Power/mains/grounding question..

[cloudscapes]

Hi there,

I'm going to build a spring reverb unit. And I'm (probably) going to use this power supply (or one like it) to get a bipolar +15/-15 going:

http://www.generalguitargadgets.com/wp-content/uploads/ggg_bipolar_ps.pdf

Planing phase at the moment. Up to here has been pretty simple, found the right transformer (20VAC 1.5A) and regulators/heatsinks. Though I'm wondering if I need to ground 0 to chassis and/or mains earth. Or just mains earth to chassis, and 0 to signal ground. Or all 0/ground/earth together. And which one of these two I'll need to get:

http://smallbear-electronics.mybigcommerce.com/115-v-ac-power-inlet-chassis-mount-3-wire/
http://smallbear-electronics.mybigcommerce.com/115-v-ac-power-inlet-chassis-mount-2-wire/

I've never worked with mains/transformer stuff yet, and it scares me a little so I want to be 200% sure I'm doing things right.

[GibsonGM]

You'll want a 3 wire connector, and to use a grounded cord.  Many use the common IEC connector that you find on many/most amps, but that's up to you.  Be sure the chassis has a grommet for the cord to pass thru to prevent chafing.

http://www.don-audio.com/bilder/produkte/gross/IEC-Connector-with-ON-OFF-switch-1-Pole.jpg

You want to attach your safety ground right next to where your connector enters the chassis, with a star washer or other security fastener to assure it will NEVER come loose.  You want that one wire to always be grounded, and it will save you in the event of a freak short.  Most important connection you will make on a project using a mains trafo.

Your zero volt point, ground for the dual polarity supply you show and your reverb circuitry, should also be connected to the chassis in the same/similar manner.    I'd keep the dual supply ground 'near' where the safety ground is (NOT connected to it, that one stands alone as it is so important).   In this case it won't matter much, but I try to keep the power "over to one side" and my audio stuff to have ITS ground on the other side of the chassis.

The grounds are common and are nothing to be nervous about (there is no current flowing in it unless there's a problem, which will blow the fuse near instantaneously), and by having your circuit plus power supply, and jacks (meaning whatever you plug into the 'verb, your guitar etc) all tied to the chassis at ground will keep them all at the same potential.    Many other 'schemes' are designed around minimizing noise, such as how you layout your tube gain stages, to minimize their exposure to ground currents.   Keeping your power transformer away from audio is going to be most important for you.

Don't forget the FUSE!!!     Many IEC connectors have a holder for one built in, which I really like.

[cloudscapes]

Thank you for such a detailed reply! If I understand correctly, something like this?



Sometimes power transformers have a ground output on the secondary. To chassis also? though if I did that, it would essentially be hooked up to the lower of the two transformer outputs in this power supply schematic.

[GibsonGM]

YW, this is a great place if you're bored and like to see what others are up to, ha ha! 

Your ground plan looks good.  Run a bare wire, or trace, around the bottom of your PCB and make that a ground bus, and you're golden.

The idea is to keep the 'sensitive audio ground path' away from the power supply one, which may have noise on it.  There are many ways to go about this, of course.  I don't think you'll have any issues whatsoever with what you're planning.   You'll have nice filtering along with the regulators! 

You won't use the center tap on the transformer for this design...it's not a ground, it just 'cuts' the transformer coil in half.  It's a "10 -0 -10" transformer, right?   (BTW, You should measure 1/2 the resistance from ctr to either end as the reading you get across the whole coil with your DMM, to confirm it is in fact the center tap if there is any doubt at all)    Roll it up, tape it off solidly so it can never touch anything, no problem. You might re-use it someday, so I wouldn't permanently ruin it (like by cutting it off or anything). 

Looks good!  I see you put the fuse in   What current rating are you planning on using?

[cloudscapes]

Understood about the transformer ground.

I don't know yet about the current rating, I'm just still in the planning phase. I intend to combine the power supply from earlier with this:

https://images.campyent.com/sites/default/files/tech_corner/reverbtank8.svg

..where I have that actual reverb tank (8EB2C1B) getting shipped out to me. Plus a bit of extra opamp stages for wet/dry blending, reverb feedback, and a filter probably. From what I can tell, the 1.5A 20VAC transformer I'm looking at is more than enough.

[GibsonGM]

Way, way more than you need   Won't harm anything to use it, but I bet a 1/2A or even less would have been fine.  Looks like the circuit draws  far less than 100mA, from a casual glance.   You may wish to revisit the specs on the transformer, to save weight and money.     

