Using 2 MC34063 for CMOY virtual ground

[KungFuPanda]

Here's an interesting idea to use a cheap (~40 cents a piece) switching voltage regulator such as MC34063 for CMOY power supply. Main advantage, in addition to super low cost is it's ability to use a wide range of input voltages from 3v to 40v, which means that it can be easily setup to power CMOY amp from a single 3.6V LiIon battery or 3 AAA batteries.

For the background - Cmoy amp requires a dual power supply (+,-,ground). The most basic design uses resistors divider to create virtual ground from a single 9v battery. That design, while simple and cheap, has many shortcomings, such as uneven virtual ground split, etc.

More elaborate designs using virtual ground circuits are described here:
Virtual Ground Circuits
Downsides with those designs are that they use somewhat expensive chips such as TLE2426 and BUF634 that cost about as much as the opamp itself (~$5 each), and they still require relatively high input voltage (at least 9V, or possibly as much as 18V for some high-impedance headphones).

An interesting alternative is to use an inexpensive switching voltage regulator such as
MC34063. A single MC34063 can be used in inverting supply arrangement to give -9V from a 9v battery thus giving -9V,0,+9V which is much superior to -4.5V,0,+4.5V that TLE2426 circuit could provide.

More interestingly, 2 MC34063 can be used in dc boost and negative voltage dc boost arrangement to power CMOY from a single LiIon ion or LiPo battery. You can easily salvage old LiIon and LiPo batteries from old laptops and cellphones.

So going off about 3.6V nominal voltage of a LiIon battery, first MC34063 could be setup to provide -10V and the second MC34063 +10V with the ground tied to battery's negative terminal.

Even if you didn't want to use a LiIon or 3 AAA batteries, there is still significant cost incentive to use MC34063 in comparison to TLE2426 or BUF634 based virtual ground circuits even in a typical 9V battery arrangement. MC34063 in package of 10 sells on Ebay for less than $4 including shipping, that comes to about ~37c each delivered. A package of 5 retails for $2.29 including shipping and handling, that also works out to 46c each, which is a lot less than $5 a piece for a virtual ground chip.

Incedentaly, MC34063 is found in many 12V car adapter phone power supplies and can be salvaged out of them for free if you have some laying around. Here Dave from Eevblog goes in to details on their usage in phone power supplies and how to put a basic power supply circuit together:
http://www.youtube.com/watch?v=qGp82xhybs4&feature=player_embedded

A simple online calculator/design tool to find values of a few resistors, capacitor and inductor needed for MC34063 based circuit for the given input voltage and output voltage can be found here:
MC34063A design tool
And a more elaborate design tool is here that allows you to play with some more esoteric settings:
MC34063 Step Up/Down/Inverting Switching Regulator - Calculator

I already got 5 mc34063 chips on-hand along with the other components so I'll be putting it together in the next few days. I looked around to see if anybody else built such circuit before, but could not find anything, which is surprising, so I'll welcome your feedback and ideas before I go ahead and put it together.

Peace
KungFuPanda

 

[!]

The TLE railsplitter and BUF chips do not require 9V to 18V, it's the opamps themselves that need such a high voltage swing when driving high z cans. TLE2426 minimum is 4V. BUF634 is 2.25V.

Opamps have decreasing PSRR with ripple frequency increase. Having +-4.5V rails from a railsplitter is superior to having +-9V rails from an inverting regulator. Voltage boosting is another thing to be avoided unless you don't care about the audio quality and if you don't, why build the headamp at all?

Salvaging old laptop and cellphone batteries seems undesirable to me. They are worn down, you'll be putting a lot of work into a project with short life batteries, potentially dangerous as they are degraded, and still need to implement a charging circuit.

When building only a few or especially only one headamp, where is the cost savings relative to time spent? MC34063 isn't an entire switching supply circuit, it's just the chip to which you add capacitor(s), inductor(s), resistor(s) and testing. You need a larger PCB, possibly larger amp case, the charging circuit too. A lot of time spent for what may or may not be cheaper and seems a worse result.

If you don't care that there's switching going on, instead of using a switching power supply for an analog A/B amp chip, skip all of that and just use a class D amp chip. The result will be better, with fewer parts, lower current, better sound.

I digress. The point of a CMOY is a simple beginner circuit to learn about opamps, but in the end people build more elaborate amps because they sound better, not putting effort into making a CMOY sound worse instead.

Electrically what you aim to do seems like it would work, but to me it misses the objectives of hi-fi sound and with not much reason, generally if it needs to be portable it's not used with very high impedance cans so a single 9V battery will work with all but a few picky opamps that need a little more voltage. If not portable then an AC-DC adapter capable of more than 9V is easily obtained or built.

Sorry if my reply seems very negative, it is only one opinion...

 

[sgrossklass]

While some commercial portable amps do use step-up converters, those need to be competently designed in order not to constitute a major EMI problem. Getting such a circuit to work is one thing, not causing major interference (sometimes up to the FM broadcast band) is another.

