MAX1044 / ICL7660 some questions.

[nguitar12]

MAX1044 / ICL7660 voltage doubler should be very well known here. However I have some problem regrading to the circuit.

1. What type of capacitor is preferred?  So far I am going to use tantalum rather than electrolytic capacitor because of smaller size (SMD). Just asking if a non polarized ceramic cap can even be used (even smaller)?

2. ICL7660 is much cheaper than a MAX1044 just wonder how they actually perform compare to each other? Does MAX1044 worth the extra money?

Thanks in advanced.

[armdnrdy]

Check the data sheets for capacitor type and sizing.

Here's a thread from the search function. http://www.diystompboxes.com/smfforum/index.php?topic=108490.msg988958#msg988958

There are plenty where that came from. You will find that the MAX1044 has limitations such as max voltage. This IC has been used in commercially available stompboxes and has not fared well in some cases. Best to go with a charge pump with a higher max voltage.

[R.G.]

LT1054 offers much better performance; higher input voltage, higher output current.

The ICL7660 has issues with the frequency of oscillation being in the audio range, making whining possible unless the decoupling and wire routing are just right.

The data sheets on charge pumps usually specify aluminum electrolytics. Aluminum NP types are problematic because of the internal construction, although they may be OK. Pedal use is not particularly taxing on a charge pump, usually.

[Transmogrifox]

Ceramic caps should be fine. 

Depending on your application the charge pump could work ok with about any kind of capacitor as long as the losses are relatively low compared to the power demand in the load.  Most likely symptom of an overly lossy capacitor will be lower output voltage than expected.

[Phoenix]

Ceramic caps are a poor choice for bulk decoupling, as they have a high voltage coefficient (capacitance value drops with applied voltage), to the point that many types of MLCC's (multi-layer ceramic capacitors) will be only around 10% of their nominal capacitance rating at their full rated voltage (ie. a 10uF 16V X5R cap may only be 1uF with a full 16V across it). This means that if you want a large capacitance (for bulk decoupling), you will either need to use a much higher voltage rating, or a much larger value, or some combination of both, in order to achieve your required capacitance under operating conditions.
MLCC's are excellent for local decoupling (directly on the pins of an IC), as they have very low ESR and high bandwidth, so they are most often seen in high-speed and digital circuits, or can be used in parallel with bulk decoupling with the heavy lifting done by aluminium or tantalum electrolytics, and the MLCC's improving the high frequency performance.

Electrolytics (both aluminium and tantalum) don't suffer from this problem with voltage coefficient, so they are used when a high value is needed in a small package. Aluminium can be used up to its rated voltage, and even handle brief over-voltage conditions. Tantalum must never exceed its rated voltage or they risk short-circuiting and catching on fire. Most datasheets will recommend de-rating applied voltage by at least 50% for reasonable life expectancy, up to 70% de-rating for higher ambient temperatures. Tantalum does have lower ESR, better high frequency performance, and has better energy density (more capacitance in a smaller package) than aluminium, but is more expensive.

As RG notes, most charge-pump IC's are specified with aluminium electrolytics, so they'll almost always be the easiest, cheapest, and often the best option. If you want improved high frequency performance, you could parallel MLCC's with the aluminium electrolytics. If you're short on space tantalums may be your best bet (so long as you make sure to follow the datasheet regarding voltage de-rating). MLCC's probably won't cut it by themselves.

If you want to do a bit of reading about the voltage coefficient of MLCC's, here's a good starting point.

Good luck.

[samhay]

>2. ICL7660 is much cheaper than a MAX1044...

To save yourself much headache, stay away from the 7660 for pedal work. You want the use the 7660S instead (note the S) as the oscillator frequency of this model is above the audio band and thus your pedal probably won't whine.

The LT0154 is a better choice again, but may not be cheap or easy to find in your corner of the world.

[italianguy63]

>2. ICL7660 is much cheaper than a MAX1044...

To save yourself much headache, stay away from the 7660 for pedal work. You want the use the 7660S instead (note the S) as the oscillator frequency of this model is above the audio band and thus your pedal probably won't whine.

The LT0154 is a better choice again, but may not be cheap or easy to find in your corner of the world.

+1

Yes, LT1054, TC1044S, or ICL7660S are the best choices.

I use the TC1044SEPA from Mouser.  DO NOT use the TC1044SCPA from Tayda-- I KNOW them to be dodgy.

MC

[Mark Hammer]

I'm starting to see more power bricks with 9VAC outputs.  I gather this is  partly a reflection of the choice by some pedal manufacturers to simply provide onboard bipolar rectification and regulation, and sidestep the nuisance, noise-risk, and guesswork associated with charge pumps.

[italianguy63]

I'm starting to see more power bricks with 9VAC outputs.  I gather this is  partly a reflection of the choice by some pedal manufacturers to simply provide onboard bipolar rectification and regulation, and sidestep the nuisance, noise-risk, and guesswork associated with charge pumps.

I think that is just so we can inadvertently plug 9VAC into our 9VDC pedals making them inoperable, requiring us to buy more products. 

[R.G.]

I'm starting to see more power bricks with 9VAC outputs.  I gather this is  partly a reflection of the choice by some pedal manufacturers to simply provide onboard bipolar rectification and regulation, and sidestep the nuisance, noise-risk, and guesswork associated with charge pumps.

You've seen more than one brick with a 9Vac output??   

I actually think that pedal manufacturers go to AC supplies when they need a lot of DC power. I've only seen 9Vac as a power input level in some of the bigger, dsp based pedals.

There is an issue with 9Vac power bricks. They're forbidden by regulatory practice in the USA except on special dispensation, on the grounds that they waste too much energy. Basically, USA energy efficiency standards have made the 60Hz AC plug in transformer illegal. I'm not sure what the pedals that need 9Vac will do in the future.

[Phoenix]

There is an issue with 9Vac power bricks. They're forbidden by regulatory practice in the USA except on special dispensation, on the grounds that they waste too much energy. Basically, USA energy efficiency standards have made the 60Hz AC plug in transformer illegal. I'm not sure what the pedals that need 9Vac will do in the future.

We've got MEPS (minimum energy performance standards) in Australia too, AC-AC wall warts need to be registered, but are exempt from no-load requirements. This has meant mostly the end for really small supplies as transformer losses are proportionally higher in low VA transformers, but they're still legal if they meet efficiency under load.
I did a bit of quick searching around energy.gov and it seems that in the US, AC-AC EPS's (external power supplies) are required to meet efficiency ratings at 0%, 25%, 50%, 75% and 100% of name-plate current, which of course is madness and is simply not possible. That really seems like a misstep by a well-intentioned bureaucracy who didn't properly understand the details (as is so common with bureaucracies).

[Mark Hammer]

I'm starting to see more power bricks with 9VAC outputs.  I gather this is  partly a reflection of the choice by some pedal manufacturers to simply provide onboard bipolar rectification and regulation, and sidestep the nuisance, noise-risk, and guesswork associated with charge pumps.

You've seen more than one brick with a 9Vac output??   

I actually think that pedal manufacturers go to AC supplies when they need a lot of DC power. I've only seen 9Vac as a power input level in some of the bigger, dsp based pedals.

There is an issue with 9Vac power bricks. They're forbidden by regulatory practice in the USA except on special dispensation, on the grounds that they waste too much energy. Basically, USA energy efficiency standards have made the 60Hz AC plug in transformer illegal. I'm not sure what the pedals that need 9Vac will do in the future.

So what do products requiring a bipolar supply do?  My assumption was that 9VAC was expressly for those.

[Transmogrifox]

MLCC's probably won't cut it by themselves.

Don't be discouraged from trying it.  As Phoenix pointed out there are properties of the capacitors one should take into consideration, but these don't represent insurmountable obstacles. 

Now you have all kinds of ideas and advice -- time to order some parts and try stuff.  That's where you begin to see what combination optimizes the size, price, layout, performance stuff. 

By the way you can make voltage multipliers with a CMOS inverter IC and some diodes.  SMT 2-inverter IC's are pretty small and cheap.  Then you can pick your own switching frequency, which may allow you to get capacitor sizes down significantly:
http://www.gyraf.dk/schematics/Voltage_multipliers_with_CMOS_gates.pdf

Just another angle to consider.

[R.G.]

So what do products requiring a bipolar supply do?  My assumption was that 9VAC was expressly for those.

Good question.

As far as I know, all the volume-production pedals with 9Vac (ish) power supplies are high current ones. I'm reasonably sure that there are numbers of one- or two-off boutique pedals that may use a trannie for bipolar, but those folks are going to be in trouble for power supplies, as those are not supposed to be available any longer. AC mains-frequency wall warts cannot comply with the energy efficiency standards, so are basically outlawed for new uses. The regulations do not say "do not use mains-frequency transformers", they just say "external power supplies must meet this list of energy efficiency standards", and the standards are such that wall-frequency magnetizing current is too large.

Pedals that really, really need bipolar are going to be stuck with two-output switchers, or two outputs of an isolated brick supply.

There's likely to be something of a hue and cry in the pedal world as this gets better known. I suspect that almost no one has noticed yet, and may not until their power supply shipments get turned away by customs, or however it eventually gets enforced.

The high-current digital pedals can just go for a dedicated low voltage switcher, as they tended to take the 9Vac, rectify it to DC, then use an on-board switcher to make their internal power for it. Their pedals don't need a very clean DC supply anyway, as they make their own. The problems with these pedals on non-dedicated supplies center on EMI reflected out on the shared DC lines, and how they treat "ground".

The regs came in through a sneaky path. They were the Department of Energy's "Energy Star" ratings stuff, which was pretty innocuous and easy to meet when it started, as well as being voluntary. Over the years, the energy efficiency "guidelines" got tighter, but were still voluntary until the last one, early this year. It was not only quite stringent, enough to disqualify many switching power supplies, but also was no longer voluntary.

The most recent level of efficiency is difficult to meet even for a switching power supply, and requires techniques that make getting a quiet DC output from the switching supply very difficult.

[R.G.]

it seems that AC-AC EPS's (external power supplies) are required to meet efficiency ratings at 0%, 25%, 50%, 75% and 100% of name-plate current, which of course is madness and is simply not possible. That really seems like a misstep by a well-intentioned bureaucracy who didn't properly understand the details (as is so common with bureaucracies).

I hope that is true. In the USA, the opposite is true. The knowledge of what would happen was clear. The desired result was removal of mains-frequency trannies, but the means was cloaked in the language of eco-ness to get it accepted.