Safer power supply

[wicked1]

Well, it was a ground loop.. sorry, I'm an idiot. Rewiring it w/ dual secondaries, two rectifiers basically made me clean up the grounding scheme. I rewired it last night as center tap w/ careful attention to grounding order, and it's quiet.

The unbalanced load definitely makes the volts coming out of the two halves of the transformer uneven, but the regulators fix that and it shouldn't cause noise.

New problem though... My 12v transformer along w/ my sagging home power (it drops about 10v when the air conditioning turns on) doesn't leave enough voltage for my 12v regulators, and they are noisy! When the AC is off, everything is quiet.

I guess I need a 15v transformer.

Edited August 1, 2012 by wicked1

[Shuriken]

Well, it was a ground loop.. sorry, I'm an idiot. Rewiring it w/ dual secondaries, two rectifiers basically made me clean up the grounding scheme. I rewired it last night as center tap w/ careful attention to grounding order, and it's quiet.

The unbalanced load definitely makes the volts coming out of the two halves of the transformer uneven, but the regulators fix that and it shouldn't cause noise.

New problem though... My 12v transformer along w/ my sagging home power (it drops about 10v when the air conditioning turns on) doesn't leave enough voltage for my 12v regulators, and they are noisy! When the AC is off, everything is quiet.

I guess I need a 15v transformer.

Yeah a 12V Transformer is cutting it very close (12V * 1.4 - 1.4 = 15,4V Rectified) That's almost exactly the 3 Volts the regulator needs to properly regulate. And that calculation doesn't take into account small fluctuations and deviations. My home power is not 230V but about 222V, the transformer output will also be slightly less. So a 15VAC transformer is what i use for a 12V linear regulator.

Edited August 1, 2012 by Shuriken

[technobreath]

Yeah, shuriken is right. Using 12v secondaries for 12VDC isn't leaving much room for shit to happen.

I didn't catch how much capacitance u got before the regulator. Adding more can help a little, but it's more like patching up a thing that really needs stitches. A couple of things to keep in mind when adding to the cap value before the regulator. Inrush current will increase. Either you tame it (can be done in a pretty crude way using a ntc thermistor) or you make sure that the rectifier circuit and the transformer will handle it in the long run.

Please be aware that this is based on my theoretical knowledge. I used to tinker around with simulations on linear supplies, and that is basically where this statement comes from - it could be wrong - or irrelevant.

[Hawkeye]

If you have four of them lying around, you could try schottky diodes instead of a standard bridge rectifier... then there would be less voltage drop there...

and additionally use a low-voltage drop regulator like the LM2940, which will also only need about 12.5V to regulate to a stable 12V/1A... it might be cheaper than a new transformer...

but this is all patchwork... so +1 on Shuriken´s words!

Many greets,

Peter

Edited August 2, 2012 by Hawkeye

[wicked1]

Thanks everyone.

What's odd is I just finished the 9090 and used a 2x15v tx for it, and don't seem to have any problems w/ it regulating 15vdc. But, it's a different transformer, so irrelevant, I guess.

My only concern w/ a 15v tx for the 6582 is, I'm also using a 9v regulator, and that's a pretty big drop for it.

Is there a down-side to using LM2940's instead of the LM79* regulators? I'm thinking there must be a reason the LM7* series is still the standard.

[NorthernLightX]

Just wanted to let you guys know that Reichelt sells some Mean Well PSU's now, among which the D-60A, which delivers +12v and +5v at a 54KHz switching frequency. Disclaimer: I have no experience with this particular model. It's about €17, might be worth a try.

[Hawkeye]

@wicked1: the LM2940 is a bit more expensive, harder to obtain and not as known as the 78XX... It has stricter requirements regarding the output capacitor (should be low-esr), otherwise it is an improved model, which helps save energy as you do not need so high input voltage that needs to be burned.

@NorthernLightX: 54KHz switching still may be a little bit low, for those people analyzing their 96khz samples in 1:1 mode - don´t wanna fuel the holy switcher/linear war here :)

Many greets,

Peter

Edited August 2, 2012 by Hawkeye

[Shuriken]

Thanks everyone.

What's odd is I just finished the 9090 and used a 2x15v tx for it, and don't seem to have any problems w/ it regulating 15vdc. But, it's a different transformer, so irrelevant, I guess.

My only concern w/ a 15v tx for the 6582 is, I'm also using a 9v regulator, and that's a pretty big drop for it.

Is there a down-side to using LM2940's instead of the LM79* regulators? I'm thinking there must be a reason the LM7* series is still the standard.

Using 15VAC to make 15VDC is fine (15V * 1.4 - 1.4 = 19,6V Rectified) that's about 1,6V of head room.

Edited August 2, 2012 by Shuriken

[technobreath]

Hehe hawk

don´t wanna fuel the holy switcher/linear war here

I haven't realized before now that there was one, but now that u say it :P. I have always been part of the linear camp myself. But switchers is probably the future I guess, but 54khz I also expect to be a bit too low.

But, it's a different transformer, so irrelevant, I guess.

It's like this - there's no magic, only math (in which I suck btw) but it's like shuriken lays it out:

VAC * 1.4 - 1.4 = Rectified voltage.

Also the 7809 has a pretty good input range up to 35VDC; so the vreg is fully capapble of handling the drop from the 18-20vdc to 9v. BUT a large heatsink is required - I don't know by hand how to calculate the size, but the vreg can't operate in higher temps than 125 C. Then again, it might be easier, less bulky, less work to just use a second transformer for that :P

[m00dawg]

Also the 7809 has a pretty good input range up to 35VDC; so the vreg is fully capapble of handling the drop from the 18-20vdc to 9v. BUT a large heatsink is required - I don't know by hand how to calculate the size, but the vreg can't operate in higher temps than 125 C. Then again, it might be easier, less bulky, less work to just use a second transformer for that :P

A transformer would be the way to go, but an alternative to running all that power through the regulator is to use power resistors to step down the voltage before it gets to the reg. I did that in my first MB6582 PSU and it worked quite well.

As far as switch-mode versus linear, I say right tool for the right job. Switchers will likely continue to gain in popularity, and that's generally a good thing. But some applications will always demand linear. The latter is also a lot more simpler.

[technobreath]

Yep u are right, I didn't think of the resistors to step down the voltage. I have seen some psu designs use this method.

[ilmenator]

Yep u are right, I didn't think of the resistors to step down the voltage. I have seen some psu designs use this method.

But that is basically just the same as the above - you convert all that energy into heat, only that you don't do it inside the VREG but inside the resistor...

[technobreath]

But that is basically just the same as the above - you convert all that energy into heat, only that you don't do it inside the VREG but inside the resistor...

Sure it is, no doubt about it - just another patch :). I'm just saying, it's a pretty crude and simple way to burn off a little heat outside the vreg if it's too hot already. But I would definetly stick to doing it the right way myself. Proper voltages / proper heatsinks. I'm just agreeing there are alternatives if a rebuild / new transformer isn't sounding as nice as a power resistor. :P

Edited August 3, 2012 by technobreath

[m00dawg]

But that is basically just the same as the above - you convert all that energy into heat, only that you don't do it inside the VREG but inside the resistor...

Right but the heat is being dissipated over a much larger area. I used 2 resistors to pull down the voltage and that ended up being an area around 4x2cm. Much larger than trying to pull all that heat out of just the regulator a quarter of the size. I had a heatsink on the regulator too, of course, but it didn't have to be near as large doing it this way and added some flexibility.

But I still think that using a transformer is still the best option.

[wicked1]

Personally, I think I'll try the schottke diodes and LDO regulators. If they work and it's burning less energy, that's all good!

