I wrote some posts earlier in the thread (can't find them now), but if you want to tinker with it, I recommend playing with the values of C5 & R9. This will modify the nested feedback characteristic of the amp and do a lot with the sound.Question for Hugh @AKSA I really like my A39, thank you for sharing this with the community. But...like any audio addict I need my next fix, I need to tinkerI know I could throw some cash at it with some naked foil resistors and pricey German capacitors, but I can't help wondering if there are gains to be had with something like CCS's for the VAS transistor or input LTP? My EE skills are lousy but I'm willing to have a go. What do you think?
The resistor value will affect the 'detail/softness' or 'presence' in the sound, and the cap will alter the bass character.
No stability problems with these mods as long as you don't go crazy with the values. You can also try to disconnect them and listen to get a reference. There is maybe some small risk with stability when you do this, but my amp was stable with them disconnected. From memory I ended up using 470k & 1n5 in my amp.
You can also connect the C4 bootstrap cap directly to the source of M1 instead of output (other side of source resistor). This will lower the output impedance of the amp, and also affects sound signature. I did this on my amp.
Easy things to try.
Having lived with this amp for almost a year and comparing it to my other builds it plays well with the standard BOM components. While it may visually look prettier with the expensive components it is my opinion a well engineered amplifier does not require fancy parts to play well. I am also not one of those that contend if it measures well it will sound good as measurements are covering a very narrow function of what an amplifier does. Also most designers will voice an amplifier to give specific harmonic content ect. and you will to either like or dislike their choices.
Bill
Bill
@bigaudioscotto
Just remember I have a very dispensable income where I can spend where I want when I want with regards to my diyaudio hobby and that by no means is a requirement for great sound. Use X’s recommendations on the 1st post as your template cause the rest is just audio jewelry.
Best,
Anand.
Just remember I have a very dispensable income where I can spend where I want when I want with regards to my diyaudio hobby and that by no means is a requirement for great sound. Use X’s recommendations on the 1st post as your template cause the rest is just audio jewelry.
Best,
Anand.
Sorry my post on #2406 probably sounds haughty (should've said, "My discretionary spending with diyaudio is pretty high, etc....").
Regardless, this is a very good circuit, and I usually spend a little more on circuits that I feel "might" benefit. It's entertainment anyway, whether visual or auditory.
Regarding vehicles...my vehicle is electric and lacks some of the 'fun' in driving that I had with ICE vehicles. So probably not your cup of tea wirewiggler.
Best,
Anand.
Regardless, this is a very good circuit, and I usually spend a little more on circuits that I feel "might" benefit. It's entertainment anyway, whether visual or auditory.
Regarding vehicles...my vehicle is electric and lacks some of the 'fun' in driving that I had with ICE vehicles. So probably not your cup of tea wirewiggler.
Best,
Anand.
Last edited:
I was kidding too, while I admire expensive cars I consider them somewhat a waist as you cannot drive them as intended on public roads. I prefer something I can enjoy on a daily basis and that audio provides. While I don’t spend a fortune on discrete components, I make up in quantity of amps.
Bill
Bill
It is quite interesting the way X has placed the pos side of the power supply on one side of the amp and the neg on the other. I would have been concerned about hum from the ac but he seems to have it under control. Nice looking amp, I consider myself lucky to have an amp that performs as well as his.
Bill
Bill
Attachments
As long as you use carbon film on the feedback network and low inductance KOA BPR source resistors, the amp will work great with the BOM specified. Basic Vishay carbon film 5% 1/4w work well for the feedback circuit. With modest components, you can afford 100 count ammo belt to pick and match if you want.
I'm with Anand. I spend most time on the circuit design, not on the components. I use Beyschlag metal films, occasional carbon compositions and Nichicon/Rubycon electros. Apart from them, I use nothing more than quality, industrial components.
I own an expensive (but old) Lexus but prefer to drive a 2003 Corolla, which is an exceptionally refined, small car, and very economical. In Oz we pay $USD4.30 a gallon and the Lexus returns appalling consumption in the city.
Simon,
Replacing the bootstrap on the VAS and the stage current to the input stage to CCSs will ruin the wonderful harmonic profile of the AN39.
It will return superior 'performance', but far less engaging and not as musical.
