Preamp bipolar power from unipolar power

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello. :). First, I'm really sorry if this thread is not in the correct section; I wasn't sure if I should post this here or in the power supply for amplifiers section. I hope this is ok.

I'm making an audio processing thingy from an axoloti, which is a neat microcontroller board for digital audio processing. It uses a DC input of 7 to 15 volts. It works a lot better with line level audio input, so I want to add a mic pre to the thingy, inside it's box. This is to avoid the hassle of carrying a separate preamp or a small mixer with the apparatus.

The preamp I'm building is the one in the datasheet of the 1512 IC from the THAT corporation. This chip is very similar to the INA127, from the infamous $5 preamp.

Everything until here seems pretty straight forward to me. Then, I realize that I need a low noise bipolar supply for the preamp to work. I've seen some schematics of +/-15v power sources that feed from 110v AC, but they are kind of big and, in an ideal world, I'd like to be able to power the whole thingy from only one source, preferably from a very common ac/dc multivoltage adapter.

The only option I've found that I, with little knowledge about power sources and converters, understood, is this. It's a dc-dc, unipolar to bipolar converter based on a LT3467, which is a switching regulator. It seems that it should of work, and it's nice that it can take different input voltages. What worries me, though is that this design is made for SMD components, and it might be important, the datasheet of the LT3467 says that it's better to minimize the area of some traces and that the layout of the PCB is very important, because it has a very fast switching speed.

I don't have the stuff necessary to make a PCB for SMD components, with solder mask. I was thinking of soldering the whole thing with through-hole components, using a SOC to DIP adapter for the regulator chip, but now that I read the datasheet I'm not sure it wont add to much noise. Do you think it would work with through-hole components? Also, it says that it provides around 80ma per rail, do you think it's enough current for this kind of preamp?

Would it be easier to build a +/- 15v source and then some kind of circuit to extract an unipolar current from it to power the axoloti? Are there other alternatives?

I want to thank you, very much, for reading this. This newbie to the community really appreciates it. I hope to be able to contribute what I can in the future.
 
Last edited:
Uh, I pick up 32 v 3 amp transformer supplies salvaged from printer/copier/fax machines for about $3 at the charity resale shop. that plus two 5 W 15v zeners, a couple of wattage limiting resistors in series, electrolytic caps in and out, and connector in the box, make a power supply. No nasty switching howl in the RF band to filter out. All the components will solder nicely to a cinch type solder terminal strip, that screws to the case side with a machine screw and elastic lock nut.
Of course, in some countries, resale of used electric components is banned by the athorities. Here my goto shop is Salvation Army resale.
Actually I find for my purposes a +18 single supply gets +- 7.5 v supplies, which gives me plenty of headroom to take in a earphone signal from a radio or CD player, without clipping.
 
Uh, I pick up 32 v 3 amp transformer supplies salvaged from printer/copier/fax machines for about $3 at the charity resale shop. that plus two 5 W 15v zeners, a couple of wattage limiting resistors in series, electrolytic caps in and out, and connector in the box, make a power supply. No nasty switching howl in the RF band to filter out. All the components will solder nicely to a cinch type solder terminal strip, that screws to the case side with a machine screw and elastic lock nut.
Of course, in some countries, resale of used electric components is banned by the athorities. Here my goto shop is Salvation Army resale.
Actually I find for my purposes a +18 single supply gets +- 7.5 v supplies, which gives me plenty of headroom to take in a earphone signal from a radio or CD player, without clipping.

Thank you, indianajo. I really appreciate it.

Yeah, I was actually worried about the noise switching regulators could make; the only reason I got interested in the article about the LT3467 regulator is that it says that it has such a fast switching speed (in the Mhz range) that any noise introduced would be outside of the range that humans can hear. Though that same speed is what makes me worried about the layout and the size of the pcb traces. Truth is I don't really get how and how much the size of a trace affects the inductance and the noise.

I hadn't thought of digging through old machines. Scavenging old stuff sounds actually kind of thrilling. I guess it could make it worth it using the bipolar power supply as the first and primary source of power. Am I right in thinking that is what you are suggesting? I wonder, do you think it would be hard to then extract 15 volts unipolar power from the bipolar source to also feed the axoloti? I imagine it could be as simple as using only two of the tree terminals, is that right?

I suppose +-7 v should be enough to elevate such signals to line. I mean, I do believe you; it makes sense. From what I understand, dynamic mics have even lower signals and need anywhere from 30db to 60db gain to get them to line. I imagine that's why +-15v or even +-18v is usually suggested to power mic preamps.

Thank you, again, really.
 
I'd suggest a Charge Pump solution using a dedicated ic.
The industry 'Standard' 7660 will give IIRC about 20mA and has max 12V input.
alternative devices have greater current and Voltage range.
I'm currently using TC962 (Microchip) - 80mA 18V max.

