[rlackey]

Hello! This is my first post in this forum, so hopefully I've done my research well enough and this post will make some sense to someone.

I am designing a modular in-line 8 buss valve console, each input channel needs to have a valve mic preamp and eq stage. The EQ doesn't have to be fully parametric, but I need more control over the mids than the standard Pultec curcuit affords.

I plan to build modified versions of both a "Gyratec IX" mic pre amp, and the SRPP Pultec passive EQ. The starting point for both of these are the standard schematics available from Jakob's website.

This is all well and good, but I am not very good at leaving things well enough alone so I want to add the MEQ filter stages into the Pultec EQ.

Now, I've read through a very helpful thread on this subject here. "Pultec filter section question" - http://www.groupdiy.com/index.php?topic=2534

It seems that much progress was made and resulted in some very useful schematics that I have also printed out.

I have a couple of unresolved questions, which I am sure would be answered by studying NewYorkDave's EQ, however I totally understand his reasons for taking down the schematics.

The first question is, due to the 50-some dB drop that results, how best can I modify the gain stage to compensate? I seem to remember JC commenting on two possible solutions, splitting the filter sections, each with it's own amplifier, or putting an amplifier before and after the passive filter.

This also links into my second question.

I have been in contact with Per Lundahl about ordering suitable transformers, as I was unsure that the LL5402 called for on the input of the Pultec was correct. The LL5402 is actually a output transformer, and Per noted to me that it was in fact unsuitable for use as a input. He recommended a LL1531, LL1540 or LL1527. He also recommended the LL1538 as a mic input transformer for the mic pre amp rather than the LL1528 which is on the schematic.

This lead me to find another thread entitled "Pultec IN/OUT transformer impedance, in which the use of a higher impedance input transformer was discussed, and Jakob commented himself that the Pultec requires a low z input transformer because it can load down to 75ohms (in a worse case scenario). Is the Gyratec IX likely to handle this situation?

Ok, to summarize both of my questions into one (long) sentence.

Can anyone help me understand how to implement the extra filter sections with either of the options JC mentioned? (i.e. a gain stage before and after the filter sections, or separate the filter sections, each with its own amplifier... I picture the options kind of like running them either in serial or parallel).

Both of these options also leave me a bit lost when it comes to choosing suitable input and output and transformers.

Please also let me know and correct my interpretations of these two options if I am wrong.

I hope I'm not asking questions that have been asked a million times before.

Thanks!

Rich Lackey  :grin:

[drpat]

...

[rlackey]

To help you out with the makeup gain on the EQ section, putting an API 2520 (or similar) opamp after the filter section seems to be the easiest way to do this. That's the way that Pultec originally did it, so it would get you closer to that original sound.

Welcome to da lab!

Thanks! That's a great start, but I am hoping to find a tube solution. I'll start running some simulations (I don't expect everyone else to do all the work for me), I was just hoping for a bit of a kick start in the right direction. There is an awful lot of knowledge in this forum!

Cheers!

Rich

[drpat]

...

[rlackey]

Well, you may well be right, I haven't heard a real tube or a real solid state Pultec before, so I can't really comment on the difference between them. However, I know that I want to stick with the tube topology, but compensate for the gain loss brought about by the extra filter sections. The idea of this project for me, is to design and build an all-tube console for studio use.

The project in whole, is more complicated than just this, as the idea is to create a real hybrid. I have also been working on DAC and ADC modules on each input channel (AES/EBU I/O on each channel plus ADAT I/O on every group of 8 channels... plus an array of digital I/O options on the master section). The console will also feature midi control (incl motorized faders) for software based multitrack recording and editing.

The digital side of things has come together far more quickly and easily than the analog side of things. I am new to designing and implimenting valve audio curcuitry, which is why I finally decided to post a question after lurking around for a long time.

The challenge is in really keeping the audio path purely analog, as high quality as possible in terms of components, and if possible, purely tube driven.

Back to the question at hand, I agree with your comments about the transformer, I don't really want to mess with the transformers at all if I don't have to. The more variables I try to mess with the more it's no longer going to even resemble a Pultec anymore.

The idea that I have in my head, and that has inspired at least a couple other threads I have found when searching, is to combine the PEQ1A and MEQ5 in a way that changes or adds as little as possible to the original design.

Rich  :grin:

[drpat]

...

[mattmoogus]

Hey Rich,

Its great to set your sights high and I admire your enthusiasm but youve really chosen a hell of a project to dive into...

The words 'all tube' 'mixing console' and 'digital' ring major complexity alarms for me.  In fact any two of those put together is a serious project even if you have a lot of experience with designing that kind of stuff, let alone all of them in the one box.

Heres my humble and relatively non technical opinion of what you need to think about before you take this on:


Design Philosophy

Firstly, you need to be clear about your reasons for wanting to make it an all tube desk.  You havnt alluded to what these are so I wont assume.  My own belief is that you can build just as good sounding (perhaps better, more sonicly versatile) a console with solid state circuitry and it will be less expensive and require less maintainence.  However there is the 'cooool' factor with all tube consoles and perhaps also the appeal of the challenge of making such a thing work.  
Either way I think perhaps you should state your reasons for wanting tubes, if for no other reason than it will make them clearer in your own mind.


EQ

Similarly, if you havnt used a pultec before, what made you decide on that for the channel eq?  If you are building one into a channel it will be impractical to build an exact clone with the same transformers etc, and I would question the usefulness of this aswell.  If I had a pultec type eq in a console channel Id definitely want a lot more freqs and bands than the standard pultec and this would mean so many controls in the eq section as to make it impractical...
..or at least make me think theres got to be a better way to do this...

I would seriously consider a circuit like this one (on Fred Forssell's site)
http://www.forsselltech.com/Evolution%20of%20an%20EQ%20Design2.pdf

That way each band has single pot for cut and boost, you can have more bands with more freqs, you can have a bandwidth control if you want, you can switch any band from bell to shelf if you want, the whole eq uses one single amp stage which you can build with tubes or transistors AND (and this is the most important) it will sound as good as (and indeed similar to) the pultec idea while being much more useful during mixing.  
Of course youll need to source inductors for this eq, but you need to do that for pultecs anyway and there are lots of options around for this.


Amp Circuit

You also need to (in my view) design or decide on a single tube amp circuit that you can use throughout the console.  It needs to be able to provide enough gain for the mic pre (two stages could be stacked if nessesary), it should be able to be used as the opamp for the eq section, and it must be able to drive low impedance, either directly or through a transformer for your direct outputs, mix buss and your A-D convertors (which will be very low Z in).  Of course it should also be quiet, simple (not many tubes) and fairly cheap, not run extremely hot and not require very expensive transformers or other parts.  This is because there will be at least 2 of these stages per channel so any problems like excessive heat or cost will be multiplied many times over the whole console.
If you can come up with a single amp that can do all these things you will save youself a lot of time, money and hastle when you have to build this console, and maintain it.  
There are some excellent and very experienced tube circuit designers here who may be able to help you with this, a lot more than I can anyway!


Transformers

Seeing as youll need a whole lot of them for this console, I would pick a brand that you can get fairly cheaply but that still performs well.  That means not Lundahl, unless of course you just inherited and oil field somewhere.  Also I think if youre making a tube console you may aswell pick transformers that have a bit of color, and Lundahl dont IMO.
Im sure everyone here is sick of hearing me rabbit on about what transformers I like so Ill spare you the repetition.  Also I work for a company that sells transformers, and if I was doing this Id use them cause I think they fit the task well, but that doesnt mean you have to! :razz:
You need to pick a mic in transformer (Im assuming this console will be used for recording aswell as mixing) and a suitable output transformer.  The criteria I would use to pick them would be;

Input:  
-should have a nice tone that works on a broad range of sources
-should have a good extended freq response without the need for much zobel trickery
-should be high ratio ie 1:4 - 1:10 to get as much level as possible into the tube amp stage
-should be inexpensive
-input headroom isnt all that important as you can pad the level down to whatever you like.  The line input can be through the same transformer with a suitable pad in front.

Output:
-should be able to handle above +20 @20hz
-ideally shouldnt be a large stepdown, if you amp circuit can drive low Z so you dont loose too much gain
-should have a fairly flat repsonse, especially in the low end
-should be inexpensive


Theres a hell of a lot more I could write but for the moment, see if any of that makes sense.  Once again, what youre describing is a mamoth project and really needs some serious thought before you proceed unless you want to waste a lot of time and money.  But good luck and I hope you get to build something cool and impressive!


[email protected]

[mattmoogus]

Oh and the A-D side, making that work in an all tube console will be a serious feat, and I would strongly advise you to build the digital part into a sepperate case with sepperate PSU etc, and keep it the hell away from the hi Z circuits in the console.  

*pant*

 :wink:

[email protected]

[rlackey]

Wow! I think that is the single most helpful and broad range of advice I have yet had on this project! Thanks so much!   :grin:

Yes, I've thought about a lot of these issues before as well. Let's see, where to start.

I like the way you organized your reply to me, so I'm going to break things down that way too.

Why Tubes?

The decision to go all tube has a lot to do with the challenge of the design, especially in integrating the ADC and DAC into each channel as well, and to be honest, the fun too.  :razz:

In addition to these two important factors, I also think if it's done well, and I select the right curcuits for the desired results of each stage it should sound pretty good. I heard a recording and mix through a TLAudio VTC and I fell in love, however I want something that suits the way I want to work.

Design Philosophy

I don't like the fact that to bridge the analog/digital divide and use software multitrack recording, I have to have three separate consoles, a analog desk (either tube or solid state... but something with some character), a digital desk (or at least outboard ADC / DAC), and a midi control surface.

This kind of thing is actually so subjective, but I started thinking, why not design a single console with all this functionality built in?

There was immediately a fire lit under me by the challenge of it, I knew that just because it hasn't been done commercially (and there are probably many good reasons why not), there was no reason it can't be done.

Modular building blocks

I started (as with any design, for anything) deciding and writing down a project brief, and defined objectives that could be used as a reference, against which to measure results at various stages of the design.

Working from large blocks, to smaller and smaller blocks, I began to map out a huge block diagram of the whole console. This started to define individual curcuits which form specific "modules". It was also shaped by how I want to be able to use the console in the end. Each "block" or curcuit is responsible for one task only and this greatly simplified things because I could concentrate on one small task at a time, but with a constant reference to the bigger picture.

I speak in past tense, because a lot of this is done. In fact a lot of the digital side of the console is done, or at least schematics are done, and for the moment should do what they are supposed to.

Completed Curcuits

So far, the following building blocks or modules are pretty much there.

Channel ADC / DAC (AES/EBU & ADAT) - The actual ADC and DAC and AES/EBU and ADAT encoding are complete, each module handing digital I/O for 8 channels. This is almost entirely based around Wavefront Semiconductor's AL1101G (8 channel ADC) AL1201G (8 channel DAC), AL1402G (ADAT Rx), and AL1401AG (ADAT Tx). The AES-EBU on each channel is encoded by Cirrus Logic CD8406 (Tx) and CS8416 (Rx). This gives me 24bit 48kHz digital I/O with 107dB dynamic range.

This same module also provides digital out for the 8 group channels, and stereo master channels for the console master section.

Digital Volume Control - A interesting problem arose from the need to have a single motorized fader do two mutually exclusive jobs.

The first job is to provide a reference signal for the microcontroller providing midi output for software control, and in feedback for the pulse width modulation of the drive current to the fader motor when receiving midi data to automatically position the fader.

The second job is to actually attentuate the level of the live audio passing through that particular channel.

The first idea was simple multiple resistance tracks, a linear track with touch track to drive the midi controller, and a audio log track to handle analog audio. However, I was afraid that switching noise from the pwm driving the motorfader would affect my audio. I could push the pwm frequency suitably beyond any audiable levels, but this was not possible with the microcontroller I chose.

So, the two jobs have to be physically and electrically separated completely, yet still be driven from a single physical fader. The solution came with having the fader itself only provide a reference for the microcontroller, this would only call for a simple and inexpensive single linear track motorized fader (with a touch track).

The microcontroller then feeds a "digital volume control" made by Cirrus Logic, the CS3308, which provides eight channels of analog volume control in 0.25dB steps. This resolution should be more than adequate. However, I don't want audio to pass directly through the IC, I would rather somehow have the IC control a tube gain stage for the audio.

Midi Controller - The midi controller is complete, based on the "MIDIBox" and MIOS software that can be found at www.uCApps.de

It's an 8 bit microcontroller with up to 64 analog inputs (10 bit resolution ADC) and between 64 and 128 digital inputs depending on how it's configured. It ouputs standard midi data across standard midi ports or usb to a host PC and will drive any of the popular multitrack recording software packages.

I actually have a number of these microcontrollers doing specific tasks, one microcontroller is dedicated to every 8 channels of motorfaders and digital volume control, while another microcontroller only handles the midi I/O. These microcontrollers also drive various graphical LCD modules which will provide channel labelling from the host software directly on the console, and a large LCD panel for the master section of the console for menu navigation.

It sounds complicated, but its not really.

So that pretty much covers everything digital in the console, there is no digital signal processing at all taking place, just the most transparent ADC and DAC I can manage to implement in order to best capture all that tube goodness into streams of 1's and 0's, or feed digital audio from software multitrack, samplers, or outboard ADAT, synths etc. back through the console's tube stages.

Ok, now you have the background and the work done to date, I'll get back to the points you made.

Pre Amp and EQ choice

I "decided" (nothing is really set in stone yet) on the Pultec for channel EQ because I couldn't find much useable info about much else, and most everything anybody said about the Pultec was good.

