Here's a nice simple one for you.... If my design has 5 CMOS serial ports, and I use something like 3 x MAX232, which have 2 drivers/receivers on each chip, then what should I do with the unused inputs/outputs on the 3rd chip?
Should I leave the CMOS inputs floating or tie them up or down? What about the RS232 inputs? Should I leave them floating or tie them to VSS or VCC?
Tie the CMOS inputs high through a moderately large resistor. The MAX232 inputs may be left unconnected, since "[t]he inputs of unused drivers can be left unconnected since 400k? input pull-up resistors to VCC are built in (except for the MAX220)."
"lighthouseman" schreef in bericht news: email@example.com...
You'd *never* leave unused logic inputs undifined. Usually CMOS are tied to GND, TTL should be pulled high. Ancient TTL needs a pullup resistor in the range 1-10k, for LSTTL this resistor can be omitted though I prefer to use it. Some components have internal pullups or (pulldowns) but if you are not sure you'd better use an external.
FWIW, if you should encounter the entirely pin-compatible MAX3232 (which allows operation down to 3volts) BE WARNED that for some reason Maxim have omitted the 400K pull-up resistors that are on the CMOS inputs of the MAX232. If your driving circuit needs them you'll have to add them externally.
David C.Chapman - (firstname.lastname@example.org)
First thing you should consider is you often don't even need 232 chips, especially for short connections. While the spec allows the receiver trip point to be between -3V to +3V, and technically a 0V trip point would be optimal, most receivers trip at a bit over a volt. Thus you can drive 232 receivers with CMOS. You only need to go between the supply rail and ground.
Now rolling your own receivers is another story, but basically it is done with voltage dividers, a voltage clamp, and CMOS inverters. If you traced out the receiver used in most 232 devices, that would be what you find.
Receivers have a resistor built in, pulling them to ground. You would need to consult the data sheet regarding transmitter inputs. If the device has pull ups on the transmitter input, then float it, otherwise ground it. [Grounding pins pulled up wastes power, but you knew that. ;-]
The reason I mention you don't really need a 232 chip is every 232 chip on the market adds jitter to the signal. It is hard to keep the charge pump noise out of the receivers and even transmitters to some degree. I've measured jitter such chips (well, my designs) and the peak to peak is about 1uS. Not a big deal in 232, but something to think about. I don't think any of the data sheets spec the jitter.
There are some 232 type chips with a negative trip point on the receiver. This was used in AutoShutDown mode. It should be on a handshake line, not data. [I can't possibly know ever variation of 232 out there, so I need to be a bit vague about this.]
-------- Did the Maxim police cut you off at mid-sentence?
How do you insure the 1n4148 doesn't hog all the current, preventing the NPN from turning on. Does this just depend on the ratio of junction area between the diode and transistor?
There are devices that only need to receive 232. In that case, a home brew receiver is fine. Rolling your own transmitter is another story. There are cases where the output is shorted. You really can't controller the external world. But a bit of resistance in series with the output would do the trick.
I made a data slicer that didn't use a 232, though I've seen designs swearing you need one. By skipping the 232 chip, I made the design much easier to phantom power from the transmitter. If you ever do this, shoot for single supply use since most of these 232 chips just charge pump invert the boosted signal, so the negative supply is not as beefy.
[Given your website and hobbies, I suspect you know what a data slicer is.]
On a sunny day (Thu, 7 Jan 2010 19:05:08 -0800 (PST)) it happened " email@example.com" wrote in :
That is a remaining line from your own posting, did you not recognise the text?
I use the MAX232 a lot too. Just not the 3.3V version, and that only because it is not in the shops here. I am not Joerg, who can only afford 2 cent transistors. Poor guy ;-)
They are reverse connected, the diode conducts when the input is negative, protecting the be junction from too much revere voltage, and the be junction conducts when the voltage is positive.
In this case the current limit is set by the DTR and RTS lines... So it is always in spec if i connects to a real RS232 port.
Actually electronics is not only a hobby, also made a living with it.
To 'do the RS232 thing right' there are many ways. In an other design I made negative with a simple sine wave oscillator, rectifier and pot core. That then also powers other stuff. Creating a negative supply almost never is a problem, here are some details of what is one directory level up: The generator: ftp://panteltje.com/pub/8052AH_BASIC_computer/dc_dc_converter_detail/5V_to_30V_converter.jpg The EPROM programmer switch: ftp://panteltje.com/pub/8052AH_BASIC_computer/dc_dc_converter_detail/EPROM_programming_voltage_generator.jpg The RS232 interface: ftp://panteltje.com/pub/8052AH_BASIC_computer/dc_dc_converter_detail/RS232_interface.jpg Hope this helps:-)
My point about trying not to use the negative supply is not that it can't be done, but it is higher impedance and lower efficiency IF it is derived from the boosted positive supply. I did a 232 with shared inductor rather than caps. In that case it didn't derive the negative supply from the positive supply, but still a boost DC/DC is beefier than an inverter since the boost just adds to the input supply.
I don't have the 232 spec handy, but I don't recall a current limit on the transmitters when I was designing those 3V parts you don't buy. ;-) The lack of a current spec made driving a serial mouse a roll of the dice. One of the datasheets for my chips had some blurb about how it could drive 10 (or something like that) popular serial mice. It didn't mention the one mouse it couldn't drive. Prior to the design, I went to Fry's and bought what serial mice we didn't have handy, then measured the current the mice needed. That produced an internal spec, but as luck would have it a newer design was more power hungry.
On a sunny day (Fri, 8 Jan 2010 13:43:44 -0800 (PST)) it happened " firstname.lastname@example.org" wrote in :
Yes, indeed, we discussed that here a while back I think. This because of my temp_pic and freq_pic designs,
those just limit the DTR voltage with a zener to about 4.7V. From what I have measured, the zener current stays below the max value on my USB to RS232 adaptors, and on the serial COM1 and COM2 ports of my PC. The currents are different though, resulting in a different voltage. It does make some difference on temp_pic in the calibration. Also depends on the type of zener used I guess. I can live with it, it all stays in one piece., and saves MAX232 chips or dedicated hardware, doing it 'the mouse way'.
It never occurred to me anyone would be hanging a zener directly on a transmitter. Most of these 232s want to do about 8V out the transmitter. I think some resistance would be nice before hitting the zener. I think the mice use regulators, but now that I think about it, I never curve traced the mouse.
On one of my 232 chips, the apps guy found it wouldn't start up if 3 transmitters were shorted to a negative supply. Value escapes me, but it was a legal voltage for the pin, even though in the scenario used, it was very unlikely. My boss said "I bet it doesn't start up if you attach it to the AC line. This is nuts! I had to tweak ESD, so I made some changes to the transmitter, but really, you can't build chips for every conceivable customer abuse. Given an emission microscope, it was possible to make the transmitters work under these conditions, but such antics are really a waste.