[funch]

Seeing those big heatsinks reminded me that I had several tiny 'sinks left over from another project, so I stuck them on my current builds. I got them from Amazon, and they are self-stick.

[/URL][/IMG]

 

[macky112]

Seeing those big heatsinks reminded me that I had several tiny 'sinks left over from another project, so I stuck them on my current builds. I got them from Amazon, and they are self-stick.

ooooh may we see some pics please?

 

[funch]

I would love to post pics, but Photobucket is not working atm. I'm still trying, but ......... .

 

[funch]

Finally got them up. It may have been a problem with Head-Fi. On my end, it's been difficult to
load pages lately. If I try to load a page, it just goes to a blank screen. I have to go back and do it
again to get it to load.

 

[funch]

These: https://www.amazon.com/Aluminum-Chi...e=UTF8&qid=1497304608&sr=8-14&keywords=rhs-01

 

[macky112]

Here is link to desktop amp.

http://www.diyaudio.com/forums/group-buys/302859-xrk971-pocket-class-headamp-gb-38.html#post5014544

Here is schematic:

R7 is the 66ohm source resistor in question. It serves as the "constant" current sink for the class A operation. You need something to draw the heavy bias load through the MOSFET. So for a 125mA bias at about 11v is about 1.4 watts. These run continuously hot so best to reduce the max load so spread over two 3w resistors so they can run forever without stress.

Hi X, does higher bias current means higher total power output? If I want to tweak bias current, do I just change the resistance value at R7?

 

[xrk971]

Output power is a combination of voltage swing and current. Power equals current x voltage. But to get an amp to make more power isn't as simple as increasing bias or voltage. Sure, higher voltages and higher current gives the capability to make more power. But you have to simulate the amp in LTSpice and see what happens when you drive it hard to make more power. With this SE design, there is a limit as distortion and clipping will take over at some point. With the simulation I can predict that with a 50ohm load, the output voltage swing is about 5.5v peak to peak before clipping occurs. So to increase this one can increase bias. But making R7 smaller is not how you do it. You have to adjust R3, R4, R5, and R7 to balance higher bias with achieving symmetric clipping and the correct harmonic distortion profile. It's a lot of interplay between competing variables.

You are welcome to download LTSpice and run he model and tweak it to see if you can get more power. One thing that can help is using MOSFET for a CCS instead of the resistor R7.

 

[macky112]

Output power is a combination of voltage swing and current. Power equals current x voltage. But to get an amp to make more power isn't as simple as increasing bias or voltage. Sure, higher voltages and higher current gives the capability to make more power. But you have to simulate the amp in LTSpice and see what happens when you drive it hard to make more power. With this SE design, there is a limit as distortion and clipping will take over at some point. With the simulation I can predict that with a 50ohm load, the output voltage swing is about 5.5v peak to peak before clipping occurs. So to increase this one can increase bias. But making R7 smaller is not how you do it. You have to adjust R3, R4, R5, and R7 to balance higher bias with achieving symmetric clipping and the correct harmonic distortion profile. It's a lot of interplay between competing variables.

You are welcome to download LTSpice and run he model and tweak it to see if you can get more power. One thing that can help is using MOSFET for a CCS instead of the resistor R7.

I'll put together this second attempt when pcb arrives. Just making some notes for faster referencing later. So if the new 19V desktop variant build bias current is off from your suggested 125mA, do I start the troubleshooting with changing R7 value? I.e. If I only get 115mA bias current, do I lower R7 value to bring it closer to 125?

 

[xrk971]

Macky112:
You should use the resistor values shown in this figure here:

Don't change anything at first. You want to make sure it works as designed first and then do your tweaking. Changing R7 will move the bias a small amount. The bigger one is R3, but that must be balanced by changing R7 and R5 to keep voltage above R7 (MOSFET source pin) approximately 1/2 of Vcc to ensure that clipping is symmetric otherwise you lose head room.

Two 33R 2W or 3W in series work well for R7 as that is 66R and about the same as 67.6R spec'd in schematic above.

 

[macky112]

Hi @xrk971 , could you tell me what is the purpose of C4 and C5 rail bypass caps? any harm for not populating them? any harm for replacing them with a single 1uF cap?

 

[xrk971]

They are bypass caps for the large 2200uF main power rail caps. Bypass caps present an ultra low ESR path for high frequency noise to get shunted to ground. Having two low ESR caps in parallel cuts ESR by half so. Even better. It will work without them and will work with 1uF cap there. Just make sure it is a X7R MLCC cap. It helps to reduce high frequency RF noise from having a chance to reach to MOSFETs.

 

[macky112]

Output power is a combination of voltage swing and current. Power equals current x voltage. But to get an amp to make more power isn't as simple as increasing bias or voltage. Sure, higher voltages and higher current gives the capability to make more power. But you have to simulate the amp in LTSpice and see what happens when you drive it hard to make more power. With this SE design, there is a limit as distortion and clipping will take over at some point. With the simulation I can predict that with a 50ohm load, the output voltage swing is about 5.5v peak to peak before clipping occurs. So to increase this one can increase bias. But making R7 smaller is not how you do it. You have to adjust R3, R4, R5, and R7 to balance higher bias with achieving symmetric clipping and the correct harmonic distortion profile. It's a lot of interplay between competing variables.

You are welcome to download LTSpice and run he model and tweak it to see if you can get more power. One thing that can help is using MOSFET for a CCS instead of the resistor R7.

Could you explain more about using a mosfet instead of R7. How to solder the mosfet pins? What spec mosfet should I use?

 

[macky112]

Macky112:
You should use the resistor values shown in this figure here:

Don't change anything at first. You want to make sure it works as designed first and then do your tweaking. Changing R7 will move the bias a small amount. The bigger one is R3, but that must be balanced by changing R7 and R5 to keep voltage above R7 (MOSFET source pin) approximately 1/2 of Vcc to ensure that clipping is symmetric otherwise you lose head room.

Two 33R 2W or 3W in series work well for R7 as that is 66R and about the same as 67.6R spec'd in schematic above.

Hi X, I have a noob question too embarrassed to ask on the DIYA forum, could you please go over each of the resistors and what their function are?
For example, what is RL? From first post on the PCA thread on Diya BOM list, you mentioned RL can be adjusted to high current draw, could you explain more? And also for R1, R2 etc

Thanks!

 

[xrk971]

Could you explain more about using a mosfet instead of R7. How to solder the mosfet pins? What spec mosfet should I use?

Please look carefully at photograph and see how the pins on the MOSFET are bent 90 to form a small "foot" that rests on the SMT solder pad. The pitch of the pins on the T220 is same as SOT223 (0.100inch).

 

[macky112]

One thing that can help is using MOSFET for a CCS instead of the resistor R7.

My bad for not being clear, I am curious about using another mosfet instead of R7 as you mentioned here for CCS