Here's another question, for the power supply portion of the circuit I can go two ways:
1) Use one of those cute PCB mount, encased toroidal transformers (Talema and Acme are common brands). I'd have wire to board terminal blocks to bring in the AC (obviously I'd have connections for both primaries to wire it up for 120/240V).
2) BYOT: Bring your own transformer. I'd have a terminal block to accept the output of a center tapped 18V-18V transformer that would be located off-board. The diode bridge and regulators would be located on-board, of course. You could also power the board from a DC lab supply this way as well.
Chassis mount transformers are readily available if you don't have one on hand, so I'm inclined to go with option 2 (I wouldn't have to deal with creepage/clearance, fusing, etc. and it *would* save a lot of space; a 15VA PCB mount transformer is about 60x60mm).
That said, I *can* do the first option if you guys really want a complete self contained package.
2) Use horizontal trimmers and locate them near the edge.
2) Use horizontal trimmers and locate them near the edge.
You cannot predict the type of chassis the user may be placing their PCB in. Look at GK's metal frame with the front panel. If you place a horizontal trimmer there, the user would need to remove his PCB to adjust that trimmer. Some chassis may have walls around all 4 sides, meaning, anytime a trimmer needs adjusting, they will need to remove their PCB from it's chassis, while functioning. For this project, it is best to have all trimmers top vertical adjustable no matter what. As aforementioned, this may be made and assembled by students and we want to keep thing safe and easy access to probe and play.
Here's another question, for the power supply portion of the circuit I can go two ways:
1) Use one of those cute PCB mount, encased toroidal transformers (Talema and Acme are common brands). I'd have wire to board terminal blocks to bring in the AC (obviously I'd have connections for both primaries to wire it up for 120/240V).
2) BYOT: Bring your own transformer. I'd have a terminal block to accept the output of a center tapped 18V-18V transformer that would be located off-board. The diode bridge and regulators would be located on-board, of course. You could also power the board from a DC lab supply this way as well.
Chassis mount transformers are readily available if you don't have one on hand, so I'm inclined to go with option 2 (I wouldn't have to deal with creepage/clearance, fusing, etc. and it *would* save a lot of space; a 15VA PCB mount transformer is about 60x60mm).
That said, I *can* do the first option if you guys really want a complete self contained package.
Use #2, I would not trust high voltage mains on home etched PCBs due to possible poor etching leaving a conductive path for the mains & I fear, for some, I would rather see them use an approved AC out wallwart keeping them away from any exposed mains. Others who know or are capable will just wire their own transformer, just make sure there is enough heat-sinking & power filtering caps to deal with an AC transformer between 15v and 18.6v.
+, this solution allows for a dual lab power supply set to +/-18v to power this project as well.
I'd say stick with GK's regulators.
Then it's all discrete transistor level (no ICs).
I'd say stick with GK's regulators.
Then it's all discrete transistor level (no ICs).
+1 from me.
Beside, the builder, I assumed none are totally noob with zero experience say in electronics, and remember, an oscilloscope is mandatory.
Vote for originality, as this is becoming a fun/nostalgic/aesthetic project rather than pure electronic project anymore, imho.
That's cool. I actually have a design I've used in the past for a fully discrete linear regulator that uses only three transistors, is short circuit proof and implements current limiting.
The design is actually built upon the simple pass regulator GK used, so with some clever layout, I can actually make it so the builder can implement either design. If you want a current limit, improved regulation and short circuit protection, just add two more small transistors, a couple of resistors plus a cap and you're set. If not, just populate the four components of the original design and solder a jumper down.
And both options are still 100% discrete. (FYI, that's the only part of the circuit I'm modifying. I've been burned quite a few times in the past by shorting part of a circuit out while probing, which leads to vaporized traces and damaged parts. All it takes is a slip of the springy ground tip on the scope probe to short something out, or a component placed backwards. The current limit on your lab supply doesn't help here, since the input filter caps of the local supply still hold significant energy.)
Hopefully this is an acceptable compromise to everyone.
That's cool. I actually have a design I've used in the past for a fully discrete linear regulator that uses only three transistors, is short circuit proof and implements current limiting.
The design is actually built upon the simple pass regulator GK used, so with some clever layout, I can actually make it so the builder can implement either design. If you want a current limit, improved regulation and short circuit protection, just add two more small transistors, a couple of resistors plus a cap and you're set. If not, just populate the four components of the original design and solder a jumper down.
And both options are still 100% discrete. (FYI, that's the only part of the circuit I'm modifying. I've been burned quite a few times in the past by shorting part of a circuit out while probing, which leads to vaporized traces and damaged parts. All it takes is a slip of the springy ground tip on the scope probe to short something out, or a component placed backwards. The current limit on your lab supply doesn't help here, since the input filter caps of the local supply still hold significant energy.)
Hopefully this is an acceptable compromise to everyone.
That sounds really great!
I wanted it to remain all discrete (no ICs), so that it **Could** have been made in the 1950s/1960s etc, in principle. Authentic vintage retro (or whatever the terminology is suppose to be).
I like the idea of having the extra pair of transistors (optionally). As current limits are useful. E.g. With so many transistors/components to insert/solder, you can accidentally have a shorting solder bridge go unnoticed or have one or two components put in the wrong way round etc etc. So it would minimize the damage.
I understand your temptation to just put in LM317's. But by having it all discrete, it makes it that much more fun.
E.g. Some people make large 555 timer **IC's**, out of discrete transistors (there are kits available), just for fun!
Yeah, I've got one of the discrete 555 timer kits I picked up at Bay Area Maker Faire in 2014. I even made a little adapter lead that's wired into the board and terminates into an 8-pin DIP footprint, letting you plug the giant discrete board into any circuit that uses a normal DIP 555. It's a great teaching aid to help explain the 555 and let people probe the internals in working circuits, and when not in use makes a good conversation piece!
Yeah, I've got one of the discrete 555 timer kits I picked up at Bay Area Maker Faire in 2014. I even made a little adapter lead that's wired into the board and terminates into an 8-pin DIP footprint, letting you plug the giant discrete board into any circuit that uses a normal DIP 555. It's a great teaching aid to help explain the 555 and let people probe the internals in working circuits, and when not in use makes a good conversation piece!
I was tempted to get one of those too but haven't yet. There's a discrete 741 too which looks interesting.
I think a discrete 78xx regulator in the same style would be cool too, maybe some other classic ICs.
I've made sure the TH pads on both the connector *and* switch footprints are large enough for 22AWG wire and the rings are wide enough for easy soldering.
I've made sure the TH pads on both the connector *and* switch footprints are large enough for 22AWG wire and the rings are wide enough for easy soldering.
Choose a cheap standard terminal strip as a component for these. This way, the user can either soldier the wires directly, or, purchase cheap terminal blocks to have the ability to screw on and off the wires.
Go with these. I know they are large, but, this means the pad holes on the PCB you make will also be large to soldier wires directly with clearance & these terminals are also strong enough for the AC power input as well.
For 7.5mm pitch:
https://www.digikey.com/products/en/connectors-interconnects/terminal-blocks-wire-to-board/371?k=ED365
Or, for 3.5mm pitch:
https://www.digikey.com/products/en/connectors-interconnects/terminal-blocks-wire-to-board/371?k=ed555
Stick with 1 size for all. 5mm should be easy to work with and you got room for a quality trace inbetween.
3.5mm will be a little tight for hand soldiering wires if someone decides not to use terminal blocks to save the few $$$.
Nice terminal blocks. Once I used high quality Phoenix Contact blocks for a project, with press-fit technology, which doesn't need soldering. Now they have even simpler connectors, didn't know this SKEDD technology:
But I guess they are expensive, and they look too modern But you need to connect it only once for this project, so for me just soldering wires to the board would be ok.