Some stupid LA2A questions

My opto limiter project is doing fine - I'm on holidays for a week, so I can spend a lot of time soldering.

Electrolumiscence photo coupler finished.

CV path was planned to be like LA3A with step up transformer, but in the end im using a high supply voltage and a high voltage MOSFET (IRF820) instead of a tube in a LA2A-like circuit.

Audio amplifier is also finished. I had planned to copy the LA3A transistor amp; but now I bought some tubes and built a copy of the LA2A circuit. I just tested it - works great.

(As I hadn't intended to use tubes, my transformers weren't chosen for heater voltage range. So used the 26V supply that was intended to feed the transistor amp for a series connection tube heater circuit instead. A funny circuit, with a LM317-based 340mA source, and two TL431-based shunt regulators for 12.6V (12BH7 series connected) and 6.3V (ECC83 parallel connected). Not exactly straight forward solution, but it works, and I don't have to buy new power transformers. (;->)


All right, and here come the stupid questions:


What are the data for the input and output transformers of the LA2A?
I think I've read about the output transformer having a 4:1 ratio.
10k : 600R - is that right?

And the input transformer? 1:2 ? or 1:2 ??

The potentiometers for gain and compression, 100k,
Are these linear or log / audio taper types ?

JH.

Stock input was a UTC A10.
This is a 600:60K , or 1:10.
Better resulys with todays high line signals might be had with a smaller ratio, but keep the impedance high tomatch up with the grid.

[quote author="cjenrick"]Stock input was a UTC A10.
This is a 600:60K , or 1:10.
Better resulys with todays high line signals might be had with a smaller ratio, but keep the impedance high tomatch up with the grid.[/quote]


1:10 - that's a lot!

And it also means that there is no way to keep that voltage gain and get rid of the low 600R input impedance ...
So what am I supposed to do in a modern / cheap environment that can't drive 600R easily? An active input stage before the LDR attenuator?

And at what voltage level is the LDR supposed to work? For comparison, the FET-controlled 1178 needs a pretty constant (and small) level across the FET (to avoid distortion), so there is an input attenuator potentiometer that sets the compressor threshold indirectly. On the LA2A, as on many modernd VCA-based compressors, the threshold is adjusted in the sidechain, which will result in widely varying voltage across the LDR or VCA respectively. This is possible because VCAs and LDRs are more linear that FETs. But even a LDR will get a little nonlinear at several volts, so I was wondering about the typical voltage range (when the compression sets in, of course).

Anyway: thanks for the transformer info!
Is the output transformer 4:1 right ?
And do you, or anybody else, know if it's lin or log pots ?

JH.

You can use a 1:1 bridging transformer for the input, 10K:10K.
This mod is recomended by Jensen.
There is plenty of gain in the line amp, at least the stock one, so losing the 1:10 step up won't matter much.

The Gain Reduction and Vol. contriol on the LA2 is 100k log.
I used Alps conductive plastic, 25 a pop!

> at what voltage level is the LDR supposed to work?

The audio path from grid to 600 ohm output has a gain of about 60. Nominal output would run just over 1 volt, transient peaks to 7 volts.

With the Gain knob set all the way up, the level at the LDR is about 0.016V nominal, 0.1V peak. LDR distortion is very low at these levels, in the 0.01% range.

With the audio-taper Gain knob set "halfway", levels are like 0.2V nominal 1V on peaks. LDR distortion is still low (it rises very slowly compared to most other gain cells) but maybe over 0.1% on peaks.

Used between normal-level gear, and using the original 1:10 transformer, the LDR does get a strong level. If you switch to lower ratio iron like 1:1 (10K:10K), you don't turn the Gain knob down so far and also the LDR does not get such strong signals. So the low-ratio input iron is lower LDR distortion. With 1:1 input transformer the LDR distortion will be insignificant.

Note that when the system goes into limiting, the level across the LDR does not rise (much) because it is reducing its resistance to keep level (sensed across it) constant. This is true of many FET limiters, but not at all true for the classic vari-Mu tube limiter.

LDR audio design is a compromise between distortion and noise (and GR depth). Practical LDR networks have to idle at fairly high impedance (about 27K in the LA-2A) so have high voltage noise. You could overcome the noise with high signal levels, but LDR distortion does become non-insignificant at several volts. You can reduce the required depth of maximum GR and work at lower impedances but deep GR is nice to have and also working with thin margin means sloppy control. You could try a lower-resistance LDR but that interacts with speed of response. You could try to find a large-area LDR (and light source) but they don't seem to come much larger than what the LA2 used. I have not really calculated it, but I think LDR techiques will give about 70-80dB signal-to-(noise or distortion), noise for low inputs and distortion at/over threshold.

When I design LDR limiters, I prefer to put the passive gain pot ahead of the LDR, so it has low audio voltage across it. But LDR distortion is so low that I don't obsess about it. In a PA crossover/limiter I let the LDRs get 2.2V RMS at limiting level and then trimmed that down to 1.3V-1.6V, the power amp sensitivities. It was just easier that way. While I try to avoid pushing that system to the limit (it is in a very genteel concert hall), it doesn't sound harsh when I do.

OK, I put togeather a few things on UTC transformers here:
http://www.vacuumbrain.com/
click on the obvious.
input was also HA-100X.
Output was usually an A-24


:guinness:
Sunglasses may be required! I will work on the dissections next week.
:guinness: :sam: :guinness: :grin:

You can also get there by clicking on the UTC logo at the transformer META thread:
http://www.groupdiy.com/index.php?topic=467&start=0

Wow, what a wealth of information! Thanks to both of you.

JH.

Hey guys,
Just out of curiosity, would't it be much cheaper, flexible and compact to use a current to current amplifier(impedance transformer) for interfacing with low impadance and voltage amplifier for the 10k:10k line interface?

thoughts?

Are we talking about cheap, flexible and compact, or are we talking about the LA2A?
:green:

:grin:

hey Dave,
Sounds like you could use a minimum wage student ghetto DIY philosophy. :green:

That's how it is for a broke ass recording artist wannabe, where the objective is to minimise the resources without jeoperdising the final output quality.
:shock: :wink:

[quote author="NewYorkDave"]Are we talking about cheap, flexible and compact, or are we talking about the LA2A?
:green:[/quote]

Well, I was talking of my little DIY project. (;->)

I'm building it step by step: first the core functions, and if I like the results, I'll "upgrade" with good transformers.

I was just wondering about the 600R input impedance (if I should choose a 1:10 input transformer), as my mixer and other gear doesn't really support this. I don't want to ruin the performance of an expensive transformer by driving it from an output with 1kOhm + 10uF impedance ...
(OTOH, when the transformer sees the right impedance on the secondary side, even _that_ might work without messing up the frequency response. (?) )

What is the general rule about using 600R input impedance devices in a "cheap" environment ?

JH.

[quote author="Learner"]:grin:

hey Dave,
Sounds like you could use a minimum wage student ghetto DIY philosophy. :green:
[/quote]

I was doing the minimum wage ghetto punkrock thing twenty years ago... and I STILL refuse to pay $25 for a pot. So don't preach to me :green:

I've made some pictures of my opto compressor project.
Schematics, too.

http://www.oldcrows.net/~jhaible/opto_compressor/jh_opto_comp.html

Don't laugh at the tube heater circuit - it was a cheap workaround to use the 26V supply that was intended for a transistor based output amplifier at first. So at least I didn't have to buy another power transformer with 6 or 12 Volts ...

JH.

Dang, thats good work there!
And a killer web site too!
I like the EL thing you did.
Would you mind if I stole those pics for my T4 page?
See ya,
cj :guinness: :guinness: :guinness: :guinness:

>Dang, thats good work there!
>And a killer web site too!
>I like the EL thing you did.

Thanks!


>Would you mind if I stole those pics for my T4 page?


No Problem. (;->)

JH.

I notice that you used the output stage from the LA2A pretty much verbatim. Just thought I'd point out that the drive capability of the White cathode follower can be improved markedly over the implementation used in the LA2A, if it's biased richer and the plate resistor value is optimized. See the schematic I posted in the "Tube Line Amplifier" thread for an example of a 12BH7 WCF that's beefy enough to drive 600 ohms directly.

Also see this article on optimizing the White cathode follower:
http://www.tubecad.com/october99/page4.html
Although the author reaches the conclusion that the plate resistor must be the reciprocal of the transconductance of the tubes for best results, I found that a somewhat higher value than that actually worked better. I used a decade resistor box and dialed in the value that gave the greatest undistorted swing.

"I notice that you used the output stage from the LA2A pretty much verbatim."

Yes. And I must admit that I haven't thought about it much, either. I didn't even know it's called a White cathode follower. I just remember I have once used a similar circuit with transistors, which worked nicely. I *was* wondering about the high plate resistor, but I filed it under "tube maths are more complex than transistor maths" and never gave it another thought.
But after your mail I did. (;->)


"Just thought I'd point out that the drive capability of the White cathode follower can be improved markedly over the implementation used in the LA2A, if it's biased richer and the plate resistor value is optimized. See the schematic I posted in the "Tube Line Amplifier" thread for an example of a 12BH7 WCF that's beefy enough to drive 600 ohms directly. "

Thanks for sharing this - this *will* be interesting for a headphone amp. Sooner or later I want to have a little tube-only amplifier for playing vinyl records over headphones. I'll keep your circuit in mind.


"Also see this article on optimizing the White cathode follower:
http://www.tubecad.com/october99/page4.html
Although the author reaches the conclusion that the plate resistor must be the reciprocal of the transconductance of the tubes for best results, I found that a somewhat higher value than that actually worked better. I used a decade resistor box and dialed in the value that gave the greatest undistorted swing."

This article shows how to optimise the circuit for high output swing at a given (low) load impedance. I haven't followed all th emathematical formulas in the article, but the autor clearly says that this does compromise the [small signal] output impedance.

So I figure the LA2A designer - certainly not stupid - might have had low output impedance as a design goal more than high output voltage swing?
This is just a guess; maybe low output impedance was important for good low frequency response in connection with the transformer?

The autor of the article goes into much detail to show the difference between output impedance, which is a small signal parameter, and the drive capability, or large signal response, of an amplifier. And it is no wonder that, as in any class A circuit, increasing the bias current will also increase maximum output current, and at a given load, the maximum output voltage swing. He also shows how in an ordinary cathode follower with active load (let's simplify it as a current sink), the upper tube *could* deliver a lot more than twice the bias current, but the lower can't sink more than the bias current, so the maximum output swing on a resistive load would be +/- that bias current * load impedance. (paraphrasing)

So far, so good.

But now I wonder where the benefit in large signal response of a White cathode follower would be over an "ordinary" cathode follower with active load. (The large signal benefit over a follower with passive cathode resistor is obvious, so I'll just aim for comparison between WCF and a cathode follower with the lower tube acting as a fixed current source.)

Increasing the bias current will increase the output swing in both cases (as it will increase the power consumption), so it makes sense to set that current just a little higher than what is required.
But - when I assume class A operation in both cases- where would be the benefit of the WCF in terms of large signal response? If both tubes can provide more current than the bias current now (and not just the upper one anymore), you can possibly get a higher output swing, but ...
... but this isn't Class A operation anymore, is it ??

And as long as you're in class A operation, both tubes are connected in parallel, so the [small signal] output impedance does determine the "force" to pull up and to pull down likewise, doesn't it ?
(Of course, if one tube doesn't conduct, things are entirely different, so you will get a vastly different positive and negative slew rate when driving a too high capacitive load.)

Does this make any sense? Maybe the whole WCF thing is aiming for minimum output impedance rather than maximum output level?

JH.

Oh and there is another thing I don't understand in this articel,
earlier on page 2:

http://www.tubecad.com/october99/page2.html

I meant that part "... high transconductance that is not matched to a high current draw [...] worsens the sound. I like to use as much current as there is transconductance ..."

Excuse my ignorance (I _am_ quite unexperienced when it comes to tubes), but what does this mean? It sounds like comparing mho's and amperes, though I'm sure this is not what the autor wants to say. I figure that he means "high gm devices like to be used at high current", but it sounds as if he tries to further quantify this somehow.

JH.

..found a rack case that might do well with the DIY LA2A:

http://www.mainlinegroup.co.uk/mless/main.php?act=Card&id_element=3632

19" 3units, front panel folds down, price £37, and they have 78 pcs in stock



Mainline surplus is a great company for those odd bits..

Jakob E.

good eyes Jakob :wink:

I ordered one! :thumb: