MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

MFOS Bipolar Wall-wart power supply
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2  Next [all]
Author MFOS Bipolar Wall-wart power supply
nosamiam
Got a question:

About maybe 8 years ago I built an MFOS bipolar wall-wart power supply (+/-12v) on a home-etched PCB. I didn't end up using it at all until pretty recently when I started down the DIY Eurorack path. When I applied power to it, I noticed it was reading +11.80/-12.03. I figured the asymmetry in the readings was due to component drift from it sitting a while in a cold and hot garage for so long. Most of my modules act just fine. A couple (DIY ones) have some issues I haven't chased down. Maybe due to the weird voltages? Not likely, but who knows. I resolved to build a new one with all new components soon, if for no other reason, just to have a more solid version on manufactured PCB.

So I just built a new one using all new components, except I'm using the same AC-out wall-wart, which I assume is mostly just a transformer. I'm getting the same voltage readings. Are they normal? Can someone explain why they are asymmetrical when the circuit looks pretty symmetrical? And should it make a difference in how my circuits perform?

delayed
That is normal. The voltage dividers don't really put out 100% +12 and -12 voltages. The only way to get that voltage is building a power supply that trims the voltage. Just like a resistor does not have 100% 1k resistance. They would read within 1% or 5% of 1K.

I have etched and built up plenty of these and they are always a little off. It have never bothered anything on my end.
thx2112
I made a bunch of those. They're great because large-amperage 12VAC wallwarts are everywhere.



Those regulators aren't very accurate, but it isn't a problem.
masterofstuff124
care to share that power supply design. I love my stripboard MFOS builds. but they would greatly benefit from your amazingly compact design!


also were do you get ac wallwart supplies/ I got a 2 amp from jameco and it was pretty expensive.
daynehacks
thx2112 wrote:
I made a bunch of those. They're great because large-amperage 12VAC wallwarts are everywhere.



Those regulators aren't very accurate, but it isn't a problem.


Im diggin that design too. Sell the pcb's on your site?
flts
To expand on what others have already said, normal 7812 and 7912 regulators may, according to datasheets, put out anything from 11.5V to 12.5V (and -11.5 to -12.5V respectively). So what you're seeing is well in spec for those. There are higher precision versions of those, such as L7812ACV / L7912ACV... However, even those are usually specified for +-2% tolerances so what you see would be inside normal operating conditions for them as well.

The only way to get exact voltages at the supply (or distribution) point is to build a supply with adjustable regulators like the LM317 and LM337, and trim the output voltages manually. That's what's done in Doepfer supplies for example.

On the other hand, a lot of small euro 4HP SMPS "power modules" have +-5% specced components again, so it's not like it's uncommon... Most circuits will have no issues whatsoever with slightly imbalanced or slightly higher / lower supply voltages than what's specified. But if they use voltage rails directly as references (bad) you may have to be prepared to recalibrate them if you move them to use another supply.

As for mini halfwave power supplies in general, aside MFOS and Frequency Central Power, there's for example the https://www.tindie.com/products/dhaillant/mini-dual-power-supply/ - it has space for only one cap per side so less capacitance = less output current, but nicely compact for small systems / test setups.

I've also drawn my own PCB which includes basically the same halfwave linear circuit (with space for two big capacitors per rail) and 7 euro power connectors in one PCB. Meaning to build a test setup and a couple of skiffs with it but once I get the PCBs and test that they work OK, I'll share the design on Github.

thx2112
I get the power adapters from thrift ($1-2 used) or army surplus ($5-10 new) stores. The high amperage ones are easy to spot because they're "power brick" sized, but smaller than computer power bricks.

Whenever there's some space on a 10x10cm pcb order I try to cram in an extra design. Sometimes it's a SMD->DIP adapter, a (weird sized) protoboard, sometimes a Eurorack module pcb like a CV/Audio mixer, tube overdrive, or attenuverter. These aren't my main focus so not a lot of time is spent debugging them or improving the design.

Here is a series of power-related boards I have now:



Power filter, 1.5A power supply, and power distribution board. I also have a tile power distribution board but am waiting for tile cables.

The designs will end up online, and I was thinking of putting the spare PCBs on my site.

They were made for my own use, and because of the unusual size constrictions are a bit hokey (like those caps and heatsinks floating in the breeze, lol.)
Addam
thx2112
Shot in the dark here, but any chance you have any MTA .156 (4-pin) power distribution boards??
flts
Sorry, not meaning to be nasty or nitpick or anything like that, but I guess with 3200uF per rail capacitance in a halfwave PSU and Tayda heatsinks, 1.5A is a really really theoretical figure, and the actual clean power you can get before exhausting the reservoir in caps or overheating the regs is closer to a few hundred mA per rail max...?

Or have I understood the idea completely wrong somehow? I haven't actually measured how much power a supply like that can output before bad things start to happen.
thx2112
I too was skeptical but in practice it works surprisingly well. There were three of these in a two-row case at Knobcon and there were no problems.

I've started putting the files are up here: http://github.com/THX2112/12VAC-Eurorack-Power-Supply
flts
Yup, I've got good experience of powering single row cases with one or two of small'ish halfwave supplies depending on how power hungry euro modules there are, and how much capacitance there is to spare in the PSU. So three would most probably do really well for a 6U.

What I meant by my previous comment in practise, is that none of the 6U systems I've done so far have required more than 1A current per rail, and if you use three of those supplies in a case like that, that's going to be like 300-400mA per rail per supply. That's in the ballpark what I kind of expect them to give with no problem - 1.5A for me would be more like the domain of a full wave supply with bigger "fin" style heatsinks.

I really like these kinds of supplies because they're convenient, cheap and perform okay. The practical problem is that unlike there, 12VAC 230V/240V external power supplies aren't really available everywhere and for cheap here in Finland. They're always like 20+ euros (20-30$) a piece new and you have to order them from bigger local electronics stores or German eBay as I haven't found any neighbourhood source for new ones. I have found a couple from thrift stores for like 1-2 euros a piece but even those places mostly have DC output wall warts these days. Either that or there's someone else around who likes these kinds of supplies...
abelovesfun
I reluctantly turned my personal version into a product: https://aisynthesis.com/product/eurorack-power-supply/

I made these after Ray's death (RIP) for personal use. I had been using Ray's version in my live cases, but the stress of having an ac jack wired to the PCB made me have to re-solder it a few times, so I decided to make a version with the jacks and connectors on the PCB. A few people wanted them, but I couldn't support it without making it a product, so now it exists on the site. Of course, now Ray's PCBs are back up at synthcube, so it isn't something I really promote. It is a great simple design.
flts
I did this based on the same idea: https://github.com/vaeinoe/moraydular/tree/master/my_little_euro

Works well for lighter loads as expected, built a couple of 60HP skiffs with zero wiring that way. Unless you use huge capacitance caps (6800uF+), the maximum load will obviously be somewhat lower than Ray's version or FC Power.

Then again, that kind of design isn't meant to power anything more than a couple of modules, even though many people (myself included) have powered a big bunch successfully.
flab
flts
i might print few of those for my 3U cases,

thanks man
dhaillant
I confirm those PSUs are really for low power synths. You can get more mA with higher input voltage, but you need to be careful not exceeding the input voltage limits of 78xx and 79xx regs. Depending on quality of your wallwart, you can get more mA too...

If someone is interested in small form factor PCBs (one cap per rail, 5cm x 5cm), I still have some stock available. Check on my Tindie page and on my blog
flts
^- We've actually used a couple of those mini dual supplies of yours for some standalone instruments and "one or two module" type boxes. The form factor is very nice for that, a good alternative for little DC-DC blocks for those kinds of projects.
dhaillant
I'd love to see pictures of them in situation hyper
flts
dhaillant wrote:
I'd love to see pictures of them in situation hyper


I don't actually remember which ones I put the PSUs myself anymore very frustrating My workshop friend is currently building a couple of simple electroacoustic / piezo instruments and I think he might have ended up using the rest that I built, I'll try to remember to ask him for pictures once they're ready...
wackelpeter
If you throw in 1,5A regulators, your wallwart provides enough current and you use larger heat sinks and maybe some bigger electrolytes before the regulators i'm able to get 600mA out of one of these on stripboard...

only problem i had was when i tried the ps3100 resonator clone on stripboard and with my home rolled vactrols i consumed too much current that affected the Output voltage... with some tweaking i got the current consumption of the Resonator from 90mA down to 25-30mA and gone where the voltage drops and the oscillators in that case run clean and well without noticing me any drifts or crosstalk or something like that...

