LED Power & Standby Indicators

[Steve Conner]

Steve - assuming 5ma will light the LED, then a 68K bleeder/current limiting resistor between a B+ of 355 and zener voltage of 6.8v seems to meet the criteria:

(355v-6.8v)/.005ma = 69,640 ohms

However, that 68k resistor needs to be able to handle almost 2 watts right?

348.2v * .005ma = 1.74 watts

Yes.

Last, I don't know what the zener's specs need to be. 6.8v * 20ma = slightly less than 1 watt, right? I guess it needs to "zene" at something less than 1/4 of that...

No, the resistor limits the zener current to 5mA. You know that because you calculated it. So the required zener rating is 6.8 * 0.005 = a couple of milliwatts only. The smallest zener available in stores is 200mW.

Also, would Enzo's diode in the power indicator circuit limit the flickering?

No. I think you misunderstood, Enzo recommends the diode in parallel with the LED to protect it from reverse voltage on the half-cycles where it doesn't light. To get rid of the flicker, you'll need to rectify and then smooth with a capacitor.

Last but not least, no one seems to be concerned about having less than 68K of resistance between the power rail and ground. It just seems like a bad idea to me. Plus, whenever I forget to disconnect my 100K cap-draining resistor it seems to affect both tone and voltages throughout the amp.

Why? Think of it as a permanent "cap draining resistor" if you like. You should have that for safety reasons. The only "bad" thing about it is the nearly 2 watts of heat it dissipates. If you connect it to the first filter cap node after the rectifier, the change in voltages should be negligible, because this node can supply plenty of current.

[GainFreak]

Go look at the schematic for a PV 5150. Steal their circuit, or something like it.

This works very well for me . It's simple and and no calculations are necessary.

[6267]

You could also LED bias a preamp stage that gets disconnected from B+ at standby (assuming standby cuts all B+ in this case, not just power tubes). You would get good gain from that stage, the LED will fade on and off at power on and power off, and you get two birds with one stone (or milliamp) as it were. Since the cathode circuit is low impedance, it doesn't pick up hum easily. You can thus put the bias LED on the front panel and run wires back to whichever stage is LED biased. I've found LED selection is somewhat non-critical, just grab a red or yellow one and run its anode to the tubes cathode and its cathode to ground.

[TheTinMan]

6267 - it would be fun to use an LED for the cathode of one of the preamp stages, but not for this particular build. At least not to start with - with later experimentation anything is possible.

Go look at the schematic for a PV 5150. Steal their circuit, or something like it.

This works very well for me . It's simple and and no calculations are necessary.

Copying known-good circuits works great for me.

Is this the PV 5150 you are both referring to? At the risk of further revealing my ignorance, it looks like a 12 volt supply is used for the LEDs. Could I use the 6.3 volt heater supply the same way?

Assuming LEDs needing 2.5 volts & 8 ma, does the attached circuit make any sense? I tried to adapt the PV circuit to use the heater winding as Enzo suggested for the power indicator, but that might unbalance the heaters or cause some other problem(s) that I don't know about.

Being ignorant of solid state electronics, I searched Mouser for the 2N3919 transistor shown on the PV schematic but came up empty there and a couple of other places. Would one of you either tell me where to find this transistor or give me a substitute part number?

I know that this is going off in a slightly different direction, so please bear with me a while longer.

Many thanks,

Chip

[Steve Conner]

The 2N3919 seems to be an obscure 80 volt, 10 amp power transistor from Semelab. That doesn't make any sense for the circuit application, so I can only assume that it's a mistake. Try a 2N3904 instead.

I also see from the circuit that the LEDs are powered off the AC input to the rectifier CR14-CR18, not the DC output. This seems strange, as it'll make them flicker, though at 120Hz this time. Maybe Peavey wanted them to go off immediately when the power switch was turned off, instead of fading out gradually.

Your circuit in the PDF is wrong. The status LED will go on whether there's B+ present or not.

[Enzo]

I don;t know what they were thinking. That is an SK3919 - a house number. SImilar is a 2N3391, but more common would be an MPS8097.

I don't notice LEDs flickering at 120Hz, but couldn;t we just slap a little cap across the LED then?

Your link is to the 5150-2 model, not the 5150, but the indicator circuits are the same, so it doesn;t matter which you look at.

A power indicator across the 6v should work OK. For the status light, all you need is some relatively low DC voltage. Without a center tap on the heater winding, the heater winding might be inconvenient as a source - no ground reference. But what kind of bias do you have? if you have cathode bias on the power tubes, use that. if you have fixed bias, you will have a diode making negative for that. Add a diode facing the other way and some filteration, and you now also have positive off the same winding. Use that to power the LED.

[balijukka]

You could also put the led at one of the cathodes.

jukka

[TheTinMan]

In the attached PDF "Plan B" is intended to be a corrected version of the PV approach using the 6.3VAC heater winding. My PT does not have a bias tap or a center tap for the heaters. However, there will be a virtual center tap for the heaters connected to an elevated DC voltage reference.

I am going to try Plan A first. Transistors are a mystery to me and it worries me to put one in between the heater supplies. Also, the B+/Standby LED will serve as an indicator of capacitor charge as well if I'm not mistaken.

The LED cathode is something I want to experiment with just for tone possibilities, but not right now. Also, it would not be a B+ indicator because the absence of the tube would make it non-functional.

Thanks,

Chip

[TheTinMan]

OK, I need to step away from this but Plan C occurred to me as a way of getting a rectified DC voltage source for the power indicator light.

The circuit already had a 50 VDC reference point for the heater's virtual center tap, so I reduced the resistor values to get 5ma of current to an LED at the "bottom".

Any thoughts on this approach?

Chip

[Steve Conner]

Yep, Plan C looks fine. Except the zener diodes should be 5.6 or 6.8V, not 2.5. Plan A looks good too, not sure about Plan B.

[Enzo]

What do the zeners protect what from?


They would keep high voltage off the LED presumably. But the only way that would get there is if the B+ went to 3kV, or if the resistor shorted, of if a wire came loose and touched the LED. With a zener in parallel, the zener is as likely to explode as the LED.

With the large value resistor, if the LED fails shorted, so what? it only handled 2v to start with. The 68k or whatever won;t care if it has to drop 2 more volts, so it won;t suffer. The only downside to a shorted LED is that it won;t light.

I'd agree that the bleeder path is through the LED/zener, so just add a dedicated bleeder right across the first filter and be done with it. Now bleeding doesn't depend upon the life of any semiconductor at all. Your pilot light doesn't bleed the amp.

[balijukka]

You may have to reduce that 100uF to somewhere closer 1uF to reduce hum.
I don't know exactly why, but if that cap is too small or too large it'll increase hum. There's a sweet spot there, that yoy have to tweak.

jukka

[tedmich]

in the spirit of "many ways to skin a cat"



supertex makes a voltage reg that handles 12-450VDC, and can be wired as a CCS, providing a nice stable 5-20mA for your LED. In a TO-92 package it costs $0.72

higher B+ (up to ~700vdc) could use a similar circuit with a depletion mode N channel mosfet feeding a regular VR in CCS config. but that seems excessive...

[TheTinMan]

in the spirit of "many ways to skin a cat"



supertex makes a voltage reg that handles 12-450VDC, and can be wired as a CCS, providing a nice stable 5-20mA for your LED. In a TO-92 package it costs $0.72

higher B+ (up to ~700vdc) could use a similar circuit with a depletion mode N channel mosfet feeding a regular VR in CCS config. but that seems excessive...

Ted - Thanks for suggesting an alternative. Unfortunately, I really am ignorant about solid state electronics. I do not know what a "CCS" is but I'll search for it. Also don't know what the "LR8" represents.

That said, I've drafted a layout for "Plan C" and am not entirely happy with how it worked out in the real world (of CAD ).

I'll look into this some more and come back with questions if need be.

Thanks again!

Chip

[tedmich]

Ted - Thanks for suggesting an alternative. Unfortunately, I really am ignorant about solid state electronics. I do not know what a "CCS" is but I'll search for it. Also don't know what the "LR8" represents.

That said, I've drafted a layout for "Plan C" and am not entirely happy with how it worked out in the real world (of CAD ).

I'll look into this some more and come back with questions if need be.

Thanks again!

Chip

Someone named "Chip" should not be daunted by SS stuff!

LEDs like to have constant current as opposed to constant voltage, which is what most power supplies (solid state or not) are usually designed to make. A "constant current source" (CCS) can be handy for supplying some circuits, even the cathodes of several tube designs (see Merlin's WEB site, the "Oddwatt" amp etc.)

The 3 terminal solid state voltage regulator is a very handy (and ubiquitous) tool in electronics, removing >75db of ripple from DC supplies, which can translate from mV -> nV ripple in the DC output. While usually constant voltage devices, wired as shown they become a CCS. The LR8 is just another 3 terminal SS voltage regulator, but one that can take 450VDC input, and I looked into it for a possible regulated tube preamp power supply. Its a little light on the current though, and they can't be paralleled in a simple way to increase current....and the old LCR rails work just fine! They can be used to make other things; the second circuit I mentioned can be used in a tiny device which makes 240VAC into mV ripple 5VDC with no tranny or big caps needed.