Hi BYRTT good to see interest for all VSSA form. By my side i probaly will replug VSSA this evening to see if my perception have changed. It take me less than 5mn to swap from BG1 to VSSA...For moment i have Dire Straits "on every street" album playing...what greats pleasure time. I love track 4, 6, 8 and 11...Good work destressing time...
Marc
Marc
Hi Marc yes certainly interest, but sorry no report back before later when i get devices to play with.
You really turned me on with the SANKEN version, because i live with those devices in daily basis in form of old Sansui DC coupled VFB amps.
Yes good idea trying a swap with VSSA for confirming what you heard yearlier, smart your layout makes swap doable.
Lovely album and HQ recording agree with the tracks, do you have M. Knopfler Sailing to Philadelphia try track 2,5,8,12 especially track 12 go to mind/feelings on good gear.
Comment on groupdelay funktion of cap if tweaked is welcome.
BR Ricky
You really turned me on with the SANKEN version, because i live with those devices in daily basis in form of old Sansui DC coupled VFB amps.
Yes good idea trying a swap with VSSA for confirming what you heard yearlier, smart your layout makes swap doable.
Lovely album and HQ recording agree with the tracks, do you have M. Knopfler Sailing to Philadelphia try track 2,5,8,12 especially track 12 go to mind/feelings on good gear.
Comment on groupdelay funktion of cap if tweaked is welcome.
BR Ricky
I have 4x Nichicon 2200uf/6.3V take spare when order from mouser for VSSA as 1000µF/50V...i wil try to change FC in place.
Marc
Last edited:
Hi Marc that would be very interesting to hear if its audible, try to listen for if Lows get lower and in theory faster/tighter bass (downside in this test is that we not use FC against FC, but FC against Nichicon think you agree). I attach picture with simulated examples of different cap values (Look at yellow line, this is possible with FC 15000uF/6.3v).
One thing i care about with the slowness down there is that most speaker/box have same error Down there (even higher delays), and one error + one error = double error so could we remove one we have the haff error.
If it is audible maybe we could adopt one of those huge value (farads) storage capacitors available nowadays for making overkill-like in this position.
And last thing, at same place Andrej use 10uF bypass which i think is intended for take out the inductor effekt by large caps, so it could be a very important cap for both SQ and stability in high/ultra high band.
Attachments
Last edited:
Hi BYRTT,
So i replug LC VSSA and i can confirm BG1-BJT have better lower presentation (more weigth, more details, more dry, driver seem to be better manage by amp in these chapter. VSSA is on the brighter side, may be a little bit "airyness" but it's realy difficult to say one is the absolute winner...If VSSA had the BG1-BJT lower frequency presentation it would be the on the top.
I think to that capa arragement could be better on BG1 : 2200uf/6v3 or more with 10uf MKT (i juste order one more set of 4), maybe the 100uf must lat place to some thing bigger....
Marc
Add on note : VSSA seems to play a little bit more in back, as with BG1 when you close eyes artists seems to be nearer. Remember my listen room i small and speaker are at 2.00 o 2.50m wide from listen place. With VSSA i have a small feeling of "artsists are placed along a half ring going through speakers and behind them". In this chapter Bg1 gives more natural présentation but has a bit smaller stage deepness...One more time all this fact are on small range, and BG1 / VSSA are over what i have built in past.
Marc
So i replug LC VSSA and i can confirm BG1-BJT have better lower presentation (more weigth, more details, more dry, driver seem to be better manage by amp in these chapter. VSSA is on the brighter side, may be a little bit "airyness" but it's realy difficult to say one is the absolute winner...If VSSA had the BG1-BJT lower frequency presentation it would be the on the top.
I think to that capa arragement could be better on BG1 : 2200uf/6v3 or more with 10uf MKT (i juste order one more set of 4), maybe the 100uf must lat place to some thing bigger....
Marc
Add on note : VSSA seems to play a little bit more in back, as with BG1 when you close eyes artists seems to be nearer. Remember my listen room i small and speaker are at 2.00 o 2.50m wide from listen place. With VSSA i have a small feeling of "artsists are placed along a half ring going through speakers and behind them". In this chapter Bg1 gives more natural présentation but has a bit smaller stage deepness...One more time all this fact are on small range, and BG1 / VSSA are over what i have built in past.
Marc
Last edited:
OOOh gets boy alike and can't wait for my turn to get pakage from Andrej (because of your Words "BG1 / VSSA are over what i have built in past". To explain why i do Andrej,s VSSA first, when it arives in august, and then your BG1 is because of time issues and speaker projekt.
your "Add on note:" is lovely explanation to a music lover. I guess your narrow listening position gives headphone alike (two speakers one listener so close). I have Alpair 10.2 sealed for top and 10" bottom, and this is going to be same top and new 15" Sealed bottom. But when i run the Alpair also 2.5m alone with EQ (DigitalDomainLinearPhase) they sound very pure ala headphone, probably because of phase domain are untouched. In this senaria i melt of pure sound pleasure and like to raise volume, but the EQ for the lows would then kill the speaker. So gold with speakers is volume, and i have a fair chance to bring in the 15" and A10.2 in the same phasedomain with some untried tricks.
Seems your BG1 on top, to dig on possible reasons. Am i right all your devices are fairly/well matched, what about the Sanken. Don't think devices matched in VSSA, except Alfets if proberly produced.
If you order more caps, maybe try 15000uF/6,3v (Yellow line in post 244), or i will try later when my turn to test comes.
Sometimes talk too much, hope BORIS and other find interesting, else note me not too.
BR Ricky
your "Add on note:" is lovely explanation to a music lover. I guess your narrow listening position gives headphone alike (two speakers one listener so close). I have Alpair 10.2 sealed for top and 10" bottom, and this is going to be same top and new 15" Sealed bottom. But when i run the Alpair also 2.5m alone with EQ (DigitalDomainLinearPhase) they sound very pure ala headphone, probably because of phase domain are untouched. In this senaria i melt of pure sound pleasure and like to raise volume, but the EQ for the lows would then kill the speaker. So gold with speakers is volume, and i have a fair chance to bring in the 15" and A10.2 in the same phasedomain with some untried tricks.
Seems your BG1 on top, to dig on possible reasons. Am i right all your devices are fairly/well matched, what about the Sanken. Don't think devices matched in VSSA, except Alfets if proberly produced.
If you order more caps, maybe try 15000uF/6,3v (Yellow line in post 244), or i will try later when my turn to test comes.
Sometimes talk too much, hope BORIS and other find interesting, else note me not too.
BR Ricky
Sanken are not matched as the other are due their prices. Even in my 100xMJL3281 and 100xMJl1302 i was be abble to form 8 pair NP well matched imagine that with about 5 to 7€ the unit...NJW3281/1302 are indicate in datasheet as to be NPN/PNP Gain Matching within 10% from 50 mA to 5 A but is this NPN/PNP matching necessary??
Marc
Last edited:
Thank you Marc
I apreceate your report.
I do have the same experience and got simmilar audible differences between LATfet and good BJT..
Please do not forget Panasonic FC does not have the same weight like Nichicon Gold and the same time the second one has thinner mid area and better highs.
So capacitors are factor to when you compare the two VSSA type.
One of the reason I didn't gave up to use darligton (BJT) and go for LATfet even do I have some bias drifting issue with.
That would be an easy solution to make the amp rock stable.
To me LATfet in the bass area a bit shy compare to BJT, please do not think I'm against LAT fet..Just my experience.
And also here the rest of the system plays a big role especially the speakers.
Greetings G
This position of the FB cap is very unique since we decouple FB from output DC current to flow and at the same time injecting operating point DC bias current for the input stage and doing it that in double, splitted FB path. This proved to be the best way also considering thermal conditions, since we are getting in the sub mV output DC offset range without the need the input pair to be matched and without the need to use any invasive DC servo to come to this superior DC offset levels. Any capacitor here insuring lowest AC voltage drop across the bandwidth is the best choice. Group delay and/or mHz (mili Hertz) range can only be realized with enormous capacitor value, measuring in F range. The good thing is that it has to be declared only to 1 V DC specs and that is perfect place to use quality memory backup caps.
Exacly for that reason.
When i think i see Pictures (when scaling in) of small misallignment in reproduction waveforms compared to refence waveform, so certainly nice to have it matched. Is it audioable? it is certainly expencive. As far as i remember i have seen PMI say its worst at higher level.
Bed needs me now, thank you for nice diy learning, BR Ricky
That just proves the Lat Fets implementation also need VAS to be buffered so to use drivers and than the situation is quite different.
That is exactly the reason why the First One will be introduced.
P.S. Please don't forget why I recommended SMPS to be used together with VSSA.
Last edited:
That is exactly the reason why the First One will be introduced.
P.S. Please don't forget why I recommended SMPS to be used together with VSSA.
Don't forget that in my situation both (VSSA/BG1-BJT) share same crappy supply...
Marc
LC
What You think about putting another driver cascade after vas pair (as it is in bjt version of amp) ?
So instead 6 transistors we will have 8 transistors in signal track ?
And there is another thing going around my head - put j-fet input in the amp, or simple j-fet buffer just befor input.
Idefixes
Nice board for toshibas mosfets - i have 8 pairs just in case in my pocket.
I have some conclusion, I am measuring/testing amps (just to confirm technical side of it) but they really soud different even if they are the same on measurement - so conclusion is to build and listen to it and so on ........ uhhhhh
What You think about putting another driver cascade after vas pair (as it is in bjt version of amp) ?
So instead 6 transistors we will have 8 transistors in signal track ?
And there is another thing going around my head - put j-fet input in the amp, or simple j-fet buffer just befor input.
Idefixes
Nice board for toshibas mosfets - i have 8 pairs just in case in my pocket.
I have some conclusion, I am measuring/testing amps (just to confirm technical side of it) but they really soud different even if they are the same on measurement - so conclusion is to build and listen to it and so on ........ uhhhhh
