LA-4 Opto DIY

These babies go for up to $95 a pair. How insane is that??

Here's how to get them at 8$ a pair:

2 standard red LEDs + 2 NSL-6910

Here's my prototype made with cardboard and duct tape:



My initial test was to replace the NSL32 signal-opto in my LA-4 clone with this one and leave the NSL32 meter-opto in. That means there's two different LEDs at work and I'm sure that has at least some influence on how it reacts. But for now I just wanted to see if it would work at all.

In that crude test this opto sounds almost just like the VTL5C4 I've tested. Now, that's the opto JBL suggest as a replacement for the old ones so I'm probably on the right track here.

It's late here now so I won't be testing anymore tonight, but tomorrow I'll make another and test the circuit with two of these.

So far the best sounding opto in my tests, and the one closest to the sound in the original blackface, has been the NSL32. Let's see if that's still the case tomorrow.

In fact I'll make optos with all four 6910's that I received. That'll maybe give me a hint of an idea if there's a bit of difference in character, but to be honest I doubt it... We'll see.

Hey good luck with this,I bin doing similar tests lately

Is NSL 6910 same as NSL 5910 ??

You prob know you can get Vactec VTL thingies with one LED and two cell in package, I think theres a NSL 32 daul verion too...

In my latest experiments I for my home made LA2A. I found the NSL 5910 with EL panel to have more natural smooth sound, the NSL 32 & all VTLs I tried all sounded too crushed I lost the highs.....But thats in a LA2A.

I even sawed some vactecs in half to hear if sounded better with EL panel ....they did..
I tried different color LEDs & Panels, At one stage I had about 8 different cells in a box with EL Panel & Neon bulb light sources & rotary swithes to select the combinations of light & cell.

Look forward to reading your findings..................Gary O.

Well, I just looked at specs and made a logical choice. Red LED I was sure of. From pitures it wasnt hard to tell it's the same type of cell as in LA-2. So which exact cell? Well, I simply went with the one reacting to red light. When designing a low(er) budget piece of equipment like the LA-4 of course you want to use parts acting in predictable ways.

5910 reacts to the 550 nm wavelength which is yellow-ish green.

6910 reacts to the 660 nm wavelength which is a deep red.

So no, they're not the same.

Success!! ;D

The DIY opto sounds just great and reacts well to adjustment of the unit.

I changed the design just a little bit to make it easier to make.

I replaced the 5mm LED with a 10mm one. Still just standard specs.

I put one in the signal path and just a naked LED in where the other opto would sit. That way I can see what it does.

As a matter of fact, when diy'ing an LA-4 you could just leave out a meter and bring that second LED out on to the front panel. Like with the What comp. Save the money for the meter. The LED tells you all you really need to know.

Fitting for LA-4 into a 1U case isn't that far fetched suddenly!!

Anyways, I tested it and it sounds just as good as the NSL32 which sounds damn close to the original blackface file I have. Even though it's the silverface schematic we designed the comp from. The NSL32 and the DIY opto both sound better and more true to the blackface than the silverface file I have. However, having dug into this I'm more and more convinced that it's primarily a simple matter of adjustment. I can't say for sure until I have the chance to get my hands on an original silverface, but I'm quite sure I could adjust a silverface into sounding like the blackface. I think they simply changed the adjustment routine in an attempt to make it sound cleaner.

So, if you got one or more original LA-4 that needs new optos, just take the old one(s) apart and replace what needs to be replaced with a standard red LED and/or a NSL6910 LDR from Silonex.

I've attached more photos for you:

I thought about finding good plastic tubes and do it all fancy like...but I just ended up cutting a piece out of those small candle some-kinda-metal cups (don't know the english word for "fyrfadslys" ) just to get a rigid body for it. Wrapped it all in black insulation tape, including the bottom of the LED of course and it works just fine.

The DIY opto sounds just great and reacts well to adjustment of the unit.

Click to expand...

