USSA-5 Build with Review

Hi Danny
sorry, most builds are for USSA-5….

Just some explanations: Because you have selected a very low IDSS value for the jfet then you are building an hybrid version between USSA-3 and USSA-3B…. Ussa-3B is also using 3K for R5/6 because of darlington bjt drivers with lower bias voltage from input stage. But since you are building USSA-3 with mosfet drivers and low IDSS, you need higher bias voltage coming from input stage and since you have lower input current then you need higher input load resistors value. 2 possible solutions (without altering The PCB):
1) Since you are not that far from the intended output bias current you can try 4k7 (or even 10k) for R5/R6. This will however put more current into the small pot And the temperature coefficient of the pot is higher than most typical metal film resistor (<=100ppm). So could be slightly more sensitive to temperature variations resulting in slightly more varying output bias current and DC offset. This could be marginal though in this application.

2) The best solution is to change 2K ohms pot to next available value (5K ohms ) pot and use 3K or 3k3 for R5/R6.

However, in the mean time you can use higher R5/R6 value as per solution 1) above. No need to change thermistor or R7/R8 for now.

Your shown option 3 is not at the proper location (see option 2 for location) and would require some pcb trace or components leads rework.

I will update the manual for option 2 in case of low IDSS for version 3.

Good luck

Fab
 
Last edited:
Hi Fab,
OK, it was not clear to me that the GR jfets were for the version with bjt drivers.
I went for the GRs because it gives a higher damping factor,
is this (GR=higher damping) still valid for the version (GR, 2SJ313Y/2SK2013Y, ECW20N20/ECW20P20) I'm building ?

I will change the pots to 5K.
 
Last edited:
Yes, the GRs gives a higher damping factor.
I did a simulation in Ltspice
Upper graph: GR jfets and a 4 ohm load has DF 31 or 62 for a 8 ohm load
Lower graph: BL jfets and a 4 ohm load has DF 20 or 40 for a 8 ohm load

Time to solder the 5k pots :)
 

Attachments

  • USSA-3 DF.gif
    USSA-3 DF.gif
    43.6 KB · Views: 112
Yes Danny, you are correct. It is just that GR grade jfet are more important for the bjt output version but it will also work fine with your version and will provide a higher DF than other original USSA-3 builders have done. You will have a unique version ;)

You can also put a 2K2 in series with the existing 2K pot but it may not be a clean job on the pcb….

Fab
 
Last edited:
The USSA-3 hybrid is alive and sounds very good. :)
First impressions: very good separation and soundstage, gives me that holographic sound with my speakers.
Bass is good, no need for higher damping factor.
So far it's the best amp I have built for my hifi-system.

I've build it modular, even if there are only two modules :)
Power supply is easily inserted this way.

Build specs:
  • LT4320 ideal bridges
  • dual mono CLC PS, star ground
  • GR grade for the SJ74/SK170, this gives 50% higher damping, I used 2k7 for R5/6 and 5k pots
  • DC coupled, since I use a DC protection on the speaker out, I can live with the dangers of DC coupled and don't have the worry about the SQ of the coupling cap ;)
  • bias at 1.3A

Thanks Fab, for this design !
 

Attachments

  • PXL_20211229_100737604s.jpg
    PXL_20211229_100737604s.jpg
    765 KB · Views: 255
  • PXL_20211229_120151945s.jpg
    PXL_20211229_120151945s.jpg
    578.1 KB · Views: 257
  • PXL_20211229_120410378s.jpg
    PXL_20211229_120410378s.jpg
    552.9 KB · Views: 251
Last edited:
  • Like
Reactions: 1 user
Hi Danny
You are welcome and thanks for your report. I am glad you are satisfied with the sound and I appreciate that you indicate “…best“….:giggle:(y)
Soundstage is the forte of the USSA series.
You also have a quite good PSU. For the no DC input cap just ensure the source or preamp has DC offset less than a few millivolts..

Have you started to experiment with H2 level adjustment if you have the THD equipment to do so or have you taken the actual THD profile at 1Wrms and 4Wrms? You may also want to wait and listen more before playing with H2…
happy listening!
Fab
 
Hi Fab,
Yes, that's what I want in an amplifier, a very good soundstage and the USSA-3 delivers it :)
I like big power supplies, all in all, an amplifier is a modulated power supply :)

DC offset of source (Soekris DAC and Iron Pumpkin preamp) is under control at around 0mv (y)
if there would be an offset at startup the upc1237 based DC protection will prevent it from connecting to my speakers.

The H2 level adjustment will be for next WE, first getting familiar to the sound without H2 adjustment, I've put the H2 pots on max. resistance.
I've done FFT measurements with REW on source components but not yet on an amplifier,
I need to bolt some big 8R resistors on a heatsink and add some attenuator resistors to not fry the inputs on my soundcard.

Now, some more listening :giggle:
Danny
 
Fab
The best sounding ofcourse ;) (USSA-5 latest version perhaps)
Actually I´m using an 8Ohm Audio Nirvana wideband speaker and preparing for a horn system
BG
BG
there is no best sounding version in all USSA-3 and USSA-5 versions. They are simply different flavors. If depends on which driver and output device you can/want to use. 5B version is the less costly one. 5.0 is the more costly one. 5.1 is in between. 5.2 has a similar cost as the 5.1.

Version 5.2 may be the most different one. It has a slightly higher damping factor and have a slightly different tonal balance (maybe appear more detailed but……maybe less natural at the same time). It is a matter of preference I guess…

Any version will be fine for 8 ohms speaker. However for the same output bias current the mosfet output versions will have an extended class A power envelope ( amplifier stays in class A at higher power).
Fab
 
Last edited: