U.A.2.1.0.8. type Tri colored led circuit.

> save as

I found some screw-ups. I fixed them in-place. Save-As.. again.

[quote author="TomWaterman"]That is awesome stuff PRR - will be on the right click - save as ala Kev!![/quote]

CORRECT

and remember to give them sensible shorter names as the long WEB names don't write to CD so well

nice meter!

but if i need to extend it at 30 led segment for a large
and accurate meter?
there is a possibility about?

???
6t9r

I found this thread and had PRR reupload the schematic, but I dont fully understand it.

Is this correct?



Should the "?" pins from the LM339 be connected to ground?

Is it possible to supply it from +-5V?

That loks like a lot of parts for just a signal activity LED.  I made one for use in consoles from one opamp (while almost as many discrete glue parts as PRR).  I only provided green for signal present and red for potential overload. I never found the both on color that useful, and it made it harder to discern the red overload condition.


JR

PS: the ? in PRRs circuit are DC voltages you need to provide (with voltage dividers) to establish the desired indication thresholds.

[edit] I forgot to mention I have already posted my circuit...
http://groupdiy.com/index.php?topic=53166.msg678671#msg678671 

please don't think that I somehow prefer PMs for general interest questions.  and don't ignore the search function like I used to find my post    [/edit]

> Should the "?" pins from the LM339 be connected to ground?

Think. Then everything happens at *zero* Volts. What do the LEDs tell you? Nothing.

Look. The "A" picture shows voltages. Different voltages.

To-Ground can't be right.

> Is it possible to supply it from +-5V?

Read LM339 data-sheet. Yes, it will do-the-right-thing with supply below *single* 5V (and up to 35V total). So that part is OK.

Ponder the resistors feeding the LEDs. If you go from my 30V total to 10V total they will be 1/3rd as bright. (LED light output is very-near proportional to LED current over the range from dim-glow to over-heated.) The eye is as logarithmic as the ear. 1/3rd light may be very fine. If not, think about the resistor values and the LEDs currents.

Everybody did stuff like this decades before AOL and Al invented chatrooms and the internet.
_______________________________________

Sorry about not restoring the second set of missing images. Obviously I have a problem with my image-host. Also the second set was not on this machine. And I'm fighting a persistant head-cold and not thinking well.

http://groupdiy.com/index.php?topic=877.msg12466#msg12466

I have found that chopping up a teaspoon of raw garlic and washing it down with some juice will chase away persistent colds (and other humans). I don't know if it kills the bugs or jump starts our immune system, or both. It just seems to work more often than not.

Get well...

JR

Thanks guys! I really appreciate your help and knowledge, and I see my stupid questions now 

btw ginger is a really good health jump starter as well

take care...

2 more cents.

Usually the ? are connected to a voltage divider setting the comparator levels for each led/color, 1.23V for +4db etc.

The many parts part seems to be a rectifier which is nice but does seem like a lot of parts, although you can use a quad chip.  The rectifier is useful in case the positive half and negative half of the signal are substantially at different levels, but this thing only measures 3 levels so probably overkill.

if you plan on hanging this across a signal line... then the issue I would state is that I want a meter not to interfere with a balanced line and so ideally a de-balancing circuit AND equal impedance to ground so it doesn't mess with the CMR of whatever you are monitoring.

Elsewhere internally in some circuit on an unbalanced signal that is a non-issue.

KasperNyhus said:

I found this thread and had PRR reupload the schematic, but I dont fully understand it.

Is this correct?

Should the "?" pins from the LM339 be connected to ground?

Is it possible to supply it from +-5V?

Click to expand...

> The many parts part seems to be a rectifier which is nice but .... ....  probably overkill.

That's stolen directly from old National Semi sheets.

You need -some- form of filtering/averaging or (on speech/music) it is all flicker.

You want -some- form of semi-precision rectifier with reference levels likely around a half-volt.

Many one-opamp half-wave plans exist. Some are just incredibly clever, even cute. Most have "some" drawback; low input Z, or different input Z for one side or the other. As you say, "don't interfere".

This stolen plan is just 2 opamps and has consistent 33K input any which way. In an unbalanced world, this is "just another load".

As you say, signal conversion may be part of the total problem, but extracting signals from different sources can become a whole other book-shelf than the blinky-light books.

Just because, I'm more of a J.R. fan , Thanks !

PRR said:

> The many parts part seems to be a rectifier which is nice but .... ....  probably overkill.

That's stolen directly from old National Semi sheets.

Click to expand...

And they are trying to sell more opamps...  ;D

You need -some- form of filtering/averaging or (on speech/music) it is all flicker.

Click to expand...

Years ago  I invested some time trying to make a cheap bi-color display that didn't suck. IMO signal present (green) does not need much help, with a low enough threshold it doesn't take much audio signal present to deliver a readable display, even half wave.  The peak (red), OTOH does require significant hold  time to not ignore brief transients or marginal signals.

You want -some- form of semi-precision rectifier with reference levels likely around a half-volt.

Click to expand...

In my experience precision is not worth the effort/cost for simple A) is signal present in that path?  and B) is the signal too hot?

Many one-opamp half-wave plans exist. Some are just incredibly clever, even cute. Most have "some" drawback; low input Z, or different input Z for one side or the other. As you say, "don't interfere".

Click to expand...

I had to go back to look at what I did. My input Z is as low as 33k ohm and not perfectly linear. The input node labelled SMPLA is a common sample node where multiple diode cathodes combine together (diode or) with the individual diode anodes sampling levels at multiple circuit points. Only half wave but in my experience sampling at multiple circuit nodes "may" provide both polarities so full wave after a fashion. 

This stolen plan is just 2 opamps and has consistent 33K input any which way. In an unbalanced world, this is "just another load".

As you say, signal conversion may be part of the total problem, but extracting signals from different sources can become a whole other book-shelf than the blinky-light books.

Click to expand...

I have some even cheaper approaches for crude indicia based around discrete transistors, but at Peavey my early designs were before they had machine insertion for transistors, so the cost of a hand inserted transistor was way way more than 1/2 a machine inserted dual opamp. So while my old design looks like a lot of parts, they are all machine inserted and often fractions of a penny a piece. The opamp was relatively expensive at something like $0.13 for the dual, $0.065 per opamp.  ;D ;D

JR

PS: I invite curious circuit designers to look at this and count the silly opamp tricks..

The opamp is operating as a high gain inverting stage to detect for low level signal present, until the input pulls the + input higher than the - input and then the opamp output flips positive until the hold cap discharges. The two lead bicolor LED connects between APK+ and APK-