I'm building a mic with a condenser capsule. I am planning on using a bootstrapped 6922 as a buffer. The capsule needs to see an extremely high impedance, way more than 10Mohms for full range response.
The 6922 will probably work fine, but are there other tubes with very low grid current? (The 6922 wasn't exactly made for this use).
What other tricks than low heater voltage and high anode voltage should I use to acheive zero (close to) grid current?
The 6922 will probably work fine, but are there other tubes with very low grid current? (The 6922 wasn't exactly made for this use).
What other tricks than low heater voltage and high anode voltage should I use to acheive zero (close to) grid current?
Negative bias is one. Be sure the grid is several volts below the cathode. Otherwise, electrons from the space charge will fall onto the grid, drawing current. The space charge potential starts at about -1V for regular currents (microamps), but since you're concerned about fractional uA, keeping it under -3V would probably be a good idea.
A small cathode, low current density (towards the cutoff region) and moderate grid-cathode spacing (low transconductance) combine to produce a low leakage tube. I would guess regular old 12AX7 and 12AU7 are just fine in this regard; 12AU7 moreso, because it has lower Rp = more reverse bias necessary to set a low plate current. You'll start with a follower, so mu doesn't help you, anyway.
Another problem is positive grid current, due to stray cathode material deposited on the grid. As the grid heats up (from the cathode's radiant heat, from the plate acting like an oven, or even from load current), this material can emit electrons. This is a problem in power tubes, where the grid is likely to run hot, and harsh conditions (that can cause cathode stripping) deposit cathode stuff on the grid. Preamp tubes that haven't been abused shouldn't have any problems with this.
RF, high gain, frame grid, etc. types are probably less suitable because they have tightly spaced grids. Older types with poorer performance are probably better (and 12AU7 is pretty old). But maybe not too old, I'm sure there's a point where things get worse (avoid a poor vacuum).
What you'll do for the circuit is, cathode follower with a constant current sink. This will keep the grid-cathode voltage fairly stable throughout the signal, so you won't have any effect from the low gain or high distortion of the follower. For a 12AU7, you might go with 300V plate, 2mA sink, and with the grid biased at 88V, the cathode should sit around 100V. That's plenty of reverse bias, and enough current to drive another grid (but keep it DC coupled, otherwise you'll disturb the constant current with resistors).
If you need a *really* low leakage input, just skip tubes. There's no shame in that. JFETs alone attain picoamp leakage easily (at least when cold). Small signal MOSFETs will even do femtoamps. There are low-leakage opamps available which have so little leakage, it's not actually possible to measure: mere background radiation makes the air conductive enough to disturb the measurement!
Tim
A small cathode, low current density (towards the cutoff region) and moderate grid-cathode spacing (low transconductance) combine to produce a low leakage tube. I would guess regular old 12AX7 and 12AU7 are just fine in this regard; 12AU7 moreso, because it has lower Rp = more reverse bias necessary to set a low plate current. You'll start with a follower, so mu doesn't help you, anyway.
Another problem is positive grid current, due to stray cathode material deposited on the grid. As the grid heats up (from the cathode's radiant heat, from the plate acting like an oven, or even from load current), this material can emit electrons. This is a problem in power tubes, where the grid is likely to run hot, and harsh conditions (that can cause cathode stripping) deposit cathode stuff on the grid. Preamp tubes that haven't been abused shouldn't have any problems with this.
RF, high gain, frame grid, etc. types are probably less suitable because they have tightly spaced grids. Older types with poorer performance are probably better (and 12AU7 is pretty old). But maybe not too old, I'm sure there's a point where things get worse (avoid a poor vacuum).
What you'll do for the circuit is, cathode follower with a constant current sink. This will keep the grid-cathode voltage fairly stable throughout the signal, so you won't have any effect from the low gain or high distortion of the follower. For a 12AU7, you might go with 300V plate, 2mA sink, and with the grid biased at 88V, the cathode should sit around 100V. That's plenty of reverse bias, and enough current to drive another grid (but keep it DC coupled, otherwise you'll disturb the constant current with resistors).
If you need a *really* low leakage input, just skip tubes. There's no shame in that. JFETs alone attain picoamp leakage easily (at least when cold). Small signal MOSFETs will even do femtoamps. There are low-leakage opamps available which have so little leakage, it's not actually possible to measure: mere background radiation makes the air conductive enough to disturb the measurement!
Tim
Right. It is what I am doing in my condenser mics with tube inputs. Tubes are 6N2P (one triode for each capsule in a stereo mike). Bias resistors I use are 18 Megaohm only. Capsules are Chinese copies of Neumann ones.
How about an EF86? I believe this is what most commercial tube mic folks are using these days. Three of them, triode connected in parallel are nearly identical to a VF14 which was used in the Telefunken U47. You could, of course, try to get a NOS VF14 which costs about as much as a new kidney.
Cheers!
Cheers!
Excellent. And you heard them in PA during a home concert. I mean previous mono version with FET ins.
http://wavebourn.com/music/alisovskiy/SebeNaBis.mp3
Let me reveal a secret source of a golden fish: 6J1P
Cheap as a dirt, and widely available.
6922 is very popular for tube mics.
When you say buffer, do you mean unity gain cathode follower? Are you going to use a transformer?
