DIAGO PS09 Power Supply ISOLATOR? Easy (albeit expensive) cure for daisy chains?

[drolo]

http://www.diago.co.uk/adaptors/isolator.html

Any idea how they achieve this?
Has anyone ever tried these?

I have a mix of about 12 pedals on my pedalboard with 6 of them being digital. I have been looking for a way to isolate power supply grounds without necessarily having dedicated transformers for each pedal. I might eventually go that route, but I'm scared that all these transformers will radiate EMI and I will be left with another source of noise ...
If I could only use one big transformer with 12 individually regulated 9V outputs, it would be great. Lifting the ground of all but one output seems to help. But Some digital pedals (EHX Ring Thing, Memory Man with Hazarai, pedals using TAPLFO PIC) still bleed noise into the other pedals...

[tca]

They say:

Tech Spec
Model    Diago PS09 Isolator
Input    9V DC Centre Negative
Output    9V DC Centre Negative, 110mA Max
Isolation    >lkV

So it seems to be passive, probably a low pass LC filter.

[R.G.]

I took the "isolation >1kV" to mean it was an internal isolation set up.

I've speculated from time to time about designing a small 9Vdc to 9Vdc power supply. It's straightforward - pretty much any of the current-mode power switchers that will run from 7V-10V without going into low voltage shutdown should do it. A 1W isolated flyback is pretty easy and a small cup-core transformer should do the job.

About the only reason I haven't published something like this is that the transformer has to be custom wound, and then there will be the chorus of questions about not being able to get the cores and bobbins and wouldn't a tin can or box of corn flakes work?

Seriously, there's a max price on this kind of thing - somewhere around $20-25 will buy a whole 'nother 9V power supply that will run another hundred pedals or so.

[slacker]

Could be an isolated DC to DC converter, something like this http://uk.farnell.com/xp-power/iw1209sa/converter-dc-dc-2-1i-p-1w-9v/dp/1435397. This meets the specs they quote.

[drolo]

So it seems to be passive, probably a low pass LC filter.

I'm not sure it could be passive, at least if it does indeed isolate the supply as they claim.

Seriously, there's a max price on this kind of thing - somewhere around $20-25 will buy a whole 'nother 9V power supply that will run another hundred pedals or so.

exactly, the only advantage would be the smaller size of the device and not needing another spot on the power strip. but I can't see myself spending so much on it...

and then there will be the chorus of questions about not being able to get the cores and bobbins and wouldn't a tin can or box of corn flakes work?

pretty accurate prediction 

Could be an isolated DC to DC converter, something like this http://uk.farnell.com/xp-power/iw1209sa/converter-dc-dc-2-1i-p-1w-9v/dp/1435397. This meets the specs they quote.

that could be it indeed ... not cheap though ...

[puretube]

Saw a few similar gadgets and isolated polarity flippers & up/dwn converters last year in Frankfurt @ MusikMesse in the studio/ keyboard accessories-department (mightabeen 2 yrs ago...).
SMPS with tiny HF transormers...
Certainly interesting for conflicting ground-issues, elusive supply-demands a.s.o. ("Humphrey").
But don`t ask me about their performance with "clocked" pedals or their supplies... (coz I haven`t checked them out yet...).

"Heterodyning" becomes more and more prominent, lately...

Daisy-chaining seems to make things even more noisy, lately...


prices might come down in foreseeable future...

[tca]

I took the "isolation >1kV" to mean it was an internal isolation set up.
I've speculated from time to time about designing a small 9Vdc to 9Vdc power supply.

What I don't get is 9VDC in to 9VDC out. Exactly 9V? There should be some losses (what current?!)... put more in than out, even with a passive filter.

[R.G.]

What I don't get is 9VDC in to 9VDC out. Exactly 9V? There should be some losses (what current?!)... put more in than out, even with a passive filter.

That's another reason it's not a passive filter.  Although "advertising speech" would say that 8.8V is near enough to 9V to call it that for most advertisers.  Inductor-capacitor filters can filter with very little DC loss, but what they cannot do is isolate. If it's DC in, DC out, and is isolated, there is a switching power supply in there to get it across the isolation barrier by converting it to AC, then reconverting back to DC.

I looked up some of my old notes, took a gander at some modern power supply controller chips, and did some calculations. Yep, you could put a 9Vdc to 9Vdc isolating, switching power supply in a case that size.

But again, as a practical matter, it has to be cheaper than another whole 9V power supply to be of any use.

[tca]

^ Thanks.

... something like this http://uk.farnell.com/xp-power/iw1209sa/converter-dc-dc-2-1i-p-1w-9v/dp/1435397. This meets the specs they quote.

So you need a 12V (>9V) source for that to work.

[drolo]

The specs for this one indicate that the minimum Input voltage can be 9V. Perhaps that's why this one is more expensive than other models that specify minimum 12V (about 4.50 Euro). These might actually be interesting. There are even models with similar price that convert 5V to 9V. I wonder if these would actually work if you regulate the supply down to 5V and use the converter to get 9V isolated?

Need to read more about how these DC/DC converters work .... I wonder if anybody in our forum has experience working with such devices for power supplies ..

[PRR]

There was a module about that size which took +5V and output +/-9V (bipolar) at a few dozen mA, with large isolation. It was used on EVERY coax EtherNet card, even the $15 ones. When coax Ether went out of style, the excess was dumped for $2 or less.