The paradigm with amplifiers driving speakers is dual; the amp drives the speaker, and the speaker talks back to the amp. It's this phenomenon that actually means that the measured performance even at 1KHz into a speaker is orders of magnitude HIGHER than the figures marketed using a special, zero phase shift resistor. Recently a local engineer here measured a good amp of 100W into a special resistor and read off 0.001% THD at +20dB (12.5W). But into a real speaker of quite good quality the THD measured through a microphone 1m in front of the speaker at the same power was 0.57%. This happens because of the phase shift in the reactance of the speaker (drivers and crossover) plays merry hell with the global feedback network, and also varies with frequency and amplitude.
The marketers never really measure correctly; they try to make the figure as squeaky clean as they can.
HD
I own an expensive (but old) Lexus but prefer to drive a 2003 Corolla, which is an exceptionally refined, small car, and very economical. In Oz we pay $USD4.30 a gallon and the Lexus returns appalling consumption in the city.
Simon,
Replacing the bootstrap on the VAS and the stage current to the input stage to CCSs will ruin the wonderful harmonic profile of the AN39.
It will return superior 'performance', but far less engaging and not as musical.
The paradigm with amplifiers driving speakers is dual; the amp drives the speaker, and the speaker talks back to the amp. It's this phenomenon that actually means that the measured performance even at 1KHz into a speaker is orders of magnitude HIGHER than the figures marketed using a special, zero phase shift resistor. Recently a local engineer here measured a good amp of 100W into a special resistor and read off 0.001% THD at +20dB (12.5W). But into a real speaker of quite good quality the THD measured through a microphone 1m in front of the speaker at the same power was 0.57%. This happens because of the phase shift in the reactance of the speaker (drivers and crossover) plays merry hell with the global feedback network, and also varies with frequency and amplitude.
The marketers never really measure correctly; they try to make the figure as squeaky clean as they can.
HD
WW,
Congrats on your wonderful external PSU!
That’s got to be one of the smallest AN’s I have seen. Very cool. Super setup you have there.
Splitting +ve and -ve PSU rails like this is common. What causes hum are grounding topology and current loops. You want to avoid making a large physical loop that encircles a trafo - it basically will integrate and collect the mains magnetic field EMI.
Think of it as you never want to “lasso” a trafo with a wire. Run two lines with twisted pair next to each other if you have to but don’t split them and lasso the trafo with them. The green wire here is the chassis protective ground.
Congrats on your wonderful external PSU!
That’s got to be one of the smallest AN’s I have seen. Very cool. Super setup you have there.
Splitting +ve and -ve PSU rails like this is common. What causes hum are grounding topology and current loops. You want to avoid making a large physical loop that encircles a trafo - it basically will integrate and collect the mains magnetic field EMI.
Think of it as you never want to “lasso” a trafo with a wire. Run two lines with twisted pair next to each other if you have to but don’t split them and lasso the trafo with them. The green wire here is the chassis protective ground.
Last edited:
Hi Hugh,
I got a PM from a member who was wondering if the AN can be configured to work as a 5W amplifier and how to go about that. IMO, just build the standard amp and don’t use more power than you need but you always have the reserves.
Would it be hard to redo the setpoints for this to be a 5W amp? Would we be trying to keep LTP bias current near 2mA?
I got a PM from a member who was wondering if the AN can be configured to work as a 5W amplifier and how to go about that. IMO, just build the standard amp and don’t use more power than you need but you always have the reserves.
Would it be hard to redo the setpoints for this to be a 5W amp? Would we be trying to keep LTP bias current near 2mA?
Hi X,
Yes, quite easy to configure for 5W........
I would leave the input stage and VAS currents the same, and adjust the two stage resistors for a lower voltage. I'd suggest 15 volt rails giving an output of about 24Vpp, which is 9 watts output.
The two current source resistors would have to be amended to maintain fidelity and I'd suggest a quiescent of 1A. That would create about 30W of dissipation per channel, 15W from each output device, and the OPS source resistors would be 0.33R. The amp could then generate up to 2A into a 6R load, making it suitable for a nominal 8R load, and thermal issues would be minimised.
Hugh
Yes, quite easy to configure for 5W........
I would leave the input stage and VAS currents the same, and adjust the two stage resistors for a lower voltage. I'd suggest 15 volt rails giving an output of about 24Vpp, which is 9 watts output.
The two current source resistors would have to be amended to maintain fidelity and I'd suggest a quiescent of 1A. That would create about 30W of dissipation per channel, 15W from each output device, and the OPS source resistors would be 0.33R. The amp could then generate up to 2A into a 6R load, making it suitable for a nominal 8R load, and thermal issues would be minimised.
Hugh