|
Has anyone built a schoeps style circuit with a cascode input?
If so did you feel it was worth it?
I think Josephson may do this? , though i haven't seen a schematic.
I have a mic where I'm going for a very neutral and detailed sound and I
wondered if it was worth a try.
Are there examples out there? I haven't found any examples that I could
just try out - I'd have to learn more about designing a cascode first.
|
My brahma studio 4x is built with 25 mm capsules, each with a cascode circuits mounted on the back, driving 4 scheops style circuits mounted on a single PC.
Umashankar
Sent from my Samsung Galaxy smartphone.
toggle quoted message
Show quoted text
-------- Original message -------- From: thet <parenthetical@...> Date: 29/11/2020 04:54 (GMT+05:30) Subject: [MicBuilders] cascode in schoeps?
Has anyone built a schoeps style circuit with a cascode input? If so did you feel it was worth it? I think Josephson may do this? , though i haven't seen a schematic. I have a mic where I'm going for a very neutral and detailed sound and I wondered if it was worth a try. Are there examples out there? I haven't found any examples that I could just try out - I'd have to learn more about designing a cascode first.
|
I found the photos - that's some nice 3D printing - my printer would be
struggling with those shapes.
Do you have schematics or any more info on the circuit?
Did you feel that the cascode gives a worthwhile improvement over a a
single fet?
What fets did you use for the cascode?
toggle quoted message
Show quoted text
On 29/11/2020 01:19, umashankar via groups.io wrote: My brahma studio 4x is built with 25 mm capsules, each with a cascode
circuits mounted on the back, driving 4 scheops style circuits mounted
on a single PC.
Umashankar
|
A cascode retains the gain of a single FET common-source with the reduced Miller capacitance of a common-source (voltage-follower).
It opens the range of possible compromises. E.g. it allows using 2SK170 by mitigating its high input capacitance and benefit from its very low input noise voltage. Or minimizing capacitive loading of a SDC, thus improving the noise performance.
Ther are some very clever cascode circuits that both bootstrap the gate-source capacitance and minimize the Miller effect.
toggle quoted message
Show quoted text
Le 29/11/2020 à 00:24, thet a écrit?: Has anyone built a schoeps style circuit with a cascode input?
If so did you feel it was worth it?
I think Josephson may do this? , though i haven't seen a schematic.
I have a mic where I'm going for a very neutral and detailed sound and I
wondered if it was worth a try.
Are there examples out there? I haven't found any examples that I could
just try out - I'd have to learn more about designing a cascode first.
--
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
|
I have read that it is also more linear and has a wider frequency
response, though those might be a consequence of some of the things you
mentioned.
What I'm hoping for is that someone has done a cascode phasesplitter for
a schoeps circuit that I can copy to get started.
I'm not sure if the benefits will be audible, so I'm looking for real
world experience.
I don't think I know enough to design one from scratch with any
confidence so I'm looking for an example circuit to try first.
toggle quoted message
Show quoted text
On 29/11/2020 11:48, Jerry Lee Marcel wrote: A cascode retains the gain of a single FET common-source with the
reduced Miller capacitance of a common-source (voltage-follower).
It opens the range of possible compromises. E.g. it allows using
2SK170 by mitigating its high input capacitance and benefit from its
very low input noise voltage. Or minimizing capacitive loading of a
SDC, thus improving the noise performance.
Ther are some very clever cascode circuits that both bootstrap the
gate-source capacitance and minimize the Miller effect.
|
Or, can a OPA mirror the cascode?? Like a OPA145? AND gate treatment?? 1G to ground? 2G?? 5G?? Pimp it?? I have become a sort of Noah's ark for smd jfets...but on 25mm capsule, dead-bug wiring is also easy.? I do simplep48RCA point-to-point.?
How does cascode FETS stack against OPA1642 Alice?? ?...and Winter has not set-in here yet.? ;^)? or cascode in-OPA-ALICE? CASCODE-Sziklai-balanced.
On Sun, Nov 29, 2020 at 7:10 AM, thet <parenthetical@...> wrote: I have read that it is also more linear and has a wider frequency
response, though those might be a consequence of some of the things you
mentioned.
What I'm hoping for is that someone has done a cascode phasesplitter for
a schoeps circuit that I can copy to get started.
I'm not sure if the benefits will be audible, so I'm looking for real
world experience.
I don't think I know enough to design one from scratch with any
confidence so I'm looking for an example circuit to try first.
On 29/11/2020 11:48, Jerry Lee Marcel wrote:
> A cascode retains the gain of a single FET common-source with the
> reduced Miller capacitance of a common-source (voltage-follower).
> It opens the range of possible compromises. E.g. it allows using
> 2SK170 by mitigating its high input capacitance and benefit from its
> very low input noise voltage. Or minimizing capacitive loading of a
> SDC, thus improving the noise performance.
> Ther are some very clever cascode circuits that both bootstrap the
> gate-source capacitance and minimize the Miller effect.
>
>
|
Le 29/11/2020 à 14:10, thet a écrit?: I have read that it is also more linear That is true. the bottom transitor operates at nearly constant Vds, making it an almost perfect voltage-to-current converter. and has a wider frequency
response, though those might be a consequence of some of the things you
mentioned. It is a consequence of minimizing Miller effect. -- L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
|
i will try and find the details. i have a sprained back. The fets are bf861s i think. Two reasons i liked this circuit. the gate resistor is just plain 1 gig and no adjustment was necessary. The other is gain is set by the resistors at drain and source. The gain is unity with that phase splitter arrangement. the images are of the pcb i use, and the tetrahedral array.
umashankar
On Sunday, November 29, 2020, 08:08:37 PM GMT+5:30, ▌ P@T ▌ via groups.io <enjoybiking@...> wrote:
Or, can a OPA mirror the cascode?? Like a OPA145? AND gate treatment?? 1G to ground? 2G?? 5G?? Pimp it?? I have become a sort of Noah's ark for smd jfets...but on 25mm capsule, dead-bug wiring is also easy.? I do simplep48RCA point-to-point.?
How does cascode FETS stack against OPA1642 Alice?? ?...and Winter has not set-in here yet.? ;^)? or cascode in-OPA-ALICE? CASCODE-Sziklai-balanced.
On Sun, Nov 29, 2020 at 7:10 AM, thet <parenthetical@...> wrote: I have read that it is also more linear and has a wider frequency
response, though those might be a consequence of some of the things you
mentioned.
What I'm hoping for is that someone has done a cascode phasesplitter for
a schoeps circuit that I can copy to get started.
I'm not sure if the benefits will be audible, so I'm looking for real
world experience.
I don't think I know enough to design one from scratch with any
confidence so I'm looking for an example circuit to try first.
On 29/11/2020 11:48, Jerry Lee Marcel wrote:
> A cascode retains the gain of a single FET common-source with the
> reduced Miller capacitance of a common-source (voltage-follower).
> It opens the range of possible compromises. E.g. it allows using
> 2SK170 by mitigating its high input capacitance and benefit from its
> very low input noise voltage. Or minimizing capacitive loading of a
> SDC, thus improving the noise performance.
> Ther are some very clever cascode circuits that both bootstrap the
> gate-source capacitance and minimize the Miller effect.
>
>
|
These are great questions. So far with my OPA circuit, I think it is very clean. I love it. One of the things I was worried about was noise. It technically has 4-5db more noise than the Pimped Alice. But that is completely swamped in real work usage. I measured this using the same input on my Scarlett 18i20 with the same gain setting and then superimposed the graphs in photoshop. 47pf Cap replacing the capsule.? (Yea, I know there is a better way to do that!) And when used on a 34mm external?bias capsule with a larger signal to begin with, it is even less of an issue. Green is OPA1642 as I am using it and Yellow is J305 and the two PNP's we normally use. ?
Or, can a OPA mirror the cascode?? Like a OPA145? AND gate treatment?? 1G to ground? 2G?? 5G?? Pimp it??
I have become a sort of Noah's ark for smd jfets...but on 25mm capsule, dead-bug wiring is also easy.? I do simplep48RCA point-to-point.?
How does cascode FETS stack against OPA1642 Alice?? ?...and Winter has not set-in here yet.? ;^)? or cascode in-OPA-ALICE? CASCODE-Sziklai-balanced.
On Sun, Nov 29, 2020 at 7:10 AM, thet I have read that it is also more linear and has a wider frequency
response, though those might be a consequence of some of the things you
mentioned.
What I'm hoping for is that someone has done a cascode phasesplitter for
a schoeps circuit that I can copy to get started.
I'm not sure if the benefits will be audible, so I'm looking for real
world experience.
I don't think I know enough to design one from scratch with any
confidence so I'm looking for an example circuit to try first.
On 29/11/2020 11:48, Jerry Lee Marcel wrote:
> A cascode retains the gain of a single FET common-source with the
> reduced Miller capacitance of a common-source (voltage-follower).
> It opens the range of possible compromises. E.g. it allows using
> 2SK170 by mitigating its high input capacitance and benefit from its
> very low input noise voltage. Or minimizing capacitive loading of a
> SDC, thus improving the noise performance.
> Ther are some very clever cascode circuits that both bootstrap the
> gate-source capacitance and minimize the Miller effect.
>
>
-- Best Regards,
Jules Ryckebusch
214 399 0931
|
Cascode allows voltage gain while minimizing the Miller effect of drain-gate capacitance. But do you want voltage gain in your mic? Condenser mics are pretty hot with purely current gain. Are you aiming at a line-level mic? If so, Schoeps isn’t the output stage you want.
The input / phase splitter of a Schoeps doesn’t have (much) gain, so gate-to-drain capacitance doesn’t limit HF response much below a megahertz. The FETs we use are mostly intended for VHF RF amplifiers, after all. However, gate-to-channel input capacitance does matter in my experience. I’m just wondering if cascode isn’t the solution to a problem which doesn’t exist in a Schoeps. If you try it, let us know how it sounds.
It may be my imagination, but minimizing the wiring between capsule and FET seems to pay off. After all, the input capacitance of a J305 is about equal to a 2 cm insulated wire lying next to a chassis. A long, floppy wire from capsule to circuit board is microphonic as it moves with respect to ground. Eliminating or securing it will pay dividends in reducing stray thumps and thuds. One solution: put the FET, the 1G resistor and the source bias resistors next to the capsule.
Or try Pat’s idea and take the FET and circuit board from a small electret and tack it to your capsule.
|
What capsule are you using here Henry?
Cascode allows voltage gain while minimizing the Miller effect of drain-gate capacitance. But do you want voltage gain in your mic? Condenser mics are pretty hot with purely current gain. Are you aiming at a line-level mic? If so, Schoeps isn’t the output stage you want.
The input / phase splitter of a Schoeps doesn’t have (much) gain, so gate-to-drain capacitance doesn’t limit HF response much below a megahertz. The FETs we use are mostly intended for VHF RF amplifiers, after all. However, gate-to-channel input capacitance does matter in my experience. I’m just wondering if cascode isn’t the solution to a problem which doesn’t exist in a Schoeps. If you try it, let us know how it sounds.
It may be my imagination, but minimizing the wiring between capsule and FET seems to pay off. After all, the input capacitance of a J305 is about equal to a 2 cm insulated wire lying next to a chassis. A long, floppy wire from capsule to circuit board is microphonic as it moves with respect to ground. Eliminating or securing it will pay dividends in reducing stray thumps and thuds. One solution: put the FET, the 1G resistor and the source bias resistors next to the capsule.
Or try Pat’s idea and take the FET and circuit board from a small electret and tack it to your capsule.
-- Best Regards,
Jules Ryckebusch
214 399 0931
|
Henry and Co.,
here is one of my first attempts at the FET up in the head:
^^ where I did not plunk the SPlitter resistors in the head, JUST the PiMP-POT and 1G R and ~SMD {or TH} FET.? Cleary something like a TH 2sk596S-C and a SMD FET would fit and then no POT.? {as in Cascode-config}.?
I found this discussion:
I have yet to use a jFET ASSY out of a PUI 5024 HD series in a mike-head but I know it will work and the jFET with it's pretreated gate and ~80SNR... would? /will-be nice.? No more costly than a boutique LSK FET.? MOST of the interior gold PCB trace in the 5024 is connected to the tiny gate on the tiny jFET., I know I can solder to that.?
I'm gonna bet the resulting SNR is gonna best the TSB2590 by up to 10db, but as Jules would point out:? "can we hear that"?? Dunno.??? SO far, most of my more recent direct solderings to capsules have been the 2sk596S-C-to: brass-looking 25mm electret in SImpleP48RCA 100% NFB wiring scheme.? Some day though, I'll try the Henry -MODS to a otherwise stock Chinese mike and drop the 5204 PCB up top.? I might hold off on the charge-amp MOD or yah it has to go up on the head-PCB harvested from a 5024 HD.? a ~10pF SMD cap, I would use.? To get "upstairs" I would probably use my stock 28Ga. stranded teflon-jacketd wire, or, 30Ga Kynar coated from my wire-wrap days.? I add "perfect capacitor-type" bypass caps to the larger caps Henry transplants in..? and the DC-rail 330R or 470R...??? Or, yah, SimpleP48 variations would work with the PUI 5024? PCB up at the capsule.?
This would be the? "already 25mm" donor if anyone wants to try:
BTW here is the USA fast way to get a 34mm 12 series clone capsule and have the protection of Prime?
and the Telefunken:
Advanced Audio:
I digress.? SO, if I cascode a 2sk596S-C with a FET "above" it?? Does the second FET have to have a certain IDSS rating?? Or just above that of the Q1, aka 2sk596S-C which would be whatever 350uA let's say?? SO could Q2, or second jFET above that with gate grounded, be just about any N-CH jFET? IF it's IDSS is greater than Q1? Asks the fellow who bought 6,000 PMBFJ113,215? jFETS.? Q1 = pre-treated Gate = less complex, low-cost, and, I have them.
