Consider that I might have little prototypes of this front end along with a lot of others I have tried for a phono preamp design I am finalizing now, which in turn will come very handy once I put my hands on the Amp.
Now I always had this question: the BA3 FE of F5, can it be turned Balance by just making two of those?
I am the quesion is retorical because I know you can, BUT, what advantage would that by by bridging two of those?
wouldn't just add up in complexity and noise on the balance version instead of lowering it as expected by 3dB?
This topology is not truly differential, while on theUGS module that is the base for Pass's amps, the concept is completely different.
Just a thought....
Right and maybe I mis-interpreted the paper by NP, but he was showing a distortion measurement of one Fet biased at low current getting up to 2A showing how distortion reduces accordingly.
So The concept was, how can you get 2A on EACH Fet without exceeding power dissipation?
I know but 2A @ 40V is not possible, so maybe that was the case of the FET biased at 12V or so and test purposes only and maybe he doesn't bias his Fet at 2A each, although I recall him stating thzat for the XA200.5 somewhere.....maybe I am wrong!
Anyhow, on the power amp I want to put together I would like to have some SE bias on it too
Significant reduction in distortion.
Read the balanced BA3 thread
Added bonus of cheaper powersupply caps. Lower voltage rating required
It is very close to the UGS6 modules in concept in my eyes, but the UGS uses feedback and the balanced BA-3 frontend uses degeneration and getting loaded... to reduce open loop gain.
A friend of mine just heard the "old" BA-1/2 Frontend against the balanced BA-3 and......preferred the old one, in his eyes more room and more details.
When I used the "old" one I heard it in combination with Pumpkin(g) and had no good impression, both together had to much gain. Driving the BA-2 outputstage with pumpkin alone was much better. I did not compare later both frontends like my friend...
would you mind developing this concept further?
Accordingly to my consideration:
P_RMS = V_RMS * I_ RMS = V_RMS * V_RMS/R (V_RMS = V_PK/rt(2) )
-> P_RMS = V_PK/(2*R)
if the above is true if you are powering up you upper side with 20V your PK ouput is likely going to be like 18V, which leads to a maximum power of on 8ohm of about 20V.
Am I missing something?
If the thing I states is true, then for 200W class A power amp, a Voltage rail of about:
V_RMS = rt(200 * 9) = 40V
V_PK = sq(2) * V_RMS = 1.414 * 40 = 56.56V -> V_RAIL= =/-58V
Considering a dissipation per device of about 30W
I_BIAS/per device = P_MAX/V_RAIL(PK) = 30/58 = 517mA
The Max theorical current that the Amp will have to deliver is given by
I_MAX = P_RATE / V_RAIL = 200/58 = 3.45A
This will lead to a minimum number of output device per quarter of:
N_DEVICES = I_MAX / I_BIAS = 3.45/0.517 = 6.67
A safe number would then be 8 per quarter which leads to a total of 32 output devices per channel.
Please feel free to comment my calculation in case I have missed something becase I feel like since there are two halves the current each halve should substain is half of that but I am not pretty sure of it.
EDIT: I have just checked the picture for the XA200.5 and it has 36 output device which could be oversizing it or simply having 4 devices taking in care of the SE bias.
Last edited:
It is very close to the UGS6 modules in concept in my eyes, but the UGS uses feedback and the balanced BA-3 frontend uses degeneration and getting loaded... to reduce open loop gain.
A friend of mine just heard the "old" BA-1/2 Frontend against the balanced BA-3 and......preferred the old one, in his eyes more room and more details.
When I used the "old" one I heard it in combination with Pumpkin(g) and had no good impression, both together had to much gain. Driving the BA-2 outputstage with pumpkin alone was much better. I did not compare later both frontends like my friend...
Generg question: you meantion UGS6, is that a typo or why are you referring to the rev. 6? That is the version of the module I have inside my amp!
Dats a mighty efficient amp ........
I am not going to check BA3 front end again because I am too lazy but an F5 can swing 20V pk on 23V rails. The heatsinks have enough headroom for 25V rails but I'd rather keep it down. Another option would be to slap an X version of Juma's F5 (2sk2013/2sj313) at the front end.
Last edited:
I have never said I was going to clone anything as that would be no fun.
I was taking that as an example because I would like to make a 200W class A power amp.
Which btw I got why the voltage swing on each side is halved.
Therefore my calculation from before can be carried on by just dividing the voltage rail by a factor of two and this soldi allows double current on the midgets for given 30W power dissipation.
I am curios because the MOSFET can handle much more than 30W accordingly to the SOA and wondering if Nelson pushes the bussing over that limit to kind of reach the 2A biasing.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.