LED meter

bruno2000 said:

Would you please check your link?
Thanks!
Best,
Bruno2000

Click to expand...

It's correct. BCL2 schemo, LED meters on page 2.

PRR said:

It will be incredibly hard to beat the Neve 192 plan for cost and parts-count. Load the output with 1K, you get 0-1V for over 40dB of input range. Trim your '339 thresholds for the steps you like.

Click to expand...

Do you have any higher resolution pic by any chance?

abbey road d enfer said:

Look there, page 2
https://groupdiy.com/index.php?topic=49634.msg762968#msg762968
Almost textbook application, except R2 and R20.

Click to expand...

That's a dual supply design, isn't it?

Bill Wilson said:

For a 5 led meter this is what you want. KA2284/KA2285

Click to expand...

Maximum voltage for these is only 16V but even if I overcome this issue I'm not sure how these can be set for -20dB to +28dB range.

Thank you for all your suggestions!

warpie said:

That's a dual supply design, isn't it?

Click to expand...

No, single 24V rail.

warpie said:

Maximum voltage for these is only 16V but even if I overcome this issue I'm not sure how these can be set for -20dB to +28dB range.

Thank you for all your suggestions!

Click to expand...

As I mentioned earlier, you simply pad the audio signal down so it falls within the range of the power supply.

Cheers

Ian

abbey road d enfer said:

No, single 24V rail.

Click to expand...

On the schematic the LM339s are powered from +VD, which appears to be +15V. But they could easily be  powered from a +24V rail.
However, the 4559 (1/2U4) and the TL072 (1/2U7) are fed with +/-15V. Am I missing something? Most likely I am, so I should rephrase the question in "what am I missing?"  

ruffrecords said:

As I mentioned earlier, you simply pad the audio signal down so it falls within the range of the power supply.

Cheers

Ian

Click to expand...

Yes I realise that. What I meant is I can't see how I can scale the LEDs on KA2284 so as the lower one will display -20dB and the highest +28db. I think they are "pre-scaled", if that makes sense.

warpie said:

On the schematic the LM339s are powered from +VD, which appears to be +15V.

Click to expand...

Correct, I had forgotten about the 15V regulator that was necessary to keep a stable reference.

But they could easily be  powered from a +24V rail.

Click to expand...

Yes

However, the 4559 (1/2U4) and the TL072 (1/2U7) are fed with +/-15V.

Click to expand...

That is correct. They provide other functions than LED meteing.

Am I missing something? Most likely I am, so I should rephrase the question in "what am I missing?"  

Click to expand...

I don't know. You wrote "I want a single rail is because I want to keep the LEDs separate from the  audio rails." The LED meter I show works out of a single rail.

What is the purpose of the 4559 and the TL072? As far as I can tell, If I use only the schematic on the 2nd page, the signal is not rectified.

warpie said:

Yes I realise that. What I meant is I can't see how I can scale the LEDs on KA2284 so as the lower one will display -20dB and the highest +28db. I think they are "pre-scaled", if that makes sense.

Click to expand...

I think the short answer is they cannot. Looks like the ratio between the comparator voltages in that device are fixed. As you say, they are pre-scaled. Ypu probably need to use a discrete comparator solution.

Cheers

Ian

ruffrecords said:

Ypu probably need to use a discrete comparator solution.

Click to expand...

Yes, that's the "easy" part and I think I have figured it out. What puzzles me is to find an efficient way to rectify the signal with a single supply.

Still working on that though, haven't given up yet  8)

warpie said:

What is the purpose of the 4559 and the TL072?

Click to expand...

The full schematic is a limiter/compressor.

  As far as I can tell, If I use only the schematic on the 2nd page, the signal is not rectified.

Click to expand...

Well, you have to find a way to rectify the voltage. Think how it was done with vacuum tubes, which were single-rail.

warpie said:

Yes, that's the "easy" part and I think I have figured it out. What puzzles me is to find an efficient way to rectify the signal with a single supply.

Still working on that though, haven't given up yet  8)

Click to expand...

AC couple it to an op-amp whose input is biased up to half the supply voltage?

warpie said:

Yes, that's the "easy" part and I think I have figured it out. What puzzles me is to find an efficient way to rectify the signal with a single supply.

Still working on that though, haven't given up yet  8)

Click to expand...

I answered this once before, in this thread...

JR

JRsaid said:

In general to perform AC functions (like rectification) from a single rail voltage establish a usable v/2 (12V). +/- 12v will handle most audio signals but perhaps incorporate -6dB in the input stage to prevent clipping. 

Click to expand...

Yeah, that's what I'm trying to do actually... I think I'm getting somewhere.

I think I'm getting somewhere.

Click to expand...

Am I? 

EDIT: It's just a modification of the "Universal Bargraph Driver" circuit, designed a few years back by SSLtech if I'm not mistaken.

You can  use the leftover opamp to buffer the 1/2 voltage. Tying it to one of the supply voltages isn't optimum, either use it as a buffer or put it somewhere between V+ and ground.

Thanks for the suggestion Volker. Does it look any better now? Also, do you reckon I should leave the C6 (10uF)?

volker said:

You can  use the leftover opamp to buffer the 1/2 voltage. Tying it to one of the supply voltages isn't optimum, either use it as a buffer or put it somewhere between V+ and ground.

Click to expand...

In the proposed schemo, there's absolutely no advantage in buffering Vcc/2, since it goes only to non-inverting inputs. However, this circuit needs smoothing in order to provide the correct ballistics; this would require additional circuitry that may require a stiffer Vcc/2.
Anyway, the major issue there is that the rectified voltage is referenced to Vcc/2; in order to get a correct indication from the LED meter, one needs to get rid of this DC component. Admittedly, it could be done by rearranging the resistor string so the meter range starts at Vcc/2, or it could be done in a differential stage.
There are easier ways to get a rectified (and smoothed) signal and referencing it to 0V.

Thanks Abbey. I joined your two schematics and the simulation gave me the following plots. Do they look correct?

I will breadboard it but I'd like to come as close as possible to the final version because ordering the components and mainly the shipping costs is a bit of an issue and ideally I'd like to place one order for everything.

Vcc = +24V

Vin =1.4Vrms @ 100Hz

Vin =1.4Vrms @ 1kHz