I *finally* finished that darned stomp box power supply

[amp_surgeon]

Here we go. I will replace the other one in my post with this one.

 

Sorry I've been absent for a couple of days!

 

Okay, obvious stuff first...

 

Looking at the flat side of the 78L09 and using the names on your schematic, the pin on the left is "out", the pin in the middle is "adj", and the pin on the right is "in", correct? Just checking. The higher current TO-220 version of these regulators has the pins in the other order. Some people think the TO-92 version is going to be the same.

 

I vaguely remember the evolution of this project, but I don't have the time to dig up the thread right now. What was the reasoning for the bypass capacitors on the rectifiers? I don't recall seeing anything like this in a simple linear power supply before, and I'm just curious what the caps are doing.

 

Was the resistor added to give the power supply a load, even when it was disconnected? Was this an attempt to address the overvoltage condition? If so, then you may have "popped" the regulators early in the process of testing the circuit.

 

Check out the usage notes on page 10 of the Panasonic datasheet:

 

http://www.datasheetcatalog.org/datasheet/panasonic/SFF00005CEB.pdf

 

The charge on the 10uF cap can discharge back into the output of the regulator when power is removed, damaging the regulator. This isn't normally a problem with an embedded supply, since the load circuit will provide a discharge path for the cap. However, an "open" supply that might not be plugged into a load when you shut it off could be damaged. Panasonic recommends a reverse biased diode between the output and input of the regulator to shunt the discharge path around the regulator.

 

Got another new regulator handy? Try plugging it in (with that 1K resistor in place as a load) and see what kind of voltage you get the first time you turn it on. If you get 9V then you'll want to replace your regulators, and add the bypass diode.

[katillac]

Looks good to me, provided your board matches the schematic. Got everything laid out right, even the little stuff (diode & cap orientation, etc.)?

Yup! I pay really close attention to polarity and orientation. Those darned electrolytic caps don't work very well turned the wrong way!

 

amp surgeon, I'll put what you said/asked in blue:

 

Looking at the flat side of the 78L09 and using the names on your schematic, the pin on the left is "out", the pin in the middle is "adj", and the pin on the right is "in", correct?

 

Yuppers! To be darned sure, the original drawing had the pinouts listed according to view of the 78L09 from both the top and from view of the flat side.

 

I vaguely remember the evolution of this project, but I don't have the time to dig up the thread right now. What was the reasoning for the bypass capacitors on the rectifiers? I don't recall seeing anything like this in a simple linear power supply before, and I'm just curious what the caps are doing.

 

I think I'm going to have to dig that thread up again because I can not remember the purpose of doing that. Other than Ted, the guys I remember having input on the design of this part was you and Norm. I still have a graphic or two I named with Ted as part of the name, as he sent them to me, and neither has those caps, so if it wasn't you, it likely was Norm.

 

Was the resistor added to give the power supply a load, even when it was disconnected? Was this an attempt to address the overvoltage condition? If so, then you may have "popped" the regulators early in the process of testing the circuit.

 

I don't remember that either. Sorry. I thought it was recommended as part of smoothing out the current or voltage, but that could be the reason. I do recall testing the voltage out of a few of the boards both prior to and after adding the ceramic cap and resistor to the "end" and the voltage was the same.

 

Check out the usage notes on page 10 of the Panasonic datasheet:

 

http://www.datasheetcatalog.org/datasheet/panasonic/SFF00005CEB.pdf

 

Aw crap. I remember someone recommending doing that (adding another rectifier diode or diode of some sort) between the out and in pins to protect the IC. There is one part I don't understand:

 

2) Floating of GND pin If a GND pin is made floating in an operating mode, an unstabilized input voltage is outputted. In this case, a thermal protection circuit inside the IC does not normally operate. In this state, if the load is short-circuited or overloaded, it is likely to damage the IC.

 

Got another new regulator handy? Try plugging it in (with that 1K resistor in place as a load) and see what kind of voltage you get the first time you turn it on. If you get 9V then you'll want to replace your regulators, and add the bypass diode.

 

I will try that soon, but I first have to either hook up another transformer with rectifier diodes (basically the same circuit) or I can tap power off my bench power supply (an old computer PSU with two wires connected so it can be used without being connected to a motherboard).

 

One thing occurred to me a moment ago and I'm glad I remembered it just now, and not a few hours from now or I'd have to get up and log in again or it would bother me all night:

 

Why would all of the non-adjustable circuits read exactly the same voltage, in spite of all being powered from separate transformers?

 

Lemme draw up a basic representation of how this entire thing is laid out. I'll spare you guys the actual layout to avoid confusion, but signal paths will be exactly as they are in reality.

[katillac]

Okay, I just finished spending a few hours between temporary soldering and drawing. I still have a way to go on the drawing stuff. The soldering part left me confused and a bit out of sorts, so I thought about it for a bit then went back and did it again, two different ways and using two different multimeters.

 

It gets weird.

 

I wasn't completely sure if you (amp surgeon) meant to put the positive to the input and the resistor on the output leg of the IC and measure there, or to go hot to positive, then ground to the adj leg of the ic, the 1k resistor on the output leg and the other end of it to the ground. So I did it both ways.

 

Both results were the same. Input voltage: 11.6v DC. Output as measured off the output leg of the 78L09 after the resistor was either 2.7v or 4.7v DC. Sorry, I forget which, as I was too upset over it to write it down before I desoldered everything again and put the parts away.

 

I double-checked the datasheet for the IC. It's a KA78L09AZ. Fairchild. Pin 1 output, pin 2 ground, pin 3 input. Or, better known as, with the flat side facing me, output on the left, input on the right, here I am, stuck in the middle with you.

 

It's been a while since a project frustrated me to the point of crying. This has. At this point, I'm ready to redo everything yet again, this time with LM317T regulators and figure out resistance needed for each one to nail it right at 9V out. Or just scrap the {censored}ing thing.

[amp_surgeon]

What happens if you feed the input of the board with 12VDC from a bench power supply? What does the output voltage look like?

 

Also, do you have an oscilloscope? I'm wondering how much ripple there is on the input of the regulator.

[katillac]

Those ratings in my previous post (#28) was from a computer power supply repurposed for use as a bench power supply (the 11.6v was off the 12v rail). I used two different multimeters to measure the power. And, that was done on two different, straight off tape reel, 78L09s.

 

No oscilloscope. I'm just a basic tinkerer with an old card table with fold-up legs for a work "bench."

[katillac]

I did a search and easily found the two threads that went into doing this project.

 

The initial thread on converting AC to DC. That was the time when I thought of AC as some mythical beast to fear. Then I began to understand it a bit more. I still have some fear of it, but a whole lot of respect for both AC and DC.

 

This was the thread that started the power supply part of the project.

 

I am going to put together another circuit like the ones I put into my project, only this time with the diode between output and input on the 78L09, and I may use a smaller capacitor on the input side of the 78L09 (something like under 1000

[katillac]

Boy, do I feel like a total goof tonight! I got out the power supply and two different multimeters to double check the voltage at the DC outputs of the unit. After my previous post, I realised that I was being too trusting of some cheap Chinese-made LCD meter and I should check with my cheap Chinese-made multimeters.

 

I am glad I did that. I hooked up leads to the first output, turned it all on - including the LCD - and compared readings between the LCD and the multimeter. DMM says 9VDC. Um, what? Try the other DMM. Nine bloody volts on the number. No decimal point anything. Flat out 9 {censored}ing volts. On TWO multimeters. I had a nice Craftsman DMM but it crapped out on me earlier this year. The two DMMs are from different manufacturers but probably use some of the same parts. Nonetheless, I feel that using two different DMMs and getting the same readings is more likely the true output.

 

Okay, so that's the good news. Now the bad news is, I have this spiffy, blue-backlit LCD that doesn't display the correct values for my project. What, if anything, can be done to get it to read correctly? Should I try a trimpot, or would that work? If so, I should have a 10k trimpot around here that I can try out and test ALL of the outputs to be sure that the LCD reads the same as those. With both DMMs, just to be sure. It's more work, but to me it's worth it.

 

I feel better now. Whew.

[nos]

Great news! I saw how distressed you were earlier and didn't know how to respond (mainly because I'm a rank amateur myself). I know that feeling and it sucks! Glad to see you're on the right track. I would think adding a little resistance pre-meter might be the ticket. Kind of a hack, but hey, whatever works, right?

[amp_surgeon]

I saw your post last night and thought I'd better sleep on it 'cause it didn't make much sense. I couldn't imagine a likely scenario that would explain what you were seeing. Reading that old thread was entertaining, though. I didn't remember about the incident with the broken meter (and didn't really need to be reminded about my Volvo fiasco), but after reading about it I had a brief thought that maybe the unexplainable readings might be another test equipment malfunction. But then I thought "Nah, that wouldn't happen twice on the same project."