Ready-made DC/DC converter modules do exist, but they obviously aren't quite as cheap.

 

[KungFuPanda]

While some commercial portable amps do use step-up converters, those need to be competently designed in order not to constitute a major EMI problem. Getting such a circuit to work is one thing, not causing major interference (sometimes up to the FM broadcast band) is another.

Ready-made DC/DC converter modules do exist, but they obviously aren't quite as cheap.

I should have mentioned that I already have my first CMOY amp working with a commercial DC-DC boost converter. It's VERY cheap, readily available and adds NO noise. However I am looking in to using MC34063 based circuit because it's even cheaper (40 cents vs $5) and can also provide virtual ground.

The first CMOY I've made used resistor divider virtual ground and 2 li-ion batteries for power. Fully charged they provided ~8.2volt. This was sufficient for lower volume, but looking at the signal with oscilloscope showed that the signal was clipping at normal volume. Keep in mind that this CMOY was built specifically to drive high-impedance 150Ohm headsets that need a higher voltage signal. For comparison, typical headsets are only 32 Ohm or less.

I could have just added batteries more batteries, but it would have been ridiculous to have to make a 4 battery pack just to avoid clipping, so instead I bought the cheapest DC-DC boost converter I could find on ebay that provided high voltage gain 25v and could run from a single Li-Ion battery (i.e. minimum 3V).

Here's an example, just the first one that popped up:
DC-DC Step up Boost Adjustable Power Supply Module 3~24V to 5~25V Converter | eBay

Don't remember if I got it from the same seller or not, but the one I'm currently using looks the same and was about the same price. With the headset I'm using there is NO noise and I haven't done anything special for filtering or voltage regulation.

Now, here's a good explanation about switching mode power supplies and Op-Amp Power Supply Rejection Ratio (PSRR) ratings and how to calculate noise based on you power supply ripple voltage and switching frequency:
Op-Amp Power Supply Quality Considerations

The example from the page above showed that in one case power supply noise ended up at -121 dB below below full scale, which is clearly inaudible, while for another power supply it came out to -82db and that could warrant some additional filtering or power regulation with something like LM317. Again, with power supply I got for $5 on Ebay, there was no noise, but with other power supplies it could potentially be an issue. I haven't had the chance to work out numbers for MC34063 to see if it'll need additional filtering, so if anyone worked out the numbers before, please share!

 

[KungFuPanda]

The TLE railsplitter and BUF chips do not require 9V to 18V, it's the opamps themselves that need such a high voltage swing when driving high z cans. TLE2426 minimum is 4V. BUF634 is 2.25V.

Opamps have decreasing PSRR with ripple frequency increase. Having +-4.5V rails from a railsplitter is superior to having +-9V rails from an inverting regulator. Voltage boosting is another thing to be avoided unless you don't care about the audio quality and if you don't, why build the headamp at all?

Salvaging old laptop and cellphone batteries seems undesirable to me. They are worn down, you'll be putting a lot of work into a project with short life batteries, potentially dangerous as they are degraded, and still need to implement a charging circuit.

When building only a few or especially only one headamp, where is the cost savings relative to time spent? MC34063 isn't an entire switching supply circuit, it's just the chip to which you add capacitor(s), inductor(s), resistor(s) and testing. You need a larger PCB, possibly larger amp case, the charging circuit too. A lot of time spent for what may or may not be cheaper and seems a worse result.

If you don't care that there's switching going on, instead of using a switching power supply for an analog A/B amp chip, skip all of that and just use a class D amp chip. The result will be better, with fewer parts, lower current, better sound.

I digress. The point of a CMOY is a simple beginner circuit to learn about opamps, but in the end people build more elaborate amps because they sound better, not putting effort into making a CMOY sound worse instead.

Electrically what you aim to do seems like it would work, but to me it misses the objectives of hi-fi sound and with not much reason, generally if it needs to be portable it's not used with very high impedance cans so a single 9V battery will work with all but a few picky opamps that need a little more voltage. If not portable then an AC-DC adapter capable of more than 9V is easily obtained or built.

Sorry if my reply seems very negative, it is only one opinion...

Hey, good points. I've addressed some of them in the reply above, so here I'll just expand and offer a bit more explanation on the overall design goals.

This CMOY amp is used in an adapter that allows the use of General Aviation, Military, Helicopter and Ground-Support headsets such as this:
David Clark H10-30 Headset - MyPilotStore.com
and this:
David Clark H3310 Ground Support Headset - MyPilotStore.com

with any regular cell-phone as an ULTIMATE HANDS-FREE solution for harsh outdoor conditions and HIGH-NOISE environments.

