1
\$\begingroup\$

According to the datasheet the low power version of the 386SX comes in a QFP100 and can be run at as low as 2MHz. I'm wondering if it might be possible to prototype a system on a standard solderless breadboard using a QFP100-to-DIP100 test socket. Based on what people do with 6502's, 2MHz seems like it could be okay but I'm worried about the max values of C[IN] = 10pF, C[OUT] = 12pF, C[CLK] = 20pF on page 67 of the datasheet and maybe other things that aren't even listed. I'm guessing this is probably stupid but I just want to play with a minimal real 386 setup.

According to Dave from EEVBlog, who measured the the parasitic capacitance on several different breadboards at 1kHz, it's about 2pF between lead columns and 20pF between power rails. I see people doing Z80 projects on breadboards at 4MHz and I see this answer says under 10MHz should be okay. So it seems like it might work if I'm reading the situation right, but am I missing anything here? I'd like a bit more confidence than that.

\$\endgroup\$
4
  • \$\begingroup\$ What is wrong if an input of the 386 SX has less than 10 pF parasitic capacitance? It is not the limit for external capacitance caused by wires and connected outputs. \$\endgroup\$
    – Uwe
    Commented Oct 6, 2021 at 21:51
  • \$\begingroup\$ I'm worried that there will be more than 10pF parasitic capacitance due to the breadboard, not less. It seems like EEVBlog measured this on several breadboards and got approximately 2pF inter-lead and about 20pF on the power rails. youtube.com/watch?v=6GIscUsnlM0 \$\endgroup\$
    – Anthony
    Commented Oct 6, 2021 at 22:51
  • \$\begingroup\$ I'm guessing this is probably stupid but I just want to play with a minimal real 386 setup. Wouldn't you want to put at least a few decoupling caps directly on the QFP-to-DIP adapter board, anyway, thus requiring you to lay out a quick board either way? Wouldn't that make the "minimal real 386 setup" much easier if you start assuming a board for anything that's not explicitly very low speed? I don't consider 2×50 rows of breadboard "easy", at all. On the contrary - it's incredible hard to design that way, wheras board design has much freedom, good reliability and easy routing, great tools… \$\endgroup\$
    – Marcus Müller
    Commented Oct 7, 2021 at 10:02
  • \$\begingroup\$ … and with PCB manufacturing in the < 10€ range... really, drop the breadboard requirement. \$\endgroup\$
    – Marcus Müller
    Commented Oct 7, 2021 at 10:03

1 Answer 1

Reset to default
2
\$\begingroup\$

The specifications you mention are for the capacitance of the input pin of the 386, not the maximum allowed capacitance on that pin. You can see that the input capacitance is twice as much on the clock input. This means any circuit wanting to drive that pin needs to be 'strong' enough to drive the capacitive load and comply with the voltage level specs.

Fundamentally, a capacitor wants to resist the change of voltage. Due to this, faster signals have a harder time to change the voltage quickly enough. This explains why your protoboards aren't really good for high speed circuits due to their higher capacitance.

In short, the chip does not care how much capacitance is connected to the input pin.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.