Some basic opto comp questions - LA-2a schematic

[kiira]

I decided I'm doing the transformer. I've been studying this for hours and that seems like the best solution to me too. The only thing is I'd kinda like to try the two LEDs as well which Wayne has been talking about - but I'm not sure yet if it makes sense to use two LEDs when stepping down and rectifying from the sidechain amp to the LED. At first it seems like doing one eliminates the need for the other but I'm not 100% clear yet.

The thing about not using PRRs JFET solution is not that I insist on Tube Purity ha ha.... it's that I don't know anything about JFETs and it would be a black box solution handed to me by PRR. If I were doing that anyway I'd probably just go and make a whole SS, mayybe discrete side chain with attack, release, etc. I want to be able to understand everything I'm able about this as I go along building it so I better stick to tubes and transformers, something I at least have a chance at. :roll:

So, a transformer and rectifiers seems like the best idea. But I'm trying to also trying to figure out the two LEDs in anti-parallel. Maybe I could use a stepdown with a split output. Or is it one CT transformer to a full wave bridge off each leg for each LED inside the rectifier loop. Hokay it took four pages to throw out the "simple little opto comp" idea lol.

I will obsess some more tommorrow. :razz:

thanks everyone,

Kiira

 

[kiira]

Hi Wayne,

[quote author="mediatechnology"]kiira and Dave;

Your diagram Dave looks like what I've been thinking about as well with the addition of a current limiting resistor for the LED.

Once you get that going kiira you could experiment with back-to-back LEDs by eliminating the rectifier diodes and driving the LEDs from the transformer secondary without the center tap and the cold end grounded. Don't know what we can gain from that but we could call it the "Manley" experiment.
[/quote]

I see. I better draw this out so I'll be able to visualize it better.

But I'll start with Dave's circuit because I understand what is happening there and why fairly well. Thanks very much Dave. I should just put that in my signature line lol.

That 6DJ8 in my schem was at the max for voltage and very high plate current and I figured as PRR commented I'd better use a stronger tube if I was going to build without the xformer, like perhaps a 5687 or something.. lots of tubes to think about trying.

Or, as another experiment, use the LEDs as rectifiers and call it the "Wayne" experiment. I'll sketch out something in a bit.

Yes that would be good Wayne because I don't quite yet understand the advantage or exactly how that would work.

Oh wait, I'm having a vision. It's a product, the Kiiratronics Konvert-a-komp (Pro? 9000?). It will have replaceable "personality sidechains... the "Dave", the "Wayne", the er um gulp .... "Eveanna" ??? :shock:

But I started drawing out the turret board and I decided I will build the makeup gain part on the primary board and the sidechain/amp on another which I'll set up to be easily removeable so I can take it out and work on it easier... or even just replace side chains as needed to try out different ideas. The tube and xformer would be offboard of course. I will try and make the vactrol socketed too. OMG if I wanted to be way zoot I could build it on an edge connect card like a BA283. :razz: Just make the traces short to connect to a row of turrets. woo.

Thanks for the drawing Dave so that we are "reading from the same page" literally and figuratively.

Your comment kiira about obsessing over it this weekend illustrates your drive and reminds me of myself sometimes. When I'm working on the bench on a design the focus is pretty intense. Don't get to do that often enough.

It's my MO for doing most things especially learning new stuff. I've always been that way. "If something is worth doing then it is worth obssessing over."

thanks again everyone this is just amazingly cool, the way this forum works. Not surprising considering the amazing people here.

Kiira

 

[SSLtech]

...The "Keef":

:wink:

Keef

 

[Gus]

One thing I see is two diode drops in most of the fragments, manybe less compress and more hard limit?

I kind of like the idea of the third one but with a balance pot.

From effect building I noted the different diodes sometimes used for clipping did something some might not have noted. How some start to conduct before the diode drop voltage and then there is....

The deadband area the signal is unclippped and before it hits the doide voltage limit. Starving the current or giving as much current as the circuit allows will sometimes change the sound. Don't want to give it all away.

Now back to the optocompressor fragments this being in the side change how much does this matter? needs testing.
do we want to preshape?

 

[SSLtech]

[quote author="Gus"]One thing I see is two diode drops in most of the fragments[/quote]
Don't overlook the fact that this is a DESIRABLE characteristic. -You actually WANT a region low down with no effect.. otherwise you have an infinitely-low threshold.

If you try and design tnis away, you risk doing what the Germans did when they purified and refined germanium, noting initially as they did that better-refined germanium made a better semiconductor. -Eventually they got it so pure that the semicinductors stopped working. -Took 'em a while to figure out why.

Keith

 

[CJ]

I do not know if this will work.
I hope not.
I just replaced a xfmr withe an opamp, punishable by no less than 20 years of hard labor at Camp Sowter.
Maybe turn the diode into a bridge.
Or build a low drop opamp bridge in front of the dc amp.

 

[Kit]

CJ, which schematic software are you using?

I like it.

Very avant garde.

It will work. Infact thats how i would do it. AC in to the opamp. Opamp feeds a fullwave bridge with the led in the middle.

The problem is the supply for the opamp.

Hmmm.......feed the opamp from the heater supply.
+-3,15V centertapped AC..........rectified and smothed you will end up at about +-4,41V DC.

Thats low..........voltage doubler?

 

[kiira]

I laid out the turret board today for the line amp (I used Dave's schematic as is) and the xformer to rectifier style sidechain and amp. There is so much room... I could easily fit another line 2 bottle stage in the left over room on the turret board. I am going to leave extra room and turrets in the sidechain section. I'll just keep the extra space for modifications and whatever. But anway there is room for the transformer and even another tube in the sidechain.

Kiira

 

[kiira]

Should I start this again in the lab since I am building it now?

I was looking at two websites, vactrol manufacturers, these pages have a lot of good information.

http://optoelectronics.perkinelmer.com/content/RelatedLinks/photocellanalogoptoisolatordatabook.pdf

http://www1.silonex.com/audiohm/index.html

I'm trying to decide which device to get and the choices seem mostly to be decided by turn on time and what they call light history. I know that the V4B has a gentle turn off so I decided that a slow light history would be a good start. That means Vactec VTL5C1, 5C4, 5C9 and 5C10 and silonex NSL-39V62, maybe others from silonex. That one is a dual and I think a dual vactrol is a good idea because that way I can try the two LED experiment and also use a second LED to drive a meter as in the LA2. Vactec makes a dual which has a large light effect too the VTL5C4/2.

Will a 15K:600 work ok on the sidechain or should I try and build the amp to be able to drive a higher ratio? I'm not sure at all how to figure out what step down ratio I need to drive the LED properly.

Are there any references around which explain how to design a tube stage/s to interface with specific outputs? What are the desireable characteristics for a tube to be able to drive low loads? or which tube characteristics make a good cathode follower - all I know is high transconductance. I will read more of the M. Jones book and take more notes. The darn thing is falling apart. $40. How annoying.

thanks!

Kiira

 

[Boswell]

[quote author="kiira"]That means Vactec VTL5C1, 5C4, 5C9 and 5C10 and silonex NSL-39V62, maybe others from silonex. That one is a dual and I think a dual vactrol is a good idea because that way I can try the two LED experiment and also use a second LED to drive a meter as in the LA2. Vactec makes a dual which has a large light effect too the VTL5C4/2.[/quote]
The duals have a single LED and a split (centre-tapped) resistor.

[quote author="kiira"]I'm not sure at all how to figure out what step down ratio I need to drive the LED properly.[/quote]
You need to determine it from the current. To give you maximum flexibility in Vactrol choice, allow for up to 20mA through the LED. If the maximum current change through the valve in your design is 2mA, then a 10:1 transformer is all you need. You don't need huge power. 20mA at 1.2V is only 24mW.

 

[kiira]

[quote author="mediatechnology"]Just to clarify for some of our readers:

What boswell is saying (correct me if I'm wrong boswell) is that you need 10:1 current, voltage or turns ratio. Not imedance ratio.

So, in the 15K:600 ohm example you cite kiira, the impedance ratio is 25. The turns, or current ratio, is five. Might be enough.

N1/N2 = sqrt(Z1/Z2)

The 15K:600 is worth trying. A 60K:600 would be 10:1 current ratio.[/quote]

Hi Wayne,

Wouldn't it be easier just to use a 10:1 then? That's a reversed input xformer and is easy to find... I have several. Or would the stronger output from a 5:1 current ratio 15k:600 be better? Uh oh do I have that backwards? I better try and figure out the current through the tube first I guess. I have some really nice 2:1 Jensens outputs but an Edcor is like $12.

I screwed up the layout. I drew the tube sockets reversed, like I was looking at them from the bottom... but the layout was from the top. It was beautiful too, just wrong! Typical. I better not make a "Kiira" personality module. It would be totally non-linear and not know which direction to send the electrons. :razz:

Thanks for the help Boswell.. that dual Vactrol would only be useful for the metering I see.

Wouldn't a plate follower be a good circuit to use on this?

Kiira

 

[NewYorkDave]

Why not use a 4:1 or 5:1 output transformer and bias the output stage for 5mA (or so) of plate current? A mic input transformer is no-go, it couldn't handle the power. The One-Bottle circuit (with or without feedback) would work... At least it's a starting point.

 

[kiira]

[quote author="NewYorkDave"]Why not use a 4:1 or 5:1 output transformer and bias the output stage for 5mA (or so) of plate current? A mic input transformer is no-go, it couldn't handle the power. The One-Bottle circuit (with or without feedback) would work... At least it's a starting point.[/quote]

Yup, that's what I'm busy doing the layout for, without the feedback. I can ask edcor to wind a 15k:600 wsm. I'm just starting over again.

If Paul is reading this... would your two stage (non semi-zen) 6SN7 preamp work as a line amp with a 10k:10k input? It will drive 600 ohms input with the JT-11 right? I have two JT-11-S 600:600. If I get involved in making a new preamp of my own I'm afraid it could not work so well and then I couldn't concentrate on making the sidechain work right - so I better use something proven.

thanks!

Kiira

 

[NewYorkDave]

(Generic blank reply)

There, that worked. The thread shows up on the forum again.

 

[Boswell]

[quote author="NewYorkDave"]There, that worked. The thread shows up on the forum again.[/quote]
Yes, but some posts must have got lost in the server crash yesterday. Mine, and I think at least another one have disappeared.

 

[Kev]

[quote author="mediatechnology"]Yeah Kev and I had a couple about the Vactrols.[/quote]

:shock:

did I get deleted ! ?

boo hoo
:sad:

doesn't matter, I'm full of shit anyway
probably nothing I haven't already said in another thread
do a search
we have been over this before