I'd try a 1/2A 250V fuse, altho one much smaller would be just fine and offer better protection to the circuit (150mA, etc).    The '72 opamp can dissipate only ~ 600mW, so you're not looking at a significant current draw here.   Not quite sure how it's done with the split supply, but a differential of 30V * 20mA gives 600mW, so we must be in that range.

[cloudscapes]

There's a same-spec'ed transformer that's 0.6A instead, I might get that. Once I actually get parts and the reverb tank going, I'll be able to measure and figure out how much I really need.

I'm quite excited of this! The casing will be stained/varnished wood, about half-rack width, front and back panels will be machined aluminium made by these guys. I was happy with their results last time I got something done through there.

[GibsonGM]

That's really sharp looking, man!  If it's wood, you may wish to shield the interior with aluminum or copper ($$) foil, tied to ground. That will keep things quiet.

The .6A transformer seems like the better way to go....smaller in physical size, too!!  As I described above, you won't be going over the ~600mW the opamp can handle.   That's a design limit.  Using your supply voltage, we come up with 20mA being a limiting factor.   

Someone will come along who is better at me than sizing a transformer properly, though.  If not, ask it as a separate topic - I'd like to see how the 'big dogs' go about it.    I do it seat of the pants.   Never had a problem doing it that way, but you may well find you can get an even SMALLER trafo and save MORE money!!    350mA, or something.  I'm always over in my ampacity, too (how much current you can get out of it...I go overboard sometimes, like you started to). 

[tubegeek]

Size a transformer by Volts x Amps. Your low-voltage supply current draws, let's say it's a wildly overestimated 50 mA. That's .05 A. Your power supply will provide a rectified voltage total of over 30V, let's say a wildly overestimated 50V.

50 x .05 = 2.5 Volt-Amps (yes, they are equivalent to Watts.) Note we have designed in a huge fudge-factor for safety by those two wild overestimations. Using the wildly overestimated numbers saves us from having to get too fine-focused about transformer and rectifier efficiency - this is a seat-of-the-pants calculation at best. I used approximations that would make the numbers easy, too. Download this Hammond Transformer power transformer applications data sheet for more detailed and accurate calculations if you'd like to get a transformer that's just the right size.

http://www.hammondmfg.com/pdf/5c007.pdf

How much current will you draw on the primary?

125 V x .02A = 2.5 Volt-Amps. So that would be something like the fuse rating you'd want to use, .025 A or thereabouts.

Even a 60 mA rated transformer will be well within spec. That would be a rating of 50V x .06 A = 3 V-A and we're using an unlikely worst case of 2.5 V-A. Your 600mA transformer should be up to the job and as others have pointed out, you may even want to size it down smaller for compactness and cost reasons.

Note that the center tap of your secondary would short out the whole bottom half of the secondary if you grounded it: the bottom leg of the secondary is grounded already at the right side of your drawing, which was nicely drawn, BTW.

[GibsonGM]

Thanks, Geek, nice to see it laid out correctly     I always get a little stymied by the fact that you're drawing more on the primary side than you might think, given you're only drawing a small amount on the secondary, and thus I over-complicate things.   

Yup, same as power - thanks! 

[cloudscapes]

But wouldnt the coil/driver draw a great deal more power?

[GibsonGM]

But wouldnt the coil/driver draw a great deal more power?

It could try And if it was DC, it would succeed, but - this is not, it's AC... The max it could pull in theory even then is 37mA, which Geek accounted for by using 50mA as a 'design figure'.    Remember, the thing isn't running *at 30V differential* but at some voltage range less than that, and it's 'rotating about the zero point').   Let's get some info here!

You'll probably see your signal averaging something like +8 to -8 if you're not driving it fully.  I'll make an assumption, and call this 'the average signal level'.

Then, the current in that coil would be 8v/800R = +10mA, and -10mA, but not at the same time.   Even swinging 0 to +15 and 0 to -15, the max current in it at any instant will be ~19mA, and much less so on average, so the assumptions are safe.   That's why I used 20mA as the upper limit, in my reply above. 

Max power dissipation for the coil would be .29W; call it .3W.    Under the rating for the opamp, BTW.

Transformer:  Using the info TubeGeek gave us, 125V*.02 (20mA) = 2.5VA in the primary, and even with a healthy margin (the 50V and 50mA), we're still ok, 2.5VA.    I'd go with a smaller trafo for size and savings, 60mA to 100mA! 

You could power this bad boy with a wall wart if it weren't for the need for dual polarities...so for ease, I'd go ahead with a trafo. 
 

[davent]

Are there any advantages/disadvantages to using the non-center tapped transformer rather then a center tapped as you're shopping for one for the project and can get the transformer best suited to the project? I was under the impression that the non-center tapped was a work around to use what was available... but i really have no idea.
dave

[GibsonGM]

The GGG power supply design doesn't use a center tap.   So it's not needed.   If you (cloudscapes) find one with no CT of the right voltage and current rating, you're all set.     The only other way I see to do it, is to get one of twice the voltage and use the CT, so you're only using 1/2 the coil then (?)

[davent]

Hadn't seen that before, thanks!

[PRR]

> the right transformer (20VAC 1.5A)

Gosh, I could build a 10 Watt amplifier with that. A thousand times more than the spring needs.

> But wouldnt the coil/driver draw a great deal more power?

The spring driver?

Those things will burn-up at a small fraction of a Watt. The nominal input is 11 milliWatts, 0.011 Watts. Even if you HAMMER it, you can't get anywhere near a Watt.

We rarely need to drive the spring To The Max. If I take Mike's reasonable numbers, even if 20mA peak, this is delivered push-pull so is only 10mA from the supplies.

> Max power dissipation for the coil would be... .3W.   
> Under the rating for the opamp, BTW.

The load dissipation and the opamp dissipation can not be equated. Under ideal conditions the load power can be much higher than amplifier power (to 3X on sine if all is ideal). If conditions are far from ideal, the amp dissipation can be far higher than the power in the load. Reverb coils are a very non-ideal load. This may be good: if worse-to-worst something is gonna smoke, better a 19-cent chip than a $30 spring. (But don't bet on it.)

> 50 x .05 = 2.5 Volt-Amps

FWIW: 5VA and 10VA parts are often cheaper than smaller VA parts. Tiny windings cost more. Also there are more of them in surplus and over-run sales. So we would often round-up to the cheaper part.

> How much current will you draw on the primary?

You have to allow for transformer efficiency. A big transformer may be 95% and we can overlook it. Small iron can be as bad as 60%. I would go-up by 1.5, call it 4VA. (But again, 5VA-10VA may be your better buy.)

> the fuse rating you'd want to use, .025 A or thereabouts.

Closely fusing small iron is very problematic. There is a turn-on surge depending on the line-phase at last switch-off and this switch-on. Complementing this is the fact that the geometry of small iron means it is fairly unlikely to over-heat (large surface area for its volume). And if it does burn, it is a small fire. You MUST fuse to protect your wall-wires and your line-cord against your wiring mistakes and lifetime accidents. 10 Amps will protect against these. You might go much lower to be very sure the cord never gets even warm. However preventing over-heating on 5VA transformers may be folly. And a barely-OK fuse may fail in a few years from repeated start-up surge. And you can't get 0.025A fuses in Lodi on Sunday. 0.5A should be fine IMHO. Just before it blows, you have the heat of a 60 Watt lamp inside your box, and that's not a big fire-hazard.

> advantages/disadvantages to using the non-center tapped transformer rather then a center tapped

Large-scale--- there are three basic connections which require very different limits on the parts. The 2-diode CT conserves cathode-stuff (costly in bottles) but needs high voltage ratings (no big deal in bottles). The voltage-doubler gives the loest diode voltage for a given DC voltage, and was popular in the old days of low-volt rectifiers. There are also differences in the capacitors needed to hit your ripple spec. However, these days rectifiers and caps many-many times bigger than this project (most small audio work) cost very-very little.

The GGG bipolar uses the voltage-doubler topology. We could use much weaker diodes than they specify, but today they would not be cheaper, if we could even find weaker diodes. (They could be 1N4001 with ample safety. GGG specced 4004 because they are the same price as the 4001. Today you can get 4007 (1000V!) for the "same" price as 4001 in DIY quantity, so that's what you should stock-up on. (If you build a million, you may find that '04 is a half-cent cheaper than '07 in truckload lots.) The voltage-doubler also tends to want largish caps (at low voltage). However this "disadvantage" fades once we commit to the regulators (still large but not over-large). And the cap-cost is small in the overall budget. And Al e-caps are dirt-cheap today.