Combining LM3914 '3915 '3916 for log/lin LED display

[clintrubber]

Hi,

Not sure this is Drawing Board stuff, but here goes:

For some reason I'd like to combine a level-indication and a 'phase correlation meter' into one LED-bar.
So that you can select whether it displays level (say 'log'/'VU') or phase correlation info from two sources ('lin'),
like http://www.groupdiy.com/index.php?topic=9301&postdays=0&postorder=asc&start=0.


Easiest approach I see so far is to connect say a LM3914 & LM3915 to the same bank of LEDs and by appropriately routing each of their input-signals determine which one is active. Since those outputs are open collector I see no problem there. Am I overlooking something ?

An alternative is to use one IC and to fumble with the top of the reference divider.
Or with some log-shaping of the input-signal, but the two IC-approach seems the most straightforward.

Anyone aware of ICs that can do both right away ? So with something like a lin/log-response-select pin ?

Thanks,

Peter

 

[clintrubber]

Thanks Wayne for the response :thumb:

[quote author="mediatechnology"]Haven't really thought about paralleling the outputs too hard but one thing to think about is the fact that the LEDs are usually in a series string. IIRC the 391X-series has a programmable current sink output not actually open collector which may not matter.[/quote]
I've seen both things possible with those LM391X's: series & //-LEDs. One neat trick is to set the mode-pin for 'dot' & then connect the LEDs in series. You get a bar again like that, but have some possibilities with where to connect the supply for partial bars and reverse bar etc.
(Nice one page summary of that in Electronics World-mag BTW. March 1998, page 227. Complete article about a 'dotscope' goes from p225-230.
They still have that on-sale themselves though, so I'm not uploading it)

Have you thought about putting a THAT 2252 rms detector (linear ~6mv/dB) ahead of the LM3914 linear display and switching its' input? I do have some if you need them. You can futz around with the detector's cap value to make it more peak-responding.

PM me clintrubber...

EDIT: Just got to the datasheet. You're right about them being essentially OC outputs (w/current control) I suppose they could be paralleled. I remember other meters - and think I have seen the LM391X-series maybe in dot mode - being used with the LEDs in a series string to reduce curent consumption. I still think a log detector ahead of the LM3914 would be a good way to do it because you could set the dB/step to just about anything. Following the 2252 you would need a scaler and offset opamp to set the range and "zero."

Thanks for the suggestion. Yep, have been considering that, but since it's meant for a decent but not super accurate indication device I'm after the least amount of complexity/soldering time. But wait, still have that AD636JH sitting around here - meant for a dB(SPL)-meter. Paid top dollar for that - better use it sometime.


Let me rethink: I'm in for a stereo LED-bar, 'VU' ('log') for two channels, and switchable to 'lin' for one bar to display L & R 'phase correlation'.

#1
Using a 3914 & rms-detector for one channel and a 3915 for the other (unswitched) channel requires some effort to let both VU's match.
Or use two 3914 and two rms-detectors. Two LED-bars.

#2
Use two 3915 for VU and one 3914 for 'phase' & switch one channel.
Two LED-bars.

#3
Use two 3915 for VU, one 3914 for 'phase' & use three LED-bars.
Saves a switch, requires 10 more panel holes. Wouldn't look bad: two bars and a moving dot for the phase correlation.

I'll be considering which enclosure... if 3 * 10 LEDs would fit it'll be #3. Otherwise I don't know yet. I'll keep posted.


BTW, all this is meant to make a small portable dual mic-pre (using the THAT1510 from Wayne & perhaps some stuff from Kev's dual battery powered mic-pre thrown in) with metering and a phase correlation meter - preferrably running on two 9V batteries (& maybe some AC-supply alternative).

Bye,

Peter

 

[clintrubber]

[quote author="Crusty2"]Peter,

I posted this a while back; maybe the most simple way to switch lin/log:

http://groupdiy.twin-x.com/albums/userpics/10071/TS24_Meter_Dis.pdf

Switching R3 in/out for log/lin. Would this work for you?

Paul[/quote]

Hey Paul, thanks ! Should look in more detail, but if this works indeed as it seems then it'd be the most easy way indeed :thumb:
(in combination with switching the input of one LED-bar to the 'phase info' from a EXOR, see the recent phase meter thread).

Does this work ? Signal is injected on the top of the reference-divider and then we get another law ? As said, I need to look in more detail.
So i.s.o. the 3915 they use the lin-type (3914) and make it into sort of a 'log'-thing by that R3 & R2. Hadn't seen this trick !


EDIT: I looked for the thread in which you originally posted this TS-24-circuit, answers it already ( http://www.groupdiy.com/index.php?topic=9985&start=0 ). Nice, this will be the lowest part count solution.


Bye,

Peter

 

[clintrubber]

[quote author="mediatechnology"]Peter - Is this preamp going to be used with a stereo pair with a common gain control?[/quote]

Thanks Wayne for taking this a few steps further again. :thumb:

No, I want to have separate gain-controls for each mic-pre, but yes, one LED-bar might still be fine enough since with the solo-switch you mentioned it'd still be easy to check which channel is which.

I'm thinking of a compact box with two channels and since I want to put a MS-matrix in it as well, the gain-controls are better kept separate - since most likely the required gains can be quite different in that mode.
Hey, after having this laid out in my mind I realized I was kind of mimicing that Apogee-box Apogee-box but then with an added 'phase correlation meter'.
That two-dot on one bar sounds that cool to me that it'd be a pity not to use it.

So what could we have then (two bars, two modes):

metering-mode #1
Two LED-bars displaying L & R, nothing fancy.

metering-mode #2
Two LED-bars, one is displaying L & R in dual dot-mode (as described by Wayne above) and the other the relative phase info.

Combining this would be most logically be placed as two // bars, but for phase the suggested 'upside T' is of course nicer.

I'll keep thinking & start drawing, let's see what brings for decisions along the way. I'll also have a look at the 2252 vs lin/log-TS24 approach. I had forgotten about the need for a signal-rectifier - using the 2252 will then keep the DIP-count about equal.

Thanks :guinness:

Peter

 

[clintrubber]

[quote author="mediatechnology"]One more thing occured to me about the "dual dot" mode. When the levels are equal, the dot will be brighter because the LED will be driven with twice the current. Sort of a free "bright-up" mode.[/quote]
Ah, even better ! :thumb:
That LED-bar is getting sophisticated !

You might want to consider having a large dynamic range level display mode available to check room rumble...

Yes, the usual compromise between resolution where the signal lives mostly and seeing all what's happening. :?

I thought I had still some NE604 around, but no. Nevertheless an interesting alternative:
Audio decibel level detector with meter driver (see fig. 1 on page 2)

OK, need to check where to get '2252.
Oh oops, '3915 is about EU1.5. '3914 EU3.5 and needs a 2252 :wink: No problem.

Bye,

Peter

 

[henk]

I still do have somewhere a design with 48 leds from the old paper called Elrad , it also is put in there audio reprint called Remix and it is done with just LM324 and is very good , max 0.2dB of.(studio peakmeter)

 

[Svart]

lm339 for comparator also works well..