Effect of many i/o lines driving hi-current ?

Go To Last Post
12 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Wondering if anybody has any experience with this -

I'm planning to use an ATtiny88 to directly drive 20 ssr's, one ssr per each of 20 i/o ports.

The ssr requires 8ma to turn it on, so the tiny88 ports have no problem individually.

Concern is that turning all ssr's 'on' at once will be a 160ma draw collectively thru the i/o ports.

The tiny88's spec says 'tbd' for a sum, so I looked at the tiny861, and it says >60ma (sink or source) is a problem (as a sum from all ports) because Voh will rise as you source >60ma, and Vol will rise as you sink >60ma.

I think it's actually a typo as I would expect Voh to fall as you source more current, but nonetheless, asking the forum if anybody has experience with this (ie, how much will the output voltage vary with greater collective loads on the i/o lines).

I could add buffer drivers between the i/o ports and the ssr's, but would like to avoid cost / connections.

Thanks!

-mark

Mark
Elgin, IL (near Chicago)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

I'm planning to use an ATtiny88 to directly drive 20 ssr's, one ssr per each of 20 i/o ports.

The ssr requires 8ma to turn it on, so the tiny88 ports have no problem individually.


Stop right there, and look at the datasheet again. Mega88--agreed. Tiny88--those ports have really wimpy drive, and even the few "high current" pins are still wimpier than on a real AVR.

Trust me--you can't even sink an LED decently on the generic Tiny88 pins. And there are only a few of the "high current" ones.

Find these charts in your datasheet: "STANDARD I/O PIN OUTPUT VOLTAGE vs. SOURCE CURRENT". 5V doesn't look too bad for your needs, but check 3V. :(

Compare with the same charts for the Tiny461, and for the Mega88PA.

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I appreciate anybody taking the time to respond, but yours is not only off-topic, it reaks of arrogance - "stop right there", "trust me"....

I'm using 5v, so as I said, the individual drive on each Tiny88 i/o line is just fine, and the Mega88PA really isn't any better, other than more detail on the collective sum limit, where it also talks about a problem if the sum is >100ma (I'm using 160ma).

So.... anybody else with actually pertinent comments on this issue?

Sorry to be so blunt, but I read a lot here, and I've found so many of your responses to be worthless, antagonizing, and full of ego.... many of the people you are dealing with here are very experienced, I myself have an MSEE and over 30 years in digital/processor design, so condescention seems a little misplaced.

Mark
Elgin, IL (near Chicago)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well i don not have so many years in processor design,since i just started,but i do know a lot about electronics,and whether i do not know something,i start digging or do some tests(of course at an cost).If i ware you,i would connect 20 leds(hug in resistors to drive each at 10 mA),and do some tests and measure the currents.It can happen that the tiny88 can run weeks like this,or can blow up in a second.(i bet will blow @ 200mA sink).
I know test's can cost sometimes,but if you do it right,you can measure yourself the current and see where to stop in a safety area.
If you dont wanna test,i suggest to place a buffer between and have no worries.

Daniel

I just Love LED_AVR_GCC :)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

There is an often repeated adage.. violate the MAX specs at your own peril. A quick search at TI web site turns up this single part solution... 20 buffers in one part for $2. There are of course other buffers out there but this offers exactly 20 channels which is what you said you wanted.

http://focus.ti.com/lit/ds/symli...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

I'm using 5v, so as I said, the individual drive on each Tiny88 i/o line is just fine, and the Mega88PA really isn't any better, other than more detail on the collective sum limit, where it also talks about a problem if the sum is >100ma (I'm using 160ma).

You can bad-mouth me all you like. Our outfit has dozens of production industrial and commercial designs with Mega48/88. A low-cost app was suitable for the Tiny48. The pin drive of the Tiny48 was a surprise. An unpleasant surprise.

Hence the "stop right there". With what you described, you >>will<< have the same surprise. It ain't gonna work decently.

Quote:

the individual drive on each Tiny88 i/o line is just fine, and the Mega88PA really isn't any better,

If you think I'm being rude, wait till you get the "rude surprise". We couldn't even drive a MOSFET decently with a "normal" I/O line. Tne few "high drive" lines work OK for MOSFETs and LEDs. But there are only a few of them, and if you mix-and-match for your qty. 20 then you will have unbalance.

