janneman said:
Mr. Jan Didden:
I think not only myself wonder about this question! From this title I raised up, a lot of response from the others. That means not only myself have a question on it.
As an engineer, it doesn't matter what gender is, as long as he is capable of doing the job. Indeed, you are not only capable, Just superior and outstanding, from the info and knowledge came out of you, I think you are a professor, your understanding of theory and circuitry is incredible. I think everybody agree with me.
Back to title, in Asia, Jan can stand for Jane or Janet, also from the way you " teaches " us, very detail, like my primary lady teacher, thus may mislead us that you are a lady.
Now, All cookies are cleared, there will be no more embarrassing.
Wish everybody Merry Xmas and happy new year.
In countries like Holland and Sweden, it is quite obvious that Jan is a proper name and a male name. Simliarly, it is probably (almost) just as obvious in english speaking countries that it is an abbreviation of a female name. I wouldn't be the least surprised if there are languages where it is an abbreviation of male names etc.
The sensmorale is that it is very dangerous to make assumptions about names between countries and between languages. What is obvious in one country is just as obviously not the case in another country. Even worse, names that are common and normal in one country might even be considered an offense in another country. I know a swedish guy who had middle name that is quite common here in Sweden. When applying for an american drivers license he asked if it was possible to have only a middle initial instead of the full middle name in the license. They said it was against the rules, but after telling them what his middle name was, they blushed and allowed him to use only the initial. (Since I don't want to end up in the sin bin in christmas time, I will not tell what his name was.
).
The sensmorale is that it is very dangerous to make assumptions about names between countries and between languages. What is obvious in one country is just as obviously not the case in another country. Even worse, names that are common and normal in one country might even be considered an offense in another country. I know a swedish guy who had middle name that is quite common here in Sweden. When applying for an american drivers license he asked if it was possible to have only a middle initial instead of the full middle name in the license. They said it was against the rules, but after telling them what his middle name was, they blushed and allowed him to use only the initial. (Since I don't want to end up in the sin bin in christmas time, I will not tell what his name was.
).
Question about Jung super regulator
Hello,
i have a couple of questions regarding Jung super regulator - more specifically, the circuit attached, from this post:
1. I understand that zener D2 is there to keep the output of the opamp well below its positive supply, but is it necessary to use a voltage drop equal to the reference voltage? Wouldn't a LED (1.9 V drop) be sufficient and less noisy?
2. If the reference (D5) was biased with a CCS insted of R8, would this reduce noise level? Has anybody tried it?
Thanks...
Hello,
i have a couple of questions regarding Jung super regulator - more specifically, the circuit attached, from this post:
1. I understand that zener D2 is there to keep the output of the opamp well below its positive supply, but is it necessary to use a voltage drop equal to the reference voltage? Wouldn't a LED (1.9 V drop) be sufficient and less noisy?
2. If the reference (D5) was biased with a CCS insted of R8, would this reduce noise level? Has anybody tried it?
Thanks...
Attachments
Last edited:
Bad (old) link, but here's a new one:
Alas, the migration of my website to new hosting fried a few links....
Try this one for ref and reg info:
References & Regulators | Walt's Blog 2014
Alas, the migration of my website to new hosting fried a few links....
Try this one for ref and reg info:
References & Regulators | Walt's Blog 2014
Point 1: Yes, the zener is very useful as a level shift and allows the op amp to bias around mid-ways-to-the-rails, when selected properly. Other means of creating this level shift are possible, for example, one or more higher-forward-drop LEDs. I can recall doing this trick in the 1997: ‘Low Noise Power for Analog Circuits‘, a ‘Walt’s Tools and Tips’ column, within the ED Analog Applications Issue of June 23, 1997, pp. 65-68. This regulator article is available on my website (see the link in post 110, above).
