You can do it a few ways. First, what's the voltage at which damage could occur? For the sake of example, let's say it's 10V peak (20V p-p). You could use some 9.6V Zeners. Or you could use five red LEDs (1.7Vf) in series. In both cases, you want the diodes (or diode strings) in antiparallel pairs to take care of overvoltage in either direction.
At low impedances, the leakage currents will be pretty negligible.
At low impedances, the leakage currents will be pretty negligible.
The interface lists it's max input as 20dBu. I think that is 22 p-p. So therefore I would want a diode/combination of diodes that total 11 volts. Those would go from the output to the power supply rails in antiparallel pairs. Pointing from negative to output and from output to positive supply rail. Right?
Thanks Sy!
Thanks Sy!
I tried adding in protection diodes and found that thy really heat up the PSU regulators. These are the supply rail to input 1N4148 or 2 series 5.1 zeners to each rail from the output. I am guessing this happens because there is no current limiting resistors in series with the power rails.
If that makes sense, should I add a resistor in series with each supply rail on the PSU, or place resistors in series on each channel? Also, if each channel the current would be limited to 50mA at most, meaning about a 0.9W resistor. What rating would be best give that info?
Thanks!
If that makes sense, should I add a resistor in series with each supply rail on the PSU, or place resistors in series on each channel? Also, if each channel the current would be limited to 50mA at most, meaning about a 0.9W resistor. What rating would be best give that info?
Thanks!
Protection diodes for the inputs typically go to the power supply rails and are not Zeners. There are of course exceptions but those don't apply here. Output limiting goes to ground. These diodes might be Zeners as SY suggested. The reason I told you about limiting the input to the PC is that we _don't_ have easy access to the power supply in the PC. What we CAN do is limit the signal output. The Tascam and PC have dramatically different limits. For all practical purposes you can't develop enough signal to hurt the Tascam.
Audio directly from mics can have an amazing dynamic range. By the time we buy a CD most of that has been removed. Example: recording something very quiet and the gain is cranked up. People come in for some reason making a lot of noise. It's possible to suddenly have a very large change in level which could be large enough to damage a comparatively unprotected input like the PC. For a PC I would limit to 4 V p-p to not make the PCs protection diodes conduct and even at that there would small currents as coupling capacitors settle down.
Protection diodes should have no measurable effect on the power supply.
G²
Thanks for your reply G^2!
I am using 1N4148 diodes on the input just like in the INA217 datasheet. I found that with those diodes in place, the regulators of the PSU heated up very quickly. I just tried 330ohm 2w resistors in series with the power supply rails (limits draw to about 53mA) Nothing heats up now. The recording sounds fine also, very good actually.
I think I'll make space in the board for output to ground diodes and use them later if necessary. Thanks for clearing up that they go to ground. I would only be using this with the tascam or a similar unit, never directly into a PC sound card. Question though, do PCI cards designed for recording, like the kind with breakout boxes, have this issue? I would not think but never hurts to know.
Is there anything I can do to make this work safely and and sound good with the instrument input which have a 1M impedance and 3dBV (5.2 dBU and 4v p-p). It would be nice to get an additional 2 nice sounding inputs.
I am using 1N4148 diodes on the input just like in the INA217 datasheet. I found that with those diodes in place, the regulators of the PSU heated up very quickly. I just tried 330ohm 2w resistors in series with the power supply rails (limits draw to about 53mA) Nothing heats up now. The recording sounds fine also, very good actually.
I think I'll make space in the board for output to ground diodes and use them later if necessary. Thanks for clearing up that they go to ground. I would only be using this with the tascam or a similar unit, never directly into a PC sound card. Question though, do PCI cards designed for recording, like the kind with breakout boxes, have this issue? I would not think but never hurts to know.
Is there anything I can do to make this work safely and and sound good with the instrument input which have a 1M impedance and 3dBV (5.2 dBU and 4v p-p). It would be nice to get an additional 2 nice sounding inputs.
Thanks for the reply Sy!
I did not think that should happen either and immediately assumed I had a diode(s) backwards but check several times. Maybe there was a bad diode. I'll redo it tomorrow to see if I can reproduce the issue. For the record, the diodes from negative power rail to input had strip toward input, diodes from input to positive rail had strip to positive rail, just like the datasheet shows. 17.4V rails to ground. Adding 330r in series on each rail to the breadboard stopped the issue. The PSU has no load or current setting resistors on the output. Nothing heats up without the diodes. Shouldn't the whole amp circuit draw only the amount of current it needs, unless there is a short, oscillation, etc?
So I went ahead and made the board and populated the critical parts, leaving out only the pad and diodes. No 48 connected. I used 330R 2w resistors in series on each power rail to the board (feeding both Ina and dc servo chips). Sounds good, nomadic offset, voltages at pins as expected. The ina217 does get slightly warm to the touch, certainly not warmer than 100F (about 38C). This is after 8-10 minutes powered on and recording most of that time at gain 1000. PSU parts cool to the touch as are all other parts of amp. Im guessing this is normal but it never hurts to check.
Thanks!
