LA-3A Design Thread [Autotransformer with Pics]

hi:

Since building an LA-2A, which I'm absolutely in love with, I've been wanting to attempt a discrete opto... and the first thing that came to mind is the LA-3A.

Here's the schematic in reference so we're all on the same page:
http://www.waltzingbear.com/Schematics/Urei/LA_3A.htm

This morning I decided to start laying out the PSU just to get the juices flowing (all part numbers and values follow the schematic):


Now I don't fully understand how this circuit works, so perhaps I could start off with a few simple questions and maybe this thread could yield some great information in the long run. And in the end, we'd have another DIY project to add to the arsenal around here.

I do not have an original LA-3A, which would be nice for comparison, so if anyone does have one that would be great. Regarding the PSU, it seems there are two rails...

1. I'm guessing +24VDC at the emitter of the 2N3053 - Q12 (since there's a 30V cap right there).

2. Just after the bridge rectifier, another rail is being used in the sidechain. I don't know what power tranny was used, but do you think it's a 24V-0-24V, which would probably give us something in the mid 30VDCs at the anode of C10.

3. Anybody know what T4 is doing?

This should get things started. Please let me know if anyone else around here is interested in this project, or even if you think it's a waste of time.

Not an answer to your question but since this psu is small maybe you want to integrate it into the main pcb, just like Tommytones Forsell boards. I thought it was neat to have all the parts on one board. Also with 24V and 35V it doesn't seem to be a psu which could be easily used for other projects so it shouldn't do any harm if the psu is on the main board.

Oh and i definitely watch this, sounds like a cool project :thumb:

Flo

Yea definitely a good idea... it is very small and could easily be integrated. :thumb:

I'm working on the main signal path right now. Making good progress... but have to decide on some iron, though :roll: ...

> I'm guessing +24VDC at the emitter of the 2N3053 - Q12

What's that on its base? A 27V Zener?

> do you think it's a 24V-0-24V, which would probably give us something in the mid 30VDCs

Not using the center-tap at all. Ignore it. Bridge rectifier, so we get 1.4X the total AC volts. A 24-0-24, or 48V total, gives 70-some volts, which will burst C10.

Working backward from the apparently 27V Zener: we need lots more than 27V at C10 to ensure the Zener and Q12 are alive and regulating. 20% or 30% just to cover line variation. I pencilled 32VDC. Now what is the nearest non-custom transformer to give at least 32VDC? A 24VAC winding is super-common, gives 35VDC under load, 40-some VDC lightly loaded at high line. 35VDC would be reasonable to feed a 27V regulator. The other extreme is we do not Q11 Q9 to over-volt and die; IIRC a 3053 is rated 50V-60V, so 35V-40V is not a danger.

I'd say 24VAC.



One objection: it turns out those 1819 lamps are rated 28V, and they have two in series! That spells 56V, which is preposterous. It was normal to run meter lamps "cool", much lower than white-hot, so they lasted longer. 12V on a 28V lamp sure won't be bright, but it may be all the light they needed, and the bulbs will last decades.

It's my impression that the stumbling block to building these is that the side chain transformer, T4, can't be gotten. Maybe somebody can correct me on that.

Very nice PRR... I've used a 22VAC toroid before. I get them from Digikey, and I think they're $16 bucks. They usually run a volt or so higher than the spec (even when loaded), so I think this would be great. I could just put the secondaries is parallel. I might breadboard this guy just to make sure it behaves as expected.

Regarding the lamps, I wasn't even concerned with them at the moment, but I see what you're saying. I think I'll leave the lamp or LED powering up to the user.

[quote author="jrmintz"]It's my impression that the stumbling block to building these is that the side chain transformer, T4, can't be gotten. Maybe somebody can correct me on that.[/quote]
grrrrrrrrrrrrr

[quote author="Greg"][quote author="jrmintz"]It's my impression that the stumbling block to building these is that the side chain transformer, T4, can't be gotten. Maybe somebody can correct me on that.[/quote]
grrrrrrrrrrrrr[/quote]

..if you dont have to be a purist about it, you could probably replace the electroluminence panel in the T4B with two anti-parrallell LEDs ( for "full wave" ) and a resistor to limit the current ( LEDs dim with current, not the voltage ) and drive them directly from the G.R.-amp without the transformer..just a thought...

johan

Wasn't CJ building his on T4Bs? I know someone was for a while and selling them at a reasonable price. There is some NOS ones on Ebay....better grab them quick!

If this can be built for a reasonable price, I would love to give it a try!

Joe

no, he's talking about a transformer. there's an interstage transformer of sorts, the availability of which may be questionable these days

I have a pair of vintage units (serial nums 17x and 22x :grin: ). If you have any specific questions, I'd be happy to help, provided it doesn't require very much disassembly.

Yes, I meant the transformer, not the T4B. Sorry for the confusion. This is something SSLTech has spent some time looking into. Perhaps he'll fall by and elucidate.

i thin kev had an la3a page at g-diy,no?

I saw Kev's LA-4 inspired Opto, but not an LA-3A page. I've just about got the entire PCB layout completed, but if that one part can't be sourced, then I'll be bummed.

