[Possum]

I'm trying to fit a wall wart I rewired to 40V to work with the Apheared #42, which uses a TLE2426 as a virtual-ground driver. I'm regulating the voltage of the wart down to ~18-19V before feeding it into the TLE2426, but I don't know how much current is going into the TLE2426, but I would definitely say it's way too much, since I've fried 3 TLE2426's. I did try to use a BUF634, but that probably also won't work with current that's too high, since it got very hot when I tried its thermal shutdown protection probably activated.

So how can I limit the current provided by this regulated wall wart so I can use the ground drivers I need? Since the #42 doesn't required much current from the power supply, I'd rather not use 2 voltage regulators (e.g. the LM317 and LM337 pair) to make a + and - rail.

 

[ppl]

The most simplest way is a Resistor in series use onms law to calulate the Value in ohms and to find the required Power. I dont Know the Current your circuit uses But just off the Top of my head try about 500 ohms @ 1 watt. should be in the Balpark.

 

[Possum]

Ah, I'm having trouble remember what I learned about voltage and current. I remember that when a power supply is not hooked up to anything and you add a resistor in front of Vout, Vout is still the same voltage (potential) as the other side of the resistor because no current is flowing. But doesn't this change once you hook it up to a circuit because there'll be a voltage drop across the resistor? Should I only worry about the voltage before hooking the power supply up to the circuit?





I'm getting really confused now. I understand V = I * R, with V being the voltage source. What happens to the current I'm already getting from the transformer/regulator portion of the power supply?



P.S. If there's a better way of achieving what I want (roughly 18-24V with only a few hundred mA max), I'm open to suggestions.

 

[Apheared]

You've got something WRONG. Current doesn't work like that... it won't flow unless it's told to. (oh crap, it did exactly what I told it to!)

You could run a TLE2426 off a car battery... technically hundreds of amps waiting to nuke whatever is downstream... but it won't... it's a potential for destruction not assured destruction. It's your load that'll do the assured destruction.

Stick a wrench across the car battery terminals, basically a 1500W 0.1 ohm resistor... and then run. So make sure it's not current that's impossible to deliver to the load, like into a short.

I mean, making a BUF634 *hot*?? cause it to self-protect?? From just an opamp or two and a couple buffers? hehe uhh no. They'll push close to 300mA!

So something's shorted. And a device will happily explode itself trying to satisfy your current requirements. While limiting current isn't a bad idea, it's not gonna fix your other problem.

 

[Possum]

Interesting, I'll have to check for shorts. I'm pretty sure there weren't any though, since I've built and rebuilt the power supply portion on a breadboard a few times when switching from the TLE2426 to BUF634. If I switched out the voltage regulator's outputs with 2 9V batteries, the circuit would work fine. So what kind of short would cause a TLE2426 to glow/flame/crack and die, and a BUF634 to get hot enough to burn my finger? What's interesting is when I plugged in the AC power supply the first time, I would get the correct voltage readings from the outputs of the TLE2426 and BUF634. But when I moved the circuit over to my source and hooked it up there, the TLE's would either fry or wouldn't give the correct output (2V and -17V rails instead of +-19V for example), which the BUF634 also outputted incorrectly. Then after quickly replacing the fried TLE2426 and changing to 2 9V batteries (or in the case of the BUF634, just changing the V+ and V- inputs), I would get a working amp.

 

[Apheared]

Wait, you're tryin to get DUAL 19V rails? 38V is pressin your luck Possum. You're gonna kill both the TLE and BUF... the BUF 36V abs max, and TLEs abs max rating is 40V and that's just too close. Technically it should work, but...

I thought you meant 19V, as in +/-9.5V post-railsplitter; in which case you're safe on V, so something is trying to sink more current than the devices can deal with; 9 times outta 10 it's a short. (or a really low value resistor in the wrong place - but they usually burn in half too)

>>wouldn't give the correct output (2V and -17V rails)

Sounds like a short or your buffers are oscillating hard. But even @ 25mA/each drain, even the TLE wouldn't explode much less the BUF. Something is wrong. It's collapsing, and that's usually from current drain IMLE

When you run it off 2 9Vs, what is the Idrain? Something normal & sane? or 100+mA? They should be the same of course, wall or batts... Well, electronics suck. I mean to say your wallwart could be feeding such a screwy voltage it could be the reason it oscillates.

I still say something is shorted.

 

[Possum]

Woops, my mistake. I miswrote it (in my head I was trying to add up to 19V

). I meant to say +-9.5V. BTW, in my first few tries with the power supply at 40V, I had checked the data sheets to see what the TLE2426, BUF634, AD823, and EL2001's could handle for max voltage, and tried to run the circuit at ~35V just to see if it was better, but it still fried the rail splitters

.









I'll check the current drain of the circuit when I get home. Could a short in the input (from music source) be causing this? The type of wire I was using for prototyping with the input/output jacks and pot was a 4 conductor telephone cable (the kind people buy to route through their walls sometimes) and had a short with the output initially (replaced that cable with separate wires). But I still have the 4 conductor cable at the input.







Must recheck circuit. It seems like no matter how careful I am when putting the circuit together on a protoboard, something else is always causing problems.





About the wall wart: This I bought from Hossfelt Electronics and was rated at 17V @ 1.5 Amps. A voltmeter showed 21V out of it. When hooked up to the initial circuit that used a TLE2426, there was a loud clipping noise. 2 9V's didn't clip. So I opened the wall wart up and found this wiring scheme:





The center tap is grounded and connected to the output ground. The two ends of the transformer output were hooked up to 1 diode each (only half-wave rectified?

) with the diodes connected together to make the V+ output. Two 4700uF capacitors were between the V+ and ground to reduce ripple.

Anyways, I added a diode bridge for full-wave rectification, disconnected the center tap from the output ground, and found the output of this (from the diode bridge) to be ~43V and am now using two 2200uF capacitors between the V+ and ground (the two 4700uF's fried because they're only rated at 35V

frying capacitors = bad smell and some oily liquid spillage). So this is what I have before a LM317 voltage regulator, which in turn outputs ~19V to the rail splitter and amp circuit.

 

[Possum]

No DIY posts on here for too long! Withdrawl!

I tried to measure the current drawn by the amp, but it was EDIT: "im"measureable with precision - I must've blown the fuses on my meter for the 200mA and 20 mA settings, only the 10A worked and that showed 0.01A.

I moved the power supply circuit to a different part of the breadboard and it seems to work fine. I think the breadboard is shorted internally in some places. Phew! Anyway, no more frying TLE's.

 

[Apheared]

Note to self, and Possum, and everyone: Digi-Key carries a good selection of fuses... buy another 25 each for all three current settings on your meters... Might be stupid spending $20 on fuses you might never use, then again... it's really frustrating to have a brain-damaged MM.

 

[ppl]

Radio Shack also Carry's These types of Fuses. I go through the Lower range ones often.

 

[Tomo]

Hello,

Can't you use much a simpler approach with voltage regulators? Noise level won't be that different.

I assume when parts heat up, it is overdriven. I would think driving a virtual ground that is connected to "ground" might make a hard load. I don't find virtual ground drivers to be particularly good when I can build beefy PSUs. May be good when you use batteries ...

Tomo

 

[Possum]

I'd like to give the virtual battery supply (similar in concept to a UPS with devices run off the UPS's batteries which are simultaneously being recharged), mentioned by ppl in a thread on HeadWize, but I can't find any information about it with a web search yet.



Speaking of voltage regulators, would it hurt if I set the output of the regulator, a LM317, to be 20V and use a 18V input? The point here is to be able to use the same power supply circuit with the amp on this 18V input temporarily while searching for a better wall wart/transformer. Would the LM317 be transparent to the circuit, or would it have negative effects since the input voltage isn't ~3V greater than the configured output voltage?

 

[Tomo]

Hello,

If you have read the specs for LM317 it should be very clear that you cannot have 20V output with 18V input.

I think you overestimated the capability of ground drivers. I do not think you can just use it as substitute for voltage regulators. (Personally, I don't see much point in using ground driver for non-portable use.) I do not think you absolutely need to use ground driver when you build virtual battery supply. Also, I probably can build much more accurate constant current source with LM317.

getting transformers is pretty easy. You can get one at Digikey for 10~20 bucks.


Tomo

 

[Possum]

Tomo,

You have misread and misunderstood what I have in my power supply now. I am using BOTH a voltage regulator AND a virtual ground driver. The voltage regulator will regulate a DC wall transformer (currently I have a large, separate 24V power supply that I use with my Szekeres) to ~18-20V. I KNOW if I use an 18V input, I'm not going to get 20V out from the voltage regulator. My QUESTION had to do with the consequences of trying to use an 18V input with a voltage regulator set to regulate at 20V output. Would the voltage regulator just pass the 18V through untouched (hence my usage of the term "transparent" before)? Or would the voltage regulator completely be inoperable and pass something completely off/incorrect?

The virtual battery power supply is a whole other question, an alternative to what I have now. The virtual ground driver in both cases is to supply a ground reference point so I would have the + and - voltages for the AD823 and EL2001 ICs. I'd rather not use a LM317/LM337 pair to make the + and - when I don't really need it. Just keeping it simple: regulate the power supply's voltage and use a virtual ground driver.

 

[Tomo]

Hey,

Ok. I got it. If are using LM317 with input 18V and try to get 20V out of it, LM317 constantly ran out of power. So it will NOT be transparent in such case.

I think once you use positive and negative regulator pair, you will probably don't need virtual ground driver. That is what I meant. If you use the GND driver, grounding gets very tricky.

Personally, when I make PSU, I would like to have my PSU ground stabilized by mother earth. I believe this tactic is very good idea considering the largest electron-reservior accessible for us is Earth. Adding GND driver seems unnecessary when I can tie my virtual GND to Earth.

Of course, if my amp is portable, Earth is not accessible for my amp. So if I were go full-out-aphered (!!), I will definitely add virtual GND drivers. This will be a very good idea.

Tomo

P.S. Lets start a discussion on Virtual Battery Supply!!