[m00dawg]

Personally, I think I'll try the schottke diodes and LDO regulators. If they work and it's burning less energy, that's all good!

True but you still have to have a reasonable input voltage into the regulators. Using Schottky diodes with LDO regulators just means the output from the transformer can be closer to the desired DC voltage, if I'm not mistaken. So you still have a heat problem if you try to feed 15VAC into a 5V regulator.

[wicked1]

No, I mean sticking w/ the 12v transformer. So, then my 9v regulator is still happy (that's the one I was concerned about w/ the 15v tformer). I'll switch the 12v regulators to the low dropout, so that maybe they'll work w/ the 12vac transformer.

5v regulator is switching, so I'm good there either way.

Edited August 3, 2012 by wicked1

[m00dawg]

Aha, gotcha.

[wicked1]

Well, amazingly, the LDO regulators and the Schottky diodes (salvaged from a PC power supply) seem to be working. I've only got about 300mv to 1/2v headroom under load, but I can't hear any buzzing. And, that's worst case scenario.. For some reason, the outlets on the other side of my basement are about 8vac higher in voltage, so I've got no problems there.

[wicked1]

New question... In the end, my power supply is like the wiki, one rectifier, Center tap configuration.

Could I use the +9v regulator on the negative rail? So like, connect + to 0, and - to -12?

I know that will give me 9v, but its ground would be 12v lower than neutral, right?

Then I guess it depends on how Wilba set up the 6582... does everything share the same ground?

is my thought process correct on this? If so, do you know if it will work w/ the way wilba has the 6582 set up? It's really a shame there are no schematics for his designs!

[m00dawg]

Not sure why you would want to try and hook it up that way? Why not hook the +9V reg up the standard way? But to answer your question, yes I'm pretty sure the MB6582 shares a single ground.

[wicked1]

Only because the draw is so uneven... Almost everything is pulling from the +12 side, which is why I have no headroom. The negative 12v side has several volts headroom. It's only powering my filters.

My thought was basically the SID's themselves are probably drawing most of the power. I've got 4x9v sids, and 4x12v sids. I was thinking I could split the load so the 2ndaries of my transformer are doing equal work.

But.... like you said, it probably won't work, and it is working successfully now...... so, I suppose I'll forget about it :).

[m00dawg]

If you have the room, I would say it might not be a bad idea to just have two bipolar and non-polar PSUs to avoid that problem. Sort of wastes a lot of components for the bipolar (given the current draw) but it will definitely keep things balanced.

For what it's worth, that is the route I am likely going to go for my band's live rack (which will need +5, +9, and +/- 12 or 15V). Planning on having non-polar PSU module boards of sorts and a bipolar board. Haven't decided how many transformers I will use, but will likely use at least 2. But the modular PSU will sit in the bottom of our rack, so I have plenty of room to work with (most of the rack will have panel mount components so it will be pretty open).

[wicked1]

That's kind of what this is turning in to! I'm actually thinking of having a rack panel made for my 6582.. It would still take the same formfactor, so I could use the same control surface PCB.. just have dead space to one side to make it rack-mountable. (Mount the filter components back there, or something).

Oh, maybe I can fit the FM in that space.... hmm. Off to the drawing board. (which is actually just a different screen here at my same desk :) )

In the end, the goal is to have the core of my studio transportable... Couple of filter and FX boxes. MB Seq, SID and FM. 9090, novation synth, and the two problem bits... mpc2000xl and mixing board (those are "problems" because of their size and are obviously designed for desktop use)

Edited August 27, 2012 by wicked1

[m00dawg]

I, too, am working on an MB6582 rackmount, although not for my live rack (I will be using the Sammich baseboards and a very stripped down control surface instead). It's untested, but =sidr8tr]here is all the info on my rackmount MB6582. I have about half of the boards verified (on paper) to fit with the front panel, but that's about as far into the validation as I have gotten to this point.