It's a big topic and we should really do another thread on it, so here I'll just briefly list major benefits:

A) High-quality Noise Canceling Mic with boom mount or [FONT=Verdana, Arial, Helvetica]Noise Shielded Mic Muff[/FONT]
B) Isolating headphones with high (22-30dB) noise reduction
A) Rugged design
B) Weatherproof
E) Cheap - $20-$60 for used and NOS (new old stock) units on ebay

Over the years I've tested and used many headsets. Majority of those made for cell-phones are absolute junk, even those that claim to have noise-cancellation and noise blocking design. Cell phone and general purpose headsets are mostly flimsy, infective, unreliable, hard to hear, overpowered by outside noise, fall out easily and get in the way. There are some better headsets available, such as those made for trucking industry, that I've been using somewhat successfully, but even they still fall short on many counts. After building my first Aviation Headset adapter and using a prototype over the last few months I can attest that GA style or Ground Support headsets are THE BEST for any kind of high-noise or harsh environment, provided you can find a they to tie them in with your cellphone (basic requirements are an amp, mic adapter, and preferably an ability to monitor outgoing audio).

---

Now, back to the original question:
>>> MC34063 isn't an entire switching supply circuit, it's just the chip to which you add capacitor(s), inductor(s), resistor(s) and testing. You need a larger PCB, possibly larger amp case, the charging circuit too. A lot of time spent for what may or may not be cheaper and seems a worse result.

Components such as capacitors, inductors and resistors needed to complete MC34063 circuit are very cheap and you'd need them for TLE2426 based design as well. And TLE2426 based circuit would also take the board space.

As far as using old laptop and li-ion cellphone batteries - for reference a brand new 9V LSD (Low Self Discharge) NiMh batteries go for about $15-$20 each.e.g.:
Digimax 9v 220mAh Rechargable Ni-MH Batteries, lsd | eBay

They are actually only 8.4v, so for a grand total of 17.2v it seems pretty spendy to shell out $30-$40 dollars just for the batteries considering their paltry capacity of 220mAh.

Now, let's consider the options for mobile power if say you don't want to pay $30-$40 dollars for LSD NiMh or to have to keep buying alkaline batteries.

You can buy a BRAND NEW 1500mAh Li-Ion battery for a mere $2.50 including shipping and handling, e.g.:
Battery For SAMSUNG SGH i8910 OMNIA HD 1500mAh 3.7V Li | eBay

In case of Li-Ion batteries, I say, skip ebay and just recycle what you've got, because after disassembling and testing a couple dozen Li-Ion and Li-Po batteries from old cellphones, laptops and other consumer electronics, even after a few years about ~20% of them still showed tested at 1500-2000mAh storage capacity, while the other 40% will usually have reduced capacity on the order of about 900mAh, and only about 40% are completely dead, which is not bad if you are getting them for free to begin with.

So for the sake of comparison, say you start off with a new 1500mAh LiIon 3.7v battery that is about the same volume as a 8.4v LSD NiMh 220mAh. LiIon batteries have much higher capacity per volume, so that explains the difference. The boost circuit I've referenced earlier is also about the same size as 9v battery, and since we'd need 2, volume for volume it's about equal. As far as weight - Li-Ion battery + the booster will weigh a bit less than 2 9v batteries. To compare capacity after voltage boost, commonly cited efficiency is about 90%, but let's assume the booster we are using is only giving us 80% efficiency, still 1500/8.4/2*3.7*0.8=260mAh - still 20% ahead of NiMh at same volume, lower weight and 1/4 the cost even if buying new LiIon battery and DC-DC boost adapter. For recycled battery and MC34063 based circuit the numbers look even better.

BTW, if anybody is considering charging LiIon batteries - the best solution I've found so far is iMax B6 charger commonly found on ebay for about $25 including shipping:
eBay - New & used electronics, cars, apparel, collectibles, sporting goods & more at low prices

I actually have 3 and use them for everything from 12V Marine Deep Cycle batteries, to NiCad power tools (1 to15 cell), any kind of NiMh, RC LiPo and of course the good old recycled LiIon. It's truly one charger that does them ALL (except Alkaline - it does not do alkaline, I have another charger that does Alkaline, but not very well because afterward they often leak and get battery acid everywhere, but charging Alkaline is another story ).

IMAX b6 charger even has a cycle to revive old NiCad batteries, has delta peak detection for NiMh, and can do full charge/discharge on LiPo and Li-Ion and other battery types to determine remaining capacity. Now, if we add the cost of the charged, it would seem that LiIon is about the same cost, but the charger is obviously useful for many other things and NiMh based design would also need some kind of charger, actually something like Imax B6 would be the most sensible charger for NiMh as well. Since batteries are usually dedicated to the project, while charger is readily usable for many other things, it does not really make sense to factor in the cost of the charger, as long as it's a universal kind that can do many things.

---

>>> If you don't care that there's switching going on, instead of using a switching power supply for an analog A/B amp chip, skip all of that and just use a class D amp chip. The result will be better, with fewer parts, lower current, better sound.

That an interesting possibility that I haven't considered. CMOY was just the first preamp I came across that seemed to fit the bill. Are there any existing well known D-class heaphone preamp circuit designs, as I'm not yet comfortable designing my own circuits from scratch? Any pitfalls to look out for?

Appreciate all the comments!