• Please review our updated Terms and Rules here

A Question About Voltage Regulation

C

clh333

Veteran Member
Joined
Feb 22, 2015
Messages
1,443
Location
Cleveland, OH, USA
Long story shortened: After finding a copy of the E&L "Bugbook" I found an LD-1 "Pencilbox Logic Designer", which the book references, on eBay. The nominal voltage for power input is +5 VDC but looking at the power input section I concluded it was regulated and could accept more voltage - within reason, of course. I plan to power it with a battery of 6 NiMH AA rechargeable cells.

At first glance I thought the TO-220 transistor was a 7805 regulator but it is not; it is a TIP29A, which is typically used for amplification, usually audio. As I understand from a little Internet research this was a common setup in the days preceding the 7805. The LD-1 dates from the '70s; I don't know how long the 7805 has been around.

I don't have a schematic but it seems the input regulation is composed of four components: the TIP29A, resistor, diode and Zener. If I understand the theory correctly, the TIP29A is not performing the regulation, however: Its role seems to be current amplification to provide enough current to supply the diode which actually regulates voltage. I'm assuming that's the role of the Zener diode pictured at the left, and the other diode is just for protection against reverse bias.

But this is all just speculation on my part: If anyone can explain it better I'd welcome their input.

Thanks,
-CH-

PencilBox.jpg PowerInput.jpg
 
No, the transistor is acting as a "pass" transistor, using the zener as a voltage divider reference for setting the base bias level.

voltage-regulator-using-transistor.gif


Read the theory here

Regulation with this circuit is better than nothing, but adding feedback (sense) is even better. You can do this by adding an op-amp, such as the 741:

voltage-regulator-using-op-amp-and-transistor-circuit.gif
 
You missed the extra diode. I suspect it was for additional drop from the zener to the base to be closer to 5V. I don't have the actual diagram.
Dwight
 
Thanks, guys, for your responses. As the board is pre-wired I will leave things as they are, but I see the role of the 741 using feedback from the output. In that case does R1 = R2?
I'll have to test this out on a breadboard.

-CH-
 
Since the op-amp is essentially functioning as a comparator, you want it to trigger on the difference between the zener diode and the output. So V₂ in the above diagram would be about the same as Vz when all's right with the world.

Working from that, we can see that the voltage divider formed by R₁ and R₂ needs to have, at its junction, the same output as the zener voltage. The output voltage Vout will be what's developed across the combined resistance of R₁+R₂, and the voltage V₂ where the two meet will be R₂/(R₁+R₂), which will be the same as Vz.

So, you can compute the output voltage by taking the above and choosing values for two quantities, say, 3.6V for Vz and 10K for R₂ and solve for R₁, given the desired output voltage. Or, just put a pot across the output, run the wiper to the op-amp and adjust to whatever you want. (The pot can have a fairly high value, say, 20Kohm, as the op-amp input draws microamps.

This little circuit is the basis for most simple linear adjustable regulated power supplies.
 
Zeners below 6.7V or so have more slope to the zener band. It is best to look up their current rating and adjust R1 for that. Although, not shown, you can use a larger zener voltage and a resistor divider to the opamp. This would allow regulation to less than the zener voltage. All or part of R2/R3 could be a pot instead of fixed voltage.
741 opamps to have restrictions on their input voltage relative to their positive and negative rails. Do look these up when determining what range of voltages you expect to use.
Dwight
 
Well, not to be a spoilsport, but there are even better solutions. Consider using an LM723 for example, which includes the zener reference as well as the op-amp. I don't know when the 723 was first marketed, but it goes back to the 1970s, some time. One of those great ICs, like the NE555.
 
Do make sure to add some capacitance at the output of the regulator. With the 723 or opamp, it can be unstable if the feedback path is too high a frequency. 10 uf is usually sufficient for most loads.
Dwight
 
Thanks again for your input. Rather than take this discussion even farther from the VCFED site's main purpose, may I ask if either of you can recommend a site that will tolerate my tadpole chops as an EE? I do appreciate your suggestions but you gents are a long way from Electronics 101; discussion of this sort must require summoning the patience needed for dealing with a three-year-old.

-CH-
 
Dunno, but allaboutcircuits or eevblog might be good places to start. Electronic Design mag has a good series of introductions as well. It's been some time since I read EDN, but there's another possibility.
 
The transistor is acting as a circuit known as a Emitter follower, the emitter voltage will follow the base voltage, but the current sourced from the emitter largely comes via the collector, not the base. The ratio of those currents depends on the transistor's DC current gain, a poorly controlled parameter that could be anywhere fro 15 to 40.

So it is important that the base resistor can supply the required base current and more to keep the zener in conduction. But there is more to it:

The transistor itself is such that its base to emitter voltage has a voltage drop in the order of 0.7V (in this application) and its value decreases with temperature at about -2.1mV per deg C, similar to a diode. What this means is that if you had the base voltage pinned to an exact value, the output voltage from the transistor's emitter will increase with heating.

