Hey Everyone;
All right... my second post and this may be a long one. I want, overall, to stimulate some discussion, clarify some of my understanding and just, overall, get some ideas for a DIY mic preamp project I'd like to embark upon. This project stems from a direct need to have a very configurable mic pre that really does what I want it to do. Of course, that's what we're all looking for, in the end, I think.
In reading through all of my documention, including the Radio Designers Handbook (F. Langford Smith), Mullard Tube Circuits, Handbook for Audio Engineers, The Art of Electronics, and various other volumes covering tube and solid-state design as well as various schematics from the industry I've sort of come to an impasse with regards to mic pre's.
Let's start w/ the 10,000 foot view of these things. All camps scream "accuracy" in the mic pre. They gain this component aspect of the pre by various designs - high impedance inputs so that there is no "loading" of the mic, differential amps all the way through the signal path for high CMRR, fully "balanced" signal path from input to output, DC-coupling between stages to minimize/avoid low frequency phase shift, minimal component circuits to avoid coloration by multiple stages of active devices, high amounts of negative feedback to increase entire pre linearity, high quality components, etc.
Let's drill down... there seem to be a few different camps, in the world, with regards to inputs. The first camp says that transformers inherently color the sound of the mic and, in some cases, unnecessarily "load" the mic down. So DC-coupling, direct to your first preamp stage is the way to go. Others say that the transformer allows for better CMRR, and "proper loading" of the mic. So, the first "confusing" part of the mic pre starts here.
Personally, I like the idea of a multi-tapped transformer, at this stage. A linear unit, like some of the Jensen's that can pick up and translate signals from nearly DC to a couple hundred kHz without phase shift or amplitude distortion. But, of course, Jensen doesn't make a multi-tap transformer for this... so, who does?
The second question is: Does the transformer step up the voltage/impedance, or step it down to the amplification stages, or do we achieve better results with a 1-to-1 transfer? And how does this impact the tonality of the pre after that stage?
Once the signal is removed from the transformer and into the mic pre's active stages, there are a ton of different considerations. If you're looking for accuracy, then a higher impedance input to the first gain stage is a must to pick up all the transient information, etc. at that stage. Secondarily, there are other things to look at. In order to be totally accurate. From the higher level design references that I have, they note that this first input stage "balances" best at about 10K ohms of input imedance. Much less than that and you load the input stage. Much more than that and you increase the EIN, random noise of the stage, etc.
So the question, here, really becomes what type(s) of input impedences are we looking at to really optimize the stage, here. There is one thing that I've noticed about all the mic pre's that I worked with (Focusrite, Groove Tubes, Avalon, Millenium, etc) is that when the higher input impedances are used (both on the input transformer and on the first gain stage) the mic loading may be somewhat less, but man, they get "edgy" and "harsh". Is it just that the "noise" I create to record has that much "noise" inherent in it? In other words are the mic's picking up all the environmental noises and really transferring them to my pre's?
In all of the gain stages we should be designing to several things: 1) Frequency linearity which requires some sort of feedback to keep the stage linear, 2) Slew Rate - to be able to accurately capture the transient information WITHOUT changing it's frequency reference, 3) Phase accuracy - control of stray capacitances, inductances and the proper matching of caps, etc. when/where needed through the signal and feedback paths, 4) Amplitude accuracy - the ability to handle inbound transients without compromising the actual signal envelope, etc.
To me, if you're going to focus on the signal path, in this way - and I could be wrong, it would require several things: 1) High voltage supplies on the active elements in the circuit. The higher the supply voltage the more we can put the signal swing in the "middle" of the transfer curve for that element (whether it be a tube, transistor or FET) where the device is the most linear. Also, in the tube world, this allows for enough electrons to be built up in the device ready to move at a transients notice so that slew rate is at is level best. 2) High quality signal path components (i.e. very tightly toleranced resistors, caps, etc.) so that the contribution to overall signal "degredation" is at a minimum. 3) Either really HUGE coupling caps between stages or no coupling caps, at all to eliminate any potential low-frequency phase shift on the input.
Finally, the output stages of the thing require a few elements, as well. The first part, again, is how to couple the output of the active device(s) to the actual output terminals - do we use transformers, do we couple via caps, etc? The goal, in the better audio equipment, is to build an ultralow impedence output so that any load can be put on the thing and not have to worry about "loading" the output of the pre. What considerations would we try to account for, in this stage.
Ultimately, after having really torn down the middle section of the mic pre, here's what I'm looking for. I want a mic pre that will "capture" the depth of the low-end on my mic's. I know it's there. I can, sometimes, here it. I want a mic pre that will accurately capture, and translate, the high-end frequencies coming from the mic's without sounding overly aggressive, or harsh. I want to be able to select different input impedences from 150 ohms to 10k ohms, or there abouts in some usable steps. I want to be able to change, possibly, the slew rate of the device, as a whole (probably just the final output stage, rather than the whole device as this seems to make the most sense).
Anyway... there starts the discussion. I'd love to get into the technical details of some of the mic pre designs and start really sussing out the cooler aspects of each. Sometimes, I get the feeling that recordists, in general, are in the "vintage is better" camp without really focusing on the engineering aspects that make "better" actually better. I'm really wanting to focus on the contributory elements that really separate a great mic pre from a poor one.
Also, if folks have some designs, other than the G9, to share, that would be great. While I've worked w/ the Avalon and Millenia's, in the past, I haven't had a chance to open them up and see how they were put together - although I have enjoyed some of the sounds that come out of them. I just got a Groove Tubes VIPRE, recently (at well below cost, for a new, in box, unit). I'm going to be spending some time, when I get off the road, for work, and taking that thing apart and really evaluating how they did some of their work, as well.
Thanks for listening, would love to have some discussion come out of this.
Thanks.
