Personally, I like the mods implied by OS in #56--- it addresses excessive heat and supply voltages simultaneously.
It's an extensive revision to introduce in the midst of other existing problems, but might be palliative at the same time. A tough call, but Dan carries the burden.
It's an extensive revision to introduce in the midst of other existing problems, but might be palliative at the same time. A tough call, but Dan carries the burden.
Hot transistors instead of hot Zener diodes. No difference in dissipation so just moving the issue. The main cause: the too high Vcc+ and Vcc- that are converted to +/-15V by dropping voltages.
Sorry
Sorry
It’s easier to cool transistors than it is resistors or zener diodes. It’s a little black metal thing called a heat sink. Transistors are made to mount to them. Resistors and zeners not so much.
Sorry, what do you mean what are Vcc? They’re supposed to be a +/-15v, but are measuring +16.6v and -17.2v. The limit of the NE5532 is +/- 15v. So it looks like they are running it right on the limit. I’m wondering if the extra 3.8v between the two is enough to damage the ICs.
Dan
No the thingies called Vcc+ and Vcc- in the schematic. The amplifiers supply rails.
If you solve the issue at the root also stable +/- 15V is no issue anymore.
BTW NE5532 has a maximum voltage of +/- 22V.
If you solve the issue at the root also stable +/- 15V is no issue anymore.
BTW NE5532 has a maximum voltage of +/- 22V.
I am absolutely open to modifying the circuit. Just the ability to play around with it and learn from it makes it worth it enough for me. So what would I be doing? Would I remove the large resistor and the one Zen or diode and this would go in its place? The zener would just set the voltage while the transistors handles the current. Use a 15v zener? I’m guessing in this instance 1/2w would work? Any recommendation as to what transistor would be best here? I am guessing that a TO220 would be best, is it possible a TO126 could work? TIP41, MJE243, ???
Edit: if I went this route is there any negative to dropping the voltage a little bit to say 12v? Maybe a little higher like 13 or 14?
Dan
The main causes are the too high difference between Vcc+/Vcc- and creating the needed +/- 15V by use of solely Zener diodes. That are 2 design imperfections at the same time. The suggested solution means no very hot resistors and Zener diodes anymore but instead hot transistors 😉
7815/7915 would also be OK (IMHO better). One must know the current draw in all cases for a good solution.
OK, I won’t interfere anymore with suggestions. Ciao!
7815/7915 would also be OK (IMHO better). One must know the current draw in all cases for a good solution.
OK, I won’t interfere anymore with suggestions. Ciao!
Last edited:
A transistor has to pass only the current drawn by the circuit; a series R plus Zener has to include current margin plus Zener current, so a pass transistor is advantageous re total dissipation.
I recommend considering 3-terminal regulators. They include thermal protection.
I recommend considering 3-terminal regulators. They include thermal protection.
True but still 15V needs to be dropped in each leg. Even with relatively low current this means dissipation.
3 terminal regulator or pass transistor, you need heat sinks of 15C/W or better on each. This should be practical. It’s better not to have the dissipation, but the only way to avoid it is another transformer. If using pass transistors, drop the zener current to 5-10 mA, and use any sustained-beta driver types. Those will have a gain on 100+, over a 1 mA to maybe 200mA range. You can’t expect all old BD or TIP types to be that well behaved.
NO. Transistors are just series regulators . Zeners no longer shunt. Cold circuit , even at 1A. 500mA can even run on a non-sinked TO-126,
just get barely warm. You actually trade both a hot resistor AND zener for a slightly warm transistor. The load is the shunt... very little gets
"wasted".
BTW , if you use an oversized zener cap - you have a cap multiplier/regulator hybrid. Most high beta trannies will work @1A with just 2-3mA
zener current. If you are only using the +/- 15 to run a few op-amps , I have used just a TO-126 or even a to-92 - just a little warm...
OS
Last edited:
NO. At 500 mA the dissipation of the transistor is 7.5W. No way a TO126 can do that without heatsink.
Luckily the current is way lower but it would be a good thing to measure the actual power consumption.