Some seemed to mention that it's some kind of centre tapped inductor. I read a thread from 2004 that went into more detail about T4. To the inductor/transformer gods, is there any information that ToroRojo could easily obtain that was help me in my quest for T4.

I've also got to decide on the I/O trannies, which will definitely change the sound, but I don't care. I'm leaning towards Lundahl, but will probably need some more info before I decided.

But first things first, is T4 stepping up the voltage to the T4B?

yep there is a couple of LA3 projects but what kept things back was an obvious choice for the interstage

there is/were
in fact beyong 10 ongoing projects of some classic gear due for posting at Group DIY
Project 1 - neve 3 stage mic pre and the short form 2 stage
Project 2 - various gain block mic-pres
there were another ... 5, I think ... mic-pres of different topologies and I'm sure you can pick a few of the likely candidates

then were to come the compressors
the T4B based units were obvious choices ... stating with the LA1

and other opto releated LED LDR combos
then the FET based units and the VCA styled ... VariMu but my knowledge was low ... PRR has provided an option there and then there is the Gates style ... see Rob F

It was the VCA Project 11 that was due to go on the weekend of the crash

but to answer the question about the LA3
yes there is a couple of PCB's about ... both clones and new layouts
I may revisit them but I need to find a sensible interstage solution
having one made was and still is in the back of my mind
BUT
don't let me stop you pushing ahead with the project.


CJ
let's get that No Frills LA2 up and the Dummies guide to the T4B and then look to an LA3

Here's something I found with the 'ole search function: good info !
LA-3A Sidechain Trafo Thread

To summarize (and someone please correct me if I'm wrong): it's an autotransformer meaning the primaries and secondaries are wired in series, with about an overall 1:25 ratio, and it needs to handle at least 100V.

Since this is Prodigy-Pro, does Sowter make something that would be suitable for the task at hand. Also, does Sowter make LA-3A input/output trannies? (I know they make some LA-2A and Neve and API type trannies).

My immediate plan is to continue with the grunt work and get the PCBs and parts list finished up while I try to find something for the autotransformer. Maybe I can go poking around in my LA-2A in a couple days.

> some kind of centre tapped inductor

Unlikely. While LA2 and LA3 don't use the same module, they both use similar photoluminescent panels. LA2 has a 280VDC amplifier with a 10K resistor. LA3 starts from a 28V amplifier, and the 80uFd cap hints at 100 ohm impedance. Be dumb, and just take the 280V/28V ratio: the coil may be a 1:10 step-up (not 1:2 center-tap winding). If so, the ~100 ohm low-side impedance kicks up to 100*10^2= ~10K impedance, similar to the LA2.

I'm not saying it IS 1:10 ratio, 100:10K impedance; but in that ballpark. 1:25 seems high, but not impossible.

It is also 8 or 10 volts across the 100 ohms; almost 100 volts at the secondary. This spells about 1 Watt. Not a mike transformer.

What might work fine is a 12V:120V or 24V:240V 10VA power transformer. Clearly distortion is no big deal: the sidechain amp isn't even biased out of crossover distortion. (I'd be real inclined to replace that stage with a little speaker-amp chip, just for simplicity.) (But then, I'd replace the audio stage with a 5532 for the same reason.)

While I'm looking at the LA2, clearly the LA3 does not work the same. The LA2's sidechain output impedance IS 10K; the LA3 has feedback around the EL driver and gives the EL low-Z drive.

The LA3's output transformer T2 is wired wacky. The yel-org winding gets fed only DC! Why? To negate the DC flowing out of R16 driver stage. And why is that flowing through the output iron? To save another hundred-uFd cap. When was this drawn.... ah, 1970. We did chit like that then. It was pretty cheezy.

i could be wrong, but we have a few la3as open in the shop right now. the EL driver transformer looks like a small speaker distribution transformer. possibly with the signal going into a 25V tap on the primary, and coming out a 100V tap, with the other end grounded. [edit] or maybe as suggested in the thread linked above, going into the speaker output winding, wired in series with the 100V winding as an autotransformer.

it is definitely an inexpensive, relatively small iron core transformer; about one cubic inch. i'll have another look soon and report what i find.

ed

I think that PRR is right about the 1:10 ratio step-up as 1 volt in = 10V out measure with a scope in circuit.

[quote author="PRR"]Unlikely. While LA2 and LA3 don't use the same module, they both use similar photoluminescent panels. [/quote]
yep
this was something that we never did fully explain

an indicator was that a T4B from either an LA2 or LA3 didn't work in the other unit
also one of the FAKE LED/Vactrol T4B's that worked fine in one unit also didn't work in the other

we analysed the schematic we had and found there was somesort of issue with ground and one of the LED/Luminescent panels.

the octal sockets of the LA2 and LA3 did seem to be wired differently
sorry I don't have details at hand and it is something I have been meaning to gather re-investigate and re-post somewhere

oh !
:shock:
and while we are on the LA3
this may be of interest
can't remember where I found it
http://www.diyfactory.com/data/LA3A_mods.pdf