 

Can you give the make and model of the panel DVM, and also tell us how you've got it hooked up?

[katillac]

I saw your post last night and thought I'd better sleep on it 'cause it didn't make much sense. I couldn't imagine a likely scenario that would explain what you were seeing. Reading that old thread was entertaining, though. I didn't remember about the incident with the broken meter (and didn't really need to be reminded about my Volvo fiasco), but after reading about it I had a brief thought that maybe the unexplainable readings might be another test equipment malfunction. But then I thought "Nah, that wouldn't happen twice on the same project." Can you give the make and model of the panel DVM, and also tell us how you've got it hooked up?

I have a knack for not making sense lately

 

1. Here's the LCD on eBay. Same seller I bought from. Note the requirement for separate unit power from reading source power.

 

2. See picture below (pretty, ain't it?). Mains power to transformer. Step down power from transformer to my power board, which changes from AC to DC, then to 9v. From there, to the DC jack, with + and - running in parallel to the LCD panel meter. A 9v battery powers the backlight and digit display.

 

 

 

 

Boom shakka lakka.

 

3. Can I use a trimpot or check out various resistors on the + input to the LCD to get it to read a proper 9v for the 9v boards, instead of 12.5 or whatever it is?

[amp_surgeon]

That's a 200V meter with an accuracy of +/- 0.5%, which comes out to 1V. That's too much slop, even if it's working well.

 

There's already a small trimmer on the meter. What does it do?

 

There are also some solder jumpers. I'm wondering if it wouldn't be possible to change the range to 20V instead of 200V. Can you take a look at the model numbers on that board? In the fuzzy picture it looks kinda like "D69-NS 1576". I want to try to see if I can figure out who made it, and get more technical info on it.

[katillac]

 

That's a 200V meter with an accuracy of +/- 0.5%, which comes out to 1V. That's too much slop, even if it's working well. There's already a small trimmer on the meter. What does it do? There are also some solder jumpers. I'm wondering if it wouldn't be possible to change the range to 20V instead of 200V. Can you take a look at the model numbers on that board? In the fuzzy picture it looks kinda like "D69-NS 1576". I want to try to see if I can figure out who made it, and get more technical info on it.

 

I goofed. This is the LCD I have. It's the 20V version. The first one I bought was the 199.9v one and was defective. He uses sorta generic pictures, so you can't completely go by what is there, but the one linked directly above looks the same, other than the number you are trying to read. Mine does not have that number. The only actual number on mine is Z10891. I'm pretty sure that the picture otherwise is the same as the one I have, including the tiny trimpot.

 

When I read your post, I zipped ito the other room to pull the cover off my project (maybe I should give this sucker a name) and pull the LCD (easily done) to get a closer look at it. That's where I got the number above from (Z10891).

 

If needed, I can try to get a larger picture of the back of the unit. My camera is an old 1.3 megapixel one, but I'm able to take better shots with it than most people get off a 6-10MP one. It has a more limited range for macro mode (down to 4-6" away from object, which is close enough), so I might be able to get a good shot of it. I may try anyway.

 

Thanks for sticking with me on this!

[amp_surgeon]

Ok, I hunted all over the internet yesterday for about an hour, and I couldn't find a source that has a meter identical to that. Of course, I probably only hit about 1% of the available sources. At one point, I did a google image search and just started browsing the results as if it were a catalog.

 

Anyway, I think it may be a variation of the Velleman PMLCD or PMLCDL meter. The differences I can see are that Velleman meter isn't backlit, and it doesn't have any "V" indicator on the display. However, it DOES have many other characteristics of the same meter, including the block of jumper pads marked P0 thru P3. I also noticed the same vendor on eBay is selling other versions that seem to have the same voltage ranges as the Velleman meters.

 

You can see the PMLCDL here:

 

http://www.vellemanusa.com/us/enu/product/view/?id=350725

 

And the brief user guide (in five languages) here:

 

http://www.vellemanusa.com/downloads/1/pmlcdpmlcdlpmledpmled5gbnlfresd.pdf

 

According to the guide, you calibrate it by providing a fixed voltage which is 1/2 full scale (10V in your case) to the measurement input, and adjusting the trimmer until you get the same reading. If you don't have a variable bench power supply then just use the 12V output of your re-purposed PC power supply, measure the voltage as accurately as possible using a DMM you trust, and then adjust the panel meter to give the same measurement.

[katillac]

Did I ever tell you how much I love you, Mr Wattson?

 

I knew you were the dedicated type, but I figured I was one of only a handful of goofs who would spend an hour on the web searching for something for someone. Wow, great stuff!

 

I'm willing to bet that only one or two companies in the vast land of China produce the boards for most of these danged things and companies flip a SMD component here and there and call it their own.

 

The little trim pot did the trick!

 

Using the same multimeter that gave me the most accurate readings the other night, I tested it on a new 9v battery and a few other batteries that I knew the voltage of. Pretty good readings.

 

I pulled the top panel off the power box, hooked up a few test leads, then pulled the LCD up and out so I could hold it while making adjustments. I began by setting one of the adjustable outputs to right at 9v, then flipped the rotary knob over to position #1 to compare values. Both were 9v. Okay, off to the races. I turned on the LCD and, using the little trimpot on the back, made fine adjustments until the LCD read 9V. I tested a few other values, using the adjustable (#8 in this case) output, comparing readings between the panel LCD and my multimeter. Works like a charm!

 

I have a winner!

 

Again, thanks bunches to those who have helped me through all this, in both technical and moral support. I would give names, but with my luck, I'd miss someone and that is not a risk I wish to take because you guys are all special to me. =)

[Slaymoar]

Damn, nice project Kat

[amp_surgeon]

[katillac]

Thanks, guys!

 

I said that I would post a drawing of how the entire project is hooked up. Here it is, minus the switches. This way, if someone wants to do the same thing, they can put switches anywhere they want them. I just realised that the green wire from mains is missing in the drawing. Green wire goes to chassis ground. I hope I did the rest right and where it's understandable for everyone.

 

 

[amp_surgeon]

Hmmm.

 

Are the negative outputs from the individual supplies really all tied together like that? Was there a reason for doing it this way? Was it to simplify the switching to the panel meter?

 

The reason I ask is because one of the main reasons people like to have separate transformer type supplies is so that they can isolate the outputs from each other. This allows you to power positive ground pedals and negative ground pedals without having the power shorted out through the signal cables. By connecting the negative side of all the supplies together it's no longer possible to "float" the DC outputs, the way you could if the pedals were running on separate batteries.

 

Maybe you could use a two-pole rotary switch. Then you could run separate negative wires from each power supply to the corresponding DC jack and to the rotary switch.

[katillac]

I did it that way to have a common ground in attempt to avoid hum from so many wires running everywhere. The next time I make one of these, it'll be much different because I'll know ahead of time what I'm looking at. As well, I won't be using a bunch of various transformers from a collection of wall warts, instead using either a single multi-tap transformer or a bunch of single, small transformers. All grounds will be separate and indeed, I will need to use a 12 position, 2 pole rotary switch. Either that or I may get another wild hair (more likely scenario) and put in individual displays using segmented LEDs. On the other hand, that also would entail having to build or buy driver boards for each display.

 

What do you mean by "float" the DC outputs?

[Kwiebe]

kattilac you have quickly become one of my favorites on this board, although I'm new to this site in general. I really wish I had the patience to do anything involving power supplies. Great job on this, I'd buy one!

[amp_surgeon]

I did it that way to have a common ground in attempt to avoid hum from so many wires running everywhere. The next time I make one of these, it'll be much different because I'll know ahead of time what I'm looking at. As well, I won't be using a bunch of various transformers from a collection of wall warts, instead using either a single multi-tap transformer or a bunch of single, small transformers. All grounds will be separate and indeed, I will need to use a 12 position, 2 pole rotary switch. Either that or I may get another wild hair (more likely scenario) and put in individual displays using segmented LEDs. On the other hand, that also would entail having to build or buy driver boards for each display. What do you mean by "float" the DC outputs?

 

The hum should be taken care of by the fact that all of the pedals share a ground through the signal cables. In an ideal world, there should be no hum on the power lines. Of course, the world isn't always ideal.

 

Voltage is a relative thing. The positive outputs of your supplies is +9V, but that's relative to the negative output. Since there is no actual electrical connection between the AC and DC sides of the transformer, that 9V is "floating", just like a 9V battery. If you connect two 9V batteries in series, you've got an 18V battery pack. Likewise, if you connect two "floating" 9V power supplies in series, you've got an 18V power supply. In other words, the loose ends of the two supplies are going to be 18V difference relative to each other.

 

To take this to an extreme, you could connect a floating 9V supply in series with a 1000V supply and have a 1009V supply!

 

In terms of mixing effects pedals, you could connect one floating supply to a negative ground pedal, and the positive output of that supply would now be +9V relative to ground. You could connect a second floating supply to a positive ground pedal, and the negative output of the second supply is now going to -9V relative to ground. Since both supplies are floating, you can connect the pedals together with a signal cable, and the power supplies will continue to work normally, even though the two pedals now share a common ground through the signal cable.

 

Now, assume those power supplies were already connected together at the negative terminal, and you hook them up to those two pedals. The common negative lead is now going to be connected to ground on the negative ground pedal, and the positive lead of the other supply is going to be connected to ground on the positive ground pedal. If you connect a signal cable between the two pedals you will be shorting the negative and positive sides of the power supply feeding the positive ground pedal - not good! This is because the positive power supply lead is grounded inside the pedal (it's a positive ground pedal), and the negative power supply lead is grounded through the signal cable coming from the negative ground pedal.

 

Floating outputs can be done with transformer based supplies because the DC side of the transformer is electrically isolated from the AC side - no common ground is established through the transformer. The only thing connecting the primary and secondary is a fluctuating magnetic field. This can also be done with a single transformer, presuming there are separate secondaries for each power supply output. This CAN'T be done with a multi-tap transformer because there is going to be a tap that is electrically common between each adjacent pair of supplies.

 

I hope that sorta made sense.

[katillac]

It made some sense to me after reading it three times, but the final part confused me even more. Please remember that you're an engineer and I still have to keep a piece of paper handy that shows Ohm's Law conversions.

 

Are you saying that, if I use this setup to power any pedals that are positive ground, I'm going to have some serious problems? I was prepared for pedals that have center pin positive by making cables or adapters to flip the polarity of the output from the power supply, but not prepared for what I think you're telling me.

 

If I have to, I will rewire this so that each power board (from each transformer) goes directly to one output, with no common ground shared. I'd have to really give it some thought to running an additional eight wires to a different rotary switch (12 pos 2 pole). I'd have to remove the original rotary, keep track of each positive wire to be sure I get them in the correct order, as well as each ground tap. It can be done, but at this point, I'm getting burned out.

 

Kwiebe: thanks! I've been working on this damned thing since November. It hasn't been a constant thing, but I've put enough hours into it that it feels like I haven't done anything else since. I've already redone the power boards twice (that's nine of them, twice each), each time re-gluing them to the base. I put down craft foam as an insulator beneath the transformers and then hot-glued the transformers and power boards to that. If I have to redo this yet again, I may just rip everything out and rewire everything from the power boards (the ones that come off the transformers and send power to the DC jacks) on out. It would be a lot less of a mess. I think.

 

I think I need another break from all this. There goes my first attempt at a fret level, crown and polish. I only have enough room to do one project at a time.

[amp_surgeon]

Yeah, the polarity of the DC jack on the pedal isn't really the sticky point. As you say, you can use polarity adapters for tip positive jacks.

 

Let's see if this clarifies my point a little:

 

 

 

You now have a complete path between the positive and negative sides of the lower power supply, thanks to the fact that the negative outputs of both supplies are tied together, and the two pedal grounds are joined through the signal cable. The lower power supply is shorted.

[katillac]

Okay, I can see the need to keep the grounds separated. It raises another question, though: for people who use things like the 1-spot adapter, which is daisy chained, don't those also share a common ground? If so, then why am I not seeing any complaints about shorting pedals when someone puts a positive ground one into the mix? Do they need a separate power supply for positive ground pedals and I just haven't seen reference to that because I haven't shopped for a 1-spot?

 

I don't get why some manufacturers do pedals with negative ground and others do theirs with positive ground. It would make more sense for them all to use the same.

 

That said, how is the signal kept separate from the power supply? I mean, without power, they won't work, so if you have two pedals, one with negative ground and one with positive ground, how do they work together when connected by a signal cable?

 

Also, if you have two boxes, one negative ground and one positive ground and they bump together, don't they short out?

[amp_surgeon]

Also, if you have two boxes, one negative ground and one positive ground and they bump together, don't they short out?

 

Only if they're running on the same power supply, or if the power supplies are tied together externally somewhere. As long as the power supply outputs are floating before they're attached to the pedals, then there's no problem.

 

Think of it this way - a floating power supply is like a battery. It's not positive 9VDC, nor is it negative 9VDC. It's just 9VDC between the two output leads. If you connect the negative lead to ground in the circuit you're powering, then it becomes a +9VDC power supply. If you connect the positive lead to ground, then it becomes a -9VDC power supply.

 

Like I said, voltage is relative.