spyder with variable voltages?

[djsimmonds]

hello,

i really like the spyder a lot (saving money and landfills). now, i have some projects that run off of higher voltages as well, and i would like to build a spyder for higher voltages (which, i assume, i could do by just using a higher rated transformer and a higher rated voltage regulator, like a 78L15). i also imagine that the cap values would need to change. is there a limit to how high the voltage can be in this design, and how would the design need to be changed to accommodate higher voltages? would it be better, for instance, to use two separate transformers for a +/-15v bipolar supply, or one 30v supply?

next, i saw on the spyder page on the geofex site that there is a "dying battery" feature, which really is just a variable voltage feature. how wide can this voltage swing be (ie, could i have a variable voltage between 9 and 24 volts? 18 and 36 volts? or is it only possible to do a few volts?)?

apologies if this was already asked, but i searched the forums and could not find an answer. thanks!

dani

[JKowalski]

Everything you are wanting to do is easy.


Higher voltage, means you need a higher voltage transformer, and the right regulator. Correct. Cap values need to change to accommodate the higher voltages, but the size of the caps don't need to.

There is no limit to how much DC you can get out of your wall, you just need the right transformer and following rectified-ac-to-dc circuit. Conventional three-terminal regulator power supplies are limited to about 40VDC asfar as your standard ones go, but you also have the option of floating them to get higher voltages or using specialty high voltage regulators - however, I doubt you will need higher than 40.

Look at the LM317 for your purposes. It will output any voltage from 1.2 to 37VDC (adjustable with a potentiometer), providing you have at least 40V unregulated DC on the input. You can also give it a lower input voltage and then you will be able to adjust between 1.2V and whatever your input unregulated DC is minus about 2V. For example, if you have 16V unregulated ontheinput of the LM317, you can adjust the output from 1.2V to about 14VDC.

There are non adjustable regulators for the typical voltages of 5v, 6v, 8v, 9v, 10v, 12v, 15v, 18v, and 24v, as far as I know. They are much easier to use, if you don't want an adjustable output, but if you don't mind the extra resistors needed, it can be useful just to buy a bunch of LM317's that will work for every voltage you ever need (inbetween 1.2 and 37, of course)

For bipolar supplies, you use a center tapped transformer. Take a look at this schematic for how that is done:

http://www.aaroncake.net/Circuits/supply3.asp



Of course, doing all of this would require a multitude of transformers - You ain't gonna find another multi-tap spyder transformer for what you want     But I assume you already knew that.

When you are looking for transformers to use, remember that the voltages on them are Vac, not VDC. You have to figure out what VDC the Vac will be after rectification and filtering, and also remember that you typically need about 2-3 volts higher DC on the input of the three terminal regulator then what you want out of it.

[djsimmonds]

thank you for the reply - it is quite helpful. another quick question, just thinking ahead towards some later synthesizer projects that may have higher current consumption - at what level of current draw would i need to have a heat sink for the 317?

[JKowalski]

All your answers to every question you put forth can be found on data sheets - make liberal use of them!

For example, take a look at the preliminary description of the LM317 on it's data sheet. It notes that the LM317 can output 1.5 A maximum. It goes on to note that there are additional regulators that do the same thing as the LM317, but with higher max output currents. LM150 (3A) and LM138 (5A). Now that you have determined which one suits your current needs, you can scroll down on the pertinent data sheet and find the section labeled "Heat Sink Requirements". That section will give you formulas or data for finding the temperature dissipation at certain current draws, and it will also give you the maximum temperate dissipation for the package without a heat sink - if your calculated temperature is higher then the maximum for the package, then you need a heat sink.

For pedals, you typically are fine if you just get a TO-220 package, without a heat sink. There are very few pedals that need more then that. of course, if you can cheaply and easily add a heat sink, why not? I am currently building my own spyder with a hand wound transformer, and I designed it so that all the regulators line up in a row and they can be easily screwed into the enclosure wall itself - letting the enclosure act as the heat sink for all of them. (Keep in mind that the metal part of the TO-220 case is connected to the center pin - which on non adjustable regulators is ground - and if you connect all the metal backs to one point you have a common ground and you negate the advantages of the whole concept of a spyder power supply! Use non-conductive heat sink material between the TO-220 case and the enclosure)

[R.G.]

Learn to read Geofex. See "Power Supplies Basics".

[djsimmonds]

rg,

i've read the power supply basics on your website. in preparing to work on the power supply, i have seen a number of bipolar schematics. in particular, one i've thought about building is http://sound.westhost.com/project44.htm

however, i am also thinking about adapting this design to take advantage of the half-wave doubler technique you discuss in order to get a bipolar supply from a non-center-tapped transformer. would this just involve changing the rectifier portion? you mention you also have to double the cap values. however, in the design i have is using a 4700uF cap on each side of the supply - this seems pretty high to me. is there a reason it needs to be so high? this design http://www.aaroncake.net/Circuits/ uses only 2200uF, though the max voltage is lower.

thanks!

dani

[R.G.]

i am also thinking about adapting this design to take advantage of the half-wave doubler technique you discuss in order to get a bipolar supply from a non-center-tapped transformer. would this just involve changing the rectifier portion? you mention you also have to double the cap values. however, in the design i have is using a 4700uF cap on each side of the supply - this seems pretty high to me. is there a reason it needs to be so high? this design http://www.aaroncake.net/Circuits/ uses only 2200uF, though the max voltage is lower.

Power filter capacitors are buckets of charge. They get filled at the peak of the AC wave which supplies the power, and then dole out current until the next peak comes along.

The pulses which fill them are very brief, a tiny amount of the AC power cycle. For a full wave rectifier on 60-Hz power, the pulses come every 8.3mS, and last for a fraction of a millisecond. For a half wave rectifier, the pulses come every 16.7mS. Between pulses, the capacitor supports 100% of the load current.

So the cap voltage sags as it doles out current. The change in a capacitor's voltage is equal to the current times the length of time it's supported, divided by the capacitance. Which makes sense - the bigger the current, the faster it runs down. The longer the current lasts between pulses, the faster it runs down. The bigger the capacitor, the less it runs down.

Put another way, dV = I*dt/C. "dt" is fixed - that's the time between pulses. I is fixed too - that's the load current. So you get to decide how much the cap will sag between each recharge pulse by choosing the value of the capacitor. This change in voltage, down between pulses and up very quickly when the cap recharges, is called "power supply ripple voltage." Ripple voltage is biggest in unfiltered supplies, where it's 100% of the input DC peak voltage.

How much ripple can you stand? That depends on your circuits. I'm guessing that you really want 60- or 120-cycle ripple in your sound to be zero; you'd live with it if it was 60db down from your signal level, probably, but even that would bother people. Circuits vary in how much they reject ripple. Some opamps have 60+ db of ripple rejection just 'cause they're good. A single ended amplifier like a single transistor, FET or tube with a resistor to the power supply as a load has zero db of rejection - any ripple rides right through. So getting ripple down is a big deal.

Power supply designers generally have a target of ripple being less than 5% in the first filter cap. Getting lower than that requires HUGE caps. It's cheaper to user some kind of second filter or a regulator to wipe off the last 5%.

With that under our belts, we can answer questions.

- is it only a rectifier change to get bipolar from a non-CT transformer?
Yes, rectifiers and another capacitor for the other polarity.

- do you have to double the cap size?
No. But you may want to. Remember that doing the alternate-sides rectifier changes the charging pulses from once every 1/120th second to every 1/60th of a second (8.3mS to 16.7mS). If you don't change anything else, that will cause the ripple voltage to double because the time between pulses has doubled. Doubling cap values at least gets you back to where you were.

- paraphrasing, you've used xxx caps, which seems pretty high to you, so is more needed?
Depends. Mother Nature is a mathematician. She doesn't care what you think is enough, she just makes the Rules and you have to live with them. But She has given us math to figure out how she'll respond when we do whatever it is we are doing. In this case, dV=IdT/C defines the whole thing. She will enforce that Rule no matter what you think ought to happen. So you're free to decide how much delta-V you can stand, and then buy as much capacitance as will keep your I (load current) and dT (recharge interval) to what you've decided you need.