# Timer circuit help

#### Søren

Joined Sep 2, 2006
472
Hi,

I am sorry, I got the relationship 'upside down'. It should have been -ln(.4). -ln (10v-6v)/10v.

-ln {(Vps-Vth)/Vps}. Of course the threshold for devices will vary, I was just pointing out that it probably would not be 50% -ln (10v-5v)/10v (.693).
Eh ?

According to this, the -ln((Vps-Vth)/Vps) part (I assume you are referring to Supply Voltage and (upper) Threshold Voltage respectively?), you expect to see the same timing whether the lower threshold (in your example) is 5.99V or 0.01V

As far as the RS232 thing, yes a graph showing all the players would be great.
I think I have it in an book about interfacing from the eighties. I'll see if I can find it and scan it (or, if I get the time, draw it the hard way).

#### Søren

Joined Sep 2, 2006
472
Hi,

It hasn't happened again because ever since I used a MAX232 convertor.
That's a tall (and most likely wrong) assumption

Now if you'd said something like "Since I've been using a MAX232 (level translator) I can't tell if it would have happened without it", it would have been a bullet proof statement.

#### pebe

Joined Oct 11, 2004
626
Søren;20682 said:
Hi,

That's a tall (and most likely wrong) assumption

Now if you'd said something like "Since I've been using a MAX232 (level translator) I can't tell if it would have happened without it", it would have been a bullet proof statement.
Søren, you are nit-picking!

I wonder why companies like Maxim go to the trouble of designing and marketing TTL to RS232 level converters if you think they are unnecessary. They must have been supplied in their tens of thousands.

I encountered corrupt data once when connecting the RS232 port of a PC to the 0V to 5V of a microcontroller, and I think that on the grounds of probability it would have happened again had I not used level converters. So I think my statement was quite correct

#### Spoggles

Joined Dec 2, 2005
67
Hello:

Actually Vps should be called the highest voltage that the cap will charge to. The closer to this max voltage that the charge gets to, the longer it will take to get there. The relationship is a function of -ln((V highest-Vth)/V Highest).

I guess this is the electronic version of zeno's paradox,
as -ln(0) is not a member of the number line. (real of imaginary)

Spoggles

#### Søren

Joined Sep 2, 2006
472
Hi,

Søren, you are nit-picking!
I don't thinks so... You claimed that it would have went wrong if you didn't...etc.
It's not nitpicking to point out that this is nothing but an assumption based on biased guessing, just reality

I wonder why companies like Maxim go to the trouble of designing and marketing TTL to RS232 level converters if you think they are unnecessary. They must have been supplied in their tens of thousands.
Oh well, first of all, you have to up that number a couple of decades (at least).
Second, you have to see the difference between a hobby design with short range and a user with lots of patience v.s. a pro designer who needs to be able to give guarantees and who might be ruined if they don't hold up.

I encountered corrupt data once when connecting the RS232 port of a PC to the 0V to 5V of a microcontroller, and I think that on the grounds of probability it would have happened again had I not used level converters. So I think my statement was quite correct
You think that it would have gone wrong and this leads you to think that your statement was correct...
Two maybes doesen't yeild a 100% definitely

Let me requote your exact words:
"It hasn't happened again because ever since I used a MAX232 convertor."
A rather categoric statement to base on a double maybe I'd say
If you want to call it nit picking to point that out, then do go ahead, a rose is still a rose, whatever you call its wrapping.

#### Søren

Joined Sep 2, 2006
472
Hi,

Actually Vps should be called the highest voltage that the cap will charge to. The closer to this max voltage that the charge gets to, the longer it will take to get there.
Of course, but usually the levels is set so that it is kept in the reasonably linear portion, but you're mssing my point... You need to have the lower threshold in your formula, as you won't be charging from ground - or to the positive supply for that matter.

The relationship is a function of -ln((V highest-Vth)/V Highest).
I have to strongly disagree, as long as you don't have the two threshold voltages in the formula, it IS wrong.

I guess this is the electronic version of zeno's paradox,as -ln(0) is not a member of the number line. (real of imaginary)
Yes, but luckily, the real world works whatever the theories say

#### pebe

Joined Oct 11, 2004
626
Søren,
Earlier in this thread I said that it may be necessary to take a 1 to a ve voltage. I was in no way being dogmatic about it  I was just trying to pass on my experiences. After all, the spec for RS232 calls for a minimum excursion of the negative part of the waveform to -3V relative to ground  and your circuit didnt.

You countered by implying that circuits that would not accept limits of 0V to 5V must be badly designed. On the contrary, I consider any circuit designed to operate with signals outside their specified parameters is a bad design. Using 0 to 5V, a long line and high data rate, the received 1 signal may not get down to 0V. The laptop you mentioned whose serial signals do not go negative probably uses the RS423 spec  it is certainly not RS232.

In trying to justify your stance, you should not assume that because you have not experienced problems with out-of-spec signals, then they dont exist.

Up until now, this forum has been free of smart a***s. Not so now. Ill let you have the last word  as you no doubt will!

#### thingmaker3

Joined May 16, 2005
5,084
Gentlemen, please target your passions not toward each other, but toward the subject matter. Argumentum ad hominem is not good debate. Good debate is always welcome. I am learning from both of you.

#### ashokvalmiki

Joined Sep 17, 2007
1
how to decide the step size of the ADC?