I did not think that should happen either and immediately assumed I had a diode(s) backwards but check several times. Maybe there was a bad diode. I'll redo it tomorrow to see if I can reproduce the issue. For the record, the diodes from negative power rail to input had strip toward input, diodes from input to positive rail had strip to positive rail, just like the datasheet shows. 17.4V rails to ground. Adding 330r in series on each rail to the breadboard stopped the issue. The PSU has no load or current setting resistors on the output. Nothing heats up without the diodes. Shouldn't the whole amp circuit draw only the amount of current it needs, unless there is a short, oscillation, etc?
So I went ahead and made the board and populated the critical parts, leaving out only the pad and diodes. No 48 connected. I used 330R 2w resistors in series on each power rail to the board (feeding both Ina and dc servo chips). Sounds good, nomadic offset, voltages at pins as expected. The ina217 does get slightly warm to the touch, certainly not warmer than 100F (about 38C). This is after 8-10 minutes powered on and recording most of that time at gain 1000. PSU parts cool to the touch as are all other parts of amp. Im guessing this is normal but it never hurts to check.
Thanks!
Apparently I'm an electronics moron. I was a custom furniture and ambient maker for a long time. I built really, really high end stuff. I specialized in hand tool work, hand cut joinery, dovetails, etc. I was actually very good at what I did. I worked quickly and accurately. However, seem unable to do a simple mic preamp on a breadboard without doing something wrong. I guess I hurry and don't think ahead, but even when I double check a layout I still miss things.
Anyway, I redid a channel layout on the breadboard with the protection diodes and inhale absolutely no problems now. The psu stays cool, phantom stays cool, diodes stay cool, no phantom voltage on the input after the caps.
Guess I'm just a dunder-head.
I got the SSM2019 in the mail today and some THAT1512 are on the way. I realize these are very similar, but the companies each sent me a few free samples so why not. If nothing else, I can make a channel or two for a friend.
Now I'm off to order a few more chanelanof passive parts and select some switches.
Thanks again for all the help. I'll be back with more soon.
Anyway, I redid a channel layout on the breadboard with the protection diodes and inhale absolutely no problems now. The psu stays cool, phantom stays cool, diodes stay cool, no phantom voltage on the input after the caps.
Guess I'm just a dunder-head.
I got the SSM2019 in the mail today and some THAT1512 are on the way. I realize these are very similar, but the companies each sent me a few free samples so why not. If nothing else, I can make a channel or two for a friend.
Now I'm off to order a few more chanelanof passive parts and select some switches.
Thanks again for all the help. I'll be back with more soon.
No, a dunder-head never gets it right. You make mistakes like all of us but you stay with it and don't give up like some of us. Your next project will go better with fewer mistakes. Eventually you'll have a big bag of known good tricks. Then you'll start combining them in unusual ways to do something new. That's when it gets to be serious fun.
Remember, It's more impressive to do a modest project very well than to mess up a big one. I've done both. I know.
G²
Thanks G^2, I appreciate the kind words.
I am thankful to you, Sy, Minion and everyone else who as continued to follow this tread and lend help though my mistakes. I have learned a ton so far, and hope to continue learning more.
Along those lines, I have some follow up questions.
First, switchable phantom power. I've seen some where the 6k81 current limit resistors are left hanging when phantom is off, others where they are taken to ground with ~47K only when phantom is off. What is the proper way of doing this? It would seem that through a 47K to ground would really boost the input impedance of the amp.
To be honest, I can't quite fathom how to interpret the input impedance of the amp. Reading the THAT design notes 140 (http://www.thatcorp.com/datashts/dn140.pdf) Figure 1, they talk about how with or without the pad impedance is still around 2K. For the life of me I can't figure out how to calculate that. If phantom is off, their 6k8 is to ground, as the 1k2 always are. Since those are the in series with each other, but in parallel with the 10R (and 1k1 if pad is on), I get 975R for each input. I assume one adds those to get the overall input impedance seem by the mic, so 1950R. The two inputs are connected by ~250R when pad is on, for a total of about 2k2. But if you drop the pad resistors and leave phantom on, I only get an impedance of 10R per input, of 20 if considered in series. I've tried measuring with my meter but I think the capacitors mess it up., at least I stick the meter between the input and the corresponding input pin of the IC and the impedance steadily climbs until it becomes an open circuit.
So how do I figure the impedance the mic sees?
Thanks!
I am thankful to you, Sy, Minion and everyone else who as continued to follow this tread and lend help though my mistakes. I have learned a ton so far, and hope to continue learning more.
Along those lines, I have some follow up questions.
First, switchable phantom power. I've seen some where the 6k81 current limit resistors are left hanging when phantom is off, others where they are taken to ground with ~47K only when phantom is off. What is the proper way of doing this? It would seem that through a 47K to ground would really boost the input impedance of the amp.
