how can I accomplish this with tubes instead of chips?

[enthalpystudios]

Hey guys....

I'm just wondering what a good way to handle this will be.... i just can't find any good examples in schematic of a tube circuit that does this really....

In fact, maybe its really just a transformer issue, although I'm thinking that a bit of gain could be good to compensate for other loss.

Anyway, I'd like to take +4db referenced line level signal, and send it to 4 destinations: an echo, cue, tape, and summing bus....

echo and cue sends will be mono summing busses. tape will be like a 'direct out,' and the summing bus will be send into a pan pot, and that to a stereo 8 into 2 passive with makeup.

I'm not sure if fixed gain, stepped down with transformers, or lower gain with 1:1 tx's. Or what.

Maybe someone has an example of something similar in sch. form?

Thanks

billy

 

[solder_city]

you want a distribution amp, which can be done with tubes, transistors or chips.

if you wanna do it with tubes... select a line amp that can swing clean at least 12 db above your peak level and use a resistive splitting network at the output. youll lose 12db in this network which is why your amp should be capable of clean output at least 12 db above the peak level you expect at each output.

the best way would be to do as above but replace the resistor network with an output transformer with four isolated secondary windings, but thats tough to find.

 

[enthalpystudios]

thanks solder... distribution amp is the keyword i found just after i posted. 12db sounds cool.... should be able to do it with one tube.... although 18db would be cooler, then i could attenuate echo, cue, and to tape with 6 db, and on the summing bus the pan pot will take care of that part.

And what are the disadvantages of using the resistive splitting network?

Who knows, maybe edcor would wind a quad secondary tx.... I wonder.
Thanks a ton for the reply

billy

 

[solder_city]

this is all inside the same console? if youre careful with your grounds you dont need the isolation of a transformer. resistive splitting would work fine. if the impedances are high enough (and guaranteed to stay that way under all conditions) you could even directly parallel the destinations and not worry about an amp. its hard to say without seeing a schematic of what you have in mind.

 

[enthalpystudios]

well, i thought that perhaps I could parallel it all by being careful about everything, but I'm not expereienced enough to really know when impedance will and won't change and under what circumstances (the unexpected changes, ya know...

I may do a mic pre/eq channel before some of these, but for right now, i'm going to consider this an 8 channel line mixer. Selectable input either from tape or from a (live) line. This will get a fader and pan to send to summing, a 'to tape' direct out, and 2 mono aux sends.

so basically, I'd like to have just around +6db to push the fader, when at max... just so that I can add a little bit, rather than only being able to subtract. And actually, it would probably be smart of me to allow for more than that, just because who knows. Then again, as long as its recorded hot, there should be no need to push that level too high.

Aside from being able to boost the signal a bit at the fader, the other helpful thing will be to have an extra 6 db to the pan pot, which will lose 6db, and then on to summing, saving some gain in makeup.

Anyway, I think I'm going to build as an 8 into 2, but put the 2 busses into a separate section, so that I could add a master type section and have the option of adding another 8 channel deal some other time.

billy

 

[enthalpystudios]

well..... i've been pondering this thing quite a bit, and I really think that I could just do a passive design with makeup, so I think i'm going to look down that route.

I still wouldn't mind having a transformer with a split secondary at the input of each, so I could monitor off of one secondary, and send the other to tape while in monitoring mode... in 'mixing' mode, I can use the extra secondary for metering, multing, all kinds of things.....

Wish edcor had a quicky split secondary.

Anyway, I think i'm going to make one, and if it works as I like and all, then I'll make some more for more channels.

hers a question.... If, in 8 channels into 2 with pan level and aux, I lose around 40 or 50 db, and make that up, I'll have some noise from the makeup.

If I sum those 2 outputs together with 3 other pairs, and do another 8 into 2, for a combined loss and corresponding makeup of around 80-100db across the whole chain, how bad will it be?

Pretty bad? I'm just not sure.....

For 24->6+2(echoreturn)->2, is active a better plan?

Thanks for the help

billy

 

[CJ]

I think you want to buffer each output.
A transformer will provide isolation between outs but not interaction of the output loads. If one sec is overlaoded, it will drag the other windings down, as this load is reflected to the pri, which sags, thus causing the other channels to sag.

 

[enthalpystudios]

yeah, I suppose I should.... I'd like to it right.

could it all be done with iron instead of chips though? Maybe with iron that isnt crazy quad secondary stuff??

Cj, how hard would it be to wind my own quad secondary tx's? it would be a pretty good solution if I could do it... maybe a 2503-alike? it probably wouldnt need a gap for this type of stuff. a 2503 is 3 identical windings, if i remember your testing results from it... and flat too, i recall. maybe I could just do 5 windings?

i dunno... i have to retire from thinking this over, been at it all day... i'm supposed to be on 'vacation'

anyway, thanks

billy

 

[sodderboy]

[quote author="enthalpystudios"]
Anyway, I'd like to take +4db referenced line level signal, and send it to 4 destinations: an echo, cue, tape, and summing bus....

echo and cue sends will be mono summing busses. tape will be like a 'direct out,' and the summing bus will be send into a pan pot, and that to a stereo 8 into 2 passive with makeup.
[/quote]

c'mon. It is not an evil word. . . Say after me- connnnnnnnssssssole :green:

 

[enthalpystudios]

i know... i just want to diy it ;]

billy

 

[sodderboy]

I know you do. It seems that people want to say every buzzword except console. Someone else was designing a "passive summing buss channel with active electronics, inserts and sends" :shock: Wassat? I think it is a console group amp.
Study some console block diagrams, and use whatever op amp you wish where there is a triangle. Even a tube op amp:
http://forsselltech.com/schematics/VT%20Opamp.PDF

Although even an eight channel desk full of these will also be able to keep your chinese take away warm.
You can use 5532's for your cue and aux- who's gonna know? And use tubes or DOA's on the mix buss and direct outs.
I need some boiled, fermented, aged water, barley, yeast, hops mixture. . . I mean beeeeeeeer :sam: :green:

 

[enthalpystudios]

that forsell schem is really interesting.... little tube building blocks ;]

but yeah, i held back from the C word for fear of being yet another "i want to build a diy mixer" dudes that never does it.

but i'm pretty hell bent.

anyway, right now, i'm thinking you have a good point about the 2 aux sends.... who's gonna know is right. 5532 buffer, passive summing, opa2604 makeup, real similar to the forsell schem. besides, they're mono.

i'll use a tube preamp design for themakeup on the main group summing bus. what i believe i should do is use a single split secondary line input tx, send one to tape, and one to the summing bus, from which i can take my 5532 buffered cue and echo sends, buffer the vu's (probably tl072 or whatever, if i'm going to go chips with my buffers), and then send on to the summing network.

i guess i'm kind of stumbling at what I should do with the mix network... i wouldnt mind keeping it balanced, but I'd like to have a pan pot at the same time....

my options are pretty much to go unbalanced and save headaches and cash on the pan pot (dual) or go balanced and start scouring the surplus spots for a decent price on quad linear pots....

well, as far as advice goes on the panning, maybe somebody could help if i phrase it like this:

whats worse, unbalanced summing, or 4 gang pan pots (with most likely +- 20% tolerance) screwing up the balance in a balanced network?

i don't mind saving/searching for 4 gang pots if those more experienced than I think I'll prevent a decent amount of noise by going balanced.

but you're totally right sodderboy, I don't think using a chip buffer for the sends will be a problem at all. it'll save a lot of headaches... in fact, one just went away i do believe.

thanks

billy

 

[sodderboy]

[quote author="enthalpystudios"]
anyway, right now, i'm thinking you have a good point about the 2 aux sends.... who's gonna know is right. 5532 buffer, passive summing, opa2604 makeup, real similar to the forsell schem. besides, they're mono.

i'll use a tube preamp design for themakeup on the main group summing bus. what i believe i should do is use a single split secondary line input tx, send one to tape, and one to the summing bus, from which i can take my 5532 buffered cue and echo sends, buffer the vu's (probably tl072 or whatever, if i'm going to go chips with my buffers), and then send on to the summing network.
[/quote]

1. Do you realize that all summing is passive; a buss is only a common connection. The activity is in amplifying the sum, which for anything more than 3x1 needs to be done.
2. You are hell bent to use a transformer for buffering, but your plan does not allow for seperate TAPE OUT and MIX level control. And the rest of the signal flow does not follow any that I have ever seen. Auxes fed post summing bus? Do you need some block diagrams? I have some simplies that show flow for clarification, and then once understood can be duded-out with whatever amp blocks you want.
Mike

 

[enthalpystudios]

thanks for the thoughts... i really appreciate your opinions, as It's dangerous for me to think this out alone... so mike, thank you for the input....

some diagrams if you have them would be really cool... i'm not at home lately, so i can't really get a block diagram made of what i'm thinking ... i may have explained a little inaccurately...

i wish i had just ms paint or something to use, but i got nothing on here....

WHOA....so.. i just typed a bunch and erased it, realizing that theres no reason to do things half assed... i'm trying to shrug off a tendency to do things quick and cheap. its a real bad habit.

so, a little change....

a mic line pre first... why skip this part? call a spade a spade (a console a console, that is) i do, however, still want to keep things reasonably simple.. not too fancy or extravagant. this is still just my home setup, i'm moving out of 'enthalpystudios' which never was such a hot studio (noisy), although i did do some cool records in there. too expensive to have 2 rents when i have an extra bedroom at home.

anyway

what i'm thinking is going mic/line amp with gain and trim, and that will basically be what goes to tape. output tx on the mic/line amp with a split secondary to isolate 'to tape' from the mixing section.

the other secondary, for mix, will get aux 1+2 buffered and sent to mono summing networks, and continue on to the level, pan, and to a stereo summing network.

now, i'll have a switch for mixing a live signal or a signal from tape... should that switch the feed to the mic/line amp and pad the input, or would it be more effective to have a separate line in transformer that can feed the mix network directly....

i'd probably prefer to pad the tape return and feed it to the mic/line amp, but would padding it and bringing the gain back up be counterproductive?

thanks again, and if you have anything i could look at that would be really cool... especially if its geared towards tube mixers, old or new either way. but anything will be helpful

thanks again

billy

 

[sodderboy]

Start with this:

http://www.dself.dsl.pipex.com/ampins/mixer/mixerdes.htm#6

And go thoroughly through the mixer META. There is so much great info there.
Mike

 

[enthalpystudios]

i like your first suggestion best.... opamp where you see a triangle that is.

be it 5532, doa, or tube op amp...

i like this idea very much...

has you (or anybody) ever built one of forsell's tube op amps? also looks like another fellow has done another, with a single 12ax7...... i think freds looks like a better design.... differential input, good cmrr.

this really gets me going...

anyway, i've printed off half a ream of data, including that last link, and i'm going to go 'hit the books' so to speak and sit outside with the woman. she hates the crt's.

anyway, i'll be very curious to try building one of these valve opamps.

billy

 

[sodderboy]

Good on'ya! The METAs here are an incredible wealth of info. Also check soundcraft and yes, Mackie PDF manuals to see the flow of simple consoles. That's what you want, but with your own flavor of buffering and amplification

And just make sure you pay a little attention to her!
"Darling, the glow in your eyes is even prettier than that of an over-biased 6L6GT"
Mike

 

[enthalpystudios]

ok, so i've been reading the radio designers handbook a bunch, mostly the a-f amplifier chapter, and the introduction chapters on general valve stuff, mathematics, etc....

i still can't really find what i'm looking for, but i'm learning more than ever in the process..... very cool book

so heres what i'm wondering... the rule of thumb seems to be that loading loss from one stage to the next is neglible if the load impedance is 10 times the source impedance....

so, if i have a 60 ohm source impedance, i want it going into at least a 600 ohm load....

now if i have 2 loads in parallel, and a 60 ohm source, would I want each load impedance to be around 1.2k (or higher) so that the source sees a total of 600 ohms? Is it that simple, or close enough?

And perhaps I'd want to drive 4 loads... in that case, if i used 4 10k/600 tx's, the source would see 2.5k total, and I'd get around 12db loss, which I could compensate for somewhere in the chain?

Would that be an acceptable way to drive 4 loads.... perhaps a small, simple, gain stage with 12-18db of gain, with an output impedance of less than 250 ohms, into 4 10k/600 transformers for isolation?

Even in a book as thorough as the rdh, i'm having a hard time nailing this down.

thanks for the running help...

and mike, by the way, i was thinking about moving up to 24 channels (24 track deck), but its certainly not the time for me to be getting in over my head... i think an 8 channel will go a long way for my purposes.... i'll just have to bounce 8-2 whenever i bust 8 tracks ;] got to learn to commit to things, rather than just f-ing about with cubase undo's and eternally twaeakable plugins

billy