UDNeSS, or You don't need Semisouth's

[mkc]

UDNeSS is about using active devices. So no need to use 2SK170s.
2x 2SK209GR will replace 1x 2SK170BL adequately.

50W rms into 8R means +/-32V rails and 1.8A bias.
You will need a total of 4 matched MOSFETs, each dissipating 29W.

This is just simuilation.
So build at own risk.


Patrick

.

Thanks Patrick. I thought about using the Toshiba parts because I have them. However, I did secure a small stash of the 2SK209GR devices due to you highlighting them.

I will try and play with the sim.

Mogens

 

[marconi118]

UDNeSS is about using active devices. So no need to use 2SK170s.
2x 2SK209GR will replace 1x 2SK170BL adequately.

50W rms into 8R means +/-32V rails and 1.8A bias.
You will need a total of 4 matched MOSFETs, each dissipating 29W.

This is just simuilation.
So build at own risk.


Patrick

.

Do the upper and lower mosfets need to be matched together?
Or matching is needed only for the upper and lower pairs? Meaning 2 pairs and not a matched quartet?

 

[EUVL]

2 pairs for one channel, but that means 2 quads for stereo.


Patrick

 

[Gerrit Kroesen]

Question on the OPA 604 opamp: it is still active, but not a single supplier has any stock, so it has basically become unobtainium. I checked the TI website, and they have many other audio opamps, but almost all of them can only have +/- 18 V supply. Would it be OK to feed them from a secondary supply with a L7818 3-point stabilizer connected to the 24 V rail? Would involve making a seperate PCB for the servo..

 

[EUVL]

OPA551 is in stock.


Patrick

 

[Rush]

I bought this one recently https://www.ebay.com/itm/252046879741

From China, it works at 22.5 volts in my case.
Rush

 

[Gerrit Kroesen]

Thanks Patrick! I will order them from Mouser, together with the other components. Just ordered the pcb's from JLCPCB with your gerber and will assemble it in a week or so.

 

[Gerrit Kroesen]

And I might even give a shot at the 50 W 8 Ohm version, after I got some experience with the V2 first. Your spice model of that gives promising results.

 

[EUVL]

Good amp for very little money.
No unobtaniums.
And all FET.




Patrick

 

[Gerrit Kroesen]

For the 32 V 8 Ohm version, I may try the OPA445. But not sure if I will insert the servo into that pcb design or not, have some more studying to do..

 

[EUVL]

There is no need to install servo by default.
You should build basic amp first, and loaded it with a dummy load.
Then you see how much drift you have on your case / heatsink.
If the drift is say < +/-100mV, I would not bother with servo.

More useful to install a speaker protection circuit.
For example that from Pavel :
https://www.diyaudio.com/community/threads/output-relays.191449/post-2835407


Patrick

 

[Gerrit Kroesen]

Thanks Patrick!
Will do so. )n my F5T I used the speaker protection from the DiYAudioStore. Works nice also.

 

[Gerrit Kroesen]

just ordered all parts from Farnell (they deliver much faster in NL), including some heatsinks, and will first breadboard the amps using a lab power supply. Will compare the UDNeSS V2 and the 32V/8Ohm. To be continued in a few weeks..

 

[EUVL]

There should be some large power rail caps close to the amp PCB.
Say 10,000µ, even with lab supply.


Patrick

 

[Gerrit Kroesen]

Still have a pile of those lying around, from my F5T PSU build, so I can line up 40.000 uF per rail. Thanks for the tip!

 

[nicoch58]

or just test the offset with caps in series

 

[EUVL]

Are we talking about different things ?


Patrick

 

[nicoch58]

Are we talking about different things ?


Patrick

post 351 "There is no need to install servo by default.
You should build basic amp first, and loaded it with a dummy load.
Then you see how much drift you have on your case / heatsink."

 

[EUVL]

There is no need for an output coupling cap, in our experience.
The 10,000µF above (post #354) is meant for power supply rails.


Patrick

 

[Gerrit Kroesen]

That is what I understood: 10.000 uF over the rail voltages. I guess the audio frequency AC current in the rails (amps) needs a charge buffer that lab supplies don't have.