LA-3A Design Thread [Autotransformer with Pics]

[mikep]

I don't know, this might be useful to someone. I once designed a GR metering circuit to work with an opamp labs model 37 (IIRC) compressor module. this thing is an octal can with only input, output and power pins. there is 30 dB of built-in gain and a fixed threshold. the comp is optical and actually sounds pretty good but it is quite hard to use gently without any metering. the circuit could be modified to work with any compressor as it just measures the difference between input and output level. it worked surprisingly well.

one small drawback is if the level quickly falls to -40 or so the meter will instantly go to zero regardless if the comp is still slowly recovering. looks pretty bad on a LED bargraph but is not even really noticable with a mechanical VU. I had to fudge a bit to get enough range at the expense of accuracy. for very small amounts of GR it is slightly exaggerated (actually a beneift) and it maxes out at about 20 dB, which is about where that opto stops compressing and a VU doesn't have much resolution down there anyway. I can share the schematic, but not right now because I am away from home till next week. But I probably gave enough info to recreate the circuit.

mike

 

[craigmorris74]

Is the model 37 similar to the 325L limiter? I would be interested if it wasn't anyway. I like messing with these modules.

Craig

 

[enthalpystudios]

[quote author="Kit"]
I didnt know that there was a step up input trafo.

Interesting.[/quote]

Yeah, I spent a good deal of time reading the schematics, as well as my beginning mind can comprehend, which was mostly, hmm, this is like the 1109 schematic here, and oh here this like the 1176 rev g schematic at this part, and etc etc.

same input transformer as the 1109, which a buddy of mine has so I checked it out for a bit. plus i did a quick excel, which i lost in a hdd mistake (broke the wrong raid array), where i made a parts list for everything, and then checked it all over for availability, and found that all of the transistors and etc etc are still available in the same or similar packages, without having to use any substitutions.

Only odd parts are the t4b and the transformers, and if you don't have a problem winding the output transformer, its no big deal. just the interstage. also, most any 1:4 or lower ratio output transformer should work fine. so it could be hand wind, edcor, cinemag, lundahl, most anything really. only issue i'd see with the lundahl or similar is finding a place to pc mount it.

a neve out type transformer could also be cool, but is a larger core and such, and i dont have the knowledge to determine if the circuit could drive the big 75ei gapped stack. i bet it could though....

anyway, idea is that there can be plenty of build options, and with a decently similar pcb, it could remain the same size and be put into sweet little 1/2ru boxes.

anyway... i stare at those urei schematics all the time... really cool that they are up for free at jbl. i'm still working on figuring out if there are ancillary components i would have to change on the g1176 board to use the original transistors throughout. thats taking me a while though.

 

[Kit]

Ok, been fooling around with the sidechains a bit. The first one.

Q1, the voltage gain stage is not needed at all. Especially not if were stepping up a line level.

Screw it.

Just split the supply so youve got 10V at Q2´s base. (Remember to take base current into the equation if we want to keep the high input Z.)

Cap couple from compression/threshold pot.

Also, for maximum sidechain gain, one could hang the 100R Q3 emitter resistor on the collector instead, in series with the vactrol LED.

I havent got any EL-panel optocomp to compare with, but i think this arrangement sounds pretty good.

I have been using the VTL5C3 with good result here on my test rig.

Nice timeconstants.

 

[mkoz]

Sorry, caught this thread a bit late. Working on our album for sometime so I didn't have a chance to check in here for sometime lately.

[quote author="Greg"]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...
[/quote]

I have a pair of UREI LA-3As (not reissues). I'll get them out of the rack and photograph them in detail for you, and can take any measurement you may need as reference.

I also have 5 of UREI LA-4s that I can photograph if you are interested.

Mach

 

[SSLtech]

We've got plenty of LA-4's here. There are more revisions of that than there are of the 1176! We've got solver faces and black faces and NO two aerthe same internally...

