I am puzzled about raising the ground pin of a fixed regulator.
Documents at TI say you can do this but they only supply a tiny example and say little about it. Perhaps the TI fixed-regulator people don't speak to the adjustable-regulator people. There stuff should be cross-referenced. Even though documents are well-written, they need to provide some context to promote understanding.
Why do that when a 317 will do just fine?
I believe a capacitor is needed across R2.
The Whammy HPA build can configure the 7815 for 15vdc, 17vdc, and 21.5vdc. I see no benefit in avoiding a 317.
So TI is wary of capacitors. I think it means that if you want to produce a bunch of cheap PCBs for sale on Ebay, then don't bother with capacitors. On the other if want to DIY a few PCBs for yourself and friends, then select good capacitors for careful design.
Documents at TI say you can do this but they only supply a tiny example and say little about it. Perhaps the TI fixed-regulator people don't speak to the adjustable-regulator people. There stuff should be cross-referenced. Even though documents are well-written, they need to provide some context to promote understanding.
Why do that when a 317 will do just fine?
I believe a capacitor is needed across R2.
The Whammy HPA build can configure the 7815 for 15vdc, 17vdc, and 21.5vdc. I see no benefit in avoiding a 317.
So TI is wary of capacitors. I think it means that if you want to produce a bunch of cheap PCBs for sale on Ebay, then don't bother with capacitors. On the other if want to DIY a few PCBs for yourself and friends, then select good capacitors for careful design.
3) because otherwise you would end up with datasheets hundreds of pages long consisting mostly of cross-references.
You'd also end up having to modify all data sheets when a new part was released and needed to be included in the cross references. Modifying a data sheet is a big deal as it will often trigger a re-qual of the part at key customers, even if all that changed was the wording in the data sheet.
The intent of the data sheet is to provide the specifications for the part and its applications. An adjustable voltage regulator is one of the applications of an LM78xx regulator IC.
Tom
The intent of the data sheet is to provide the specifications for the part and its applications. An adjustable voltage regulator is one of the applications of an LM78xx regulator IC.
Tom
It is easy to set up, just a resistor to output / input from the ground pin, so it is cheap and reliable, and consistent in production.
I think you need a preset to do that on a 317, and the total parts count (= cost) is higher in case of the 317.
I think you need a preset to do that on a 317, and the total parts count (= cost) is higher in case of the 317.
Last edited:
No, you do not need a preset for a 317. The principle is exactly the same, only the reference voltage is 1.25V instead of 15V.
See the datasheet, peferably before posting.
See the datasheet, peferably before posting.
Parts count and simplicity on the shop floor also count in a factory, where most are used.
The more complex it gets, the more errors the monkeys on the floor will make.
Happens in all fields, which is why Japanese machines are easy to fix compared to German and British. They are designed to be easy to assemble, or the reverse, of course.
The more complex it gets, the more errors the monkeys on the floor will make.
Happens in all fields, which is why Japanese machines are easy to fix compared to German and British. They are designed to be easy to assemble, or the reverse, of course.
What parts? What complexity? The procedure and parts required for the LM317 are identical to that for the 7815. Only the values are different.
The TI 79xx has a more detailed example of the unfixed 7915. It has capacitors.
However, why use a pseudo-adjustable when better performing actual regulators are available.
However, why use a pseudo-adjustable when better performing actual regulators are available.
- Why choose 78xx/79xx when the better and also cheap LM317/LM337 are already in stock in your drawers? I have yet to meet a DIYer that does not have a few of these in stock (contrary to 7815/7915) as universal building blocks chosen for their good properties.
- Why “adjust” an 7815 when you don’t need its specific output voltage?
- Why not choose better regulators for an audio design without penalty or higher costs?
What tells us this about the design?
BTW it is Volt, Ampère, Watt etc. so 15V DC. Writing stuff right on a technical forum is fun. You should try it!
- Why “adjust” an 7815 when you don’t need its specific output voltage?
- Why not choose better regulators for an audio design without penalty or higher costs?
What tells us this about the design?
BTW it is Volt, Ampère, Watt etc. so 15V DC. Writing stuff right on a technical forum is fun. You should try it!
Last edited:
Meet one 😉
FWIW I always (and everybody I know) have a few LM7*15 in their parts drawers, not for any specific project, "just in case" because they are ¨so popular and widespread.
