Recommendation for 5-10W amp.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
fab

fab

Member
Joined 2004
fabaudio.online
Paid Member
e_fortier said:
Well, after some digging I found a way to test both type of lateral MOSFET. With a 24Vdc supply, I adjusted the Vgs to get a static Id of 100mA and then wrote down the required Vgs. Most 2SK134 require 0.7V to get Id of 100mA while most of my 2SJ49 require 0.8V to get 100mA.

I have found some matching pair of 2SK134/2SJ49 but I had a much easier time matching a bunch of 2SK176/2SJ56. Most 2SK176/2SJ56 needed a smaller Vgs to get an Id of 100mA.

Would it be OK to use the 176/56 instead of the 134/49 ?? They do not seem to be as popular on the web so I wonder if they sound OK.

This is a cool project, I'm learning a lot and at the end I know it's going to be a great sounding amplifier.

Thanks for any advice.
BR,
Eric
Click to expand...
Hi Eric

It is not usual to try to match N and P channel of MOSFET. It may be overkill unless The advantage could be more symmetry in gain of input stage due to same load resistors value?...

The 176/56 is the big brother of 134/49. They are essentially the same otherwise the higher voltage of 176/56.

I am pretty sure your amp will sound good since my headphone amp based on a similar design sounds great.

Good luck
 
e_fortier

e_fortier

Member
Joined 2001
AndrewT said:
And the coupling cap on the output needs time to charge up through a route that does not include the speaker.

My interpretation of this is that speaker isolation using some form of delay timer is required to give time for the capacitor to charge to approximately half rail voltage
AND
a resistor route available for this charging current to flow without creating a DC pulse on the speaker.

This is important if a very high efficiency driver were the load.
Click to expand...

Hi Andrew,

Thanks for the recommendation. I have a 1k resistor permanently soldered after the output cap (2 x 3300uF) but being such a high resistor value the RC constant is slow so to avoid damaging my speaker I will add a switch on the speaker output. I had the same issue with my ACA.

BR,
Eric
 
e_fortier

e_fortier

Member
Joined 2001
fab said:
Hi Eric

It is not usual to try to match N and P channel of MOSFET. It may be overkill unless The advantage could be more symmetry in gain of input stage due to same load resistors value?...

The 176/56 is the big brother of 134/49. They are essentially the same otherwise the higher voltage of 176/56.

I am pretty sure your amp will sound good since my headphone amp based on a similar design sounds great.

Good luck
Click to expand...

Hi Fab,

Thanks for the info, much appreciated.

Too bad for the rainy day for our St-Jean Baptiste celebration :(

BR,
Eric
 
juma

juma

Member
Joined 2005
e_fortier said:
I have a voltmeter at the output and this is what I have, I did not short-circuit the input maybe it's the reason for this...

I do not have any load yet.

BR,
Eric
Click to expand...

Short circuit the input and put 10R resistor between the - pin of the output capacitor and the ground and measure the DC voltage across that 10R resistor. If it still shows some DC voltage it means that that you have very bad (high leakage) output cap.
If there is no DC in that setup then everything is OK, you can connect the speaker but instead that 10R use the few hundred Ohms resistor.
 
e_fortier

e_fortier

Member
Joined 2001
juma said:
Short circuit the input and put 10R resistor between the - pin of the output capacitor and the ground and measure the DC voltage across that 10R resistor. If it still shows some DC voltage it means that that you have very bad (high leakage) output cap.
If there is no DC in that setup then everything is OK, you can connect the speaker but instead that 10R use the few hundred Ohms resistor.
Click to expand...

Input is now shorted.
Output has 8 ohm resistive load.
DC voltage at output is down to about 10mV DC.

Even with input shorted my load is warming up...I think I have some oscillation somewhere, my Id is up to 1.5A and it's not stable...

I haven't soldered C6 yet...could that be the issue ?

Thanks,
Eric
 
juma

juma

Member
Joined 2005
e_fortier said:
Input is now shorted.
Output has 8 ohm resistive load.
DC voltage at output is down to about 10mV DC.

Even with input shorted my load is warming up...I think I have some oscillation somewhere, my Id is up to 1.5A and it's not stable...

I haven't soldered C6 yet...could that be the issue ?

Thanks,
Eric
Click to expand...
Load that heats up with input shorted clearly shows some mischief happening and the oscilloscope would help. In this situation 10mV reading is not reliable.
Install C6=330p and see what happens. If it doesn't settle down add 330pF-1nF parallel to each 150R feedback resistor.
 
e_fortier

e_fortier

Member
Joined 2001
Thanks for the help.

I will install C6 when I receive them ( on their way from Digikey)

This is the latest :

No load, input floating : Id is 550mA, DC out 230mV
No load, input shorted : Id is 680mA, DC out 210mV

I will follow your recommendation and report back.

Thanks,
Eric
 
e_fortier

e_fortier

Member
Joined 2001
Hi Fab,

I think I made some progress.

IF I leave the input floating (not tied to gnd) and I connect an 8 ohm load there is no oscillation, the Id goes and stays at about 600mA and the DC voltage is 3mV.
As soon as I touch the input with my finger or trying to connect a music source the Id goes wacko up to nearly 1.6A

My problem seems to come from the VAS, I will try with a smaller cap. as you recommend.

Thanks,
Eric
 
juma

juma

Member
Joined 2005
Eric,
since our bodies are big conductors, touching the input means introducing the signal induced in our bodies (mostly 50/60Hz and harmonics) to the amp so the behavior you observed is normal.
Did you solve oscillation problem (heating of the load without a signal at the input) and how does the amp sound to you (are any anomalies present) ?
 
e_fortier

e_fortier

Member
Joined 2001
Hi Juma,

Oscillation problem was due to some sort of ground loop I created with my bird nest set-up, by changing 2 gnd locations solved the issue, not even sure if I need C6 but I will add it. I'm using and will be using a SMPS for this amp, it's like a laptop supply (brick) but rated 24Vdc, 5A. Worked like a charm with my ACA and JLH (floating outputs to create a bipolar supply) so I didn't see a reason not to work with a similar rated output class A amp. Your amp which I call 'mini F5-Juma' is dead silent, just as it should be. There is no anomaly with this amp.

The amp sounds quite good but it's hard to tell with a mono set-up and an inefficient spkr (EL-70 in BR enclosure), it does indeed plays nicely but once the stereo unit is built and connected to my FE-206E I will report back.

Thanks,
Eric
 
Last edited:
e_fortier

e_fortier

Member
Joined 2001
Had the pleasure to listen in stereo tonight ;-)

How did the bias of 0.7A came about, trying to understand it. My JLH which works at +/- 14Vdc (28V total) needs a bias of 1.3A to provide an output of 10W into 8 ohm, the JLH amp dissipates 36W/channel at idle but your slightly less powerful amp which is rated about 7W has a bias of 0,7A and dissipate only about 17W at idle...smaller power supply and heatsink therefore overall reduced cost for about the same power...this is great but how is this possible..I'm scratching my head since the few SS Class A I have built are normally about 25% efficient. Yours is more like 40%.

Schematic I have used is attached.

Thanks,
Eric
 

Attachments

  • 0_mini F5 Juma schematic.gif
    0_mini F5 Juma schematic.gif
    60.5 KB · Views: 502
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom