SMT mylar capacitors

[Alex Eisenhut]

Going through my TRS-80 Model 102 parts list, I saw some caps are Mylar. They are 1206 parts. I've never seen Mylar caps that small in SMT and in any case, I've only ever seen them as little SMT stacks of sheets, not a monolithic white chunk.
They don't seem to be used in anything particularly demanding, C71 C72 and C73 in the level shifting networks in the attached schematic.
VDD is +5V and VEE is -5V in the 102. I don't see anything special going on here, it's simply 5V CMOS logic going in from the right, and feeding these MC14584 hex Schmitt inverters which are powered at +/-5V, hence the level shifting.
Of course, I am upset that regular SMT 1206 caps are not white.
1) Can anyone tell me if Mylar caps are needed here? This is the machine's RS-232 serial port.
2) Can anyone tell me if there are 1206 39nF capacitors that are white?

[coppercone2]

well they won't work as ultrasonic transducers and disturb other things on the PCB. They also might have slightly more parasitics so they won't produce as big of an inrush as ceramics.

Looks like an old circuit, maybe the ceramics had a bad reputation then. I would still consider it high end to use them in this circuit, gilding the charge pump.

[Yansi]

There are none mylar/foil SMD caps in the photo.  The reason for using them (mylar/foil) is mostly linearity. Most SMD ceramics above about 1000pF are  using highly non-linear materials, making them extremely difficult to use for anything else than supply decoupling.
Of course higher capacitance NP0/C0G ceramics exist, but price goes significantly higher.

Also as you have already pointed out, there is no reason for a foil capacitor in that part of a circuit.

1206 mylar caps likely do not exist, not 39nF at all. Will be more like of a size of a typical THT foil cap, but without the leads.

[coppercone2]

for a max232 everyone is gonna by default use ceramic capacitors, the data sheet says they are acceptable and people just think 'its a fucking digital logic thing fuck it im not spending any money on that'

but, if you have capacitor nonlinearity, a consequence is that EMI profile might change over temperature range and across other drift parameters significantly with a charge pump circuit.

[Alex Eisenhut]

I guess I should add that the BOM in the service manual lists C71-C73 as "Mylar".

[coppercone2]

'The capacitor type used for C1–C4 is not critical for proper operation. The MAX232E and MAX241E require 1µF capacitors, although in all cases capacitors up to 10µF can be used without harm. Ceramic, aluminum-electrolytic, or tantalum capacitors are suggested for the 1µF capacitors.

When using the minimum recommended capacitor values, make sure the capacitance value does not degrade excessively as the operating temperature varies. If in doubt, use capacitors with a larger (e.g., 2x) nominal value. The capacitors’ effective series resistance (ESR), which usually rises at low temperatures, influences the amount of ripple on V+ and V-.

Use larger capacitors (up to 10µF) to reduce the output impedance at V+ and V-. This can be useful when “stealing” power from V+ or from V-. '

But what I said about the film vs ceramic is still relevant, especially if you use it as a precision sensor interface.

[Yansi]

Then I would call BS on that. 1) there is no reason for mylar there, 2) there is simply no reason. 

PS: MAX232 sucks.  Do not use!


PS for others (especiall coppercone2): There is NO MAX232 chip in the circuit. Hell... have at least a brief look at the post before answering!

[coppercone2]

charge pumps will vibrate your PCB. Someone on this forum posted a measured case before.

And if you change the value of a switched power storage part the EMI profile is defiantly going to change.

*i assume the circuit was a discrete version of a max-type chip

[Yansi]

NO MAX232 part in the circuit!

And please do not assume anything. It is a simple schmitt comparator with a frequency compensated voltage divider.

There is no charge pump, please stop bitching about it.

[coppercone2]

what the fuck am i bitching about? relax aspie

[Alex Eisenhut]

Yeah I wasn't super clear on that, the serial port is driven by the level-shifted doubled-up inverters. There is no MAX232, none is needed as there already is -5V on the computer.
I don't see that Mylar caps are required here, I simply want to replace white caps with white caps. I guess it's just a bit of an irrelevant obsession of mine...
I was just surprised that there was such a thing as 1206 Mylar caps. I guess not...

[Alex Eisenhut]

It is a simple schmitt comparator with a frequency compensated voltage divider.

Yeah that's pretty much what I see when I look at the schematic. I didn't see the need for Mylar here.
Boy I hope my next question about Japanese logic families will be clearer... 

[drussell]

Perhaps there was an earlier version of the circuit board that used different capacitors in those positions?  Maybe something like those little (often green) through-hole mylars that used to be in everything of that era?

Wasn't the model 100 basically all through-hole?  What did it use?  Maybe they basically copy & pasted that part of the circuitry and didn't update the BOM properly if the 102 was always SMT in those positions.