I guess it only costs you a few bucks to get a Tiny and Mega, and hook up you load to a few of the pins both high-drive and regular. True, I'm not addressing the TBD of total current limit. I'm gonna suspect lower.

Reprot the results of your test. I'd be happy to eat crow on this one, knowing how you have found a way to drive these loads reliably on a Tiny48.

I'll have to check on Monday whether that design was a 5V or 3V design. If ours is a 3V then indeed you may be OK, if the unbalance doesn't cause any problems.

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

If you think I'm being rude, wait till you get the "rude surprise".

Lee,

After his extremely rude response to good advice I'm surprised you even bothered to reply. I'd leave him to it and let him find out the consequences later when half his customers are let down.

It's clear someone in this thread is arrogant but it isn't you!

Cliff

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A fool and his Tiny88 are soon parted comes to mind! :D

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well, I certainly was "flippant" at best. And indeed I was trying to wave a warning flag. Now, I'll have to look at that app on Monday to check wheter a 3V or 5V app. All I know is that we couldn't drive one of our normal MOSFETs or sink an LED decently with "normal" pins on the Tiny48. I remember looking at those datasheet curves with our "team" and it should have been enough drive--but wasn't.

So I'd suggest the few-$ test and check it out before going much farther. Indeed, the about-$1 price point of the Tiny48 is enticing compared to the $1.50 Mega48 (qty. 100 disti prices). That is 50% less for the micro, and is significant on apps trying to get low BOM totals. Our volumes are modest, but we've built about 1000 of these industrial timers with Tiny48 so far. That is like $500 more in our pocket.

At 5V things look a little better, and mark might be fine. But try it first!

Quote:

Wondering if anybody has any experience with this -


Not exactly with 20 SSR relays, no. But pin-drive with Tiny48? Yes.

So take it for what it is worth. Report back, and I'll be happy to gnaw on a helping of crow. You know us old guys--always gotta give advice.

Lee

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

Sorry to be so blunt, but I read a lot here, and I've found so many of your responses to be worthless, antagonizing, and full of ego.... many of the people you are dealing with here are very experienced, I myself have an MSEE and over 30 years in digital/processor design, so condescention seems a little misplaced.

_________________
Mark
Elgin, IL (near Chicago)


Yes, indeed, I am the resident Surly Curmudgeon.

BTW, many of our AVR designs get packaged into controllers in Elgin, and one of our board houses is there as well for other designs. So many 10s of k of AVR/year of our designs at least pass through Elgin. Small world.

Now, that off-topic aside probably falls into the "worthless" category.

"Antagonizing"? Hmmm. Yep, must be my overall attitude. And indeed "Back the AVR truck up beep beep beep" is often used, and would correspond directly to my "Stop right there". I have to take exception to "full of ego", though, as my blusterings are a symptom of low self-esteem.

Lee

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

There's a trick I've used to get around the total source or sink current limit per port, which is on each port to connect half the loads to GND (drive them high) and half the loads to Vcc (drive them low).

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ah, written text can be so easily misinterpreted, and I thought Lee was perhaps starting a sling, like I have seen on some of his other threads. I see from his courteous and thoughtful replies that I was quite wrong, and I apologize for my overeaction.... I may be a grouchy old man.

Appreciate the responses and ideas here. I'll take Daniel's point and do a little prototyping to see if the proc misbehaves, overheats, etc, under the heavy drive load, and if so, I like the TI chip that Al found - perfect as a full i/o buffer on 1 chip, for this app. KK6gm's source-splitting idea I had already incorporated, but good to hear that somebody else thinks this may help too.

I also looked at the ssr spec more, and they show typical in-drive trip points as ~6ma, but spec 10ma, so my 8ma design target just bumped to 10ma and 160ma collective went to 200ma, exacerbating the whole problem....

Lee - is Emerson your customer in Elgin? I know they use a lot of AVR's in their small-appliance controllers. I work for an aerospace company in Rockford, IL, but we have employees from Emerson that I've had small-proc design discussions with.

-mark

Mark
Elgin, IL (near Chicago)