Congratulations  ;D I see that waveforms are looking same but can you post sound samples? It will be great to hear the result 

8:1, threshold in appr. in the middle:

Original blackface LA-4.

DIY LA-4 with Silonex NSL32 optocouplers og Cinemag output transformer.

DIY LA-4 with DIY opto og Cinemag output transformer.


20:1, threshold all the way down:

Original blackface LA-4.

DIY LA-4 with Silonex NSL32 optocouplers og Cinemag output transformer.

DIY LA-4 with DIY opto og Cinemag output transformer.

Pretty similar... but DIY sounds bit more noisy and there is some more distortion compared to blackface LA-4. . Did you compared opto with original in the Blackface LA4?

I haven't got an original. I only have the test files made for me by user Craptical.

That also means that they're done on two different setups alltogether. Thus he could be feeding more into the original than I am. That could be one explanation for the different noise level.

BUT apparently you're igoring the hum on the original which is absent on the diy version. So I assume you're focusing on hiss. Well, maybe the hiss is a bit more present because of the better opamp in the diy version. (Better bandwidth...)

The amount of distortion is as far as I can tell simply a matter of adjusting both the Ratio and Threshold trimmers. The Threshold trimmer in particular. Maybe I'm not precise enough in my adjustment. My reason for suspecting it's just a matter of adjustment, besides the fact that it seems to be the case when I fiddle around with it, is that we've made this after the silverface's schematic, but it still sounds almost identical to the blackface (to me at least).

You can compare with the silverface test files here:

Original silverface LA-4. (8:1)

Original silverface LA-4. (20:1)

Note the slow attack and much less distortion especially in the 20:1 file. I can easily adjust the diy version to have the that character as well. The "best" transient handling is largely a matter of setting the Threshold trimmer just right but both trimmers affects the character greatly.

I came up with a much easier and better-looking way to do it a few years ago:

Get your NSL 6910 and go to your local hardware store, and get a plastic "pipe-cap". Here in the USA, there are lots of semi-elastic plastic pipe end caps, which slip over the end of a pipe, to cover the ends and seal them from liquid ingress, etc.

All you do is get one for ½" diameter pipe, which should be about 1" long. -Then you simply 'poke' two holes in the end and place the diode inside it, with the legs poking through to the outside world. -If you like, you can put a small 'blob' of plastic-bond two-part epoxy in the cap tip beforehand, to seal the holes up, and to firm-up the diode mounting.

The finished result looks almost exactly like the real thing (just with a rounded end) and is COMPLETELY air, gas and water-tight, if you use the epoxy. -It slips firmly on and off the NSL if you ever have to replace it.

Easy-peasy. -I probably have some somewhere that I can take a picture of, but they look factory-made.

A note: The Silver-face LA4's have painfully slow release times, to my taste. -Certainly, the EARLIER black-face LA-4's were MUCH more usable. -With a blackface, you could compress a 120BPM drum kit overheads, and it would 'release' quite a lot between beats, with a good 'depth' of GR modulation, if you onow what I mean. -The later black-face and ALL of the silver face versions released VERY slowly, to the point where the needle would only modulate by about a dB or so, at 5dB metered GR, compared to around 3dB of modulation with the same early black-face version.

A local studio here in the Orlando area has six LA-4's: a pair of early black face units and TWO pairs of silver face. -you could set them to compress the same signal, set them to both compress the EXACT same amount with a constant tone, then replace the tone with fairly heavily modulated program material, and all o fa sidden the silver face units would all stay MUCH more deeply compressed.

Now don't get me wrong, that makes the silver face more useful for higher-ratio AGC tasks, and no mistake, but for 'poppin' Bass tracks, or drum overheads for tempos above a slow jass 'crawl', the silver face units were just to molasses-slow to ever be chosen.

That's all down to the optos though, so if you can find a fast and a slow LDR, then drive BOTH LEDs with the same signal, and switch between the two sets of optos... you'd have a very flexible unit.

But I'm fairly sure that the 6910s are the slower LDRs... You'd have to tell me if I'm wide of the mark.

Keith

Very interessting indeed!! I thought all blackface units were the same in terms of character and more in line with most engineer's preferences.

So do you recognize this blackface as a later model?

I think I can adjust it to a behaviour closer to what you describe there. Would you give it a listen and tell me if you start recognizing the character?

Thanks for the tip regarding the pipe ends. I'll check it out.

BUT apparently you're igoring the hum on the original which is absent on the diy version. So I assume you're focusing on hiss. Well, maybe the hiss is a bit more present because of the better opamp in the diy version. (Better bandwidth...)

Click to expand...

Sorry, I forgot to say "congratulation" first and then to add some criticism. So, give me a chance to correct myself 
I was adding my comment in the recording session pause so I was a bit superficial . Yes, Original has some hum and that is really ugly thing. Also, I don't have anything against some saturation, at least check my signature  Since test is done in two locations and different setups (that's a news for me ) I can say that you really did a job ! Congratulations 

I didn't take offense. Sorry if I seemed jumpy!

And thanks for your comment! It's much appreciated and I hope you read me like I want to feed you more details about the process so you can judge the result on a better foundation. That was my intention at least. ;D

Hey SSLtech,

A handful of test files for you, if you're up for some listening? ;D

What I've done here is to first run a file through with no compression so you have some sort of reference when judging the compression happening.
So here's your 120 BPM beat:

Raw beat

Then I've compressed it 8:1 to the point where my console LED meters said the bassdrum was levelled.
And here's the result with the trimmers set fully clockwise and counter clockwise:
CW or CCW doesn't mean much to you of course but you know what it means to me... )

8:1 - Clockwise
8:1 - Counter Clockwise

Same thing but 20:1 and the threshold control pot all the way down:

20:1 squashed - Clockwise
20:1 squashed - Counter Clockwise

Do you hear what I mean? The Clockwise setting on the trimmer is rendering it pretty useless for this particular signal.

I'm comparing it to the silverface test file I posted earlier and even though it's bass and not drums, it's the same tendency I hear.

With the trimmer CCW we're close to the setting I made when I wanted to match the blackface bass file which was 8:30-ish.

The level of distortion depends greatly on the Ratio trimmer setting but in this test I just left it in the same position.

Can't play those files... -Are they regular 16-bit WAV files?

Keith

No, 24 bit. I'll upload 16 bit versions in a bit.....no pun intended...

Can't play those files... -Are they regular 16-bit WAV files?

Click to expand...

I don't have a problem with files. Working nice here  ;D

16 bit versions here! I've included just part of the descriptions for reference. See the full comments in the previous post.

Raw beat

Then I've compressed it 8:1 to the point where my console LED meters said the bassdrum was levelled.
And here's the result with the trimmers set fully clockwise and counter clockwise:

8:1 - Clockwise
8:1 - Counter Clockwise

Same thing but 20:1 and the threshold control pot all the way down:

20:1 squashed - Clockwise
20:1 squashed - Counter Clockwise

Hmm... this is getting more and more interesting! Maybe time for some Silverface modding. (And I´ll look into the hum issue, haven´t noticed any when recording. Maybe caused by looping in/out of the coverter?).

/Dave

Could be, but it's also more apparent on files like this because we're actually compressing rather hard. I've started using the comp even though it's not completely finished. Still open box sitting on top of a keyboard with the pots hanging out! ;D And when I use it on material recorded well I have no noise issues at all. The file you ran through had a lot of noise on it already so test with something else first that you know is as it should be.

I would REALLY love to hear a report if you start fiddling with the trimmers in the silverface. Everybody at the danish groupdiy site just love your blackface so I certainly understand if you wanna see if you can match the silverface closer to it.

@SSLtech
I've uploaded 16 bit versions for you above. Would be great with your take on it.