MAX1044 pump charge - voltage boost question

[Rectangular]

ok, so I rigged up a voltage boost pump charger (ICL7660 in place, same thing as 1044),  anyway, it works like a charm. right now I've got it up to 26.7 volts. however, I'm using this to power a 15v dc circuit (its got a 15v regulator at the very beginning of the power input) , and all the electrolytics on the board are 25 volts. I wanted to know a safe way of bringing down this 25v the pump charge is making,  to around 20v, before sending it to the circuit , just to be on the safe side.

btw, I originally tried just using 1 voltage boost stage on the pump charge (about 18volts) but it was under powering the effect. I just need 3-5 more volts to get the right levels

heres the original article about pump charging: http://www.geofex.com/circuits/+9_to_33.htm

-rec

[petemoore]

  LM317 Voltage Regulator'll do that.

[Rectangular]

hey pete

the regulator is part of the effect circuit, so replacing it isn't really too practical. the pump charging is on a separate little daughter board,  its just my way of getting this pedal running with a 9v wallwart daisy chain setup (god-lyke's powerall). I just need a way to preemptively produce less than 25. I was wondering if using 4.7uf caps for the second boost stage would do  this? I'm not sure how tied those caps are to the voltage output of the pumpcharge circuit

[dxm1]

1N4114?

[Rectangular]

I'm no good with zeners. I figure something in parallel or series with something else would be able to do this, or a cap change somewhere in the pump charge circuit ?

[R.G.]

Charge pumps of the simple kind (the 7660 is one of these) only work in integral multiples of the input voltage. So you're going to get 18V minus a diode drop, 27V minus two diode drops, 36V minus three diode drops, etc. Just dropping something in in parallel or series is not going to change what the charge pump does. The size of the bucket cap in particular does not change the output voltage.

The simplest something you can use is, as recommended, a zener. Not being good with zeners is a really good reason to learn it now.

You could use a string of silicon diodes. Each diode eats up 0.5 to 0.7V, so four of them is 2.0 to 2.8V. Eight is 4.0 to 5.6V, which is pretty much exactly what you need. That's probably about as cheaply as you can do it.

A zener can knock the voltage off with one part. But it's limited in how much current it can let through without burning up. Even 1N4148 diodes are good for 100ma each, so your pedal is unlikely to burn them out. 1N400x series is good for 1A. But zeners are rated in watts, so you need a zener which can let through as much current as your pedal needs without burning up.

A 5.1V 0.5W zener is very common. The current a 5.1V zener can carry is 0.5W/5.1V or a little less than 100ma. This makes a 5.1V 1/2W zener about equal to a string of 1N4148's. There are 1W and 5W zeners as well.  The only problem you have is that you have to know ahead of time what current your pedal uses.

An even simpler way is to put in a 7824 three terminal regulator. The 7824 will take in anything over 26V and put out 24V, guaranteed to be less than 25V. If the input voltage to it sags to less than 26V, the 7824 will just follow it down, so that 25V in gives 23V out, 22V in gives 20V out, etc. A 7818 might do the same thing and still leave 3V of headroom for the regulator in your pedal. Neither of these require you to know how much current your pedal uses.

There's another way. Use a series resistor from +27 and a 24V zener across the whole rest of the load. The resistor keeps the current from burning up the zener, the zener keeps the voltage from going over 25.


So many ways, so many ways.

[Rectangular]

hey R.G., thanks for the advice

I tried the simple 400X diode thing, and that seems to have brought it down to a safe level (around 24-23v) however, I am very perplexed by what happens when I try and hook this up to the effect circuit. let me try to explain:

when I power the effect, using the 1 boost stage from the pumpcharger (18v) the effect circuit seems to drop it to 15v immediately, this is the measurement I get from the Input pin of the 15v regulator. the Output of the regulator is a couple volts lower, as you'd expect (around 12-13v)

now, when I use two booster stages on the pump charge (26-27v) and hook this up to the effect, the supply pin on the 15v regulator is even lower ! it measures around 12 volts.

so, how is it that a regulator, given more voltage, is dropping it even lower than when I was providing it with 18v ?  is my 15v regulator defective ?  this is very confusing, because I've got a dual voltage power supply I built for synth modules, and it runs off 15v AC, and uses a 7815 and 7915, and I get a perfect -/+15v DC for both of them.   

I should mention that the effect I'm trying to power is the Electro Harmonix Polyphase, which was originally designed to run off AC, with an internal transformer/regulator deal. so maybe its a real power guzzler ? most early EH effects seem to use a lot of power...

any enlightenment, or alternative solutions you can provide would be very appreciated

-rec

[Paul Perry (Frostwave)]

I expect you are trying to suck too much current from the 1044.
They have a considerable impedance. Yes, that polyphase is a guzzler.
I believe (but, check the data sheets!) that you can run 1044s in parallel to double the current. (someone please tell me if I'm wrong!)

[Rectangular]

should I just give up the fantasy of running this pedal off my Powerall and go with a dedicated wall wart ? as this is going to be my gig version of the pedal,  I was really hoping to make it hassle-free and not add the burden of another dedicated power adaptor (the weight adds up)

couldn't I use a large cap, 2200uf or so as a buffer,  to give a reserve, so that the current doesn't drop so dramatically ? are  the large caps what makes my AC synthesizer supply so stable ?

[dxm1]

A zener can knock the voltage off with one part. But it's limited in how much current it can let through without burning up.

I thought the 7660 was limited to 10mA?

[R.G.]

should I just give up the fantasy of running this pedal off my Powerall and go with a dedicated wall wart ?

Not necessarily. Measure how much current the pedal really needs. That will tell you how much you have to supply. If it's more than maybe 40ma, you'd need a lot of 7660/1044s so maybe you need another kind of power.

I was really hoping to make it hassle-free and not add the burden of another dedicated power adaptor (the weight adds up)

Yep, it would be nice.

There are always ways to do things if you're willing to go to the expense and trouble. There may not always be cheap, small, inexpensive ways to do something conveniently, though. Mother Nature makes the rules and expects us to figure them out.

Go measure the current the pedal needs. We can work from there.

I thought the 7660 was limited to 10mA?

It's not limited as I'd use the word. 10ma is about all that you can practically expect to get out of one under normal use conditions. Maxim's app notes say that the unregulated charge pumps like the 7660 and 1044 approximate an 80 ohm internal source impedance, so the current is bounded. But there is no sharp limit like you'd get with a current limited regulator.

[Rectangular]

Hey R.G.

took me a minute to realize I wasn't getting any amp readings, because the probe was plugged into the wrong socket of the DMM 

anyway, this badboy is guzzling at least 44-45 mA ! so I might have to roll with another power solution. I'm tempted to rip out the 7815 and just treat it like a synth module, but that defeats the gig-worthy aspirations I had in mind for it. running 4 or 5 7660's just to get it up to spec isn't a preferred option, that seems like overkill, plus 7660's aren't the cheapest chips, at least locally.

I always used to wonder why EH made all these pedals with AC power cords, it seemed so wasteful and extravagant... little did I know they're more like synth modules with a stomp switch on 'em

[Paul Perry (Frostwave)]

Cheapest & easiest, at the moment, is probably a $5 17.5 V or  higher voltage plugpack from Jameco (or Goodwill - seem to be a rash of 24V DC scanner power supplies at the moment) & then regulate it with a 78L15.
And if you are buying a transformer by mail, here's a tip - whenever you buy a one-off, buy two. It's insurance..

[Rectangular]

hey paul, I've got a bunch of high-voltage wallwarts (scrapped from telephones, scanners, etc), I was just trying to avoid hauling yet another wart around, those powerbars fill up fast

actually, I've got an 14v wart that actually runs at 18v, but that doesn't seem to provide enough power. a lot of these transformers seem to drift pretty far from the labeled spec

if my poly case was a little bigger, I'd just throw a 20-24v AC transformer in there, unfortunately I built the case to only accommodate the PCB + offboard stuff