To be honest, I can't quite fathom how to interpret the input impedance of the amp. Reading the THAT design notes 140 (http://www.thatcorp.com/datashts/dn140.pdf) Figure 1, they talk about how with or without the pad impedance is still around 2K. For the life of me I can't figure out how to calculate that. If phantom is off, their 6k8 is to ground, as the 1k2 always are. Since those are the in series with each other, but in parallel with the 10R (and 1k1 if pad is on), I get 975R for each input. I assume one adds those to get the overall input impedance seem by the mic, so 1950R. The two inputs are connected by ~250R when pad is on, for a total of about 2k2. But if you drop the pad resistors and leave phantom on, I only get an impedance of 10R per input, of 20 if considered in series. I've tried measuring with my meter but I think the capacitors mess it up., at least I stick the meter between the input and the corresponding input pin of the IC and the impedance steadily climbs until it becomes an open circuit.
So how do I figure the impedance the mic sees?
Thanks!
I'd also love to know if I can do anything to make this useable with the HiZ inputs of my interface having a 1M input impedance and max input level of 3dBV. Obviously, its only worth it if I get better sound than the interfaces built in mic pre. I've seen some mention of a re-amping box, like a reverse DI box that converts line level to HiZ. Is that a good path to follow?
Thanks!
Thanks!
First, observe that between + input of the chip and ground, there's a 1k2. Likewise, between - input and ground.
Next, observe that there's a 6k8 from each of the mike inputs to ground.
C1 and C2 are essentially shorts at AC.
Now, with the switch in the 0dB position, the 1k series resistors are bypassed and the 267R is out of the circuit. So (ignoring the 10R), the input impedance on each side is 1k2 in parallel with 6k8. That's about 1k, you can calculate it more exactly. Impedance from + to - is, then, about 2k.
Switch to -20dB. Redraw it, ignoring the 10R again for simplicity. Do the exercise the same way.
The input impedance of the next stage is irrelevant as long as it's not too low for the chip to drive. It will hardly notice the 1M you have.
Next, observe that there's a 6k8 from each of the mike inputs to ground.
C1 and C2 are essentially shorts at AC.
Now, with the switch in the 0dB position, the 1k series resistors are bypassed and the 267R is out of the circuit. So (ignoring the 10R), the input impedance on each side is 1k2 in parallel with 6k8. That's about 1k, you can calculate it more exactly. Impedance from + to - is, then, about 2k.
Switch to -20dB. Redraw it, ignoring the 10R again for simplicity. Do the exercise the same way.
The input impedance of the next stage is irrelevant as long as it's not too low for the chip to drive. It will hardly notice the 1M you have.
Thanks Sy. I understand the relationship between the 1k2 and the 6k8 to ground. I initially thought they where in series since they each where connected to ground. As in 6k8 in series with 1k2, that parallel with the resistance in the + and - input lines. I see now that the 6k8 and 1k2 are in parallel with each other, since they are referenced to the input ground.
Or, since the 267R connects bth inputs, does the mic see: 6k8+1k1+267+1k1+6k8 all in sereis, parallel to 1k2+267+1k2 in series?
But what about the voltage level? The preamp puts out much more than the interface instrument input is rated for.
Thanks!
With the pad in place, is this correct: The 1k1 and 267 are in parallel, so about 215R equivelent. That is in series with the 1k2, for 1415R. That is parallel with 6k8, for total of 1170R for total of 2340R. Is that right? The 267R between inputs throws me off.
Or, since the 267R connects bth inputs, does the mic see: 6k8+1k1+267+1k1+6k8 all in sereis, parallel to 1k2+267+1k2 in series?
But what about the voltage level? The preamp puts out much more than the interface instrument input is rated for.
Thanks!
Impedance-wise they still load the signal. They just happen to be at an elevated Voltage. That capacitor to ground on the 48 Volt source is what does it.
G²
So now I'm getting to the case work and I'm in search of some small push button switches for the pad and phantom on each channel. I've searched through mouser and partsexpress and did not find what I wanted.
Ideally, these will be DPDT and SPST push button switches that panel mount though an ~1/4" hole with a nut. I don't want toggle switches because they are to easy to bump on catch with a cord, and I don't like rockers because they are bigger (even the minis) and require a square hole. Heck, while we are at it, cheap would be nice also. These things are expensive! If push button is not an option, then a really small rocker would be great, something smaller than 1/2"x3/4". I somehow suspect I'll end up with micro toggles
Any help would be greatly appreciated. Thanks!
Ideally, these will be DPDT and SPST push button switches that panel mount though an ~1/4" hole with a nut. I don't want toggle switches because they are to easy to bump on catch with a cord, and I don't like rockers because they are bigger (even the minis) and require a square hole. Heck, while we are at it, cheap would be nice also. These things are expensive! If push button is not an option, then a really small rocker would be great, something smaller than 1/2"x3/4". I somehow suspect I'll end up with micro toggles
Any help would be greatly appreciated. Thanks!
I got an email update saying G^2 replied to this thread. Weird!
Anyway, locking toggles have been suggested. Those sound nice and look nice but require lots of scratch. They also seem pretty big. I might try a micro toggle and if it seems to long, inset it a little or shorten the toggle part.
More searching......
Anyway, locking toggles have been suggested. Those sound nice and look nice but require lots of scratch. They also seem pretty big. I might try a micro toggle and if it seems to long, inset it a little or shorten the toggle part.
More searching......
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.