Dar
All right... my second post and this may be a long one. I want, overall, to stimulate some discussion, clarify some of my understanding and just, overall, get some ideas for a DIY mic preamp project I'd like to embark upon. This project stems from a direct need to have a very configurable mic pre that really does what I want it to do. Of course, that's what we're all looking for, in the end, I think.
In reading through all of my documention, including the Radio Designers Handbook (F. Langford Smith), Mullard Tube Circuits, Handbook for Audio Engineers, The Art of Electronics, and various other volumes covering tube and solid-state design as well as various schematics from the industry I've sort of come to an impasse with regards to mic pre's.
Let's start w/ the 10,000 foot view of these things. All camps scream "accuracy" in the mic pre. They gain this component aspect of the pre by various designs - high impedance inputs so that there is no "loading" of the mic, differential amps all the way through the signal path for high CMRR, fully "balanced" signal path from input to output, DC-coupling between stages to minimize/avoid low frequency phase shift, minimal component circuits to avoid coloration by multiple stages of active devices, high amounts of negative feedback to increase entire pre linearity, high quality components, etc.
Let's drill down... there seem to be a few different camps, in the world, with regards to inputs. The first camp says that transformers inherently color the sound of the mic and, in some cases, unnecessarily "load" the mic down. So DC-coupling, direct to your first preamp stage is the way to go. Others say that the transformer allows for better CMRR, and "proper loading" of the mic. So, the first "confusing" part of the mic pre starts here.
Personally, I like the idea of a multi-tapped transformer, at this stage. A linear unit, like some of the Jensen's that can pick up and translate signals from nearly DC to a couple hundred kHz without phase shift or amplitude distortion. But, of course, Jensen doesn't make a multi-tap transformer for this... so, who does?
The second question is: Does the transformer step up the voltage/impedance, or step it down to the amplification stages, or do we achieve better results with a 1-to-1 transfer? And how does this impact the tonality of the pre after that stage?
Once the signal is removed from the transformer and into the mic pre's active stages, there are a ton of different considerations. If you're looking for accuracy, then a higher impedance input to the first gain stage is a must to pick up all the transient information, etc. at that stage. Secondarily, there are other things to look at. In order to be totally accurate. From the higher level design references that I have, they note that this first input stage "balances" best at about 10K ohms of input imedance. Much less than that and you load the input stage. Much more than that and you increase the EIN, random noise of the stage, etc.
So the question, here, really becomes what type(s) of input impedences are we looking at to really optimize the stage, here. There is one thing that I've noticed about all the mic pre's that I worked with (Focusrite, Groove Tubes, Avalon, Millenium, etc) is that when the higher input impedances are used (both on the input transformer and on the first gain stage) the mic loading may be somewhat less, but man, they get "edgy" and "harsh". Is it just that the "noise" I create to record has that much "noise" inherent in it? In other words are the mic's picking up all the environmental noises and really transferring them to my pre's?
In all of the gain stages we should be designing to several things: 1) Frequency linearity which requires some sort of feedback to keep the stage linear, 2) Slew Rate - to be able to accurately capture the transient information WITHOUT changing it's frequency reference, 3) Phase accuracy - control of stray capacitances, inductances and the proper matching of caps, etc. when/where needed through the signal and feedback paths, 4) Amplitude accuracy - the ability to handle inbound transients without compromising the actual signal envelope, etc.
To me, if you're going to focus on the signal path, in this way - and I could be wrong, it would require several things: 1) High voltage supplies on the active elements in the circuit. The higher the supply voltage the more we can put the signal swing in the "middle" of the transfer curve for that element (whether it be a tube, transistor or FET) where the device is the most linear. Also, in the tube world, this allows for enough electrons to be built up in the device ready to move at a transients notice so that slew rate is at is level best. 2) High quality signal path components (i.e. very tightly toleranced resistors, caps, etc.) so that the contribution to overall signal "degredation" is at a minimum. 3) Either really HUGE coupling caps between stages or no coupling caps, at all to eliminate any potential low-frequency phase shift on the input.
Finally, the output stages of the thing require a few elements, as well. The first part, again, is how to couple the output of the active device(s) to the actual output terminals - do we use transformers, do we couple via caps, etc? The goal, in the better audio equipment, is to build an ultralow impedence output so that any load can be put on the thing and not have to worry about "loading" the output of the pre. What considerations would we try to account for, in this stage.
Ultimately, after having really torn down the middle section of the mic pre, here's what I'm looking for. I want a mic pre that will "capture" the depth of the low-end on my mic's. I know it's there. I can, sometimes, here it. I want a mic pre that will accurately capture, and translate, the high-end frequencies coming from the mic's without sounding overly aggressive, or harsh. I want to be able to select different input impedences from 150 ohms to 10k ohms, or there abouts in some usable steps. I want to be able to change, possibly, the slew rate of the device, as a whole (probably just the final output stage, rather than the whole device as this seems to make the most sense).
Anyway... there starts the discussion. I'd love to get into the technical details of some of the mic pre designs and start really sussing out the cooler aspects of each. Sometimes, I get the feeling that recordists, in general, are in the "vintage is better" camp without really focusing on the engineering aspects that make "better" actually better. I'm really wanting to focus on the contributory elements that really separate a great mic pre from a poor one.
Also, if folks have some designs, other than the G9, to share, that would be great. While I've worked w/ the Avalon and Millenia's, in the past, I haven't had a chance to open them up and see how they were put together - although I have enjoyed some of the sounds that come out of them. I just got a Groove Tubes VIPRE, recently (at well below cost, for a new, in box, unit). I'm going to be spending some time, when I get off the road, for work, and taking that thing apart and really evaluating how they did some of their work, as well.
Thanks for listening, would love to have some discussion come out of this.
Thanks.
Dar