25W Class A amp with Lateral MOSFETs

[manniraj]

A quick update on my Juma 24w class A build, I connected the input ground to the psu ground. Then replaced the red leds with the proper 3.5v forward voltage green led on the LED1 and jumper on LED2. Now using a 24vdc laptop adapter I set the voltage across the source (S) of the laterals against the ground pin to half of the rail voltage (12vdc) very close on both of the laterals (12 & 12.4vdc) using the trimmer.

Offset across the SPK+/- was still hovering >3v. Now as the output cap is in series with the SPK+, so I put a 100R/2w resistor between the SPK+/- and the offset came down to 3-5mV and very stable as long as the bias trimmer is not touched. As no power resistors on the board I put a 0.1R/5w BPR resistor in series with the negative rail and I can see the voltage drop across being -40.8mV.

Let me know if the steps seems to be fine @prasi and if my bias and offset setup is good enough of should I tinker anything more.

Thanks

 

[manniraj]

As suggested by @prasi I did check the voltage drop across the single green led and its hovering between 3.8-4v with the second led being a jumper. Now all seems to be fine when the speaker outputs are connected with a 100R/2w resistor around 3-5mv max and seems to be stable. When I connect the amp to a sacrificial speaker, I see that initial power up their is a big woofer movement (bear in mind the green led is still off) then at the 9.5v mark (checking the dmm of the Source pin of the laterals against the ground for symmetrical voltages) the green led comes on and their is a distortion and cone movement on the speakers. So not sure whats wrong with the amp board if anyone who built the version as depicted in the post #180 with the RC coupled feedback please provide some inputs.

Thanks

 

[manniraj]

A quick update on my adventures today. Based on the @ElFishi posts I did lift the C4 pin going to the output and connected it to the collector of the Q4 (BC560). Now with this change the oscillation gone but the bias issue with a single green LED was still their, so I removed the jumper from the led2 and put another green led having the same 3.5v forward voltage. Set the symmetrical voltage across the sources of the laterals to half the rail voltage (12v) and then I put a 100R resistor on the speaker outputs so offset was hovering from 0.5mv-20mV. Then I put a 0.1R resistor in series with the negative rail and the voltage drop across it was 120mV so the current consumption came up to 1.2A. Now the laterals were getting nice warm and hot

I did play some music and it does play without any distortion. But the initial cone movement when I switch on the amp is still their which I feel is more like thump not sure how do I get rid of that as using a 24vdc laptop adapter. Need to try using a toroidal to check the full potential of this amp along with a thermistor in series with the primary acting as soft start and get rid of the cone movement.

Otherwise the amp was dead silent so far I listened with maybe low gain because of lower rails.

thanks

 

[Vunce]

Nice work Manniraj!
Maybe try a capMx based linear psu, the slow ramp up could help reduce the amount of cone movement at power up.

 

[manniraj]

Nice work Manniraj!
Maybe try a capMx based linear psu, the slow ramp up could help reduce the amount of cone movement at power up.

Thanks Vunce, right now do not have a toroidal to power this so using my spare 24v/5A laptop adapter which I had got to test the Lineup 4w amp

So for now using this adapter of 24vdc rails using 2 LEDs to get the 1.2A bias is fine but with higher rails I need to remove second led so that I will not run more bias across the laterals and minimize the long term usage. So I am thinking of putting some 2-din dip socket on the 2nd LED and put a jumper or led as the case maybe based on the incoming DC rails.

thanks

 

[ElFishi]

But the initial cone movement when I switch on the amp is still their which I feel is more like thump not sure how do I get rid of that as using a 24vdc laptop adapter.

Glad you are making progress!
The cone displacement comes from C2 charging faster towards its stable state value than C3. You can increase C2 to slow that down (I did). In the end I wouldn't worry too much. It is slow and goes away within seconds.
Edit: there will always be some cone displacement as C8 needs to charge through the speaker.

 

[manniraj]

With the 24vdc rails I do get some slight distortion in the music playback for few instruments as well as female voices. I am interpreting that to do with the lower rails. So will use my spare 22vac 6.8a trafo which gives around 32vdc to try on this amp to rule out the input stage bias being not sufficient with lower rails.

Thanks

 

[mz543578854]

Hi @juma , I read that you say the Svaja amp does not produce a switch on thump. Running the switch situation on in LTSpice with empty Cs shows a significant current flowing through the loudspeaker to load C8.
Why is there still no noise when switching the amp on or a damage to the speaker?
Thanks, Mo

 

[juma]

Hi Mo,
Cap multiplier on power rail ensures slow amp start so no thump should be heard.
Why do you think that current flow through the speaker is harmful - speaker's job is to flow current and move the cone (unless it's far beyond the speaker's power rating). Calculate the current and the pulse width that the speaker takes when bass drum is hit at 10 W of power....

 

[mz543578854]

Thanks, @juma . That makes perfect sense and with a CapMx as PS added in LTSpice, I indeed see it nicely ramping up and current through the speaker is minimal during that period.