Zener diodes in the range if 5 to 6.5 volts tend to have a close to zero tempco, meaning the voltage is fairly stable with heating. Higher value zeners have a positive tempco and lower voltage ones its negative like an ordinary diode.

As suggested, most likely the additional diode was placed in series with the zener to gain the correct output voltage, but because of that diodes negative tempco it will also help stabilize the output voltage with heating because the reference voltage at the base will drop -2.1mV/Deg C counteracting the transistors' base-emitter voltage temperature dependence.

The transistor also has an intrinsic emitter resistance, think of it like a small resistor in series with the emitter, so as the current load increases, the emitter voltage will drop. This is why a voltage regulator works better with its output sampled and a feedback loop with an OP amp and preferably a very stable voltage reference, even better than a zener. This makes IC's like the 7805 very attractive as they are very temperature stable and tend to maintain a very stable voltage under varying current loads, and they even have overheating/shutdown protection. So they work better than the emitter follower, with less parts, they do require capacitors on their terminals.

Although the 741 gets slammed as obsolete, high offset voltages, inputs and output voltages can't go rail to rail etc and that they are rubbish compared to modern OP amps, there is actually a very good case for them... because, they are totally deaf at radio frequencies, even in audio systems they tend to slew rate limit at around 20kHz. It means for low frequency application circuits they never give any trouble at all with high frequency instability. I have a large collection of mil spec 741's and they are better made (manufacturing quality) than any plastic jelly bean modern OP amp, so I always use these for low frequency work. Here is a machine I made with some a while back to stabilize the mains voltage to my vintage computers:

http://worldphaco.com/uploads/THE_CONSTANT_VOLTAGE_MACHINE.pdf

So I would always encourage a novice to start with a 741, after mastering low frequency circuits suited to it, move to an RF capable OP amp after that and the fun will begin.

Edit: One other thing, if you ever stop to test them, at least all the LM358's I have, the output stage in them has moderate cross over distortion, so I never warmed to that OP amp. One of my favorite modern op amps is the TLC 272 linmos type and a great little OP amp for video and RF work is the AD8056.
 
Last edited:
But crossover distortion hardly matters in a voltage regulator!

If you want to go really old-school, try a µA709. I believe that it was the first op-amp in an IC (Bob Widlar's invention). I still have a few.
 
Hugo,

If I ever get as far advanced as your CV machine I will be scanning my mailbox for notices from the Nobel Awards Committee.

-CH-
 
Once upon a time power supply module kits were common at electronic shops. I bought this one some years ago at the electronic markets in Tokyo, because it was a "classic".

I uses an LM723 IC for the voltage feedback and the current via the 0.5R output resistor is also monitored. I has connections to the E,B & C to parallel more 2N3055's, if you need higher current and the rectifiers and input filter cap are merely increased for the requirements.

Here is the LM723's data sheet:


https://www.ti.com/lit/ds/symlink/lm723.pdf


A good rule of thumb, with a full wave rectifier on a 50Hz power system, but close enough on 60Hz;

10,000uF filter capacity has one volt of ripple at 1 amp of current.

Or for arguments sake about 5000uF would have 2V of ripple at 1A, or if the current was 2A then the 10,000uF would have 2V of ripple etc etc.

It is always important that the input filter cap is large enough to support the current, or the voltage will trough too low between the peaks of the full wave rectified waveform and ripple will appear in the regulator's output. This is why you will see the amazingly enormous looking filter capacitors with values like 50,000uF in the power supplies of many vintage computers prior to SMPS supplies taking over.

I once built a radio with a very unusual and interesting IC, the LM372, so I just had to use one of my beloved mil spec 741's for the audio driver. These 741's have gold plated leads and I cannot bear to solder to them so I use high quality sockets with gold plated pins. I may well be the last person on the planet who still loves the 741:

http://worldphaco.com/uploads/THE_TRF-ONE.pdf
 

Attachments

  • lm723.jpg
    lm723.jpg
    63.4 KB · Views: 1
  • LM7232.jpg
    LM7232.jpg
    126.9 KB · Views: 1
Believe I mentioned the 723, complete with datasheet back in #9, Hugo.

I've built switching regulators using the 723 also, back before the NSC "Simple Switcher" line. Not a lot of external components, but you've got to be a bit fussy about the capacitors. Ripple is about 250KHz, so characteristics count.
 
Chuck(G) said:
Believe I mentioned the 723, complete with datasheet back in #9, Hugo.

I've built switching regulators using the 723 also, back before the NSC "Simple Switcher" line. Not a lot of external components, but you've got to be a bit fussy about the capacitors. Ripple is about 250KHz, so characteristics count.
Click to expand...

Yes, that is what made me remember it and look for the module, but I forgot you posted the data sheet.
 
You must log in or register to reply here.
Back
Top