Wrong component quantity computation

[set-soft]

Current kicost from git computes quantities wrongly when the multiplier mechanism is used.

Using the following schematic, Schematic.pdf

We get the following XML where the most relevant parts are:

    <comp ref="R1">
      <value>0</value>
      <datasheet>~</datasheet>
      <fields>
        <field name="Description">RES SMD 0 OHM JUMPER 1/8W 0603</field>
        <field name="digikey#">CR0603-J/-000ELFCT-ND:  5</field>
        <field name="manf">Bourns</field>
        <field name="manf#">CR0603-J/-000ELF:5</field>
      </fields>
      <libsource lib="Device" part="R" description="Resistor"/>
      <sheetpath names="/" tstamps="/"/>
      <tstamp>6064C12F</tstamp>
    </comp>
    <comp ref="R4">
      <value>0</value>
      <datasheet>~</datasheet>
      <fields>
        <field name="Description">RES SMD 0 OHM JUMPER 1/8W 0603</field>
        <field name="digikey#">CR0603-J/-000ELFCT-ND</field>
        <field name="manf">Bourns</field>
        <field name="manf#">CR0603-J/-000ELF</field>
      </fields>
      <libsource lib="Device" part="R" description="Resistor"/>
      <sheetpath names="/" tstamps="/"/>
      <tstamp>6064EF62</tstamp>
    </comp>
    <comp ref="R5">
      <value>0</value>
      <datasheet>~</datasheet>
      <fields>
        <field name="Description">RES SMD 0 OHM JUMPER 1/8W 0603</field>
      </fields>
      <libsource lib="Device" part="R" description="Resistor"/>
      <sheetpath names="/" tstamps="/"/>
      <tstamp>6064F812</tstamp>
    </comp>
  </components>

R1, R4 and R5 are 0 ohms resistors.

• R1 has a 5 multiplier
• R4 doesn't have any multiplier, counts as 1
• R5 doesn't have a manf# field, counts as 1

This is a total of 7 resistors.

Here is the BoM generated by KiBot, it doesn't support KiCost multipliers, so we get wrong values: subparts-bom_r1.xlsx. But it helps to display all the relevant fields.

Now, running:

kicost --no_price -wi subparts.xml

we get: subparts.xlsx

Taking a look at the Qty cell for R1,R4,R5 we see =BoardQty*3 which is wrong.

I'll create a PR to fix it. But we need more test cases to ensure the patch doesn't break other stuff. Additionally, my patch will add qty pre-computation. In this way the Qty column we get from xlsx2csv has values we can verify, and not just 0.

Looking at the code I see some stuff that is quite confusing to me. The code seems to support the multiplier mechanism for manf#, which generates manf#_qty. But it also supports it for all distributors, generating distributor#_qty. The code even checks the multipliers are the same. Why this? Isn't manf#_qty enough? I understand the user can then ask to buy some ammount from one distributor and other ammount from other, but I don't see why this is related to the multiplier in the fields.

I also see the code is failing to consolidate subparts that are the same as regular parts, or even repeated subparts.

[hildogjr]

The reason of coexistence of manf#_qty and distributor#_qtys was because some time I had a equivalent part, if I purchase a different quantity, in different distributors. Something like:

MPN = PART1 digikey# = 2 : STOCK_DIG1 mouser# = 3 : STOCK_MO2

digikey# and mouser# are prioritized over MPN when searching in each of this distributors.

If I remember correct (I wrote this), I check if the quantity is the same just to warning on terminal.

I also see the code is failing to consolidate subparts that are the same as regular parts, or even repeated subparts.

I think is related with https://github.com/xesscorp/KiCost/issues/266. I didn't check this yet.

[set-soft]

I also see the code is failing to consolidate subparts that are the same as regular parts, or even repeated subparts.

I think is related with #266. I didn't check this yet.

Yes, I think is related.

BTW: propagation of manf# looks like a bad idea to me. I understand people is lazy, but assuming two components are the same just because their value is the same is a very bad idea. In fact I think things like manf# and distributor# are the only reliable grouping criteria. But this my opinion.

I'm finishing a "Subparts" filter for KiBot. This filter is intended to add better compatibility for KiCost users. After some tweaks (well a lot of tweaks) I'm generating an acceptable BoM for the schematic I mentioned before: subparts-bom.xlsx

I needed to add an optional field to provide Value replacements. Using Conn_01x06 pN/5 for a sensor looks really confusing to me.

[set-soft]

The reason of coexistence of manf#_qty and distributor#_qtys was because some time I had a equivalent part, if I purchase a different quantity, in different distributors. Something like:

MPN = PART1 digikey# = 2 : STOCK_DIG1 mouser# = 3 : STOCK_MO2

digikey# and mouser# are prioritized over MPN when searching in each of this distributors.

I don't fully understand it. Isn't this something you can choose from the spreadsheet? Do you have a real life example?

If I remember correct (I wrote this), I check if the quantity is the same just to warning on terminal.

Yes, the code warns about it.

[hildogjr]

Usually pins headers or borniers (for example purchase 2 x 2 pins bornier on one distributor and 1 x 4 pins in other).