I know, 2-3 threads...too much coffee [and I want MORE]
=-=-=-=-
Cascode allows voltage gain while minimizing the Miller effect of drain-gate capacitance. But do you want voltage gain in your mic? Condenser mics are pretty hot with purely current gain. Are you aiming at a line-level mic? If so, Schoeps isn’t the output stage you want.
The input / phase splitter of a Schoeps doesn’t have (much) gain, so gate-to-drain capacitance doesn’t limit HF response much below a megahertz. The FETs we use are mostly intended for VHF RF amplifiers, after all. However, gate-to-channel input capacitance does matter in my experience. I’m just wondering if cascode isn’t the solution to a problem which doesn’t exist in a Schoeps. If you try it, let us know how it sounds.
It may be my imagination, but minimizing the wiring between capsule and FET seems to pay off. After all, the input capacitance of a J305 is about equal to a 2 cm insulated wire lying next to a chassis. A long, floppy wire from capsule to circuit board is microphonic as it moves with respect to ground. Eliminating or securing it will pay dividends in reducing stray thumps and thuds. One solution: put the FET, the 1G resistor and the source bias resistors next to the capsule.
Or try Pat’s idea and take the FET and circuit board from a small electret and tack it to your capsule.
|
well indeed, I'm asking because i'm not sure if there is any benefit.? I
don't want voltage gain.
However josephson seems to think there is a benefit, and uses a cascode
as a phase splitter like a schoeps.
and I was looking at the C414B-ULS schematic, which is way overcomplex
and I don't understand it but the input is actually a setup with 3 fets
and a BJT, two of the fets look rather cascode like - so are these and
other such examples needless complexity? Or did they have a reason?
It may well be needless complexity. I think a lot of engineers feel the
need to justify their existence by designing stuff that is clever rather
than necessary or optimum.
I agree about moving the fet near the capsule - that might well be more
productive and easy to understand improvement. I have some smd fets that
I could try for that though a regular fet might be just as easy.
Since I'd heard about them for years and never built one I thought i'd
try a cascode if I could find an example circuit to start from.
toggle quoted message
Show quoted text
On 29/11/2020 15:33, henryspragens@... wrote: Cascode allows voltage gain while minimizing the Miller effect of
drain-gate capacitance. But do you want voltage gain in your mic?
Condenser mics are pretty hot with purely current gain. Are you aiming
at a line-level mic? If so, Schoeps isn’t the output stage you want.
The input / phase splitter of a Schoeps doesn’t have (much) gain, so
gate-to-drain capacitance doesn’t limit HF response much below a
megahertz. The FETs we use are mostly intended for VHF RF amplifiers,
after all. However, gate-to-channel input capacitance does matter in
my experience. I’m just wondering if cascode isn’t the solution to a
problem which doesn’t exist in a Schoeps. If you try it, let us know
how it sounds.
It may be my imagination, but minimizing the wiring between capsule
and FET seems to pay off. After all, the input capacitance of a J305
is about equal to a 2 cm insulated wire lying next to a chassis. A
long, floppy wire from capsule to circuit board is microphonic as it
moves with respect to ground. Eliminating or securing it will pay
dividends in reducing stray thumps and thuds. One solution: put the
FET, the 1G resistor and the source bias resistors next to the capsule.
Or try Pat’s idea and take the FET and circuit board from a small
electret and tack it to your capsule.
|
|
|
Yea was thinking that. I have two but didn’t want to open mic to look at it hahaha
Best Regards,
Jules Ryckebusch
toggle quoted message
Show quoted text
On Nov 29, 2020, at 14:45, thet <parenthetical@...> wrote:
?I think it is a RK12 from micparts
On 29/11/2020 15:37, Jules Ryckebusch wrote:
What capsule are you using here Henry?
|
Jules, the capsule in the photo is one of a matched pair of RK-12s from Mic-Parts.
It’s maybe my favorite capsule. It’s not a true replica of an old brass C-12, tho it is edge-terminated and the hole drilling pattern is the same. The backplate isn’t a triple-chamber resonant design. It’s a Neumann-style single-chamber resistive damped backplate. It’s not as zippy as some "C-12s", but it has the reach and intimacy that C-12s were famous for. And it’s EQ-able, which ’67 capsules are not as much.
You’ve noticed cheaper edge terminated capsules around. Single-side cardioid-only capsules for $30, sometimes less. Double side are much more expensive because the front and back have to match. How closely they match determines how good the patterns are, and what they sell for. A matched pair of multi-pattern capsules requires getting four matched cardioids. I’m willing to pay Mic-Parts to measure and match for me, and to toss the duds into the big box under the bench.
An example matched pair of RK-12s from Mic-Parts: Note that there are front and back measurements of each capsule, sensitivity is -35.8 and -35.9dB, and frequency response is within a dB, even at the top end. If you apply 1/3 octave smoothing to the curves, these look as good as the curves marketing departments draw. (No offense to the honest mic companies out there. Y’all know the sort of spec sheets I mean.)
|
Yea, I have a pair of matched ones. ?Love em. They were a bit cheaper then lol Best Regards,
Jules Ryckebusch?
toggle quoted message
Show quoted text
On Nov 29, 2020, at 18:27, henryspragens@... wrote:
? Jules, the capsule in the photo is one of a matched pair of RK-12s from Mic-Parts.
It’s maybe my favorite capsule. It’s not a true replica of an old brass C-12, tho it is edge-terminated and the hole drilling pattern is the same. The backplate isn’t a triple-chamber resonant design. It’s a Neumann-style single-chamber resistive damped backplate. It’s not as zippy as some "C-12s", but it has the reach and intimacy that C-12s were famous for. And it’s EQ-able, which ’67 capsules are not as much.
You’ve noticed cheaper edge terminated capsules around. Single-side cardioid-only capsules for $30, sometimes less. Double side are much more expensive because the front and back have to match. How closely they match determines how good the patterns are, and what they sell for. A matched pair of multi-pattern capsules requires getting four matched cardioids. I’m willing to pay Mic-Parts to measure and match for me, and to toss the duds into the big box under the bench.
An example matched pair of RK-12s from Mic-Parts: Note that there are front and back measurements of each capsule, sensitivity is -35.8 and -35.9dB, and frequency response is within a dB, even at the top end. If you apply 1/3 octave smoothing to the curves, these look as good as the curves marketing departments draw. (No offense to the honest mic companies out there. Y’all know the sort of spec sheets I mean.)
|
"and I was looking at the C414B-ULS schematic, which is way overcomplex
and I don't understand it but the input is actually a setup with 3 fets
and a BJT, two of the fets look rather cascode like - so are these and
other such examples needless complexity? Or did they have a reason?”
It looks to me like a gilded source follower. The “extra” 2 FETs and BJT are there to force the desired operating current regardless of FET variation. I suppose that is important for mass production. No trimpot to tweak, no technician to train. I don’t think it was intended to improve response.
|
now that's interesting.
the 414BULS? (and the TLS) is a very flat neutral mic - in fact some
dislike it because it lacks the top bump of a classic C12.
I suspect that the flatness is more in the capsule than the electronics,
but I was wondering if there was an advantage to any of that overly
complex circuit.
toggle quoted message
Show quoted text
On 30/11/2020 18:20, henryspragens@... wrote: "and I was looking at the C414B-ULS schematic, which is way overcomplex
and I don't understand it but the input is actually a setup with 3 fets
and a BJT, two of the fets look rather cascode like - so are these and
other such examples needless complexity? Or did they have a reason?”
It looks to me like a gilded source follower. The “extra” 2 FETs and BJT are there to force the desired operating current regardless of FET variation. I suppose that is important for mass production. No trimpot to tweak, no technician to train. I don’t think it was intended to improve response.
|
Henry- I've always been impressed by your response graphs - both here and on your own site -- most detailed.?
Can I ask how you calibrate your source?? I've found it easy enough to generate a reasonably high quality pink noise source, but I have no way of knowing how linear the response of whichever transducer I use is... That might be a loudspeaker - or perhaps even a headphone transducer?
I can equaliser to a known mic curve to a certain extent (I generally used my new style 'black' Rode NT1 as a reference mic).? but the 'published' response for that is really only a 'marketing' curve.
What do you use as a reference, to ensure you are measuring the response of the mic under test - and not of the transducer ?...
toggle quoted message
Show quoted text
On Mon, Nov 30, 2020 at 12:27 AM, henryspragens@... wrote:
?
An example matched pair of RK-12s from Mic-Parts:
Note that there are front and back measurements of each capsule, sensitivity is -35.8 and -35.9dB, and frequency response is within a dB, even at the top end.
If you apply 1/3 octave smoothing to the curves, these look as good as the curves marketing departments draw. (No offense to the honest mic companies out there. Y’all know the sort of spec sheets I mean.)
?
|