In truth I have only begun to work on the valve stages of this console (which is really the heart and soul of the whole project) so I'm still open to new suggestions and ideas. I will check out the curcuit you mentioned.

I have been lurking around for a while (the amount of posts I have read by now make me feel like I already know some of you!) but I registered and stuck my neck out now because I need help, and advice and suggestions!

I really liked the G9 and SRPP Pultec EQ that I saw on the Gyraf site, they just seemed to compliment each other somehow, and there seemed to be lots of info and help available, so that was (and is at the moment) the starting point. I just wanted more mid control from the EQ.

Amp Curcuit

I totally agree! I'm there 100%, one tube amp curcuit that I can use throughout the console is without doubt the way to go cost wise and simplicity wise.

Anyone reading this thread who feels intrigued or interested in helping in this regard is most welcome! This is a totally "open source" project, and I am happy to share anything with anyone. The only thing truely original in this project is the way I am combining existing ideas and curcuits into a much larger animal, but the individual building blocks are nothing new.

Transformers

I haven't inherited a field of oil, but I was initially thinking using Lundahl across the board, I really would like to stick to just one manufacturer. I thought I might be able to order in a large enough volume to get some kind of discount.

Color is good, I've also contact Sowter about this project, and they might even send me some samples.

Please tell me more about the transformers your company sells, and perhaps which ones you might reccommend. For mic input I was considering the Lundahl 1538 which is a 1:5, or the Sowter 4935 which is 1:7. Both of these came out similar in price (the LL1538 was £31.48 GBP and the Sowter 4935 £31.42).

I didn't think those prices were bad, plus Lundahl were willing to give me 30% off (now they are closed for four weeks though  :roll: ). However, it adds up super quick in an application like this.

Price vs Quality is a tough line to draw, I'll have to find a balance somehow.

Don't stop there! I'm soaking up what you wrote, so if you have any other ideas, feedback or input, I can use all the help I can get.

Oh, also I've been trying to figure out exactly how I'll separate things, this has as much to do with final construction and board placement as anything else. Expecting this to be in any way compact, is probably asking a bit too much! I want to be clever with space though and make it as compact and as integrated as possible.

These are somewhat uncharted waters though I think, so like anything, I can do as much theoretical planning as I want, but a point is going to come where I have to just try it.

Cheers!

Rich  :grin:

[rlackey]

BTW, another thought as far as separating analog from digital curcuitry and so forth. I suppose I see in my head some kind of a "rack" frame for the console in which different sections plug in to some kind of a backbone. Each section can be in it's own closed housing, just as one might stack 19" rack equipment into one rack. The front panel in this case of each individual section would make up the whole surface of the console.

This however leads to the question of how to separate things. Does each channel strip have it's analog and digital sections in separate housings screwed into the "rack" right along side each other?

Or as the console is modularly separated into groups of eight channels already, should there be two separately housed larger units that plug into this backbone one containing digital and the other analog curcuitry for eight channels each. Then the front control "fascia" of the console would have to be separate from either of these as the faders/pots/switches/encoders for a single channel would have to interface with both the analog and digital modules.

Either way this involves a lot of seperate enclosures, and a very large frame or "rack". Each plug in module, I suppose (either containing a single channel's or group of eight channels, digital or analog curcuitry) would need to include it's own PSU.

Hmm, this has me envisioning now a front fascia panel on which all the controls are mounted, hinging upwards to reveal what is basically a standard 19" rack system laying down underneath it.

The hinged front fascia would be divided into groups of eight channel strips, and then a slightly larger one for the group and master section, and when opened would reveal the standard 19" rack enclosures underneath.

It's difficult explaining this without drawing it. Let me know if I am not making any sense.

This could be a good way to approach it, but if we assign 1U of "width" per channel, then a single 6U rack enclosure would still, if possible have to contain all the valve curcuitry needed for a group of eight channels plus a suitably sized PSU to drive it all. Then an additional 2U enclosure could easily contain all the digital curcuitry required for eight channels.

Of course this requires very long leads in a tight and well routed wiring loom connecting the switches and pots on the facia though a suitable plug and socket to the actual valve curcuits inside it's enclosure. I'm not sure how well this would work, as I am sure these leads should be kept as short as possible.

I hope this made sense to somebody. I'll keep thinking about it.  :razz:

I'm going to be soldering till I die!

Rich  :grin:

[mattmoogus]

Wow I wasnt expecting that so fast!

Ok...

