1590A/LM7809 regulated power supply

[Plexi]

I have a question which may prove unhelpful. if we have a transistor like the BC548, it is rated ~500mA maxx collector current and can be series-pass [?] wired to boost the current capacity of a regulator. can we wire say 3 transistors series-pass w/ one reg, and provide 3 outputs w/ higher current rating than the nude reg, or will we lose control of the regulation?

Give me a few hours to translate what you said 
So... I wait for the expert people to answer that.

[antonis]

Do I smell some guy trying to take something for nothing..?? 

You can always by-pass the regulator with current passing transistor(s) but you'll need Emitter ballast resistors (to equally share the current for common load  - not necessary for split outputs..)

If you use p-n-p pass BJT, Emitter is tied to unregulated IN, Base is also tied there AFTER a resistor (it's value is set to create a 600mV voltage drop across it for the current you want to start by-passing - take in mind the minimum necessary current for regulator operation) and Collector to OUT (load)..

In case of n-p-n pass BJT, Collector to unregulated IN, Base to regulator's OUT (via current setting/limiting resistor) and Emitter to load..
(take in mind the V_BE voltage drop - e.g. for 9V regulator, final OUT should stand on 8.4 Volts..)

[greaser_au]

I have a question which may prove unhelpful. if we have a transistor like the BC548, it is rated ~500mA maxx collector current and can be series-pass [?] wired to boost the current capacity of a regulator. can we wire say 3 transistors series-pass w/ one reg, and provide 3 outputs w/ higher current rating than the nude reg, or will we lose control of the regulation?

It depends the on the configuration.   using Antonis's two positive regulator examples (thank you!):

With a PNP series pass transistor, the base bias is derived across a resistor that feeds the regulator, so the regulator current drain is what controls the transistor. We are effectively paralleling the regulator's internal series pass transistor (is the external resistor the internal transistor's ballast resistor?  )...  That is:  the operating conditions of the transistor are intimately tied into the operating conditions of the regulator- I'd say you could not split the outputs. As the output monitors the collector, and the regulator is a closed feedback loop, the output voltage should be just about as stable as it would be from the regulator alone across the range of sensible current output.

WIth an NPN series pass transistor, all it needs is a stable base reference voltage, and it will try to hold the emitter at (about) 0.6V below that.  Just a standard resistor and zener diode regulator on the base would suffice...  so yes, a bunch of transistors with separate emitters could be used.  in this case the transistor's rE will cause the emitter voltage to droop as the current rises.

However, either way, watch out for the dissi.........................pation!    3V at 500mA is 1.5W.   a TO-92 has a junction-to-air thermal resistance of something like 200C/W, so by the time you're pulling 500mA, the device is on fire!

david

[antonis]

We are effectively paralleling the regulator's internal series pass transistor (is the external resistor the internal transistor's ballast resistor?  )... 

(with the risk of pouring oil on the troubled waters..)

Academically speaking, only if internal series pass element is also a p-n-p transistor and a resistor of equal value of the one between Emitter & Base of the external p-n-p series pass transistor is placed in front of outer pass transistor Emitter..

But yes, as long as by-pass current "triger" resistor actually works as an internal current pass limiter, we can undoubtedly consider it as a ballast resistor..!! 

[greaser_au]

(with the risk of pouring oil on the troubled waters..)

Academically speaking, only if internal series pass element is also a p-n-p transistor and a resistor of equal value of the one between Emitter & Base of the external p-n-p series pass transistor is placed in front of outer pass transistor Emitter..

Note the emoji (sorry,  I do have a perverse sense of humour)   The internal series pass device is an NPN in the usual 'reference' circuit diagram, and the 'sense' resistor  function here is completely opposite to a ballast resistor... 

david   

[antonis]

So, to shake hands we have to compromise the term "ballast" for "not_necessarily_equal_current_sharing"..

[greaser_au]

So, to shake hands we have to compromise the term "ballast" for "not_necessarily_equal_current_sharing"..