Keith

 

[enthalpystudios]

i still think that the original layout would be totally fine for the la-3a.... the output transformer is offboard, and easy to wind as well.. or approximate with a lot of different ei625 square stack tx's. input is basically a 1:5+5 mic input transformer, a cinemag could do great. and then the t4b saga of course, but aside from that all the other parts are available, thats why i've always wondered about just scanning or photoing a board and using it as a template to make a clean/to-scale layout. they already did it once, eh

 

[CJ]

i just remodeled my entire apt and found la3 xfmr docs

 

[enthalpystudios]

is it as i thought?

it was b11178, 1:3.5 on a 625EI square stack for the output (ed anderson and you helped me on that), and i was also pretty sure of the input, as i think its the same model number as the 1109 i/p

inquiring minds want to know ;]

and want to make repro pcbs ;]

but all new super front panels, with compress/limit up front eh

 

[CJ]

yes on Ed.

Did I post that somewhere?

 

[enthalpystudios]

wire gauges and turns count for a single pri and single sec are somewhere on this forum... not sure where, but i also may just have it written down somewhere.

was the input mounted on or off board in the original though?

 

[CJ]

darn Google broadcast has 8 gatways, i keep getting IP Merry go Round,
:twisted:

I have input and output, but no innerstage data.

 

[barclaycon]

The LA3A is a nice unit - much loved by, amongst others, Bob Clearmountain. I got my pair about 20 years ago and decided to completely rebuild them. I replaced the input and output transformers with Jensens and replaced the line amp with a 990.
I don't know how far you are going to go with your project, but if you are intending to use the standard Urei circuit then I would do the 'mod' which is to reduce the gain of the output amp. by 12dB. This reduces the output noise and means the Gain pot is more or less at centre for normal operation.
I fitted IEC mains sockets instead of flying leads and also XLR sockets instead of the tag strip. Some people also move the Compress/Limit switch to the front panel.
If you are building a clone, what are you going to do about the optical cell? I would imagine that apart from the T4 inductor, this would be the most difficult part to source.

 

[Kev]

[quote author="barclaycon"] I would imagine that apart from the T4 inductor, this would be the most difficult part to source.[/quote]
T4 inductor ?
:? ... interstage transformer
mmm ... but
I think cj may already know a little about the T4B .. opto cell and dual LDR assembly.

perhaps someone will allow cj to play with their LA3 interstage transformer and an hacksaw
:green:

 

[barclaycon]

If you look on the circuit diagram, do you see a secondary winding for T4 ?
T4 is there for isolation, but strictly speaking I think it's an inductor rather than a transformer.

 

[edanderson]

T4 is a transformer connected as an autoformer. something like 1:10.5 ratio. someone did some pretty detailed measurements on it already and posted the info here.

i did post some simplified winding info for a home wind of the b11148 output. if this project actually gets off the ground, i can supply some repro b11148s. it isn't the "greatest" transformer, but if you want to duplicate the circuit exactly, you need this specific transformer.

for the input, a cinemag 1:8 would be a good substitution.

in addition to the copious knowledge that cj has compiled about the t4b opto gain reduction cell, i think dripelectronics is already making some version of them to go along with his la2a pcb.

note that the output mod changes the feedback in the output amp, which in addition to lowering the gain also changes linearity and overload character of the amp.

the output circuit is basically an opamp; if you wanted to simplify, the whole output section could be replaced with a socketed discrete opamp and a more conventional output transformer.

ed

 

[gevermil]

Anyone make a move on this kit Idea ?
Ive got everything except the t4 :wink:

 

[Kev]

find cj's page on diy'ing the T4B
http://www.vacuumbrain.com/docs/t5.html
but you need to ask cj where all the pictures went
he may have a new page

there is also a LA3 mods PDF at the DIY Factory
... data section
Ed Evans Circa 1991, update 2003
http://www.diyfactory.com/data/LA3A_mods.pdf

and from the dim and distant past
http://recording.org/users/kev/T4B.htm
no idea how much of the page will still work and link correctly

 

[CJ]

good god, when was your last post?
good to sse you didn't get et by a GW!
:thumb:
working on the pics(see Brewery to help!))

 

[Kev]

yeah it's been a while
but I do a quick scan of the forum as often as possible when my net access allows for it

currently visiting my parents in Queensland and there is no surf ... dead flat
so I'm surfing the net instead ... with their much quicker access

it's all good