Some years back I used a module "like" that, rather like the Farnell $10 lump, to jazz-up a 6V battery to higher voltage. IIRC I found +5V to +/-18V(36V) and stacked two to get near 86V. The 6V was a half-Volt over the module's rating but it worked for some hours.

As the master says, it is trivial (if you own a coil shop) to wind-up an inverter/rectifier to transform DC across an isolation transformer and back to DC at a voltage of your choice. Either proportional to input, or regulated. The module is a few bucks, the case and plugs could be more, and the *promotion* to get the word out and get enough sold to be worth-your-while could be more than the module and case altogether. (Plus you need to be in Banjo World but BW takes 30%-60% of the retail as their profit, so you won't get much.)

[R.G.]

I messed with this a bit last night. Looks like an EE8 or EE 10 core will do the job, or an 1109 pot core. Getting the gap and turns right for a regulating flyback design is touchy, but if you do that you can use a single 8-pin chip as the controller and get an integrated internal switching MOSFET to do the hard work.

A better design if you can afford the extra parts would be to do a "dumb" switcher - just chop to AC, transform it across the isolation barrier, rectify to DC, and then regulate with a 78L09 or LM317. This is "better" in the sense that the hard part, the transformer design and building, can be done with off-the-shelf parts more easily, and is not as touchy in terms of getting everything right to make the output low noise.

Either way, it's got to run at about 100kHz or more to be usably small, and if your part layout and current paths are not skillfully done it will make lots of RF interference. It will, of course, need to be tested to be sure it doesn't violate your particular country's laws on RF emissions and testing for those emissions - as EH knows very well.

[drolo]

A better design if you can afford the extra parts would be to do a "dumb" switcher - just chop to AC, transform it across the isolation barrier, rectify to DC, and then regulate with a 78L09 or LM317. This is "better" in the sense that the hard part, the transformer design and building, can be done with off-the-shelf parts more easily, and is not as touchy in terms of getting everything right to make the output low noise.

Looks like an interesting project. What kind of transformer would be used in this method?
Just asking out of curiosity. I'm still reading up "flyback converter for Dummies" and still not sure I understand the difference between that and what you described above ...

[R.G.]

Looks like an interesting project. What kind of transformer would be used in this method?
Just asking out of curiosity. I'm still reading up "flyback converter for Dummies" and still not sure I understand the difference between that and what you described above ...

Flyback tranformer converters are not really transformers in the normal sense. A flyback is a simply load/dump of an inductor. Isolated flybacks just take the output from a different winding than the input.

For a simple inductor flyback supply, one charges the inductor up to some current, storing energy equal to LI²/2. When you try to stop the current, the inductor reacts by inverting the voltage direction across the inductor and raising the voltage magnitude until the same current flows that was flowing before you tried to stop the current.

Quick illustration: You have 12V and want 24V; you have a 1H inductor, a power MOSFET, a fast diode, and a filter cap. So you connect one end of your inductor to +12, and pull the other end to ground with the MOSFET. The current increases by V = L di/dt, or di/dt = V/L; which means, the current rises at a rate (change of current per unity time) of the voltage divided by the inductance. So the current rises at a rate of 12V/1H = 12A/Sec.

If your inductor saturates at 1A, you have to stop there or burn up things, so you only turn the MOSFET on for 1/12 second, when the current is 1A.

When you reach 1A, the energy in the inductor is E = LI²/2.

When you turn off the MOSFET, the magnetic field cannot change instantly, so the voltage across the inductor inverts and it PULLS until it can keep the same current flowing in the winding. So the MOSFET drain end flips up above the 12V supply and the voltage rises to **ANY** voltage that makes the same 1A flow. This can easily be 100s or 1000s of volts, and is in old-style car ignition coils. But you have cleverly arranged your diode to let current flow from the low end of the inductor to the filter capacitor, so when the inductor end gets to a voltage higher than the cap, the voltage is clamped by the cap and 1A starts flowing into the cap through the diode. Current flows into the cap until the energy in the magnetic field of the inductor is exhausted and cannot keep the voltage across the diode up. So the cap energy goes up by the same amount the inductor magnetic field was holding, or

E = LI²/2 = C*V²/2
and you have moved a packet of energy into the cap which raised its voltage a bit. How much the cap charges depends on the size of the capacitance.

If you do this over and over, you transfer a power P = F*E, where F is how fast you do it (the switching frequency) and E is the energy in the inductor per packet.

For isolation, you simply recognize that it's not the copper windings that hold the energy, it's the magnetic field, and you can have two coils, one to fill the magnetic field up, and a second where it comes out. Both coils must have the same volts per turn, so if the second coil can let the energy out at a lower voltage per turn than the first one, the energy comes out there.

To make all this work, you have to know pretty exactly what the inductance is and what the saturation current on it is, and stay under that. You also have to optimize the inductor for maximum energy storage if you're doing a bang-up job, so making it small AND tolerant of DC currents AND with good energy storage means a tricky job for designing and feeding it just the right frequency/time/duty cycle and so on. Also regulating. The output voltage is not related to the input voltage in any simple way, and will depend on the load, frequency, and primary duty cycle.

ACK.

The transformer-only way is much easier. You simply have - a transformer. You arrange to feed this transformer AC on its primary, and its secondaries get a voltage equal to the turns-ratio. The only trick to this is that you have to make sure your AC on the primary is really AC, no DC offset. That means taking some care in how you feed it chopped DC.

[puretube]

non-digital: