads) I've already had to repair it once more due to my use.

I wonder about other solutions.
For example, a board with a snap-off cutout.
Assemble the board as normal, with wire jumpers between the two
sections, then simply put into a jig that snaps the board edge off, and bends
up to the appropriate angle.
I've no idea if this would be cheaper.

 

[Steve S]

Forget the front panel and put your controls and indicators on the 1
inch by 5 inch strip at the front of the top of the unit. You could
step this part down if that offers an advantage (use Courier font:

Upright or sloped
\ ________________
\------>| / |
| / |
Controls |/ | Side view
_\______| |
|_________________________|


That'll be USD 15000 for design consultancy. Used notes only, please.

Thanks for the suggestion John, check in the mail (sorry, just out of
notes!)

Definitely fix the rear panel sockets firmly. I suggest 12 mm bolts.

The original design uses a 12mm countersunk machine screw to fix each coax
connector to the aluminum rear panel, plus the connectors have plastic lugs
into the circuit board which I guess are meant to absorb much of the
insertion force. The DC socket (standard 2.1mm concentric type) has no
fixing hole so that relies on its through-hole soldering - it does not
require much force from the user to insert the power jack into it, but I
realize it could be abused. I could move to a DC socket with a fixing hole
or two. I wonder how many consumer items are returned for repair due to this
kind of stress on solder joints of connectors/switches?

Steve

 

[John Woodgate]

I wonder how many consumer items are returned
for repair due to this kind of stress on solder joints of
connectors/switches?

I've seen a few, and I 'don't do repairs', so I guess it's very common.

BTW, I meant 12 mm DIAMETER bolts. (;-)

 

[John Fields]

I'm trying to decide the best approach for a new design (consumer audio
product). For a similar product in the past, I used two PCBs:

1) Main PCB, mounted horizontal in the unit, containing 90% of the
circuitry, double-sided with components mainly surface mount on one side. It
also includes right-angle coax sockets and a DC socket that go through and
attach to the rear aluminium panel.

2) Front Panel PCB, mounted vertical in the unit, attached behind the front
panel using glued standoffs glued. It contained 3 x 6x6mm tactile buttons
and 4 x 0.7" seven-seg LED displays. This is a double-sided PCB, with all
components through-hole. FWIW the front panel is quite small - 1" high, by
5" wide.

I want to simplify and reduce the manufacturing costs for a new design of
product similar to this (quantities of 2-5k per year). So I'm considering
using just one PCB, attaching right-angle 7-seg LED displays and tactile
buttons to it for use as the front panel.

As I see it, the pro's of this approach would be:
a) Only one PCB
b) Less components - ribbon cable and header
c) No need to mount to front panel

The con's would be:
a) Harder to source right-angle displays and switches, and more expensive
b) Excess force on a connector on the rear panel would push front panel
switches and displays forward into the front panel

The rear connectors are screwed to the rear panel by one screw each, but I
think that last point may still be a concern - coax connections can be very
tight and attract a lot of user force sometimes.

Does anyone have any thoughts as to the best approach in order to minimize
manufacturing costs without causing problems down the line?

---
Use an adhesive overlay with membrane switches (or tactile domes) and
transparent areas for the displays to shine through as your front
panel and mount your PCB behind it, coplanar with it, on standoffs
molded into the rear of the front panel or on standoffs secured to the
rear of the front panel with flat head screws which will be covered by
the overlay when it's stuck down. Provide a pigtail on the rear of the
overlay which goes through a slot in the front panel (which makes it
invisible from the outside of the device) and plugs into a header on
the PCB, and you're done!

 

[Steve S]

---
Use an adhesive overlay with membrane switches (or tactile domes) and
transparent areas for the displays to shine through as your front
panel and mount your PCB behind it, coplanar with it, on standoffs
molded into the rear of the front panel or on standoffs secured to the
rear of the front panel with flat head screws which will be covered by
the overlay when it's stuck down. Provide a pigtail on the rear of the
overlay which goes through a slot in the front panel (which makes it
invisible from the outside of the device) and plugs into a header on
the PCB, and you're done!

Thanks for all those suggestions, John. So far I've avoided the use of an
overlay for reasons of appearance, although I guess if done well it can look
very good. Also great for hiding a multitude of "sins" such as the
screwheads uses to mount the PCB behind it, as you suggest. I currently use
a bespoke piece of tinted acryllic pushed through from behind the front
panel - it looks great but using clear/tinted areas of the overlay would
save money, which is what I am looking for.

Personally, I never like the feel of the membrane switches I've come across
so far, but the cost saving when done in large enough quantities could be
significant I expect.

Steve

 

[Steve S]

I've seen a few, and I 'don't do repairs', so I guess it's very common.

BTW, I meant 12 mm DIAMETER bolts. (;-)

Of course, I should have realized! "Gonna need a bigger drill!..."

 

[John Fields]

Thanks for all those suggestions, John. So far I've avoided the use of an
overlay for reasons of appearance, although I guess if done well it can look
very good. Also great for hiding a multitude of "sins" such as the
screwheads uses to mount the PCB behind it, as you suggest. I currently use
a bespoke piece of tinted acryllic pushed through from behind the front
panel - it looks great but using clear/tinted areas of the overlay would
save money, which is what I am looking for.

Personally, I never like the feel of the membrane switches I've come across
so far, but the cost saving when done in large enough quantities could be
significant I expect.

---
As far as looks go, you can get stuff that's absolutely gorgeous if
you get the esthetics of the artwork right, since there are no limits
on colors or finishes, or basically anything else.

If you don't like membrane switches you can always go to
tactile-feedback dome switches, But they're more expensive. What I do
for applications where I can't use domes and I need feedback from the
keypad/switches is to use a cheap magnetic beeper which gives me a
100ms or so beep whenever a key gets made. So far, clients like it.

 

[John Woodgate]

I read in sci.electronics.design that John Fields <jfields@austininstrum
ents.com> wrote (in <[email protected]>) about
'Separate front panel PCB, or use right-angle switches and displays?',

What I do for
applications where I can't use domes and I need feedback from the
keypad/switches is to use a cheap magnetic beeper which gives me a 100ms
or so beep whenever a key gets made. So far, clients like it.

I don't, for one. I can't hear the beeps. If you want beeps, make them a
nice alto frequency. 400 Hz is good.

 

[John Fields]

I read in sci.electronics.design that John Fields <jfields@austininstrum
ents.com> wrote (in <[email protected]>) about
'Separate front panel PCB, or use right-angle switches and displays?',


I don't, for one.

 

[budgie]

(snip)

The DC socket (standard 2.1mm concentric type) has no
fixing hole so that relies on its through-hole soldering - it does not
require much force from the user to insert the power jack into it, but I
realize it could be abused. I could move to a DC socket with a fixing hole
or two. I wonder how many consumer items are returned for repair due to this
kind of stress on solder joints of connectors/switches?

Laptops by the dozen. It's the most common electromechanical failure item on
them IMOE.

 

[Spehro Pefhany]

(snip)


Laptops by the dozen. It's the most common electromechanical failure item on
them IMOE.

I did an emergency repair on a good friend's FSR radio set the night
before he headed off to Bangladesh on a charity mission with his
daughter (I generally refuse such "opportunities"). Pressure on the
volume pot had broken a solder joint. It was a decent Japanese brand,
too.

Customers can find ways to abuse stuff that you'd never even imagine.
One product I designed was targeted at higher-end bakers- talented in
their craft. More than one of them decided it was an excellent idea to
operate the membrane keys with the sharp tip of a big knife...10 or 15
thou of polycarbonate is no match for a big knife.


Best regards,
Spehro Pefhany

 

[John Woodgate]

I read in sci.electronics.design that John Fields <jfields@austininstrum
ents.com> wrote (in <[email protected]>) about
'Separate front panel PCB, or use right-angle switches and displays?',
Only just. When last measured, I had 45 dB loss at 2 kHz.

Just LOWER THE FREQUENCY!

 

[Keith Wootten]

In message <[email protected]>, John Fields

If you don't like membrane switches you can always go to
tactile-feedback dome switches, But they're more expensive. What I do
for applications where I can't use domes and I need feedback from the
keypad/switches is to use a cheap magnetic beeper which gives me a
100ms or so beep whenever a key gets made. So far, clients like it.

Never tried it, but how about whacking the panel from behind with a
solenoid every time a key is pressed? Might work.

Cheers

 

[John Fields]

I read in sci.electronics.design that John Fields <jfields@austininstrum
ents.com> wrote (in <[email protected]>) about
'Separate front panel PCB, or use right-angle switches and displays?',

Only just. When last measured, I had 45 dB loss at 2 kHz.

Just LOWER THE FREQUENCY!

oooooooohhhhhhhhkkkkkaaaaaaaaaaaaayyyyyyyyyyyyyyyyyyyy!!!!!!!!!!!