There's no need to compromise, I was messing around... (EDIT for clarity:  we can shake hands here without any need for compromise. your usage was essentially correct).

I first used the PNP 'boost' circuit that you outlined in a high school electronics project back in 1980, right out of the data section in the back of the Dick Smith Electronics catalogue (no doubt Duckpops will now tease me again about no longer being able to find my Dick, even though I'm sure it's in a box I left at my parents' house ).

david

[duck_arse]

I did say may prove unhelpful, and greaser, you've only yourself to blame for the following:

I do know where my dick is, and looked at it last night just before bed. that's when I realised I had a few legs in the air (emitters, in my proposed/dreamed configuration) that were doing nothing.

antonis: not so much something for nothing, but more for less.

[antonis]

antonis: not so much something for nothing, but more for less.

Sorry Duckie, I perstistently forget that your sink drains CCW... 

[Plexi]

Update


The caps will preferably be ceramic, and other stages will be added aside (copy / paste only).

About 78Lxx and 78xx, I was thinking: what about A's if I use, for example:
- 2 78xx
- 2 78Lxx

If I use only the 78Lxx, and not the 78xx stages, both are supporting 1A ...?
Or have a "relief" from other stages?
To avoid overheating.

[antonis]

78LXXs can handle a maximum current of 100mA (don't even think about 1A..)   

Most important is their power dissipation capability which is about 200^oC/W (as for all TO-92 devices and as above well said by greaser_au)
e.g. a 78L09 powered from 12V unregulated PS (which is the minimum working voltage difference) may handle 100mA continuous current but, for sure, can't do the same powered from 18V PS.. 
(at least, not without sufficient cooling - a junction temperature of 160^oC is marginally OK for a relatively little working time..)

To put it another way, 300mW should be considered maximun power dissipation for an "open air" non-smoker 78LXX regulator..!!

As long as you'll not avoid possible groud loop issues (by feeding them from a common rectifier..) I can't see the reason for copy/paste and not a larger PCB with a common ground plane..
(it will save space, cooper & wiring lengh..)

[anotherjim]

The regs have thermal shut down too.

Maybe it would be helpful to know what your amperage needs are?
A while ago I noticed I had a large stock of 10R 1% 1/8 watt resistors, so on a narrow offcut of strip board, I soldered 10 of them in parallel, fitted some scrap test probe cable and crocodile clips - made a 1ohm resistor. You can do better and give it 2.1mm output plug and a 2.1mm socket. The 1ohm goes in series with the + wire and just wire the 0v straight through.. Include a test clip hook at each end of the 1ohm.
Power each pedal with this thing in line. Measure the voltage across the 1ohm with the pedal working. The volts reading is directly also a current reading, so 10mV means you have 10mA.

[Plexi]

Good morning people!
Going back to this :-)

Antonis:
So, my idea of add some 78L to have one/two 100mA regulated will make no sense with the idea of use a power supply "whatever I found from 12 to 24v, with 1A or more".
Even if I use without one LM78xx..
BUT! it can be IF I place them after one LM78xx stage... that would be correct, but very redundant...avoiding again the 1A.

>As long as you'll not avoid possible groud loop issues (by feeding them from a common rectifier..) I can't see the reason for copy/paste and not a larger PCB with a common ground plane..
(it will save space, cooper & wiring lengh..)

Sure, you're right.
But now I ask: at least two or three regulators feeding, in any way... can stabilize more the signal? Helping each others?
Maybe betters if one day I upgrade the power supply to 18v/>1A to add one 18v slot.


Jim:
My amperage needs are low... at least if I'm planning to feed two delays and some OD with it.
Maybe I'll use the space left to give bipolar power, using a LM7908 and re-drawing the schematic (last minute idea).

About the 1ohm resistor in series before the pedal, direct to +: it's like a fuse?

[anotherjim]

About the 1ohm resistor in series before the pedal, direct to +: it's like a fuse?

