My Experience With The 723
m (New Article moved to My Experience With The 723)
Revision as of 19:26, 23 May 2007
My Experience With the 723 Voltage Regulator IC
I came across the LM723 in the 1980s and thought it was much too complicated to use. I liked the 3-terminal regulators and stayed with them until I had to make a power supply that would handle more current than they could supply. In the mid ‘80s I’d come across a copy of an article called “The Many Talented 723" by Glen Prescott. I don’t know where this was published and I’ve been unable to locate it through Google. However I still have my copy and I decided it was worth scanning and cleaning up the figures, so I can share it with the world.
There are a couple of companies who make the device, that describe it in their data sheet titles as a “Precision Voltage Regulator”. This is what it is. If you were to design a voltage regulator from scratch, using discrete components. You’d want to end up with something like the 723. I’m not going to go into detail about it, because the data sheets and Mr. Prescott’s article do that very well already. I just want to share some of my experiences in the use of it.
My First Mixer Power Supply:
I built a little microphone mixer while I was in electronics school and I needed a power supply for it. I had a pair of LM723 ICs in a drawer, but I’d never worked with the chip. I decided that I should learn about it. The National Semiconductor Linear data book showed a lot of different ways to use it and decided to build a plus and minus 15 volt supply with external pass devices. I included trimmers for fine tuning the outputs and I had a little aluminum case that I could fit the transformer, rectifier, filter capacitors, and PC board into. I used a sharpie pen and drew the layout by hand except for the IC patterns. I etched the board and assembled it. It powered my mixer very nicely and I got an A in that class.
Last year I started experimenting with an Analog Devices AD7329–a 13-bit, 8 channel A to D converter which requires ±15V as well as 5V supplies. I dusted my little power supply off (it is over 20 years old now) tuned it on and checked the outputs. I thought I’d probably need to adjust the trimmers some, but even after sitting for the last 20 years, it was measuring +14.998V and -15.01V. I was very impressed and decided it was too nice to leave in a box. So it is out of retirement and I use it for my experiments on my bench whenever I’m working with op-amps that need ±15V.
Second Mixer Power Supply:
I service several recording studios here in Honolulu. One of my clients has a console whose power supply failed. It had failed before and I’d replaced the regulators, which were high current (7 amps) devices in TO3 packages. But the last time it failed I could not find replacements. I’ve forgotten the part number of the device, but it does not seem to be made any more and no one has produced a substitute.
My client was understandably upset. Recording studios need to have working equipment–especially the mixing console. I told him I’d need at least 48 hours if not 72 to build replacement regulators. He said to hurry.
I used the same circuit as in my little mixer supply above. I hand wired it on pad-per-hole perfboard and had the regulator cards complete in 30 hours. I got the positive regulator working with a 2N3055 transistor as the NPN series pass device. But when I tried to use the PNP compliment–2N2955 transistor, I could not get it to produce 15V at 7 amps. I finally figured out that the chip, since it uses the Vz pin in the negative configuration, just doesn’t have the same drive capability as it does when using the output pin in the positive configuration.
The clock was ticking and I was sweating, when I thought of trying a Darlington transistor. I had just purchased some a few days before, so I had them on hand, at midnight with no parts stores open. That proved to be the solution and so I switched the 2N3055 on the positive side for the NPN Darlington compliment of the transistor I had on the negative side. The devices are sold as a complimentary pair at a couple of local electronics stores. This was one of my design objectives, so in case of emergency, the parts are available locally.
The TO3 packages of the external series pass transistors went into the same holes on the heat-sink as the original regulators. A couple of stand-offs support the cards with the 723s. I connected the regulator inputs to the filter caps and the transistor outputs to the fuses, and the studio was back in business.
The 723 has built in current limiting, but it is designed as a positive regulator. So there is no provision for current limiting in a negative configuration. Also current limiting is dependent on a resistor which must handle the maximum current the regulator is designed to supply. In the case of large currents >1 amp, this produces significant heat. I left the current limiting out of my designs. I didn’t want to have only the positive half of the supply getting limited while the negative half operated with no limit. I reasoned that since the supply outputs were fused, there was already protection in place.
This is the case with all the regulator designs I present here. If you need current limiting and you just need a positive regulator, the 723 can easily accommodate you. Just figure out how much current you need to supply and how much over that you’re willing to tolerate, then calculate ILimit = Vsense / RSC.
The regulator is also capable of foldback current limiting where there is a “knee” region before full shutdown occurs. Personally I have never used the current limiting features, so I cannot write from experience. If the supply is to be used in an application where shorting or over currents are possible, I use fuses.
Low Voltage Supply:
When I began experimenting with embedded processors, I found that many needed supplies in the 3V to 4V range. By this time I had become fairly adept at using the 723 and since I had some still, I decided to use one, rather than investing in a fixed low voltage regulator. I went with a positive output using an external PNP transistor. Again, I added a trimmer pot, so I can adjust from around 2.75V to 4.25V. This little board (the transistor has a TO92 package) measures only 1" x 2" and has a power on LED indicator, and two–2-pole screw terminal blocks. It has been very handy.
“Big Bertha” Bench Supply:
In the world of professional audio, the Jensen 990 discrete opamp is a famous classic. Designed to work with ±24V power supplies, it is considered the “holy grail” of opamp designs. I have worked on MCI’s JH500 series consoles which have ±36V power supply rails–they use a high voltage opamp like the TI OPA544 and add a NPN/PNP output buffer to each. The whole console is filled with these–talk about an air conditioning bill. But having these high voltage rails means the audio signal can reach large levels of voltage swing and go onto the tape as hot as you can stand it. Of course this is analog. With digital there is no need for this. But there are audio people who prefer analog, and a high voltage opamp can add to the sound of the console.
I have another client who records to digital media, but mixes on an analog console. He had me modify the console master stage with discrete opamps running on their own separate ±24V power supply. He felt that it made the mixes sound better. I can’t say for sure, I was not exposed to listening to it every day, but I have no reason to doubt it. He gets a lot of work coming into his studio.
But the story here is about “Big Bertha”. I decided I want to design a modern version of the Jensen 990 and I might as well make it able to operate at ±36V. To do this effectively I needed a suitable power supply. I built one using two 36V transformers in series. I regulated it with a pair of 723s and 20A Darlington transistors. I don’t plan on drawing anywhere near that amount of current, I just had them on hand. In fact everything I used to build “Big Bertha” was residing in drawers and boxes at my shop. (One of the design goals was to make it without spending any money.)
Actually the unregulated supply is around ±56V. This is beyond the 723's input voltage limit. I could have made floating regulators (more on those later), but I decided to try “amplified zeners” instead. I came across the “amplified zener” in “The Power Supply Handbook” by the editors of 73 magazine. I’ve used it over the years, when I’d get stuck for a solution. It has several clever ideas that have proved very useful, the amplified zener being one of them. Basically you take a zener diode and a transistor and connect them so the zener’s cathode is at the transistor base, and the transistor collector is the zener’s anode. No resistor needed as long as you avoid the knee region of the zener. I used this trick to knock the unregulated supply down to around ±38V (the 723 max input is 40V). I have these squarish heat sinks that can hold TO3 devices and I stacked 4 of them together; 2 on top of 2, so my amplified zeners and the Darlington pass devices are contained in an area that is the perfect size for a 5" fan, should I ever need to draw so much current that things get warm. So far that hasn’t happened. I can adjust the output from around 12V to just over 30V. I know it isn’t up to the ±36V of the MCI JH500, but I decided that wasn’t really needed. If someone wants an op amp to handle that, I’ll use the OPA544.
I haven’t really had a chance to put “Big Bertha” to the test. Mostly I’ve used her to power breadboard circuits at around ±18V. This is what Amek and SSL consoles operate at and those are what my remaining clients have. The computer revolution put most of the recording studios in Honolulu out of business.
Still, I like the idea of having a supply that can give me that kind of voltage range with a rock solid regulation that doesn’t drift or sag.
The 723 is a versatile voltage regulator. According to National Semiconductor’s data sheet: “The LM723/LM723C is also useful in a wide range of other applications such as a shunt regulator, a current regulator or a temperature controller.” And “Can be used as either a linear or a switching regulator”.
I have only used it as a linear series regulator and I’ve been pleased with the results. I have just started experimenting with switching regulators. I built my first one with a TL494 PWM controller. I will probably try out the 723 in that mode of operation when I have some time on my hands and no other projects waiting. But that may be a while, so if anyone has actually done it, I’d appreciate it if you’d share the info.
I hope this article is useful and I welcome comments and ideas.
RAF 19:04, 23 May 2007 (PDT)