Any way to improve the sound of my Acurus.

Status
Not open for further replies.
Member
Joined 2010
Paid Member
The "Leach Amplifier" title refers to a college professor, Marshall Leach's landmark design originating in the 1970s which became a popular student DIY project - more recently being made available as a set of PCBs for the growing mass of students and DIYs globally, who still wanted to build it with the benefit of readymade simplicity - perhaps moreso since the professor's passing a couple of years ago: Remembering W. Marshall Leach Jr. (and his low-TIM amp) | EE Times
I can't say what version or even if his PCBs are still available, though.

Present day texts like DIYaudio member Bob Cordell's, also use it and another precursor, the Locanthi T output stage design, as the main basis for BJT designs . Feel free to get a glimpse inside here: Designing Audio Power Amplifiers: Bob Cordell: 9780071640244: Amazon.com: Books
 
Last edited:
Present day texts like DIYaudio member Bob Cordell's, also use it and another precursor, the Locanthi T output stage design, as the main basis for BJT designs . Feel free to get a glimpse inside here: Designing Audio Power Amplifiers: Bob Cordell: 9780071640244: Amazon.com: Books

Thanks, I have this book, I'll look into it.

So Leach is not a commercially available amp, is it really that good? I looked at the schematic, it's very simple, quite similar to the Acurus. I consider Acurus is "so so" amp. I am trying to improve the Acurus just as a stop gap amp so I can take my time to research for a new tube amp build.

I am new in audiophile amps, Just been learning in the last few days. I discovered a lot of SS amps are DC coupled and can burn the speaker if something goes south. That is a no no to me. That would be very expensive for me to blow the speaker.

I finished modding one channel of the Acurus, just very nervous to put it back in. I am going to use one of the 12" Celestion guitar speaker to power up the amp!!!
 
Last edited:
"The tail resistor will follow the input and the tail current change with the input swing."

Ignoring the current in the other half of the diff input decreasing?

Diff pairs usually run class A.

Most SAE, GAS, Accuphase, etc. use resistors for current sources for their cross-coupled dual-diff input designs, I wonder why that might be?

An Introduction to Electroacoustics and Audio Amplifier Design
 
BY DJK-The v4.5 Leach was the result of over 20 years of refinement, it's hard to improve without starting with a blank sheet of paper and a whole different approach.
__________________

IN 20 DAYS , powered by a hawksford VAS (spookyamp).

The professors circuit was quite inspirational , but easily beat. (below)

This is actually tested by builders and the H/K 990 <.001%thd20K 50w.
The professor's EF3 is the best OPS short of an error correcting one.
OS
 

Attachments

  • spooky.jpg
    spooky.jpg
    92.6 KB · Views: 176
They seem to like it ? The HK990 got stellar reviews as well. Leach V4.5 uses a
zener based cascode only , I use zener + cap multiplier.

SOME swear by the even faster slewing CFA input stages. :scratch:

PS - I like the fact that this design has to same 10ppm @ 100w as it
does at 1w. It also has the most :eek: semi's (18).
OS
 
I used to think it required the fully-comp topology to sound good, but I changed my mind in 1986 after doing some cap mods to a McIntosh MC2120.

The MC2120 has a single-ended diff-pair with a resistive tail to a regulated voltage, a Vas with a resistor to the same regulated voltage, and a bias transistor.

Four transistors, total.
 
Acurus bias current

:hot::hot:
It is a 3 channels power amp. There is no model number other than 200X3!!! I just want to see whether there is any way to improve the sound quality.
Thanks
Hi Alan!
Most Aragon and Acurus amps are current starved in the output transistors from the factory. So you will get a big improvement by turning up the bias current. You might even go class AB, by turning up the current till you reach a reasonable temperatur on the heatsinks, and run in class A for the first one or two Watts. ( where most of the music lies )
Thorsten
 
:hot::hot:
Hi Alan!
Most Aragon and Acurus amps are current starved in the output transistors from the factory. So you will get a big improvement by turning up the bias current. You might even go class AB, by turning up the current till you reach a reasonable temperatur on the heatsinks, and run in class A for the first one or two Watts. ( where most of the music lies )
Thorsten

Thanks, that's what I am wondering too. I just touched the heat sink of the remaining two power amps that are still working in the chassis after running for 4 hours, it's not even warm!!! I don't run at high volume, so what I feel is the idle power, it must be less than 5 deg above room temperature at worst, my guess is less than 3 deg.

I already reassembled the amp back with new thermal pads ready to install back to the amp. I am going to put it in and adjust the bias.

Thanks
 
Member
Joined 2010
Paid Member
The principle of increasing bias is simple enough and there certainly is a control for it but, there are often less obvious reasons for keeping it at some particular level that maybe doesn't seem right. Many power amplifiers using the Toshiba 2SC3281/A1302 pairs get excellent sound quality with just 50mA current through each complementary Emitter Follower pair. Similar products using traditional (Then Motorola) parts were leaning to 120mA and above, and really needed it.

A more practical issue is stability. Some bias controller designs just don't work properly or drift too much when set away from their sweet spot, so keep your eye on the current setting when current steadies at its new setting - after it has been running with some decent output power that is, and before the heatsink has a chance to cool.

With old parts, watch out for breakage of the plastic - it can cause disasters if bias then skyrockets up to the max.
 
Ha ha, from touching the heatsink, I double the transistor are running 50mA each idle. My amp has 4 complementary pair, so the total will be 0.2A. The rails are +/-80V. That will be 2X16=32W. The temperature of the heat sink cannot be more than 3 deg above the room temperature after using for 4 hours. I don't think my amp is close to 50mA per pair. Most like 50mA the for all 4 pairs of transistors.
 
I power up the amp and measure all three power amp. For each power transistor, there is a 0.5 ohm resistor to help balance different transistors. The voltage drop is 3.5mV on two and 2.4mV on one. Each transistor conduct less tham 7mA at idle!!!!:eek::eek:

I am adjusting to 25mV which gives me 50mA per stage and 200mA total idle current.(4 transistors)
 
Last edited:
That is definitely a bit miserly! :( You often see a theoretical figure of 26mV drop across each emitter resistor for lowest crossover distortion. It is always 26mV irrespective of the resistor value. With 0.5 ohms that works out to 52mA, which should be quite safe.

I suggest leaving the meter hooked up and monitoring the idle current for several hours to make sure it is stable at the higher setting. Some commercial amps have quite bad thermal tracking.

The 26mV figure isn't set in stone though. The last BJT amp I built gave its lowest HF crossover distortion at something like 13mV per emitter resistor, 130mA through 0.1 ohm resistors.
 
Last edited:
Yes, I am button it up and let it run for a while first. But I can tell you, I don't think it's going to drift that far, it took a good 1/4 turn to get to 25mV!!! I tried the max, I only got 70mV or so highest.

Ha ha, I am going to use the best distortion analyzer......my ears!!!. I just adjust to the best sounding.

What is the reason of 26mV on the emitter resistor?
 
Last edited:
Update

I hooked up to the system and compare:

1) between idle at 7mA and 50mA(25mV). I can hear slight improvement with 50mA per stage.

2) I compare my beefed up connections and added filtering, it sure seems to improve a little when all run in 50mA idle. I switch the speakers back and fore to eliminate the speaker location effect and the stereo input variations, sure seems like the better sound follow the amp.

The sound seems to be a little more forward, ever so slightly brighter, more solid with the higher current and beefed up amp. This is no scientific as I just play the same few lines of music over and over, and judge by memory. I am sure psychosomatic comes into play. But I guess I few there's an improvement and that's all I can ask for.

I am going to modify the second channel and test it again because I don't think I can hear the sound stage and imaging with one side at a time.
 
What is the reason of 26mV on the emitter resistor?
I guess it's a misunderstanding that was posted by somebody and got quoted and repeated by millions unverified .... the way "knowledge" gets distributed over the Net.

I guess so (might be wrong but in that case would love to see Math proof, not just "xyz Audio Guru says so") because of 2 reasons:

1) I can't find much reason in a statement that 26 mV across the ballast resistor is the important parameter ... no matter the resistor value or transistor current :eek:

2) it's strikingly similar to the 26 mV used to calculate internal Emitter resistance, which absolutely is not the external Emitter resistor , no matter how confusingly similar both phrases sound.

Fact is that internal Equivalent emitter resistance in ohms is
26/Ie , Ie being the current in mA passing through said emitter.

It's a very useful formula to calculate important parameters such as maximum transistor stage voltage gain and input impedance.

Read the general explanation in "Transistor Characteristics" in:
ESP Amplifier Basics - How Audio Amps Work

A much more detailed (albeit very complex) explanation (Ebers Moll transistor model) in:
Bipolar junction transistor - Wikipedia, the free encyclopedia

By the way, that's where the transistor models used in simulators come from ;)

and the basic Physics explanation which in fact shows where those d*mned 26 mV come from ... in any and all silicon junctions by the way.
Yes, even in humble 1N4002 diodes and such.

Read the "Thermal Voltage" section in:
Boltzmann constant - Wikipedia, the free encyclopedia

So it's understandable that somebody read about transistor equivalent resistance (which is an internal parameter) and the 26mV used to calculate it and somehow thought that the same "magic value" applied to plain resistors.

No way, an absolutely nonlinear semiconductor junction and a linear resistive element are intrinsecally different.
 
Status
Not open for further replies.