new bus board design

[EATyourGUITAR]

I designed a passive bus board for supper low impedance 0v. the power entry is designed for 8 AWG wire with crimped copper bolt on terminals and an american #10-32 machine screw with nut and lock washer. the mounting bracket is 3d printed in under 2 minutes. any mounting height is possible. the board also accepts drilling on the sides for any size screw hole but you don't need to drill anything if you use the 3d printed feet. NO CV GATE BUS. deal with it

doepfer does not publish a standard for keyed bus board headers??? I looked. all I can see is third party vendors for shrouded headers. someone please check this and tell me my shrouded header is correct.

[pjbulls]

The Doepfer standard is literally 'unboxed headers only, red stripe down'. With the negative rail (red stripe = pin 1) pointing downwards the slot should be on the right with the header on your side of the board, so yours is correct assuming the headers go on the side that's shown.

e:
note that this assumes that your power cable was properly assembled at both ends, i.e. red stripe marks pin 1, which is why Doepfer specifically tells people not to use boxed headers (I still use them though).

[Gandalf]

The Doepfer standard is literally 'unboxed headers only, red stripe down'. With the negative rail (red stripe = pin 1) pointing downwards the slot should be on the right with the header on your side of the board, so yours is correct assuming the headers go on the side that's shown.

e:
note that this assumes that your power cable was properly assembled at both ends, i.e. red stripe marks pin 1, which is why Doepfer specifically tells people not to use boxed headers (I still use them though).

Doepfer changed to boxed headers on their busboard a while back......

[Altitude909]

you should wire up the +5, better to have it and not need it vs the other way around

[pjbulls]

Doepfer changed to boxed headers on their busboard a while back......

Very possible, I've never seen anything of his up close. This is what the documentation says tough:




Seems like unshrouded headers were favored quite a bit until about a year ago

[seismic]

Arrr, mister schreiber is not going to like it..

I personally prefer unshrouded headers.
My little opinion: it should be possible to align a connector the right way for a eurocrack user.

I’d consider making bigger vias for the plane stitching, that could lower impedance even further.

I like the solid copper connect of the headers pins over thermal reliefs. Maybe less fun to solder but nowadays no problem.

[EATyourGUITAR]

added 5v in such a way that you can do it if you really need to but you can only use one module for each 5v plugin power source. I only expect 1 module per bus board to need 5v power so the idea is that you use one slot for the removable 5v supply and the adjacent slot marked "5V LINK" to connect the module that requires 5v. I really didn't want to cut my massive 0v plane in half. that would completely ruin what I am trying to do here with low impedance 0v being the absolute top priority.

[EATyourGUITAR]

Arrr, mister schreiber is not going to like it..

why? how can I make it better?

[J3RK]

Arrr, mister schreiber is not going to like it..

why? how can I make it better?

It looks good to me. I always add some electrolytic caps near the inlet, but I don’t see much I’d change.

[EATyourGUITAR]

Arrr, mister schreiber is not going to like it..

why? how can I make it better?

It looks good to me. I always add some electrolytic caps near the inlet, but I don’t see much I’d change.

this is for a 5A per rail linear system being loaded at 3.8A on the +12v. I wanted this system to be zero maintenance for the next 40 years so I intentionally avoided any electrolytic caps on the bus board. 100uf on 12 bus boards adds up to 1200uf per rail. not really good for the power supply. 470nF seems to be the perfect value here.

[Gandalf]

Doepfer changed to boxed headers on their busboard a while back......

Very possible, I've never seen anything of his up close. This is what the documentation says tough:




Seems like unshrouded headers were favored quite a bit until about a year ago

Doesn't really matter what historical documentation is produced, the fact is that as of last year Doepfer have replaced the A100 with open headers with a boxed header version. So yes, Doepfers documentation is not current but you only have to look at the accessories page where the busboards are listed to see that boxed headers are IN and open headers are OUT. Don't know if they intend to do the same with the modules (which they should). The responsibility 9assuming modules and busboards have properly orientated headers) is on the cable and the user. if you use 'right' busboards with 'right' modules with 'right' cables then there is no issue. 'wrong' cables should be simply binned and replaced with the correct type

[J3RK]

Arrr, mister schreiber is not going to like it..

why? how can I make it better?

It looks good to me. I always add some electrolytic caps near the inlet, but I don’t see much I’d change.

this is for a 5A per rail linear system being loaded at 3.8A on the +12v. I wanted this system to be zero maintenance for the next 40 years so I intentionally avoided any electrolytic caps on the bus board. 100uf on 12 bus boards adds up to 1200uf per rail. not really good for the power supply. 470nF seems to be the perfect value here.

Ah! Gotcha. I should have read the details. I was just looking at the board

[seismic]

Arrr, mister schreiber is not going to like it..

why? how can I make it better?

See here:
viewtopic.php?f=16&t=194257&hilit=Libb+busboard

[EATyourGUITAR]

Arrr, mister schreiber is not going to like it..

why? how can I make it better?

See here:
viewtopic.php?f=16&t=194257&hilit=Libb+busboard

on page 3 I need to stop reading this before it gets too deep. Genus LIBB is definitely a better bus board. But my bus board is DIY. It is a balance between price vs performance. this is why I settled on 2oz copper. it was designed for bus bar power distribution in very very large cases 1800HP. when you have 1800HP the proposition for getting your own bus board PCB manufactured at cost in china is very attractive compared to $3,120.00 + tax + shipping at current LIBB prices for 24 (passive) bus boards. I could buy a slightly damaged 1992 honda civic + my low cost bus boards for $3,120.00. no car no power supply. just poverty and really nice (passive) bus boards. I have to pass up things when they get to that level. did you know you can get a linear 5A per rail power supply for $200?

https://www.digikey.com/en/products/det ... AG/1634169

I like nice things but I also don't want to pay too much. much respect to paults, niel, hinton all of them. I have learned from all of them and I would even say that paults is my friend and teacher. every forum pissing match has gift wrapped lessons about diplomacy, electrical engineering, and business.

P.S. I will not be doing any whitepapers. it is 2oz copper with 470nF film caps. you can see the size of the ground pour. cross sectional area for 0v top and bottom combined is 1.6 inch by 2oz copper by 6 inch from edge to middle worst case scenario. the total length is 11.9 inches. multiply the cross sectional resistance by 11.1 inches if you do not have redundant power connections at both ends. I got 198 micro Ohms edge to middle. probably close to 300 micro Ohms for +12v edge to middle and 350 maybe 400 micro ohms for -12v edge to middle but I only did precise calculations for the actual 0v combined copper planes 0.000198 Ohms. good enough for me. as long as it is better than the cheap overpriced super basic $40 passive bus boards while also being cheaper than the LIBB (which I admit is better). that and bolted copper termination.

[seismic]

I didnt even got to page 3...
But i got carried away on another discussion about it...

anyways:

id say a good psu, proper cabling with decent connectors are as important as a busboards itselfs.
I even isolate digital modules on a 2nd psu.

And another thing is many people spend much on modules and not a extra dime for power&cabling&distribution.

[Altitude909]

have you worked with 2oz copper before? its a nightmare. unless you plan one getting all the pth parts wave soldered, i suggest you run a small test run and see how usable with will be. Soldering headers to a 2oz gnd plane with no thermals will require pretty hardcore soldering kit, not something the normal DIYer would have

[Gandalf]

have you worked with 2oz copper before? its a nightmare, unless you plan one getting all the pth parts wave soldered, i suggest you run a small test run and see how usable with will be. Soldering headers to a 2oz gnd plane with no thermals will require pretty hardcore soldering kit, not something the normal DIYer would have

All the above and, because it is DIY you can guarantee a substantial fail rate due to people applying excessive heat to get the 0V pads to solder properly and consequently deform the moulding and the alignment of the pins in the PCB (ie no longer 90 degree to the PCB, or moved up/down relative to their manufactured position). There is a good reason why proper passive busboards cost more and that is in part due to the additional design that goes into making such a design. Not really an issue for production processes but definitely an issue for DIYers

[EATyourGUITAR]

JLCPCB said they do PTH but all assembly services are 1oz copper only. then I called a united states based company who suggested that everything be surface mounted and loaded into reels or trays. they offer 4oz copper as an option. is anyone interested in getting in on some kind of group buy or something? these only fit into 150HP with two bus boards per row and wires going up the middle of the case. maybe if you angled them you could make them fit anywhere. if I order 25 PCB's for myself that will be the end of this product forever. if you want to stack these up for a rainy day now is the time to say something.

[cackland]

if I order 25 PCB's for myself that will be the end of this product forever.

[EATyourGUITAR]

if I order 25 PCB's for myself that will be the end of this product forever.

yeah because I don't want to pay to stock this in my store if it doesn't sell. 150HP per row, high end passive bus boards. I doubt this will sell if I buy extra. I would love for someone to prove me wrong. do you want some bus boards?

[EATyourGUITAR]

after talking to a fabricator I decided that I needed to switch to surface mount. the cap is acrylic film 1uF. fancy stuff. I figured I should probably put caps everywhere since they have to load the reel into the robot why not YOLO. panasonic ECP-U1C105MA5 for the cap. the shrouded header is amphenol FCI (not amphenol RF) 52601-S16-8TLF. T is for tray good for pick and place robots.

[Gandalf]

after talking to a fabricator I decided that I needed to switch to surface mount. the cap is acrylic film 1uF. fancy stuff. I figured I should probably put caps everywhere since they have to load the reel into the robot why not YOLO. panasonic ECP-U1C105MA5 for the cap. the shrouded header is amphenol FCI (not amphenol RF) 52601-S16-8TLF. T is for tray good for pick and place robots.

passivebus7.gif

That will surely push the price up. For SMT they will need a stencil and they are not cheap and with a run of only 25, the cost of the stencil and SMT setup costs will be high........... You also need to be careful with SMT headers, especially if they don't have any locking pins or stabilisers, as they can be easy to rip of the board when a user is trying to remove an IDC cable while there hand is inside a shallow boat and, inevitably, end up just yanking on the cable rather than gripping the IDC socket (one reason we always recommend using the strain relief)

[Altitude909]

idk about stencil costs being high, framed ones are $15 at JLC but I agree about the SMD headers. The ones that are the through hole ones with just the legs bent on the one side to make them SMD in particular are pretty bad in that regard. I remember seeing ones that were lower profile that sat flatter against the PCB with the pins out the sides, those would probably be a better choice. Boxed headers like that are not really designed for constant on/off use so IDK if anyone even makes a ruggedized one with reinforcement

All that being said, if you must insist on 2oz, SMD IS the way to deal with it. When i was working on a SMPS design that needed the heavy copper for cooling, I couldnt get the damn regulators with the big tabs soldered down by hand (I resorted to board heater) and the ultimate solution was to simply do everything in the reflow oven and there were no problems there, no additional heat, schedule, etc since the high heat transfer works to your advantage in an oven vs against you by hand