It’s not as simple as you wish. Copyright law as a long history, and the rule is that “… Copyright owners have the exclusive statutory right to exercise control over copying and other exploitation of the works …” History of copyright - Wikipedia Without a statement no assumptions can be made. It would be dangerous to equate dissemination with permission, e.g. somebody takes the novel you were writing and publishes it. Has it been published to the public? Yes. Did you give permission for this? No.
Bear in mind that copyleft is based on copyright law so the fact that you are enjoying open source is because of the law.
I can open both the schematic and the layout with no problem at all. The libraries are not accesible, of course, but the symbols and footprints are embedded into the schematic and layout files.
Thanks a lot for letting me know. Finding out was my objective in “hiding” my libraries yesterday but there is always some doubt.
@Lutz_Muller This tells me that there is not a problem with my posted zip file. If you cannot open it, I think you need to open a new discussion thread about being unable to fully read my zip file. I will soon get back to editing my file to add the “conventional” component values such as K and M ohms, and pF and uF. I have already added a column using “edit symbol fields”.
I tested gnumeric’s sort capability on a column with the entries [1.2M, 2.2k, 10k, 7k, and 3.01k] in that order.
I just used gnumeric’s top-bottom sort order and got a column [10k, 1.2M, 2.2k, 3.01k, 7k]. I got the same sort order if I put a space between the numeral and the M or k.
When I removed the letter suffixes, and resorted the column I got [1.2, 2.2, 3.01, 7, 10]. So the only way I know to correctly sort the numeric values into the proper order is to apply the SI prefix as a scale factor of 10.
The next step, I’d suggest might be googling “sorting numbers with SI prefixes in excel”. Somebody has probably done this already.
I stock components in bags each of which brackets a range of values. Having the values in numerical order (and sorting the BOM similarly) reduces the need for me to go back and forth between bags when I assemble boards and just plain seems logical. But with spreadsheets sorting in a way similar to what you describe; I think that such a sorting operation seems to be on the far side of useless.
For me, I think in engineering notation. You are pretty much doing that whenever you use K ohms or picofarads; etc. But I have come to accept that many people (including some engineers) do not like my method of designating Rs and Cs with engineering notation only. IMHO this disagreement is easily solved by simply adding another column of information. I only need to have the one column in the BOM with these straight numerical values. I have already filled out another column using the “edit symbol fields” function; a quick glance at the schematic seems to show that this worked as desired.
During the last year or so I have worked on customer BOMs with many columns, much of which contained what I felt was useless information. So adding or reserving one column for numeric values seems like an easy thing…
Many people who are more software-savvy would tell me something about a macro or a batch file for sorting alpha-numeric component values, but that seems much more complicated; far out of my “wheelhouse.” The schematic already contains the added column of information; however I am expecting many more edits over the course of any collaboration. So I am not “beating it to death” right now. I will put the new project file (new filename) and probably new license into a new zip file and post it soon.
In fact; here it is. I invite anyone interested to check it out.
Generated from Protel (1997…2017) and from KiCad (2017…) *.csv file I copy into second page (tab) of my spreadsheet. First page contains all elements I use. The full description of elements (from first page) is copied into second page. It is done automatically. In Protel I used element name for it. For KiCad I modified my spreadsheet to use name+footprint to find element. I have one column destined to sort all elements in my BOM and its contents is also copied from page one to second.
Then I (manually) sort whole bom by that column and by name column (In LibreOffice you can sort by 3 keys at once).
That sort column has probably many historical accretions. For ICs to be placed at beginning it starts form 1 (ATXmega controllers have there 1D4, HVD72 have there 1P). For resistors I have there 6C1 for 0…<10 ohm, 6C2 for 10…<100 and so on. Sorting with that column and with name (2 key sorting) gives me sorted all elements as I want.
Hope it can be useful for someone.
Thanks for the update. I had been thinking about this thread but wasn’t sure I’d be able to find it easily. Am looking forward to reading more on this project.
I finished building one of my boards and began testing it on Thursday. Today (Friday) I have it running reasonably well. This “doubler” build of the board can produce from about 4V to 60V output. The coarse + fine digipot adjustments can set Vout within 2-3 mV of a desired voltage, and that voltage seems to be stable. (That is a big improvement over my cheap commercial bench supply which output voltage tends to move around.) So far I had a solder splash and have changed a few component values. No cut traces or tacked-on components so far. I need to look more carefully at the output current capability.
I have had some nice results today with the 60V maximum doubled output version. Anyone interested; please check out the attached .zip. This is still a work in progress. I may go for a modest increase in power from the present approximately 17W. Also I have not yet completed the “singled” output version. Depending upon the choice of coupled inductor, that one will be lower power.
