Voltage Regulation. Diodes or Voltage regulators?

[Little Ricky]

Next step in  my quest to increase power in Vox PF. I purchased a Signal transformer 241-8-36 specs are :36VCT (18-0-18) - 2.8Amp - 100 VA. Rectified 18v should bring me to 25-26v . I'm currently looking at the voltage regulation down to 15V. Initial thoughts were to use 7815,7915 regulators.  This is still an option but chassis space is limited and running leads to  the "regulator board/heatsink" has me concerned, each of the 4 leads will need to be at least 5 inches long. i've compared some of the small Vox ss power supplies and was going to use a regulator set up based on the AD30VT Then noticed that the AC30VR and VR30 use 3W 15V diodes (1N5352BG) with 3W-10W resistors. My plan is to power an LM3886.



Researching further is see that Fender uses a similar setup on the Frontman 65R and 212R amps.

Is the combination of higher wattage diodes and resistors "safe"? It seems to be a fairly common practice.

[Roly]

Zeners are popular in manufacturing because they are cheap.  In my experience they are also somewhat unreliable (since they generally run fairly hot, not something desirable in a semiconductor).  They also almost always fail shorted and take their series resistor with them.

Three pin regulators are (almost) indestructible due to their internal protection against over-current, over-temperature, and zombie attack.

There is one trap however; notice on the Vox AD-30VT circuit C20, 21, 22, 23, and 24.  I always fit these right at the regulator since they can be prone to RF oscillation otherwise.  If you take care of that (and the surrounding diodes) you will find they are childsplay.

Since the preamp current doesn't vary much you can make them run a bit cooler by selecting a resistor on the infeed to drop the actual input voltage a bit to something over their rated output voltage, say plus about 5 volts (don't go too close 'tho) e.g. R39 and R40.  These are sometimes called "economising" resistors but what they are really doing is moving heat from a more fragile component to a more robust one.

Note that the tab on the -ve one is not ground as it is on the +ve one, the pinout is different.

HTH

[Little Ricky]

Thanks. So as long as the regulators and all associated bits are together it doesn't matter how far away that separate board is?
I noticed the PF and AD30VT switch the final capacitor diode pair. Does it matter which comes first?

[Roly]

So as long as the regulators and all associated bits are together it doesn't matter how far away that separate board is?

I noticed the PF and AD30VT switch the final capacitor diode pair. Does it matter which comes first?

1) No.  The only important thing is that C20-C23 (not C24, my error) are as close as possible to their respective regulators.  I often mount them right on the regulator leads but that's a bit of overkill; as long as the total lead path isn't more than an inch or two should be fine.  Once you do that you can run feet (or metres, depending on your location  ) of wire to them; they just have to have this high frequency bypassing nice and close.

2) No.  They are all in parallel and the diodes will perform the required clamping function anywhere along the lines.


I'll just mention that short and solid low resistance wiring is required from the power transformer, through the rectifier and main filter caps because there is considerable AC ripple around this path, but after there just thick enough for the current flow (one circular mil per peak amp, minimum).


With a supply of +25V and a regulated output of +15V, you need about 5 volts across the regulator so it can function, leaving about 5 volts excess to drop in the "economising" resistors.  To work this out you need to know what the current draw will be on your 15 volt lines, then apply Ohms Law;

R = E / I,

R = 5 / preamp current (if I is in mA then R will be in k ohms)

Really this is only needed if your regulators are getting uncomfortably hot in normal operation.

HTH

{now I'm off for a late lunch}

[Little Ricky]

Thanks Roly.

[Enzo]

I have a question.  If the existing 15v supplies work now, why change them?   Your LM3886 doesn't use them and won;t care how they are regulated.  You have an existing circuit, you can measure the zener draw plus circuit draw by measuring voltage drop across that resistor.  If you increase the main power supply voltages, then adjust the resistor value.

Yes, three-terminal regulators regulate "better" and don't have the same weaknesses as the zeners.  Bur all that is relative.   If I design something from scratch, I use IC regulators.  But if I have an existing circuit, I rarely find it worth the trouble to rip out the old zener circuit and cram Vregs in their place.   Just my opinion.

DC power supplies do not care how long the wires are - I mean within reason.  Don't coil up three feet of wire in the corner or anything.  DC doesn;t radiate anything, and the low impedances of power supplies will not pick up anything.  It DOES however matter how and where you ground them.

[Roly]

Good points @Enzo - if it ain't busted, don't fix it.  :tu:

{Back in the pleistocene era I was building a (7400-series TTL) logic system for some industrial controller and it was going simply bananas.

With the deadline looming I eventually rolled out the big guns, a 100MHz CRO, and discovered that one of my un-bypassed three pin regulators was taking off at around 50MHz(!) and very effectively modulating the 5 volt supply line, invisible on the modest bench CRO.  I remember one of the Hams on the crew wondering out loud if this could be exploited as a six-metre power amplifier. ( :duh)

Ever since I have always fitted a couple of 0.1uF disk ceramics o.n.o. right on the regulator pins, VHF style.  There is actually a note in _{tiny print} at the very bottom of the NS datasheet about this where "more than a short distance" quantifies to more than a six inch loop, i.e. three inches away.}

[Little Ricky]

Right. I  am looking for different opinions . I understand the voltage Regulators  would be the correct way to do the job. As mentioned my other option was to replace the low spec resistor/diode circuit with higher value parts.  it seems like they are both viable solutions to the problem