First up I work for JLM Audio http://www.jlmaudio.com.  The transformers we make are a 1:4 mic input and 1:1:1 line output.  I love them both and thats what I/we would use in a similar project.  But of course we would say that...
How many channels are you thinking this desk will have?  16-32 with a $100 (aussie dollars Im talking) input transformer per channel...it adds up quickly!  Then add an output transformer to that...

The TL Audio valve gear (and Im assuming their consoles are the same as they only have 1 or 2 tubes per channel) isnt what I would call 'real' valve gear.  Its basically a tube stuck in the middle of some bad sounding solid state circuitry for marketing reasons.  It all sounds harsh and...harsh to me.  Not what good tube or solid state gear should sound like.  
If you want a console to sound like that its easy, you dont even need a tube, just a B class transistor amp surrounded with SSM chips.

Real tube gear sounds nothing like that, generally smoother and slower, with lots of detail in the mids and often not the tightest low end.  Of course its possible to make a tube amp thats the opposite of this but then why would you bother using tubes if you want clean and fast :?:

I seriously, seriously, really and truely think you should make it solid state, using 990 footprint descrete opamps.  Unless youre already experienced with tube circuit design you wont be able to get a practical circuit by simply combining existing DIY projects into a channel amp.  Even if you could come up with a layout that physically worked, without the understanding you need to design it from scratch I dont think youd overcome the myriad of problems youll encounter when troubleshooting it once its built.

I know this sounds negative but its the truth and I think many here would agree.

Building it around 990 opamps will give you just as good a sound as a very well designed tube circuit (I think better) and will let you 'plug and play' with existing mic pre, eq and even compressor designs.  It will also allow you to swap opamps to change the sound of the console, or even individual channels, and give you many more tonal options and much more control than youd get with a tube design.

You also then dont have to worry about the extra complication of building a high current, high voltage supply for the tubes and having high volts floating around inside the console.


Volume control

I think by far the best way to do this is with a VCA, like SSL do in the K series, using several THAT 2181s in paralell to reduce distortion.  Even if you just use one its a pretty good solution to a difficult problem (giving your DAW control of your audio faders).  There is a way to do it with tubes but its so much more difficult that its not even worth thinking about.


Digital

I dont see a good reason for putting the ADA convertors in the same case as the audio path, and many reasons for keeping them sepperate.  For one, you can build them and get them going before you finish the console.  Keeping sources of digital noise away from your audio path is important.  Also, the life of whats considered a good convertor today is much shorter than whats considered a good mic pre or eq, so being able to upgrade the digital part of the console sepperately to the analog parts would be good!

Im not even partially across what makes for good digital design so I cant say much more than that.  I dont know of the chips you mention but I do know theres much more to a good convertor than the chip you use.  Clocks, power supply and surrounding analog circuitry all have a huge influence.


Theres so much more to say...think about the 990 idea for a bit though.



[email protected]

[rlackey]

Ok, I think I've got a lot of good information to process now.

I would be interested to know more about the tube method you mentioned for volume control (the one that isn't even worth thinking about... that's just about the worst thing to say to me, then I'm guaranteed to think about it  :razz: ).

I'm not quite ready to give it up yet, but I won't hold onto the all tube idea if it's really not going to work, or simply isn''t worth it. I just want to arrive at that conclusion for myself too.

I'm enjoying the learning process just as much as trying to put together a coherent design, so even if I throw it out, I want to walk all the way down the road because I started, and I will learn a lot. I'm not in any rush.

I am going to keep reading up as much as I can though.  :grin:

Rich

[rlackey]

Check out this beast!
http://www.vintec-audio.de

This is one channel!
http://www.vintec-audio.de/vintec/index_E.htm

Now that's cool. I would be hoping to get away with far fewer tubes per channel, but the packing density and board design are very inspiring.

The engineering behind that boggles my mind  :shock:

I give up.  :?

Matt, to answer your question about channels, I was planning to start with just the master section, with it's 8 group channels and the master output, and then one section of eight input channels. The idea is to have 32 in total, but I want to design it in such a way that it could be expanded to up to 96 or whatever was called for.

I'm kidding, I don't really give up.  :razz:

Now at least I know it can be done... now where to find the schematics...

Rich  :grin:

[W DeMarco]

If you don't like the sound of digital, save your money and get a 2 inch deck!!   Oh yeah and a 16 trk stack for it.

Pultec PEQ with MEQ

Started by craigmorris74 The Lab

Yet another Royer Mod based Mic

Started by bruce0 The Lab

Yet another Profire 2626 mod question

Started by kingkorg The Lab

Quick Pultec MEQ-5 output transformer question.

Started by fallout Filters/Equalizers