Yes, in-line in the + feed. With it in series with the feed to a single pedal, all the current that pedal takes must pass through the 1R. That current will cause some voltage to be dropped across the 1R. If you measure that voltage with your DMM and use Ohms law, you can calculate the current. The magic thing about 1R is it makes the voltage reading the same numbers as the current - no need to calculate. I=V/R. If R=1 then I=V. So 100mV reading means 100mA is flowing.
If you put the 1R in the feed to a daisy chain lead, you get the total current of all of the pedals on the chain.

Reading current is one area where I especially wouldn't trust a cheap/average quality DMM. Even a good pro quality DMM needs really good clean connections/test probes to be trustworthy reading milliamperes. Also switching the DMM to current can be bad for it (or whatever you connect it to) if you forget to change it back to voltage. On current ranges, the DMM is a very low resistance between the test probes - pretty much a dead short.

I like to use 10R (I=V/10) or 100R (I=V/100) value in the RC power input filters in my pedal builds - makes it easy to know the pedals current draw. Nothing to stop you doing it if the R is 47ohm or any other value. 1R happens to be a no math no-brainer, but 1R is unfortunately too small value to act as a filter with the input capacitor.

[Plexi]

Apply all the advices... thanks!
I added a bi-polar segment too.

Ok.. I need a second eye around here.
I think it's done:

[antonis]

Don't even dare to power it up...!!!   
(polarity-mind items are ready to declare you their official public enemy..)

Second eye sees the 220μF electro placed with wrong polarity...!!

Second mouth says you have to take a look - from time to time -  at components datasheets (79XX pin-out is ADJ - IN - OUT)

Despite the above, this isn't the way for getting negative output..!!
(or else, we would use 79XXs as voltage inverters..)

You can get it from a center-tap secondary with GND on middle lug and full wave rectifier's (-) connected to 7909 IN (pin 2) or from an "ordinary" secondary of twice the voltage you want for (+) & (-).

In any case, you have to "split" positive & negative voltages (with reference to GND) before any regulation..

P.S.
To hate me even more, I should point you that, despite any current circuit arrangement, you can't get -9V from a -8V regulator - like DC-DC converter..

[Plexi]

You got it, Antonis!
As I assumed, a total disaster 
I followed this layout:


You're right, I didn't check the LM7909 pinout, just used as reference in the layout...a HUGE mistake, my bad.

Sorry, at least, my typo: must be 7909 

[Plexi]

I'm in the right path?
Ignoring the voltage, huge mistake in pinout, and my last layout 

[antonis]

As for the center-tapped, it's OK..

As for wall-wart supply, it has to be AC..
(don't get tangled up in the usual DC wall-warts supplies for pedals or similar..)

I can't see any reason for R1 & R2, other than for "adding" their 5mA current draw to regulators quiescent current (?) (and maybe for C10 & C5 bleeding resistors when power off) so I don't take a stand.. 
(maybe a more experienced person in PS design could enlighten us..)

C1,2,3 & C6,7,8 power bank of 10mF each is  - maybe - an overkill..
(of course, in your scheme, it comes to smooth the great voltage ripple due to half-way rectification in combination with output current..)

If you want a current of 1A, CR1 & CR2 should be rated at 3A/50V or greater..
(I don't like designs where diode's maximun continous current rating is marginally the same with circuit's normal need..)

P.S.
Considering price of center-tapped & "ordinaly" transformers the same, the cost of 2 additional diodes is much less than the cost of elevated filter capacitance need - double diode voltage rating cost difference (if any) for half-wave rectification is negligible..

[Plexi]

Antonis...thanks a lot!

Update, sad update: 
I built it, everything was fine... until every sound got muted; check and the LED was on, without light almost off.
When I check for voltage coming out: 2,4v.
Nothing was burned and not even hot.

Changed the LM7809, and it start to run again at 9v.

What happened!?!?