Hi all,
I'm going to design my first mic preamp and I have some tech question for you.
First, I'll use a 1:2.5 or 1:5 input transformer (most probably LL1538XL or Sowter 9045) with a AT4040 mic. I measured today about 40mV RMS as mic out on pin 2/3 so I'll have about 120mV peak (differential). And this only singing not so loud! Explosive sounds will have about 100mV RMS and maybe more.
Now I'm going to choose the right JFET for this application but I have no experience with these devices, so I started to look around to gather some info. I see that a lot of devices have a very "small" cut off voltage in the range of -0.4:-1V and this will not allow me to use a 1:5 transformer with 120mVp (600mVp) input without going into cutoff region or going positive with Vgs at max input. What device could I use which is quite easy to get on the market today and usable for this application? 2N5457?J310? And why is this parameter (Vgs cutoff voltage) so variable? I mean, for example, in the 2SK30 data sheet, I can see that Vgs cutoff can range from -0.4 to -5V...is not easy to design a circuit properly with such a variability.
Thanks to everyone who could help me!
Stefano
I'm going to design my first mic preamp and I have some tech question for you.
First, I'll use a 1:2.5 or 1:5 input transformer (most probably LL1538XL or Sowter 9045) with a AT4040 mic. I measured today about 40mV RMS as mic out on pin 2/3 so I'll have about 120mV peak (differential). And this only singing not so loud! Explosive sounds will have about 100mV RMS and maybe more.
Now I'm going to choose the right JFET for this application but I have no experience with these devices, so I started to look around to gather some info. I see that a lot of devices have a very "small" cut off voltage in the range of -0.4:-1V and this will not allow me to use a 1:5 transformer with 120mVp (600mVp) input without going into cutoff region or going positive with Vgs at max input. What device could I use which is quite easy to get on the market today and usable for this application? 2N5457?J310? And why is this parameter (Vgs cutoff voltage) so variable? I mean, for example, in the 2SK30 data sheet, I can see that Vgs cutoff can range from -0.4 to -5V...is not easy to design a circuit properly with such a variability.
Thanks to everyone who could help me!
Stefano
Stefano
Just abandon that path entirely and go to the Jensen site where you will find the best collection of microphone preamps used in pro audio mixers worldwide. Painless, reliable
and satisfying !!
Just abandon that path entirely and go to the Jensen site where you will find the best collection of microphone preamps used in pro audio mixers worldwide. Painless, reliable
and satisfying !!
Why abandon that path? I don't want to copy trivially, I want to understand designing things and trying to get what I'm looking for! Maybe I'll end up drawing the exact schematic that I can see on the Jensen website, but at least I'd have learnt something...😀
One important concept to realize is that you're generally not going to operate a JFET without any sort of feedback whatsoever, so your transformer output level is not necessarily going to be the Vgs across your JFET. For example, if you used a common source amplifier with some amount of feedback to the source, it will tend to bootstrap the source and raise its voltage level, thus decreasing Vgs.
To drive this point home, look at an op amp with 120dB of open loop gain, wired as a follower. This circuit will handle signals that span the power supply voltages because of feedback that reduces the voltage across the two input terminals. That gain of 1 million doesn't restrict the input to the power supply voltage span divided by 1,000,000!
JFET Vgs cutoff varies a lot because it depends upon the doping level of the device, and that doesn't seem to be controlled so tightly as is the geometry of the device. To make life simpler, JFET manufacturers will sort and grade parts post-manufacture into 2 or 3 different selections, usually coded with a color, such as blue, red or green. So, while the Vgsoff range is still pretty wide within a grading level, it's a bunch lower than for all devices of the same basic type. Also, sorting devices by Vgsoff also sorts them for forward transconductance, so the grading allows you to control the open loop transconductance to some degree by specifying one of the grading variants when you purchase.
One final concept is that if you're using a high output mike with a loud source, you may want to have overall preamp gains as low as 0dB without a pad. For example, a high output condenser placed next to a snare drum pretty much puts out line level, but you want a preamp to power it, remove common mode noise, load it properly, and provide a gain trim, something you still want a preamp to do. So, consider lower ratio transformers, such as a split primary/secondary 1:1 transformer, sometimes also referred to as a repeat coil. These usually have excellent bandwidth, low distortion, and superior balance to transformers with a ratio other than 1:1, since the windings can be multifilar and thus tightly matched and coupled by design. And, most importantly, you won't have excess gain from the transformer that's hard to get rid of later without compromising the noise performance of the completed circuit.
Best of luck!!
To drive this point home, look at an op amp with 120dB of open loop gain, wired as a follower. This circuit will handle signals that span the power supply voltages because of feedback that reduces the voltage across the two input terminals. That gain of 1 million doesn't restrict the input to the power supply voltage span divided by 1,000,000!
JFET Vgs cutoff varies a lot because it depends upon the doping level of the device, and that doesn't seem to be controlled so tightly as is the geometry of the device. To make life simpler, JFET manufacturers will sort and grade parts post-manufacture into 2 or 3 different selections, usually coded with a color, such as blue, red or green. So, while the Vgsoff range is still pretty wide within a grading level, it's a bunch lower than for all devices of the same basic type. Also, sorting devices by Vgsoff also sorts them for forward transconductance, so the grading allows you to control the open loop transconductance to some degree by specifying one of the grading variants when you purchase.
One final concept is that if you're using a high output mike with a loud source, you may want to have overall preamp gains as low as 0dB without a pad. For example, a high output condenser placed next to a snare drum pretty much puts out line level, but you want a preamp to power it, remove common mode noise, load it properly, and provide a gain trim, something you still want a preamp to do. So, consider lower ratio transformers, such as a split primary/secondary 1:1 transformer, sometimes also referred to as a repeat coil. These usually have excellent bandwidth, low distortion, and superior balance to transformers with a ratio other than 1:1, since the windings can be multifilar and thus tightly matched and coupled by design. And, most importantly, you won't have excess gain from the transformer that's hard to get rid of later without compromising the noise performance of the completed circuit.
Best of luck!!
One final idea: If you do require matched Vgs, for example, to make an input differential pair for an amplifier, it's best to choose a monolithic dual JFET. The Linear Systems LSK389 is a good choice for low source impedance amplifiers that require low voltage noise, and for this device, you get a Vgs match of 20mV or better.
While the absolute doping level is still going to vary, two devices diffused next to each other on the same wafer will be very similar to each other, making the difference in Vgs between the two devices very low. Like discrete JFETs, the monolithic LSK389 pairs are graded into A, B and C levels to allow you to choose among the various absolute Vgs and Gfs levels of the pair.
I'm not sure what sort of amplifier design you want to use, but in general, a differential pair as a front end can work well with an additional stage to provide extra gain for feedback. These designs in general allow the front end pair to largely control the characteristics of the completed amp, but still allow for feedback to minimize the inevitable variation of device behavior due to manufacturing variability, time, temperature and other necessary evils of the real world.
While the absolute doping level is still going to vary, two devices diffused next to each other on the same wafer will be very similar to each other, making the difference in Vgs between the two devices very low. Like discrete JFETs, the monolithic LSK389 pairs are graded into A, B and C levels to allow you to choose among the various absolute Vgs and Gfs levels of the pair.
I'm not sure what sort of amplifier design you want to use, but in general, a differential pair as a front end can work well with an additional stage to provide extra gain for feedback. These designs in general allow the front end pair to largely control the characteristics of the completed amp, but still allow for feedback to minimize the inevitable variation of device behavior due to manufacturing variability, time, temperature and other necessary evils of the real world.
Are you against IC chips? If not, you can use either opamps, or instrumentation-amps (instamp) or discrete opamps. You can use something like a low noise opamp or one of the IC mic preamp chips from THAT Corp, put your choice of mic input transformer in front and voila... mic preamp! Of course, need to add a -20dB pad, polarity reverse switch, and 48V phantom powering (via pair of 6K8 resistors). Low noise, lots of gain (up to +66dB). Hang a steel or nickel output transformer (from Jensen, or other companies, wire for 1:2 ratio, and bump up your total gain to +72dB).
The nice thing about mic preamps (unlike the audiophile world where faithful reproduction of a recorded sound is the goal), is you can add as much or as little "flavoring" as you want and it will still be "correct." i.e. because you're the one producing the recorded music, that audiophiles will try to play on their home system.
The nice thing about mic preamps (unlike the audiophile world where faithful reproduction of a recorded sound is the goal), is you can add as much or as little "flavoring" as you want and it will still be "correct." i.e. because you're the one producing the recorded music, that audiophiles will try to play on their home system.
Thanks a lot! I'll need to study more deeply this topic. I'm interested in distortion contribution to "the sound" I'm looking for, so I'll experiment with different circuits...
Ok, got it...
Ok, thanks again! I think I'll look into 1:1 or 1:2 (or something like that) transformer for now. But usually they are designed to be used with low input impedance stages, if I undestood well...
I'm not against IC 🙂 I'd like to experiment starting from simple things on. Op amps are a life saver for a lot of things but I need to deeply understand the different sonic "imprint" related to every design I can try, so to choose the one which is the best for me. I'm only starting this way, maybe I'll end up with a couple opamps and no transformer at all, who knows? 😛
Exactly! And I want to know where this "flavor" comes form, at least for what I can know/try.
Thanks!
- Status
- Not open for further replies.