Recommendation for 5-10W amp.

I made one channel of the mini-F5 for testing, with component values according to the attached schematic. Except that I used a 19.5V notebook smps to power the circuit while testing, since I'm still waiting for the toroid to be delivered.

The amp seems to work, I fed it with a 2.8Vpp sine wave from my scope meter and it's amplified (in fact, it clipped but the generator output from my cheap scope meter is fixed at 2.8Vpp so cannot lower it.)

However, I have an issue. I noticed that Vgs of Q4 is only 0.89V and Vgs of Q3 is only -0.53V. That correspond to around 140mA of bias, too low actually. So I want to know whether I should increase R3 and R5 to get higher Vgs, or this is simply due to the use of the 19.5V supply, and will be fixed when a 24V supply is used (I would like to have around 750mA of bias current).

Thanks and regards.
 

Attachments

  • Mini_F5a.jpg
    Mini_F5a.jpg
    630 KB · Views: 604
  • Mini_F5b.jpg
    Mini_F5b.jpg
    634.2 KB · Views: 558
  • 1K_out.jpg
    1K_out.jpg
    278 KB · Views: 532
  • 20K_out.jpg
    20K_out.jpg
    259.7 KB · Views: 503
  • Schematic.jpg
    Schematic.jpg
    58.9 KB · Views: 571
  • Vgs_Id.JPG
    Vgs_Id.JPG
    184.7 KB · Views: 375
Nice looking PCB.

In my amp R3, R5 are 1K1 (with a 24V supply) but that is because my NOS LatMOS seems to be lazy.

Before changing R3 and R5 and would wait to use your 24Vdc supply.

I think your low bias is mostly due to a low supply. Do you own a small bench test 24V supply ?

Rgds,
Eric
 
It sings

It'll take another 2 weeks for the 200VA dual 18V toroid I ordered to be delivered, but I took an undersized 35VA dual 9V from a linear regulator to give the 24V dc I needed for testing.

At 23.8V Vcc, I got a bias of 500mA for each MOSFET, so I increased R3/5 from 680R to 820R. However, it was too much to ask for the tiny 35A toroid and Vcc dropped to 22V, anyway I got a bias of around 600mA with that.

Next I connected a 5R 20W resistor to the output as dummy load, and I saw oscillation from the scope meter when test signal was present. I removed the 4.7uF output bypass capacitor and the oscillation got even worse. It turned out oscillation was due to the nest of crocodile lead wiring mingled altogether, but was resolved by separating the input and output wires with proper distance.

Finally connected the test channel to a cheap Denon speaker, and it sings !! Source is just a Samsung smartphone's audio out, but even with a single channel the sound is quite nice, it has a taste of sweetness resembling that of tube amps :cool:
 

Attachments

  • Mini_F5c.jpg
    Mini_F5c.jpg
    412.9 KB · Views: 218
Help please.

Hi all,

Since the amp has an output cap (C3) I always get a turn-on an off thump.

I decided to add a speaker protection module from e-bay that includes a upc1237.

It's much better, I do not have any turn-on thump but I still have a turn off thump if the following set-up. Please have a look at the attached diagram.

The amp is powered via a SMPS (24Vdc, 5A) and now it has a connection for a 12Vac input for the speaker protection module, both power are going through a DPST switch on the back of the amp. Both supply are connected into a single power bar strip.

If I close the power of the amp via the power bar switch (SW1) I do NOT get a turn-off thump.

If I close the power of the amp via the switch on the back of the amp (SW2) I do get a turn-off thump...:scratch::scratch:

I shorted both inputs and it doesn't change the outcome.

I'd like to avoid having to close the amp by turning off the whole power bar.

Anyone knows what can be causing this odd behavior.

Thanks in advance for any suggestion.

BR,
Eric
 

Attachments

  • odd.pdf
    381 KB · Views: 140
Last edited:
I am already aware of this, I'm looking for a solution with my current set-up. I do not want to rebuilt the amp, it's already finished. Your amp is different than the one I have built.

Why does it work with closing SW1 and not with closing SW2 is the question. Very tough question to answer, I know.

Thanks,
Eric
 
I would hazard that your amplifier power rails collapse too quickly compared with the supply to the protection board - so the protection board doesn't cut off the speaker on power down before the amplifier rails collapse.

If this is the correct diagnosis then one option is to reverse engineer your protection circuit - that chip has a pin that detects when the power goes off so maybe you can have that pin monitor the amplifier power rails directly ?

Another options is to slow down the collapse of the amplifier power rails - not sure on your options there other than a lot more rail capacitance, which your SMPS may complain about on power-up. And you'll need a lot of capacitance for a Class A amplifier - yeah, scrap that thought.

Actually, I'm surprised the protection circuit doesn't cut off faster - is it working properly ?
 
Last edited:
Member
Joined 2012
Paid Member
I am already aware of this, I'm looking for a solution with my current set-up. I do not want to rebuilt the amp, it's already finished. Your amp is different than the one I have built.

Why does it work with closing SW1 and not with closing SW2 is the question. Very tough question to answer, I know.

Thanks,
Eric

Zen Mod gave the answer. The 12 vac is not the lowest current loop. It has 10 times the current of the 120 vac feeding it. You want to switch the 120 vac side (of both PS's) like you are doing with the power strip switch. If you do not like that switch, make your own in a setup where you can reach it. BTW, one way to not get a turn on thump with the standard ACA is to unplug the PS from the wall outlet, instead of using the switch on the DC side. But, of course, the outlet might wear out eventually.
 
Sorry, Eric, I was not clear enough. Cap multiplier slowly rises voltage unless capacitor gets full.
Your smps also has capacitor at the output. If you turn on SW1 after SW2, voltage rises until capacitor gets full. In this case time is much shorter compared with cap multiplier due to much higher frequency and smaller capacitance.
If you turn on SW2 after SW1, you don't have voltage rise, you get nominal voltage at the moment when you connect load, so you get huge voltage peak. That peak goes thru amp to speaker and you hear thump.
Now, when you turn off your amp you have the same situation, but reversed. Voltage goes from nominal to zero, capacitor gets empty. If you turn off SW2, you divide amp from capacitor, voltage on amp goes to zero in one cycle, again you get voltage peak, only reversed, and, do I need to say, thump.
Edit: While I was writing, smarter people already answered.
 
Last edited:
I would hazard that your amplifier power rails collapse too quickly compared with the supply to the protection board - so the protection board doesn't cut off the speaker on power down before the amplifier rails collapse.

If this is the correct diagnosis then one option is to reverse engineer your protection circuit - that chip has a pin that detects when the power goes off so maybe you can have that pin monitor the amplifier power rails directly ?

Another options is to slow down the collapse of the amplifier power rails - not sure on your options there other than a lot more rail capacitance, which your SMPS may complain about on power-up. And you'll need a lot of capacitance for a Class A amplifier - yeah, scrap that thought.

Actually, I'm surprised the protection circuit doesn't cut off faster - is it working properly ?

BINGO !! Problem solved. You were correct, I simply added a 6800uF, 35V cap across the 24Vdc and now everything is great, no more pop of any kind. I'm thrilled !

Thanks to everyone, what a great forum.

Time to go enjoy some music :D

BR,
Eric
 
Sorry to resurrect that thread

But I am on the search for a small amp to power my tweeters from 3Khz upwards. They are planars with 3,6 Ohms flat impedance and around 90dB sensivity.
I have the lateral mosfets and an enclosure with heatsinks capable of around 10W dissipation per channel.
So I simulated a version of jumas amp from post #217 https://www.diyaudio.com/forums/pass-labs/255523-recommendation-5-10w-amp-22.html#post4000309 which biases the laterals around 0,2A, so it will be Class AB, dissipating 5W per mosfet. It simulates fine...

But since I am not an experienced amp designer at all, I would like to ask you guys if that approach is valid?
Is there a similary low parts count amp schematic out there with the 2sk1058 and the 2sj162 (which are the parts I use), that could work better with the low bias approach?
Thanks for your help.
 

Attachments

  • 0_mini F5 Juma schematic w vol.gif
    0_mini F5 Juma schematic w vol.gif
    59.4 KB · Views: 261
  • Screenshot 2021-05-01 at 15.31.24.png
    Screenshot 2021-05-01 at 15.31.24.png
    107 KB · Views: 260