Luckily the current is way lower but it would be a good thing to measure the actual power consumption.
We already know current draw is less than ~ (30-15)V/180R = 83mA, corresponding to less than 1.25W.
The actual load current will likely be well under half an amp. If it is up at 500mA you still get a very hot transistor as JP states. The load current is probably in the 100-150 mA range, which necessitates a very hot shunt regulator or a moderately hot series one. The moderately hot 3-ish watt series one is tolerable with a 2 inch square heat sink that has a couple fins on it.
More likely around 50 mA. That is still 1.5W of heat so 0.75W per transistor. Way better than just with the Zener diodes. I would definitely use either TO220 transistors or 7815/7915 as these have adequate tabs to screw on heatsinks.
My suggestion is more expensive but 7815/7915 and 2 heatsinks (both needed anyway, extra MCFM50/15 transformer or not) cost next to nothing. Nice solution compared to the original circuit.
My suggestion is more expensive but 7815/7915 and 2 heatsinks (both needed anyway, extra MCFM50/15 transformer or not) cost next to nothing. Nice solution compared to the original circuit.
Last edited:
I don’t like zener shunts because even three quarters of a watt conducted ONLY through the leads will eventually oxidize the solder joint to failure. With a transistor on a finned heat sink at least there is another thermal path for most of it and the solder joints have a chance at lasting a decade or more.
Looks like there is a way to stack another transformer on there. Sometimes there isn’t.
Looks like there is a way to stack another transformer on there. Sometimes there isn’t.
Yes it is a plain bad design practice but applied a lot in cheap(er) devices. I once worked at a Japanese audio brand and we had amplifiers with JFETs regulating the +/- 70V rail supply voltages to lower voltages that desoldered themselves because of the heat (despite pretty large heatsinks). I replaced quite some Zener diodes in the described resistor/Zener diode scenario too. When the Zener diode goes to semiconductor heaven the load will see about full rail voltage leading to catastrophe. It may be cheap, solving the catastrophe usually is not.
Some manufacturers took the effort to have 2 x 12..15V wound extra on the main transformer. Practically never such issues. Try to find such a multivoltage transformer today 🙂
Some manufacturers took the effort to have 2 x 12..15V wound extra on the main transformer. Practically never such issues. Try to find such a multivoltage transformer today 🙂
Last edited:
This post is my recommendation, TO-220 with clip-on heat sinks. Under 1.25W expected dissipation in each, built-in current/thermal limiting for catastrophes.
Here is the original schematic, I measured the current at both places marked with a red “X”. Right behind R79 and R80.
Here are the results, the 289 is measuring the +15v and. The 87V is measuring the -15v. At power up they were at about 75 mA each and settled to where they are after about 10 seconds or so. So 70 mA is roughly where it’s at. I hope this is accurate. Clearly im still having an issue and hope all of the ICs are drawing.
There seems to be some disagreement on how it should be converted, but if I were to use the methods that was pictured in post 56 is this how it would go? I only drew out the positive side. So the same for the negative side.
I eliminated R79 (80R) and wasn’t sure if I should take out the following 470 uF cap (C27). If that looks appropriate I guess I would just need to know what values? There are two capacitors and a resistor there with unknown values. Would I use a 15v zener, if so is 1/2w okay? Someone mentioned possibly lowering the voltage a little 14v, 13v, 12v thoughts? Just make it a little safer for all of the silicon. Lastly, appropriate transistor. Looks like TO220 is recommended. What I currently have on hand in multiples for TO220, as far as BJT.
TIP41c/TIP42c
KSC2073/KSA940
MJE15032/MJE15033
MJE15034/MJE15035
2SC5171/2SA1930
KTC4370A/KTA1659A
2SC3298B/2SA1306B
2SC4883/2SA1859
If it’s helpful at all I do have a bunch of 7815s and 7915s, a bunch. If it is felt the regulator is a better way. I hope something above will be useful.
Also, does anyone think it’s possible that the +16.6 and -17.2 volts may have damaged the NE5532s?
Thanks again,
Dan