Winston OBoogie
Well-known member
...
trafoless microphone preamp wish list
- Thread starter bcarso
- Start date
Help Support GroupDIY Audio Forum:
pstamler
Well-known member
I'd like to air one of my prejudices and say that I would like the gain of the electronic circuit to not be variable, at least not via the usual routes. The reason is that most of the time the gain is varied by varying the amount of feedback around the circuit, and in effect you have a completely different amplifier at low gain and high gain. Different distortion for sure, different output impedance, probably different bandwidth and/or noise performance.
I've just been testing out a piece of commercial gear, and its output Z is about 110R at minimum gain, and 2.5k at maximum. Harmonic distortion spectra are completely different.
I don't want that in a preamp; it's one variable too many for me. Controlling how much gain goes between the input and output is necessary, obviously, but controlling it this way doesn't appeal to me. In my own designs I've used a low-gain input amp, designed to be very hard to overload, then a level control, then another gain stage. Both gain stages retain their characteristics regardless of where the level control is set. There are other ways to do this too, obviously.
Peace,
Paul
I've just been testing out a piece of commercial gear, and its output Z is about 110R at minimum gain, and 2.5k at maximum. Harmonic distortion spectra are completely different.
I don't want that in a preamp; it's one variable too many for me. Controlling how much gain goes between the input and output is necessary, obviously, but controlling it this way doesn't appeal to me. In my own designs I've used a low-gain input amp, designed to be very hard to overload, then a level control, then another gain stage. Both gain stages retain their characteristics regardless of where the level control is set. There are other ways to do this too, obviously.
Peace,
Paul
bcarso
Well-known member
Thanks Paul---that's exactly the sort of input and information we need. Thanks as well to all others who have made suggestions. This undertaking may generate more than one design, as I hardly expect unanimity!
One intriguing thing would be to see just how big a fixed gain stage's output swing could be at a gain sufficient for the vast majority of input levels, while still not spoiling the S/N ratio for hotter sources.
Another alternative to address your concerns would be a design in which the loop gain was kept more or less constant while the closed-loop gain was varied, while still preserving the other attributes like noise. Some distortion mechanisms would have to be controlled other ways, for example common-mode distortion, which if present gets worse at lower gains and hence probable larger input signal swings.
The problem with output Z could be handled with a separate buffer, although this would add parts.
One intriguing thing would be to see just how big a fixed gain stage's output swing could be at a gain sufficient for the vast majority of input levels, while still not spoiling the S/N ratio for hotter sources.
Another alternative to address your concerns would be a design in which the loop gain was kept more or less constant while the closed-loop gain was varied, while still preserving the other attributes like noise. Some distortion mechanisms would have to be controlled other ways, for example common-mode distortion, which if present gets worse at lower gains and hence probable larger input signal swings.
The problem with output Z could be handled with a separate buffer, although this would add parts.
Apparently tubes are out, sounds like discrete sandstate, Massenburg flaunts the superiority of +/-28v shall we go there? Power transformer is easy to find, LM317 and LM337 pair? Maybe a dedicated line driver chip like the SSM2143 or servo out. Can we eliminate the use of electrolytics in the signal path (blocking phantom pwr etc.) Should we use economics to determine the sound quality or the end result, whichever comes first?
Analag
Analag
Svart
Well-known member
I haven't been here in a couple of days and you guys are doing this without me?!
Ok, here are my criteria.
1. Clean and clear. It's easy enough to make a circuit colored by sticking trafos or tubes on it, but it's hard to find a truy clean preamp due to the problems that have been noted in postings in this thread.
2. Addressing the problems(again) with gain vs. noise/output impedence, Distortion, etc.
3. It fetches me beer on command.
4. It uses FETs..
5. 1,2 and 4 are optional. 3 is mandatory..
:green:
Ok, here are my criteria.
1. Clean and clear. It's easy enough to make a circuit colored by sticking trafos or tubes on it, but it's hard to find a truy clean preamp due to the problems that have been noted in postings in this thread.
2. Addressing the problems(again) with gain vs. noise/output impedence, Distortion, etc.
3. It fetches me beer on command.
4. It uses FETs..
5. 1,2 and 4 are optional. 3 is mandatory..
:green:
Samuel Groner
Well-known member
That's possible with FET switches, isn't it? The company I worked for managed an approach with relays, but it is hard work and pricy.
Another wish: Skip those 'lytic input caps. So 100 uF WIMA caps, lower-value film with JFET frontend, floating PS, high-voltage design or whatever you like!
Probably incompatible with the "cheap" constraint...
Samuel
Svart
Well-known member
not true, i think cheap is a state of mind. with a proper sandstate design and leaving out hard to source or expensive switches and such i think the actual parts costs could/would be very little.. I'm betting a pair would be much less than 100$.. to me, that is CHEAP. I've just built some SSL9k preamps and so far they have been wonderful. If i didn't have to buy 20$ a piece switches for gain they would have been less than 20$ a channel..
Hi John,
Single ended CFB kicks ass on this end too!
Cheers,
Tamas
Samuel Groner
Well-known member
That's the point I want to make: both remote controlled gain and skipping 'lytic input caps will very likely require expensive "and such".
Samuel
If you spend say $44.17 odd bucks per channel on a pre that should perform on par but that's being modest...should outperform pre's that cost $2357.00 then it's worth it without a doubt. Go to www.tubesandmore.com for Solen
fast caps these should replace the lytics needed.
Analag
fast caps these should replace the lytics needed.
Analag
+/- 28V is still in the realm of easily attainable, so we're still good there.
I'd say if the per channel cost could come in less than $100, we're on the high end, but still doing good.
To summarize where we've headed:
A tranformerless mic pre with a gain of roughly 70dB.
The input stage must have an adjustable input impedence ( 600-3K is the range we've loosely settled on, barring no stronger opinions)
The input stage is looking like the BJT cascoding J-Fet's ( do we call this single ended CFB? )
The gain control will be implemented in a way such that output impedence will not be changed (or at least changed drastically?) at different gain settings. This most likely will mean at a minimum two gain stages ( which was likely anyway, given the gain sought)
Minimize or eliminate the use of electrolytics in the signal path.
Other requirements? are there areas we're missing?
Regards
ju
I'd say if the per channel cost could come in less than $100, we're on the high end, but still doing good.
To summarize where we've headed:
A tranformerless mic pre with a gain of roughly 70dB.
The input stage must have an adjustable input impedence ( 600-3K is the range we've loosely settled on, barring no stronger opinions)
The input stage is looking like the BJT cascoding J-Fet's ( do we call this single ended CFB? )
The gain control will be implemented in a way such that output impedence will not be changed (or at least changed drastically?) at different gain settings. This most likely will mean at a minimum two gain stages ( which was likely anyway, given the gain sought)
Minimize or eliminate the use of electrolytics in the signal path.
Other requirements? are there areas we're missing?
Regards
ju
Svart
Well-known member
still looking at using cheap(er) solutions for the gain switching, like normal pot values instead of some reverse log dual deck kind with some crazy value..
do i sound bitter? :green:
do i sound bitter? :green:
magicchord
Well-known member
withdrawn by poster
[quote author="fum"]The input stage is looking like the BJT cascoding J-Fet's ( do we call this single ended CFB? )[/quote]
No, that is a cascode. Global feedback from the output of the amp to the source of the inpput FET is CFB. The Hiraga is fully push-pull. I was referring to John's described circuit.
Tamas
No, that is a cascode. Global feedback from the output of the amp to the source of the inpput FET is CFB. The Hiraga is fully push-pull. I was referring to John's described circuit.
Tamas
chrissugar
Well-known member
In my opinion if you want transformerless and cheap we have enough options like the Green Pre, the SSL9k or the AMEK2500, the Graeme Cohen preamp, Fred's Jfet based double balanced pre and many others.
I think this project should be in the direction of wire with amplification and "money no object" kind of project. When I say "money no object" I mean in resonable limits (no exotic snake oil).
1)variable input impedance
2)enough gain for most of the applications ( 60-70dB )
3)low enough noise to be usable for critical classical music recording
4)no change in sound with amplification
5)imunity to external RF perturbations
6)solid output buffer to drive low impedances (it can be single ended).
7)high voltage PSU for high headroom
chrissugar
I think this project should be in the direction of wire with amplification and "money no object" kind of project. When I say "money no object" I mean in resonable limits (no exotic snake oil).
1)variable input impedance
2)enough gain for most of the applications ( 60-70dB )
3)low enough noise to be usable for critical classical music recording
4)no change in sound with amplification
5)imunity to external RF perturbations
6)solid output buffer to drive low impedances (it can be single ended).
7)high voltage PSU for high headroom
chrissugar
Svart
Well-known member
Would super high headroom really benefit us in such a magic box? If you look at it this way, i would say that most DIYers are using some kind of inferior console and thus the reason for DIY.. soooo stuffing a badass signal into a box that is only going to bottleneck it kinda negates any money that we spent on it.
[quote author="chrissugar"]In my opinion if you want transformerless and cheap we have enough options like the Green Pre, the SSL9k or the AMEK2500, the Graeme Cohen preamp, Fred's Jfet based double balanced pre and many others.[/quote]
Interesting that you called out these because they are all using the same design as the Cohen amplifier, and that is not a variety by definition.
OK, so what sort of budget would you consider expensive enough?
Interesting that you called out these because they are all using the same design as the Cohen amplifier, and that is not a variety by definition.
OK, so what sort of budget would you consider expensive enough?
pstamler
Well-known member
[quote author="chrissugar"]In my opinion if you want transformerless and cheap we have enough options like the Green Pre, the SSL9k or the AMEK2500, the Graeme Cohen preamp, Fred's Jfet based double balanced pre and many others.
I think this project should be in the direction of wire with amplification and "money no object" kind of project. When I say "money no object" I mean in resonable limits (no exotic snake oil).
1)variable input impedance
2)enough gain for most of the applications ( 60-70dB )
3)low enough noise to be usable for critical classical music recording
4)no change in sound with amplification
5)imunity to external RF perturbations
6)solid output buffer to drive low impedances (it can be single ended).
7)high voltage PSU for high headroom[/quote]
I go along with Chris on almost all of these design criteria; my one real exception is with 6). There needs to be a balanced output option as well as unbalanced.
Two more:
8) Should be able to accept peak input level of 0dBu or higher without overload at lowest gain setting.
9) In a harmonic distortion test, no harmonics above the fourth should be detectable.
Peace,
Paul
I think this project should be in the direction of wire with amplification and "money no object" kind of project. When I say "money no object" I mean in resonable limits (no exotic snake oil).
1)variable input impedance
2)enough gain for most of the applications ( 60-70dB )
3)low enough noise to be usable for critical classical music recording
4)no change in sound with amplification
5)imunity to external RF perturbations
6)solid output buffer to drive low impedances (it can be single ended).
7)high voltage PSU for high headroom[/quote]
I go along with Chris on almost all of these design criteria; my one real exception is with 6). There needs to be a balanced output option as well as unbalanced.
Two more:
8) Should be able to accept peak input level of 0dBu or higher without overload at lowest gain setting.
9) In a harmonic distortion test, no harmonics above the fourth should be detectable.
Peace,
Paul
[quote author="tk@halmi"][quote author="fum"]The input stage is looking like the BJT cascoding J-Fet's ( do we call this single ended CFB? )[/quote]
No, that is a cascode. Global feedback from the output of the amp to the source of the inpput FET is CFB. The Hiraga is fully push-pull. I was referring to John's described circuit.
Tamas[/quote]
What I'm trying to put a description on is:
[quote author="Winston O"]
single-ended version that I built and this does not have the same issues with matching. Again, this amp consists of a parallel set of 2SK170BL's (4 in this case for 6dB theoretical noise improvement) with a single low-noise transistor as a cascode - a Fairchild KSC1845. The standing current through this compound stage is hovering around 14mA.[/quote]
There's a second half to the description, that we may or may not need to use in this design, no?
Regards
ju
No, that is a cascode. Global feedback from the output of the amp to the source of the inpput FET is CFB. The Hiraga is fully push-pull. I was referring to John's described circuit.
Tamas[/quote]
What I'm trying to put a description on is:
[quote author="Winston O"]
single-ended version that I built and this does not have the same issues with matching. Again, this amp consists of a parallel set of 2SK170BL's (4 in this case for 6dB theoretical noise improvement) with a single low-noise transistor as a cascode - a Fairchild KSC1845. The standing current through this compound stage is hovering around 14mA.[/quote]
There's a second half to the description, that we may or may not need to use in this design, no?
Regards
ju
Similar threads
