Another 100W hifi amp

[Aniket]

What are the voltages this amplifier can drive into 8r0 and 4r0 test loads without clipping of the signal?

I had a PC based signal generator, now as the amp has proper heatsinks and power supply, i took a quick test without frying the drivers.(I don't have a dummy load and oscilloscope)

I played 37, 45, 50, 1000 Hz tones until I could hear possible distortion and got the voltages at the output driving a speaker with 4 ohms impedance and 3.5 ohm DC resistance as below:

37 Hz: 31.1V AC rms
45 Hz: 31.4V AC rms
50 Hz: 31.8V AC rms
1000 Hz: 32.6V AC rms

Also, the supply rails were +/-49.4V DC unloaded.

Regards,
Aniket

 

[vzaichenko]

Roughly 30V rms @ 4ohm load gives us 225W output power... must be not easy to listen to, huh? ))

 

[Aniket]

Roughly 30V rms @ 4ohm load gives us 225W output power... must be not easy to listen to, huh? ))

Yeah, my households were rattling like hell, that too with both channels driven

 

[AndrewT]

you need to borrow a scope and make up some 2r0 4r0 and 8r0 test loads.

A pair of 4r0 200W loads does as both a 2r0 400W and 8r0 400W

I have 4 off 4r0 225W and they do for all my testing.
Pics in another Thread.

 

[vzaichenko]

yeah, it would be good to see the wave form while clipping, as well as square wave handling - there may be some "ringing", over-compensation, under-compensation visible... you never know, unless you see it on the screen

 

[Aniket]

Yeah, I am planning to buy one.
Rigol, Agilent and Tektronics are reasonably priced here

 

[tinitus]

I actually use even the "bigger" one - with the main power switch

I have one where the switch have gone faulty
and I cannot find an exact replacement that fits

I prefer them seperate from now on
fuse drawer is still handy though

might be practical to have it all on seperate small alu plate, and then mount as a unit
out of necessity I did this not long ago

 

[vargasmongo3435]

Nice Aniket !

Aniket I did check the design and it does respond really well here are some

results:

20Hz 1Vp signal at 8 ohms load give about 48W and 0.001% THD

20Hz 1.5Vp signal at 3.8 ohms load give about 231W and 0.002% THD

20KHz 1.5Vp signal at 3.8 ohms load give about 158W and 0.151% THD

I did exaggerate a bit the speaker load "3.8 ohms" because is not always exactly 8 or 4 ohms, and I usually check only with 20Hz and 20KHz, go beyond 20KHz will serve not purpose lol in my opinion
20Hz looks pretty clean and 20KHz




Regards
Juan

 

[vzaichenko]

20KHz 1.5Vp signal at 3.8 ohms load give about 158W and 0.151% THD

20kHz THD looks a bit high for such schematic, meaning the open-loop bandwidth is not high enough, which is sometimes a result of over-compensation.

Try to reduce C3 in simulation and you'll see the improvement. Now, the trick is to find the optimal value for enough stability and rather low distortion at higher end...

 

[CBS240]

It seems like for those bias currents that C3 and C18 at 100pf are a bit large. Perhaps it is being overcompensated? Perhaps reducing the lag compensation and adding phase leading compensation? such as a small pf cap in series with a resistor (a zero) placed in parallel with R29.

 

[vargasmongo3435]

20kHz THD looks a bit high for such schematic, meaning the open-loop bandwidth is not high enough, which is sometimes a result of over-compensation.

Try to reduce C3 in simulation and you'll see the improvement. Now, the trick is to find the optimal value for enough stability and rather low distortion at higher end...

