Double FR4 board with ground plane, pic please ?

[Al (astral highway)]

Hi folks,

Can anyone post a both-sides-of-the-board photo of *any*circuit you’ve built on the FR4 board that has a solid ground-plane on one side and circular pads on the other side? I have bought some that does *not* have a conductive pathway between the pads on the top and the ground plane on the other side?

I could probably think this through but would be pleased to see an example where someone has used this for an RF (HF only ) or IC project ?

I am only used to vereboard (don’t like), traditional PCB and Manhattan style construction

Thank you

[gingpeakin]

Here's a few old half finished circuits. I use this stuff all the time - you can stick SMD components on the underside between the pads. Made by Roth Elektronik and sold by RS 518-4046. Its quite expensive but saves hours and you can end up with a decent prototype. These aren't particularly high speed but I have made some RF bits in the past

Ging

[Al (astral highway)]

Thanks Ging, nice views of the top. It doesn’t have to be particularly fast, it’s only about 220kHz. What isn’t clear is what you have done with the IC pins on the ground plane (other) side. If you were to solder the pins they’d all be shorted to ground - what do you do ? As I say, on mine you need to make a physical connection between the two sides... sorry if not clear !

[ms660]

So far as I can make out the ground plane is topside (components) Underside is pads:

http://docs-europe.electrocomponents...6b80c64a13.pdf

Lawrence.

[Al (astral highway)]

Ahah, so we have to cut ‘islands’ around the pins on the ground plane side ...
Thanks Lawrence

[gingpeakin]

It's the top that is the ground plane. You can either solder your legs directly to it or use track pins like these -
http://uk.farnell.com/harwin/t1559f4...500/dp/1143874
to join one side to the other.

[ms660]

Originally Posted by astral highway

Ahah, so we have to cut ‘islands’ around the pins on the ground plane side ...
Thanks Lawrence

So far as I can make out from Ging's pics, no cutting/removal of any of the ground plane is required to isolate any component leads or feedthrough pins from the ground plane, the isolation "circles" are already done.

Lawrence.

[gingpeakin]

The legs of most components go straight through the holes to the pads on the bottom as there is a larger hole in the top ground plane than the hole itself leaving a gap - you have to blob solder between your leg and the ground plane to make contact.
I say most components go straight through... some have shoulders that are wider than the gap.
You might be able to see the odd pull-up resistor on the underside in the first picture. I bet you can't see the 0603 decoupling caps. I have to wear 2 pairs of glasses to see them! The other picture hopefully shows the blobs grounding some of the pins on the top of the board. It's all wired up with this -
https://uk.rs-online.com/web/p/hooku...-wire/2094849/

[MrBungle]

Originally Posted by astral highway

Hi folks,
I am only used to vereboard (don’t like), traditional PCB and Manhattan style construction

I've gone for dead bug and PCB. I prototype deadbug over ground plane then move it to a PCB (OSHpark or hand drawn).

I did try using single sided copper clad and using the top as a ground plane i.e. drill holes and then use a large drill bit to clean around non-ground connected pins. That was marginally successful. By the time I'd piddled around with that I could have knocked out a prototype and a board.

Veroboard (matrix or stripboard) is vile.

[David G4EBT]

Originally Posted by astral highway

Hi folks,

Can anyone post a both-sides-of-the-board photo of *any*circuit you’ve built on the FR4 board that has a solid ground-plane on one side and circular pads on the other side? I have bought some that does *not* have a conductive pathway between the pads on the top and the ground plane on the other side?

I could probably think this through but would be pleased to see an example where someone has used this for an RF (HF only ) or IC project ?

The 'G3TSO' Multi-Band transceiver which featured in Radcom in the 1980s used that technique extensively - the IC pins that weren't to be grounded had the copper from the edge of the hole in the ground-plane cut away with a countersink drill bit, or where the pin needed to be grounded, it was left as-is and soldered to the ground plane. I etched and built all of the boards, but didn't take any pictures.

I think the PW Robin Frequency counter PCB used that technique too. (Goes up to 1 GHz).

The only example I can really illustrate is of the 15 Watt HF transmitter power amplifier which featured in the defunct 'Radio and Electronics World' magazine many years ago. The components were all mounted on the top side of the board with their leads soldered to the tinned 'lands' and the underside of the board was a ground plane. If you look at the PCB mask which shows the upper-side to which the components were soldered, you'll note small holes dotted around on the grounded areas of the board, which were for pins to short the grounded areas to the ground plane on the underside. On the component overlay picture, you'll see the heads of Veropins where the shorting holes are. The design used four CB transistors, (long since obsolete). It was popular at the time and Cirkit did a kit for it.

I built the amp (noting some errors in the article along the way, which will have caused considerable heartache to any constructors that didn't), and installed it into a 5 Watt SSB 80 Metre 'Epiphyte' transceiver that I built. I never got around to connecting the amp to the transceiver because basically, I lost interest in amateur radio. What finished it for me was when I had contacts on air and said I was using a home-brew transceiver, the response was invariably: 'Sounds excellent - we're strictly non-technical at this end' Says it all really - end of conversation, end of hobby. The transceiver complete with the (untested) PA gathered dust till I sold it to another forum member.

Pics 3 &4 show the PA installed on the rear apron of the 'Epiphyte'.

Hope the pics are of interest, if only of tenuous relevance.

[Herald1360]

If the topside groundplane is really solid (except for the punched grid of holes) just use a Vero "spot face cutter" or a 3mm 1/8" drill bit to clear around any holes where the component leg isn't going to be soldered to the groundplane.

[jimmc101]

As gingpeakin has said, in general there is sufficient clearance between the groundplane and the holes not to require any additional clearance.
If there is any doubt about about possible shorting, Kapton tape can be used under the component. A sharp scriber will easily puncture the through holes.

Attached is a pic of a similar type of board where the holes were misaligned with the ground so I used tape under the IC sockets to play safe.

I've used this board with ECL line receivers used as linear amplifiers at 50MHz plus and down to a few millivolts.

Jim

[Al (astral highway)]

That’s a cracking good example, David - thanks for the detailed post and pix- and I love your accompanying story, although sorry to hear that experience took you out of amateur radio -a big loss to the community.

Thanks also Mr Bungle, Chris and Jim, all illuminating points and tips and I now have a very clear idea what to do. It’s reassuring to hear an example of a VHF application, for future reference. The builds all look very neat, too, which is appealing to me.

[Al (astral highway)]

So, just to be clear,

1) components are placed SMD style one the side with copper pads/ islands. To avoid shorting them, you remove the islands where necessary with a small twist drill.

2) the ground plane and islands in the other side (RF ground ) are dialectrically isolated. If you push a lead through to the RF ground plane side it needs to be soldiered

3) pins can be placed to lift them SMD side leads away from the copper pads/islands

Have I got this right ? Thanks

[Herald1360]

Sounds like that's "upside down". If you have a board with isolated pads one side and a ground plane with clearance round the punched holes on the other you would mount through hole components on the ground plane side and wire up in whatever way suits on the pad side. SMD parts can be mounted on the pad side where the size and layout permits. Through hole leads or pins that need grounding are just "blobbed" with solder on the top (groundplane) sidewhere they go through.

[Al (astral highway)]

HI Chris, thank you .... I’m normally pretty good at working things out intuitively . but because I have usually used SMD (Manhattan style ) construction methods , this mode of construction with FR4 board is a potential source of confusiopn me .I don’t have the board in front of me at the moment, but the ground plane is continuous, with no clearance, except the holes drilled throughout the other side- so it’s a short circuit on the ground plane side and isolated pads on the other side ...

[Herald1360]

Same as I said then, BUT you'll need to open out any holes in the ground plane that non grounded leads/pins go through. Spot face cutter or 3mm drill as above.....

[Al (astral highway)]

Thank you Chris, that’s all clear now!

[David G4EBT]

A lot of ground-plane designs use single-sided board, leaving as much copper around the tracks and pads as possible as the ground-plane. The attached artwork is from the 'Mini Mod', designed by Ian, G4JQT. The tracks and pads are where component leads are soldered beneath the board - the holes in the ground-plane are where component leads are grounded as necessary. The second pic shows the components on the top of the board. (Same component layout as Ian's, but my own version of the artwork on the underside).

Is there any particular reason for using a double-side board Al, with the ground-plane on the underside?

[Al (astral highway)]

Hello David,

Thank you, that's another really useful illustration.

Originally Posted by David G4EBT

Is there any particular reason for using a double-side board Al, with the ground-plane on the underside?

Yes: I am designing a pretty beefy half-bridge with four IGBTs and some big old RIFA low ESR 330uF caps, or larger. Stray inductance needs to be designed out in whatever ways possible.

It makes sense that if I devote one side of board of the board to the -ve side of the DC bus and other side of the board to the +ve side, I can reduce stray inductance to a minimum.