LM317/337? ,,, the few times I actually needed one, I had to special buy it on my next trip to parts supplier.
But maybe you rub shoulders with the top of DIY class, congratulations 😉
He’s a manufacturer. Buys what the BOM calls for (Or designs around what he can get at reasonable prices).
But I’m a DIYer and I seriously doubt I have any LM317/337 in stock. LM7812’s - yep. Mostly for running fans and 4000-series logic. I also don’t use 3 terminal regulators for audio power supply design at all. I was cured of that when one oscillated due to distributed capacitance and took out 44 op amps I had to unsolder and replace. The discrete design that replaced the chips didn’t oscillate.
But I’m a DIYer and I seriously doubt I have any LM317/337 in stock. LM7812’s - yep. Mostly for running fans and 4000-series logic. I also don’t use 3 terminal regulators for audio power supply design at all. I was cured of that when one oscillated due to distributed capacitance and took out 44 op amps I had to unsolder and replace. The discrete design that replaced the chips didn’t oscillate.
Replacing 44 opamps would change one's point of view ... I get it. I am sure that discrete power supply circuits would be fine in most audio designs. But extra filtering might be needed.
I believe the regulator manufacturers are trying to have it both ways. In one hand, they claim you don't need capacitors for stability, on the other hand they use them in most application examples. These regulators are sensitive devices and the 317-337 are most sensitive.
The problem was the local bypassing sprinkled all over the board to keep the op amps stable and not talking to one another over the power supply lines messed with the negative regulator. Zener diode and pass transistor to the rescue. No feedback, no oscillate. Just one little 2 ohm base stopper and it’s unconditionally stable. And starts up properly, every time with no tricks or finagling required.
The 337 may be the worst of the bunch. Try using one for a -Vg supply for a depletion mode RF power FET and see what kind of trouble you run into. KHz oscillations, latch up - and total loss of regulation when the power FET goes into rectification (positive gate current). Those RF power devices cost more than 44 LF353’s too, and just as dead.
The 337 may be the worst of the bunch. Try using one for a -Vg supply for a depletion mode RF power FET and see what kind of trouble you run into. KHz oscillations, latch up - and total loss of regulation when the power FET goes into rectification (positive gate current). Those RF power devices cost more than 44 LF353’s too, and just as dead.
ST has different style in their L78 series regulators, an op Amp and pot are used to set the output voltages.
But like above, everybody has their end use and purchase style, and many devices are sold at OEM level for lighting use with non standard voltages.
But like above, everybody has their end use and purchase style, and many devices are sold at OEM level for lighting use with non standard voltages.
One learns a new thing everyday. Almost all regulators need caps and so tell the datasheets one way or another (but it requires reading them on a per type, per manufacturer basis). Of the classic regulators the negative versions always are weaker and while one can try to use LM317/78xx without caps the negative versions will oscillate for sure. Dinosaurs LM317 and LM337 outperform dinosaurs 78xx/79xx for many decades now and sure they all need caps. In the higher priced (mid end) audio gear one nearly always could see LM317/LM337.
A serious audio design today would use better regulators though like LT3045/LT3094 which are an unbelievable amount better in many aspects. Bonus is that transformer voltage can be lower and losses into heat will be lower as well. Using 78xx/79xx, certainly with lifted GND pin, today in high performance audio simply is a poor design choice.
Last edited:
The 7815 devices are quieter than the 317 without caps - about 150uV wide band noise vs about 350uV. Where possible, I go for 7815 - less hassle.
Old school 3 terminal regs don't oscillate if you follow the application guidelines. I have never ever had a linear regulator chip fail since c. 1978 when I started using them (LM723 to begin with and then 3 term regs not long after that.) We used them all over the place in industrial applications - a lot of it in hazardous areas where there was no room for error.
Old school 3 terminal regs don't oscillate if you follow the application guidelines. I have never ever had a linear regulator chip fail since c. 1978 when I started using them (LM723 to begin with and then 3 term regs not long after that.) We used them all over the place in industrial applications - a lot of it in hazardous areas where there was no room for error.
I have a cheap 6VDC adapter that uses a 7805 with one resistor on the ground pin. The manufacturer was counting every component.
My audio designs use discrete transistor regulators.
Ed
My audio designs use discrete transistor regulators.
Ed