Project update February 16 2022: Favorable data & confession…
Today I tweaked the converter output to 50V @ 600 mA. That is 30W. The converter will produce up to 60V and I think it will do 60V @ 500 mA. However my e-load has 60V output transistors and quits at <60V so I do not have good flexibility to adjust loading @ 60V.
The converter ran with 84.9% efficiency (measurement was somewhat crude) when delivering 22.5 Watts. That efficiency is reasonbly good and indicates that there are no problems that are seriously degrading efficiency.
I also performed a dreaded test: I intentionally shorted the output while it was delivering full power. This did not damage anything. The power converter output came back up when I removed the short.
Confession: Fortunately I had reported 30 Watts as a design target when I started this thread. But I erased or misplaced my own spreadsheet in which I had done some of the underlying calculations. This is not big problem as I have several copies of my spreadsheet with the formulas. But I did not have my data entry points. But yesterday I noticed my earlier post where I had mentioned 30W, so I decided to go for that.
After getting 30W out of the doubled SEPIC version, I am now working to complete the “singled” SEPIC which will produce 30V maximum. I need to decide on the coupled inductor; that decision will affect output power.
The non doubler board (my second board) is working well (except for ONE THING) and producing up to 30V at 30 Watts. I find it surprising that the board produces this power level using a 22 uH MSD1278 Coupled Inductor
There is ONE THING on this board that has me puzzled. The dual digipot at U5 is supposed to draw about 1 uA through Vdd pin 4, but the actual current drain is closer to 1 mA when the voltage is +5V. The only reason why I noticed this is due to the blocking diode arrangement through which I power it:
J7 is a connection for three 1V5 alkaline cells (or small low voltage power supply) to keep the chip (U5 pin 4 Vdd) powered when there is no main power supply. The idea is to keep the U5 memory alive so that it does not forget the potentiometer setting when the main supply is shut down. Use of J7 is optional (and I was not using it.)
The digipots need a Pin 4 Vdd power supply voltage that is close to or above the highest voltage applied to the potentiometers. With excessive current drain at U5 pin 4, the voltage on that pin was closer to +3V instead of +5V. With +5V applied to the high end of the potentiometers and roughly 3V at Vdd, the digipots were not working correctly. Right now I have applied a “band aid” to this issue by changing R86 to 3.01K ohms. This would be perfectly OK if 1 mA were the normal expected supply current. (Except that it would drain alkaline cells much faster.)
I know that this is not an issue with my board or circuit design. I know this because the first board works as intended with R86 = 100K ohms.
I have cleaned the board in the suspect area, and I have replaced the ceramic bypass capacitor. I cannot think of how I might have damaged the chip. And this fact: DC resistance to ground (without power applied) at U5 pin 4 is about 21K ohms. The resistance to ground at that pin on the other board is much higher. It is my opinion that new ICs from major manufacturers are very seldom defective. Replacing the chip will be a last resort. And if I do replace the chip without resolving the problem, I will be mad as heck…
It will be one or two days before I can get back to the lab to look at this further.
The right hand side of D17 has >5V0 on it. On posting the above, I realize that with nothing plugged into J7, leakage across both D16 and J7 could cause the problem. This seems like an unlikely situation, but I think I need to check it out. It would be easy to remove D16 and then clean that location.
Today I had an hour or two in the lab. I literally put the two boards side by side and powered up both of them simultaneously. I changed my dropping resistor back to the original 100K (actually the same chip which I had removed earlier) so that both boards would have the same resistor value.
The LM317 regulator voltages were within 1-2 mV for the 5V which I applied to the potentiometer high end (and also power 5V logic.) The only pins which were not at essentially the same voltage between the two boards were:
The +5V Vdd which I know is a problem. The first board had 5V109 (indicating that roughly 24 uA = 2V4/100K ohms is plenty of current for Vdd) and the second board had 3V846.
The potentiometer wiper output voltage which seems to get screwed up because Vdd on that chip is much less than the 5V05 on the potentiometer high end. At power up the potentiometer setting should be mid-range. Board1 was good at 2V53 but board2 was at 4V477; pretty much the same for both potentiometer wipers from that dual potentiometer chip.
It was sort of a “hail Mary” attempt; but removing that diode as I mentioned did not help. Maybe this was more of a “sleet Mary” attempt?
I feel like I have ruled out everything other than the chip itself (I replaced the bypass capacitor earlier.) But the board works OK with a 3K dropping resistor so I am still reluctant to replace the chip.
