Capacitors C3 - C6

Information and announcements on updates to the Megajolt firmware and PC software

Moderators: JeffC, rdoherty, stieg, brentp

Post Reply
aarc240
Posts: 65
Joined: Mon Dec 18, 2006 10:54 am
Location: Mid North South Australia

Capacitors C3 - C6

Post by aarc240 »

OK so we use a MAX232 and 1uF or we use a MAX232A and 0.1uF.

Having had some experience with these chips when pushed for speed, my debugging results may help someone.

1. use polarised tantalum capacitors installed the correct way around regardless of which value is needed. Ceramic capacitors are not really right even though they work ok with shorter serial cables in a low noise environment.

2. use the MAX232 and 1uF tantalums if you are going to use a lllooonnnggg serial cable.

Brent has done an outstanding job with layout on the V3 board so it's just a matter of installing C3, C4 & C6 with +ve on the capacitor toward the MAX232 and C5 with +ve toward pin 16 on the MAX232. Unfortunately I'm not familiar with the V2 board.




MartinM
Posts: 433
Joined: Sat Aug 27, 2005 12:08 pm

Interesting. When you say

Post by MartinM »

Interesting.

When you say "pushed for speed" are you talking about very high baud rates? MJLJ only uses 4800 baud - so do these considerations still apply?

And how long is "lllooonnnggg"?

Do you have any experience of the SP232ACP from Sipex where (the datasheet at http://www.sipex.com/Files/DataSheets/sp232_312a.pdf says) "the charge pumps will operate with polarized or non-polarized capacitors ranging from 0.1 to 1μF"?...
...but there are no indications of any limitations that may come into play depending on the type and value of capacitor chosesn
...and the SP232ACP is a lot cheaper than the MAX232A as well!

aarc240
Posts: 65
Joined: Mon Dec 18, 2006 10:54 am
Location: Mid North South Australia

Pushed for speed

Post by aarc240 »

Correct in that we were using very high baud rates, less critical with lower baud rates.
Some of our cables are in excess of 200m and in a very noisy electrical environment (lots of RF!).
I have looked at Sipex but no experience - no Aus outlet & not enough motivation to import.

Appears that the voltage swing on the data lines is more consistent with the tantalums. The ESR of tantalums is probably more significant than being polarised.

btw, a larger capacitor (within reason) in a charge pump type voltage stepup will nearly always give better response to transients on the load. One of those rare ones where bigger is usually better!

Post Reply