I will try out that sir reducing C3 and C18

 

[Aniket]

Aniket I did check the design and it does respond really well here are some

results:

20Hz 1Vp signal at 8 ohms load give about 48W and 0.001% THD

20Hz 1.5Vp signal at 3.8 ohms load give about 231W and 0.002% THD

20KHz 1.5Vp signal at 3.8 ohms load give about 158W and 0.151% THD

I did exaggerate a bit the speaker load "3.8 ohms" because is not always exactly 8 or 4 ohms, and I usually check only with 20Hz and 20KHz, go beyond 20KHz will serve not purpose lol in my opinion
20Hz looks pretty clean and 20KHz




Regards
Juan

Thanks Juan,

Multisim is a very conservative simulation software and it's simulation results are very much close to the actual real world performance.

On the other hand I am getting a lot better THD performance in simulation.

1kHz 1.7Vp input signal on 8 ohms load gives 131W output, 0.001% THD

20kHz 1.7Vp input signal on 8 ohms load gives 131W output, 0.005% THD

1kHz 1.6Vp input signal on 4 ohms load gives 233W output, 0.001% THD

20kHz 1.6Vp input signal on 4 ohms load gives 232W output, 0.014% THD

pics attached

20kHz THD looks a bit high for such schematic, meaning the open-loop bandwidth is not high enough, which is sometimes a result of over-compensation.

Try to reduce C3 in simulation and you'll see the improvement. Now, the trick is to find the optimal value for enough stability and rather low distortion at higher end...

I already checked various values of C3 in simulation. For 47pF THD20 falls to 0.007% for 4 ohm load at 232W output.
In fact I reduced the value of C3 in the working model and the amp was stable upto 47pF( i checked 100, 82, 56, 47, 39pF ), below than this it was oscillating.
So I left 100pF as a fail safe value in the place of C3.

It seems like for those bias currents that C3 and C18 at 100pf are a bit large. Perhaps it is being overcompensated? Perhaps reducing the lag compensation and adding phase leading compensation? such as a small pf cap in series with a resistor (a zero) placed in parallel with R29.

Yep, I agree it's a little(not much) over compensated. The optimal value for C3 would be 68pF, maintaining stability and gaining some high frequency performance as well. I didn't had it in stock. would buy 68pF 1kV ceramics on my next parts purchase.

Regards,
Aniket

 

[vzaichenko]

On the other hand I am getting a lot better THD performance in simulation.

1kHz 1.7Vp input signal on 8 ohms load gives 131W output, 0.001% THD

20kHz 1.7Vp input signal on 8 ohms load gives 131W output, 0.005% THD

1kHz 1.6Vp input signal on 4 ohms load gives 233W output, 0.001% THD

20kHz 1.6Vp input signal on 4 ohms load gives 232W output, 0.014% THD

OK, these are the values I would expect from the schematic

 

[vargasmongo3435]

I'm not an expert but I think this design will perform really really well, remember we are testing in a virtual simulator and we are pushing to the limits and still making good lower THD with 20KHz so I imagine that is a really quiet amp, and I betcha in the real world will outperform even better that is for sure.

Regards
Juan

 

[Aniket]

I'm not an expert but I think this design will perform really really well, remember we are testing in a virtual simulator and we are pushing to the limits and still making good lower THD with 20KHz so I imagine that is a really quiet amp, and I betcha in the real world will outperform even better that is for sure.

Regards
Juan

DIY 120W amplifier sound test - YouTube

Have a look on the video on the above link, it sounds brilliant, very open and detailed

 

[vargasmongo3435]

Hell yeah ! I will check that out man

 

[vargasmongo3435]

Aniket those Pionner woofers are 4 ohms ?

 

[Aniket]

Aniket those Pionner woofers are 4 ohms ?

Yep, those are TSW-1208D2 400W, each with dual 2 ohm voice coil, connected in series to make a 4 ohm load.

 

[vargasmongo3435]

oh that explains I think they are way too stiff "rigid" for your amp if you get paper cones woofers at 8 ohms even 10" woofers size man that would sound magnificent

I have two Alpines type R woofers and they barely move jejejejejeje love your video bro it does motivate me

Regards
Juan

 

[radio]

PCB for power supply in pdf

Hi Aniket!
Give show PCB-botton layer for power supply in pdf format (in the ratio1:1).
thank you!