There is an oscillator frequency associated with these circuits so you do have to ground them correctly to avoid noise but it's really just a question of taking the chip ground to power connector / chassis rather than via the mic amp circuit.

I don't really understand the remark about 15V or 18V rails being used for mic preamps with 60dB gain. The voltage affects the maximum output level and not the gain.
Taking Line Level to be +4Bu on a Sine Wave gives a peak voltage <1.75V so there's no problem getting that. But the extra voltage gives the headroom. 7.5V probably gives max out voltage peak around 6V (depends on circuit) so <15dBu.
Pro Spec kit typically has +21dBu max.

The voltage will also affect circuit performance to some extent - it sets currents in the transistors etc - so that's all part of the design equation.

Handy Calculator:

http://www.analog.com/en/design-center/interactive-design-tools/dbconvert.html
 
Thank you. I'll look into the TC962. I'm not really familiar with switching regulators, but if you say that eliminating their noise is as simple as proper grounding, it sounds like it can be worth to just try it with that or another charge pump.

Do I understand that the trick is to get a trace directly to the chasis/power source ground to both the charge pump and the preamp? I realize this must sound really dumb, and I apologize, but what do you mean by "via the mic amp circuit"? Would it be like taking the ground for the charge pump trough the preamp chip without it being connected to the power source?

From what I understand and what I've been told, please bear in mind that I know that I understand little, the reason such voltage levels are usually recommended in examples of mic preamp designs is that the lower the input level, the more it is affected by noise, or the worst its signal to noise ratio. I suppose that most of the noise should be taken care by common mode rejection, but supply rejection is something apart, and I imagine that an easy way to reduce supply noise is by using higher voltages. Though, maybe this days it is possible with lower voltages. Or maybe I didn't get the supply rejection thing right. I'm sure the reason I was told had something to do with it, though.

I'm don't really know for sure, hehe. My approach to electronics is to tweak others designs so that I can quickly return to programming things for them. Hardware is a strange land for me. That's why I really appreciate your advise, rmaudio, thank you. Also, that calculator is nice, it made somethings clearer for me.
 
Last edited:
Well, I don't know about A/D converters, they may require 10v. Lots of those in pro mixers these days, lots of digital signal processing goes on.
My disco mixer is all analog, taking 50 mv mag phono signals, 2 v line output from CD player signals, and ~5 vac earphone jack signals from an FM radio, mixing them and producing 2 vac out into a 3 m long cable to the power amp. Reason I mix the signals, I don't have to walk around the table and switch the input every time I change source the way I had to do with a PAS2 preamp.
Op amps get weird if the input gets near the rail, there is a chart for that on the datasheet usually. Volt or Volt and a half cushion appears to be adaquate. I had an 18v race car wall transformer, built a +-7.5 v zener regulator; the disco mixer sounds waaay better than it did before with +- 15 v and hummy +-17v transformer mounted right next to the 50x gain op amps.
Cool thing about my plan, you can build it all with a $50 soldering iron in the time it would take to install and learn about a layout program. Plus I don't have to update the layout program when microsoft or apple forces you to update your op system. I don't have one. The soldering iron doesn't get revised. Layout programs are useful for people building more than one system.
If you've got more than 4-6 op amps you may need regulator IC's instead of a simple pass zener diode. You calculate the series resistor for the zener from the power rating. I used 5 watt zeners, so I figured out how much current I wanted in the zener to produce 2.5 watts. Then V=I*R I took the voltage drop from the supply (9) to the zener (7.5) so V-1.5, I is so many milliamps, that tells you what R to use. I used 22 ohm series resistors, two of them to make the voltage drop symmetric and put the middle of the zeners right inthe middle of the supply. Middle of the two zeners is analog ground.
 
Last edited:
Thank you. I'll look into the TC962. I'm not really familiar with switching regulators, but if you say that eliminating their noise is as simple as proper grounding, it sounds like it can be worth to just try it with that or another charge pump.

Do I understand that the trick is to get a trace directly to the chasis/power source ground to both the charge pump and the preamp? I realize this must sound really dumb, and I apologize, but what do you mean by "via the mic amp circuit"? Would it be like taking the ground for the charge pump trough the preamp chip without it being connected to the power source?

From what I understand and what I've been told, please bear in mind that I know that I understand little, the reason such voltage levels are usually recommended in examples of mic preamp designs is that the lower the input level, the more it is affected by noise, or the worst its signal to noise ratio. I suppose that most of the noise should be taken care by common mode rejection, but supply rejection is something apart, and I imagine that an easy way to reduce supply noise is by using higher voltages. Though, maybe this days it is possible with lower voltages. Or maybe I didn't get the supply rejection thing right. I'm sure the reason I was told had something to do with it, though.

I'm don't really know for sure, hehe. My approach to electronics is to tweak others designs so that I can quickly return to programming things for them. Hardware is a strange land for me. That's why I really appreciate your advise, rmaudio, thank you. Also, that calculator is nice, it made somethings clearer for me.

I think from your post after the one above that you're clear about the voltage rails now. Lots of words there but basically it's headroom and some circuits / ics perform better at higher rail voltages.

wrt the Charge Pump ground - just think of it as the ground current from the Charge Pump circuit needs to return to the power source without going through the ground connection for the mic pre.
Detail will depend on your physical circuit.
eg I had a Charge Pump circuit on a separate board but for test lazily took a convenient '0V' from the audio board = a buzz.
Moving the wire from the Charge Pump board to connect directly to chassis = no buzz.
btw while a Charge Pump circuit does indeed have an oscillator a lot of people wouldn't refer to it has a 'Switching' or 'Switched Mode' supply. They are a relatively simple circuit not involving inductors / transformers etc hence simpler to build and 'manage'.
 
These DC-DC convertors are widely available from different manufacturers and are extremely useful in situations like this.

Yeah - I use those type (Murata / XP) on several of my day job boards (non audio). They are pretty good and trouble free. Some years ago they had a reputation for unreliability. 'Newport' were the big name in these til they were taken over by Murata.
For audio work back in the last century we had boards with two of these type to give +/-15V from a single 5V rail. Used the type with internal linear regulator on the output and plenty of LC filtering. Was good enough to run 20 bit Crystal ADC and DAC boards (measured with AP2).
Still - I'd be wary of running them without a dedicated Ground Plane (I generally have 4 layer boards with these).
Be aware of radiated as well as conducted noise so layout critical.
 
For all my line level circuits I recently switched to using DC-DC converters powered by a USB adapter. I use CL(R)C filters and low noise LDO regs after the DC-DC converters. My PSU design is shown here. It saves a lot of space compared to the conventional PSU with dedicated transformer. The final populated board is shown here. In principle, DC-DC converter with other input voltage can be used making it possible to power the PSU from an existing single polarity supply inside of existing equipment.

Regards,
Oleg
 
I want a thank you all very much, it's been very illuminating.

You're right rmaudio: I kind of get the voltage levels thing now. I guess I probably don't need that much headroom, the Axoloti takes a consumer level signal that, from what I understand, at most should be 2v. So if a used a charge pump to generate a -12v supply from the 12v supply, I kind of think it should be enough. Thank you also for the information about grounding; it's not hard for me to put this on a separate board, to make it easier to think about the ground.

I think I lumped the charge pump with the switching regulators because of this article. But I get it know, I think. Charge pumps work with capacitors and the things that most people call switching regulators need inductors. I'm sorry for all this confusion; there is just so much information and terminology for each very specific area that I get easily lost.

The charge pump circuit seems easy enough then. I didn't know about the dc-dc converter packages that Mooly posted about, though. They seem very practical. I kind of get the idea that they are more noisy than a charge pump, is that right?

Dear xrk971, the steep slope filter that you write about, could it be something like a small capacitor and a resistor in paralel? Also, a CRC is two capacitors in parallel and a resistor in series in the middle, right? Could you point me in a direction about how to calculate the size of the resistor? I'm betting on the supply rejection of the amp to be enough in order to not need a super stable supply with some special linear regulator; though, I don't know really.

These seem like very practical options. It might sound silly, but what I like the most is that I can make it with through hole components. It's not only easier for me to build but for other people too; I want to make a tutorial on youtube about the whole thingy I'm making. If the preamp part of the thingy can be made on a perfboard, it is more probable that other people will find it useful and make it instead of just using an external preamp.

Finally, thank you for sharing your design OlegSh, it was illuminating. I didn't know such things as LDO chips existed. From what I understand, you are using them for their ability to stabilize the voltage more than their conversion ability, am I right?

Thank you all, very much. I'm genuinely thankful for your help. Your information continues to help me sort and make more sense of a whole lotta confusing information on the Internet. I really appreciate it.
 
Administrator
Joined 2007
Paid Member
The DC convertor has the potential to be more noisy than a charge pump particularly with regard to radiated interference, although in practice they are well shielded.

The charge pump can be run at relatively low frequency as long as the caps are big enough, however these convertors are usually poor in efficiency and the rails can not support much current demand (thinking here of the 7660 types).

The DC convertor maintains constant output voltage over a wide range of input voltage and loading, the charge pump output voltage increases or decreases with the supply voltage.
 
Finally, thank you for sharing your design OlegSh, it was illuminating. I didn't know such things as LDO chips existed. From what I understand, you are using them for their ability to stabilize the voltage more than their conversion ability, am I right?

Typically DC-DC converters have good but not excellent output voltage regulation, and they have significant noise and ripple up to 100's of mV at the output. LDO regulators help to improve both: voltage stability and noise filtering in the range up to 1MHz. Above 1MHz I rely on the CL(R)C filters right at the DC-DC converter output.
 
Founder of XSA-Labs
Joined 2012
Paid Member
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.