There are many suitable, readily available tubes for condenser mics. Here are a few examples from my file of tube mic schematics:
6072 - used 1 section only in the venerable AKG C12
6922 - used in Rode condenser mics and others
EF86 - triode wired, used in the Neumann U67, Brauner
EC92 - used in East German Neumann (Gefell)
EF732 (5840) triode wired - used in a lot of modern mics, Royer, Brauner
7586 (Cathode Follower) - AKG C61 etc. pencil condensers, C12A (in 414-like body)
6111WA (CF) modern Neumann tube mics
These are all used with grid resistors in the range of 50M ohms to 400 M ohms, depending on the capsule, plus fixed bias applied to the cathode,often derived from the DC heater supply. The tubes are selected for low leakage, so best to have a few on hand. Royer is the only circuit I know of that uses leak bias (cathode grounded), The rest have fixed bias and ca. 150M grid resistors.
Below-spec heater voltage is not something I've seen in tube mics, but I think the EC92 is run at ~5.8V in the Gefells. I'll need to measure one to be sure.
Anode voltages are generally under 100V (6072 at 60V is common). B+ ranges from 120V to 230V and there is usually a ca 100K load resistor. The tubes are run at sub-1mA current to deal with grid leakage.
The anode is cap coupled to a parafeed step-down transformer at about 6:1 to 10:1 ratio to get about 200 ohms effective output impedance, balanced output.
This is pretty consistently the tried and true formula for tube condenser mics. There is a lot of DIY mic activity on the Prodigy Pro board.
Oh yeah, I think the capsules from Peluso are basically a selected and possibly slightly tweaked chinese capsule that are about $100 and work OK. I also have been lucky enough to pick up some vintage Neumann and AKG capsules on Ebay from time to time.
Cheers,
Michael
When you say buffer, do you mean unity gain cathode follower? Are you going to use a transformer?
There are many suitable, readily available tubes for condenser mics. Here are a few examples from my file of tube mic schematics:
6072 - used 1 section only in the venerable AKG C12
6922 - used in Rode condenser mics and others
EF86 - triode wired, used in the Neumann U67, Brauner
EC92 - used in East German Neumann (Gefell)
EF732 (5840) triode wired - used in a lot of modern mics, Royer, Brauner
7586 (Cathode Follower) - AKG C61 etc. pencil condensers, C12A (in 414-like body)
6111WA (CF) modern Neumann tube mics
These are all used with grid resistors in the range of 50M ohms to 400 M ohms, depending on the capsule, plus fixed bias applied to the cathode,often derived from the DC heater supply. The tubes are selected for low leakage, so best to have a few on hand. Royer is the only circuit I know of that uses leak bias (cathode grounded), The rest have fixed bias and ca. 150M grid resistors.
Below-spec heater voltage is not something I've seen in tube mics, but I think the EC92 is run at ~5.8V in the Gefells. I'll need to measure one to be sure.
Anode voltages are generally under 100V (6072 at 60V is common). B+ ranges from 120V to 230V and there is usually a ca 100K load resistor. The tubes are run at sub-1mA current to deal with grid leakage.
The anode is cap coupled to a parafeed step-down transformer at about 6:1 to 10:1 ratio to get about 200 ohms effective output impedance, balanced output.
This is pretty consistently the tried and true formula for tube condenser mics. There is a lot of DIY mic activity on the Prodigy Pro board.
Oh yeah, I think the capsules from Peluso are basically a selected and possibly slightly tweaked chinese capsule that are about $100 and work OK. I also have been lucky enough to pick up some vintage Neumann and AKG capsules on Ebay from time to time.
Cheers,
Michael
Last edited:
Thanks for all the good answers
The reason for using the 6922 is that I have over a hundred laying around, or actually in a box, but still, no matter where they are, one is going to be in a mic...
I was thinking 2mA coming from 125volts. Now I will try 1mA instead. I am considering a CCS, but that requires a negative supply as well. Besides the whole point of a tube mic is to get tube sound, and alternative 1 is better for that.
Shown are 2Meg resistors, b/c I have those in my favorite nonmagnetic Vishays, but I can order higher value resistors. The bootstrapping should provide enough high impedance, and the low value resistance will help with noise and prevent problems if the grid starts leaking. (Not that I worry about noise in a CF, since it's not amplifying anything, but still it's something I can tell those who don't know that).
I've been thinking about something like this:
The reason for using the 6922 is that I have over a hundred laying around, or actually in a box, but still, no matter where they are, one is going to be in a mic...I was thinking 2mA coming from 125volts. Now I will try 1mA instead. I am considering a CCS, but that requires a negative supply as well. Besides the whole point of a tube mic is to get tube sound, and alternative 1 is better for that.
Shown are 2Meg resistors, b/c I have those in my favorite nonmagnetic Vishays, but I can order higher value resistors. The bootstrapping should provide enough high impedance, and the low value resistance will help with noise and prevent problems if the grid starts leaking. (Not that I worry about noise in a CF, since it's not amplifying anything, but still it's something I can tell those who don't know that).
I've been thinking about something like this:
Attachments
Good idea, a new niche...Will look up some of those expencive high ohm resistors for medical/avionic equipment.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.