That said, it isn't entirely straightforward as to what op amp will work (best) within this circuit. Why? Both the input as well as output CM ranges must be adequate. The mentioned article goes into this, and describes methods to test for proper behavior. Recommended reading for all the op amp swappers out there. One reason the AD825 is so useful in this circuit is that it is free from input stage CM quirks (unlike older parts like the TL07x series, unfortunately).
If all of this sounds a bit intimidating, just use the designer-recommended parts, and stay out of possible troubles.
I'll address Point 2 in another post.
Walt Jung
Last edited:
The short answer is, no, CCS biasing will not reduce the noise.
The LM329, all by itself, generates around 75nV/Root-Hz, and has a 1/f noise corner below 100Hz. See "Zener Noise Voltage" from the device data sheet. This is true whether or not it is driven from a resistor, or a JFET. This specified level of noise is equal to or better than any other reference now readily available. This point was/is a large part of why this part was originally chosen. And, it is still technically the best all around part in this application. Hand selection of more esoteric devices might possibly get noise down some, but that isn't for everyone.
I have seen numerous regulator schematics using JFETs (even cascoded JFETs!) to replace the simple 5k R4 from Fig. 1 of "Improved Pos/Neg Regulators". The PRO of this is that one then won't need to worry about changing a feed resistor for other output voltages. But, I don't know anyone who does such a thing, as the output voltage is set once, and once only. So, this is a moot point. The CON is that you are wasting money on unnecessary parts using a JFET, and even if you do insist on using one for this function, you will then likely need to select a current set resistor for the FET, to make it supply the desired 1-2mA.
The low pass filter R1-C7 provides good filtering of noise, for any type of reference device.
Hope this helps clarify how to bias an LM329 in this circuit.
Walt Jung
Walt, I've had several instances of the Super regulator (both positive and negative) which would regulate without load, but when loaded with 100 ma the output voltage would collapse. The solution turned out to be decreasing R8 to get 4 ma because of marginal gain in the particular D44/45 devices ( sourced from Digikey so hopefully the correct original devices).
I wouldn't expect a current source to lower the noise of the reference, but would it improve input rejection?
I wouldn't expect a current source to lower the noise of the reference, but would it improve input rejection?
Last edited:
Are you *really sure* it is low pass transistor gain?
I just looked at the D44 series types as spec'd by Fairchild, and the low gain charge seems questionable to me. Yes, you can find grades with low gain at 2A, but at 100mA, it is higher. Look at the D44H11 for example. Did you *measure* the gain for a couple of parts, and find that they do have lower gain than typical at 100mA? The 5mA from the current source as configured will provide up to 200mA with a worst case gain of 40 with D44C8, even more with a D44H11. So, I don't agree that the current source needs to be changed to get 100mA. So long as the pass xstrs meet above specs, that is. I just had an exchange with a user on my blog (see link below), who was trying to get 360mA.
As for a current source feed to the ref diode improving input rejection, no it won't. It isn't connected to the input, it is connected to the output.
I just looked at the D44 series types as spec'd by Fairchild, and the low gain charge seems questionable to me. Yes, you can find grades with low gain at 2A, but at 100mA, it is higher. Look at the D44H11 for example. Did you *measure* the gain for a couple of parts, and find that they do have lower gain than typical at 100mA? The 5mA from the current source as configured will provide up to 200mA with a worst case gain of 40 with D44C8, even more with a D44H11. So, I don't agree that the current source needs to be changed to get 100mA. So long as the pass xstrs meet above specs, that is. I just had an exchange with a user on my blog (see link below), who was trying to get 360mA.
As for a current source feed to the ref diode improving input rejection, no it won't. It isn't connected to the input, it is connected to the output.
Thank you for the new link Walt. Walt's Blog 2014 | Walt Jung's 2014 Blog and Info Archive Have for 30 years plus found your writing insightful and very useful. Regards
Thanks for those very kind words, Triodethom. Makes feel a bit old, however. In fact, this year marks the 40 anniversary of IC Op Amp Cookbook.
... of which I have a 1st edition. How old does that make me!
Jan
Really, really old. I used to read your articles when I was a kid.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.