Tag Archives: DIY Stuff

Material prominent in DIY content.

GLED431: An Ultra Low Noise LED Reference Cell

Looking for a low noise reference circuit for an audio regulator, at 2.5V? But you’ve found bandgap circuits too noisy? Read on!

This GLED431 circuit is very simple, and works as a 2.5V shunt reference with ultra low noise.

Consider the simple circuit to the right, which I call the GLED431. Just 3 low cost parts, all easy to get. It acts like an extremely low noise 2.5V zener. On my setup, noise measures around 2nV/√Hz, so if you take out the measuring system noise, the actual noise is likely below 1nV/√Hz. Really quite good. We’ll have more on this later on, in 2016.

While the GLED431 performance is very high for noise, you will need to apply about 5mA (or more) to make it work. Yes indeed, this current threshold is much higher than that of the TL431. But, it also has around 1/100 the noise! Caveat(1): The voltage won’t be as tight as typical bandgap ICs, nor as low for temperature drift. Those are conscious tradeoffs.

Here are some Vout measurements on a sample set of 5 LTL-4231Ns, in the lab prototype shown, after 1 minute warmup:

#1: 2.5094V, #2: 2.5093V, #3: 2.5069V, #4: 2.5019V, #5: 2.5062V

Not too shabby! In the schematic, the leftmost R values are just as shown from lab tests, as trimmed for the  2.500V target Vout. Obviously, just use a single 150Ω RN60D unit for this R. Note that the forward voltage of the LTL-4231N green LED (LiteOn) and the Vbe of the ZTX951 (Diodes Inc.) conveniently add, producing the desired Vout of 2.5V. Caveat(2): These two parts should not be changed if you expect to get close to 2.500V!

In use, if you are building say, a 5V regulator, select a series resistor so that 5mA is supplied to the GLED431 cell (499Ω). With this, also be sure to select a very low noise op amp, and reduce all the surrounding resistances, so as to minimize their noise contributions.  Finally, be careful to minimize capacitive loading.

I am now releasing this simple version, as a Christmas present to the readers. Stay tuned for more, have fun with the GLED431, and have a great holiday!

Walt Jung

December 24, 2015


Gary Galo AudioXpress preamp series

We are pleased to note that Gary Galo has recently made a very large Guest Contribution to this website. This is in the form of a series of his notable preamp modification articles, that appeared in AudioXpress. The articles are listed below in the sequence they appeared. Click the individual link for a given article. A ZIPfile with all the articles is available as well, at the end.

The four (+) part series on modifications to the Adcom GFP565 preamp:

Part one, from the 11/03 issue of AudioXpress.

Part two, from the 12/03 issue of AudioXpress.

Part three, from the 01/04 issue of AudioXpress.

Part four, from the 02/04 issue of AudioXpress.

A follow up, from the 12/04 issue of AudioXpress.

Miscellaneous letters and corrections.

Another preamp article, on Gary’s design of a music library preamp, for the Crane School of Music, SUNY Potsdam.

A zipped file with all seven articles (~12megs). 

It is a real pleasure to be able to make this landmark series of preamp articles available to the readers of this website.

Our thanks to Gary for making all of this possible, and for executing some superb examples of true DIY craftsmanship!

Walt Jung




Klaus Noll’s ‘Showcase: A Headphone Amplifier’

In Grayson King’s ‘Valkyrie’ Preamp Guest Contribution post, a related piece was mentioned. This one is Klaus Noll’s Showcase: A Headphone Amplifier, which appeared in AudioXpress in May of 2003. We are pleased to add this work by Klaus as another Guest Contribution.

Klaus Noll’s article is similar to Grayson’s in that it describes a line-stage type preamp with a wideband composite amplifier using the AD744/AD811 pair, but optimized for headphone use. It also differs with regard to the power supply regulators used, which are similar to the Improved  Regulators from Audio Electronics of 2000.

In response to my query on posting his article here as a Guest Contribution, Klaus said:

“Thank you very much for your flattering letter which arrived today a little belated. Of course you can put the article on your website, after all it is your intellectual property which I used to build what I think must be one of the five best headphone amplifiers in the world.  Thank you again and kind regards.”

Well thanks to you Klaus, first for building such a fine preamp/headphone amp, and for sharing the details via AudioXpress. And of course, for offering it now, as a Guest Contribution here.

Both Grayson and Klaus have created worthy DIY audio projects, with full PCB patterns and part details, and it is great to be able to reprise them here.



Grayson King’s ‘Valkyrie’ Preamp

Since I’ve been running this website, I have had numerous emails about past articles. These have increased of late, which is a good thing. One of the more interesting developments have been the “By Request” and the “Guest Contributions” categories. Here’s an excerpt of a recent email one from Vladislav Polur, which falls into both. I have made some minor edits, for clarity:

“I am sending you a link about a review of a preamplifier made based on your article about using video op amps in audio. I could not find the author of the preamplifier. The reason I am interested in this design is because I had a chance to listen to an amplifier based just on signal op amps, within my friend’s system. It was a great experience; it sounded better than any amplifier I had a chance to hear before, tube or solid state.”

I think Vladislav may have given me a bit too much credit for the entire preamp realization, even it was/is in part based on my work appearing within Gary Galo’s POOGE-5 article. That box insert did in fact use video op amps, and it was titled High Performance Audio Stages Using Transimpedance Amplifiers. That preliminary piece sets the stage for the highlight item of this post, as follows below.

The item that matches up best with Vladislav’s listening experience and his cited link was Grayson King’s preamp project. This work was from the 1994 series of The Audio Amateur, and was entitled Valkyrie: A Line-Stage Preamplifier. The preamp was Grayson’s senior project while at Clarkson University in pursuit of his EE degree. It was developed under the tutelage of Gary Galo, and it did indeed make not just a fun university project, but also a great example of a worthy project for others. Grayson did do a fine job with his preamp! After graduation he took a job with Analog Devices in Boston, working in the same department as I did then, applications engineering.

It is a pleasure to help Grayson’s preamp to find some new friends here, and I want thank him for helping to make it all available once more, as a “Guest Contribution”.

A final note for those pursuing line stage performance. Another such preamp was described in Klaus Noll’s Showcase: A Headphone Amp, in AudioXpress, May 2003.

Improved Pos/Neg Regulator Load “Problem”

I recently received an email from a reader, Anton, residing in Moscow, Russia. He had built a version of the Improved Positive/Negative Regulators from Audio Electronics, Issue 4, 2000. While his circuit was functional, he was concerned by the fact that when he applied a 360mA load, the output voltage dropped, an apparent loss of regulation. This sounded to me like the regulator didn’t have enough drive under load, so I suggested some things to check:

Is the D7 LED ON, under all conditions? If it goes out, it means the pass transistor is starved for current, and the output then falls. If you have a pass transistor with a Beta of 100, a 360mA load means the current source must provide 3.6mA. But note that this is close to what it can do with R19 (Fig. 1) at 249Ω. You may need to drop this resistor down some, if you really need 360mA of output. Note: there was a discussion on this very topic, on page 11 of the original article, under the topic “Change of Current Source Resistor” (i.e, R19). The upshot here is that the older R19 value of 100Ω provided for more current, so I suggested that Anton lower this resistor value, to see if it fixed the voltage dropping under heavy load. He later reported back changing R19 to 160Ω (just for test), and it worked correctly with the 360mA load.

I was glad that he got it working OK, and pointed out should drop that R19 value down some (if he does in fact need 360mA). I left that for him to determine. If you need much less current, then leave the value as it is.

The moral of the story is that the circuit is robust, but it does have fixed limits. So, if you find the output dropping unexpectedly, check the load current carefully, both on board and external.

Thanks for the report, Anton!


Sources 101

This was a great article and I actually built a test circuit based on Figure 16a in part 2 of the article.  However, I wanted to use the source for a mosfet in a headphone amp circuit.  The circuit is designed to operate the mosfet at 150 mA.  I used a more robust transistor in place of the pn2222a.  It worked just fine and I was able to dial in 150 mA of current.  However, I don’t have the test equipment to measure performance.   I’m curious if anyone has tried to get more current out of this circuit and what type of performance resulted.