Well maybe not class AAA specs but for my needs it's okay and sounds good enough...

The more important Problem for me is to get cheap 15VAC wallwarts with 1-1,5A Ratings. Seem to become more and more rare, because not many devices use them anymore... I have a few 9V DC and 12V DC wallwarts, more than enough and can have acess to 24V AC 2,5A wallwarts but those are a bit to much voltage, so i would need to drastically reduce the voltage before the regulators and need large heatsinks...
abelovesfun
Amazon has them: http://amzn.to/2pmbxv8
dhaillant
The problem with this kind of PSU is the ripple.

Larger capacitors will be able to smooth more the input voltage of the regulators, but to some extent. Then you have two possible solutions: either increase the AC voltage or increase the current coming from the wallwart.
dhaillant
wackelpeter wrote:
I have a few 9V DC and 12V DC wallwarts, more than enough and can have acess to 24V AC 2,5A wallwarts but those are a bit to much voltage, so i would need to drastically reduce the voltage before the regulators and need large heatsinks...


24VAC is too much. It gives almost 34V peak. That's just 1 volt under the absolute maximum input voltage. Even if it's not DC and you substract 0.7V from diode drop, it's not good for your regulator's health. And don't forget the bad behavior of those wallwarts when there's no load: voltage goes up a lot. Dead Banana
wackelpeter
dhaillant wrote:

24VAC is too much. It gives almost 34V peak. That's just 1 volt under the absolute maximum input voltage. Even if it's not DC and you substract 0.7V from diode drop, it's not good for your regulator's health. And don't forget the bad behavior of those wallwarts when there's no load: voltage goes up a lot. Dead Banana


I noticed that, on my 600mA case i have a 18V or 20V AC wallwart (can#t remember exactly and don't want to pull it out now). There were needs for really big heat sinks in this psu... I also thought about using some 8-12 ohm resistors before with high w rating, to reduce the voltage going into the regulators a bit...

But with these heat sinks i'm fine until this beast of a PSU will blow up in smoke...
wackelpeter
dhaillant wrote:
The problem with this kind of PSU is the ripple.

Larger capacitors will be able to smooth more the input voltage of the regulators, but to some extent. Then you have two possible solutions: either increase the AC voltage or increase the current coming from the wallwart.


If you're using that for +/-15V with a 15V AC wallwart then Things should be fine to some extent.
15x1,41 is in safe regions within the at least needed 17V for the Regulator.
Then 2 2200 or 3300uF between diode and regulator, possibly 2 tantalum caps as close as possible to the regulator on its input and output and that thing can be stressed a bit...

If it turns out not to be enough for your want's you can replace it with a bigger and more solid PSU and keep it for some other purpose... that's how i approach the whole thing... Only thing is you have to keep it to be exchangeable.
dhaillant
Yes, it's not only dangerous for the regulators but is also a lot of heat to dissipate in your case. Remember that the voltage difference between input and output is lost as heat. And it's a bigger problem with larger currents.


Quote:
15x1,41 is in safe regions within the at least needed 17V for the Regulator.

For 15 V out, you need between 1.5 to 2 V dropout for the regulators to work. If you provide only 16 V to a 7815, then, you won't get 15 V but something like 14.5 V.
Now when you say 17 V is enough, it is, but if you're able to provide 17 V DC and stable enough.

It's important to think about the waveform provided by the half wave rectification. Even with big capacitors, when you start to draw current, ripple appears.

If you have 15 V rms, 17 V peak is your maximum. Between each "conducting time", the voltage drops as you draw current from the capacitors.

Vr (voltage ripple amplitude) can be calculated as follow:
Vr = I / (f * C)
Where I is current drawn, f your mains frequency (50 Hz or 60 Hz) and C your total capacitance.

So, if you draw 0.2 A with 4700 µF in 60 Hz country, your ripple will be around 0.7 V.

If V peak (the maximum) is 17 V, the minimum is 17-0.7 = 16.3 V.

If you substract the diode drop, you end up with 15.6 V

Your 7815 won't be fed with enough input voltage and its output won't be 15 Volts. Dead Banana
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2  Next [all]
Page 1 of 2
Powered by phpBB © phpBB Group