Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Solid State Talk all about solid state amplification.

My Sony TA-5650 VFET amp adventure.
My Sony TA-5650 VFET amp adventure.
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 15th January 2018, 11:51 AM   #11
ilimzn is offline ilimzn  Croatia
diyAudio Member
 
Join Date: Feb 2005
Location: Zagreb
The problem is that you keep comparing two things that are different and then somehow concluding the problem is the same, and then going one step further and introducing a cascode as a solution (which is not even mentioned in the text you referenced in any way).

1) How do you know it's the output stage (VFETs) that are oscillating?
2) The Sziklai connection as drawn in your quote is NOT correct (and yes, the Prof does make mistakes). It would be a Sziklai connection if colectors of Q18, Q16 connected to emitter of Q14, and also collectors of Q17, Q19 connected to emitter of Q15 (with appropriate scaling of R34/R35) The way it is drawn. the NFB loop is around BOTH Q14 and Q15.
3) Granted, the driver circuit of the TA5450 is similar in some ways but not the same, it is NOT to be dismissed that the source of the VFET goes from one power supply, and the driver part goes from another. At the very least this means that instead of a 'R45' as on your referenced schematic, there are impedances of a whole lot of withing and two power supplies.
4) Replacing the output BJTs in your referenced schematic with VFETs (including the change to the power supply situation) changes a LOT. A VFET is a completely different beast compare to a BJT. One thing which is wildly different are input capacitances (MUCH smaller), voltage rather than current drive and vastly lower transconductance, and also internal impedance on the order of 10 ohms, not kiloohms like the BJT.
5) VFET is indeed completely different than any other semiconductor in use. It is only notionally similar to a JFET for being depletion mode, but it is not in any way to be approximated with a current source like all other FETs and all BJTs. The current through a VFET depends on voltages on ALL of it's electrodes, and there are no real treshold voltages as such.
6) gate to gate bias is actually fairly irrelevant, what is relevant is gate to source, also source to drain.

I would urge you to first take the schematic of the original 5450 power amp and simply remove the protection parts, to see there is far more going on in that driver, and try to figure out why. This is the key.
  Reply With Quote
Old 15th January 2018, 03:12 PM   #12
maxlorenz is offline maxlorenz  Chile
diyAudio Member
 
maxlorenz's Avatar
 
Join Date: Oct 2003
Location: osorno , Chile
Yes, I am aware of what you state. Thank you for pointing me that Vds is also very important. Unfortunately, I cannot make another trafo with higher V to compensate for the consumption of the cascode.

Sorry, I keep making the same mistake. I meant base-to-base voltage between opposite drivers, which I understand bias these to pass current so that +/-76V reduce its value and become the V bias (together with source V), Vgs, for the VFETs. At least that's how I see it.

Removing the protection didn't help. I already tried that.

I don't know what is starting the oscillation, but the whole output section is involved, and the VFETs are at risk. The paper provided a clue to as it can be a case of Miller capacitance doing wrong. Maybe a Colpits resonator? WHo knows...not me certainly.

Anyway, you may think cascodying is no magic but certainly it does something magical to the sound: it just sounds more like music...that is why I was already modifying some of my outputs, because I heard significant imrpovements everytime, not because I am an incurable stuborn...

Someday I will hopefully understand this complex circuit, with your help I hope

Best wishes,
M.
  Reply With Quote
Old 15th January 2018, 10:43 PM   #13
ilimzn is offline ilimzn  Croatia
diyAudio Member
 
Join Date: Feb 2005
Location: Zagreb
The bias of the VFET comes from base to base difference of Q310/311. However, the result is the actual VFET bias Vgs, and it need not be the same for the P and N part as they are not ideally complementary and do vary quite a bit even within the same rank.
The bias is developed by droping voltage from the +-75V supply with respect to source of the VFET, but consider what this means - any change on the +-75V line and any change at the source (normally +-44V line) changes the bias voltage. Also, any change of the +-44V line changes Vds. Both affect the VFET current. In your case the source voltage is also affected by the output voltage, so one more variable.

Note C310/C313. Have you tried to stop the oscillation by changing values?
  Reply With Quote
Old 16th January 2018, 09:41 AM   #14
maxlorenz is offline maxlorenz  Chile
diyAudio Member
 
maxlorenz's Avatar
 
Join Date: Oct 2003
Location: osorno , Chile
Yes. I tried 39pF without success.

Now it is sounding so good that there is no way back. Though I never heard the stock amp in stereo (only mono), I like this sound better. Now not only it has good basement but musical detail and nuances, instrument's textures and colors. all are coming through nicely. Low level detail and ambient sounds are much pronounced, which gives that "live event" impression so welcome.

Now my focus will be to recuperate bias regulation. Given VFET's response curves, with my present Vds, I need much lower Vgs. I will swap some resistances and experiment. The bias pot will be increased to 2K. I can also decrease the emitter resistors for the pre-drivers and see what happens...I bet it will sound more dynamic with that move. I have now higher value for VFETs gate resistors (130R if I remember correctly) that could also influence VFET's behavior, I imagine. Ah! The joys of experimentation and learning.

I even have spare NEC VFETs which I could use on another project for an "all VFET cascode output", as time allows...

Bet wishes,
M.
  Reply With Quote
Old 16th January 2018, 08:49 PM   #15
juanitox is offline juanitox  France
diyAudio Member
 
juanitox's Avatar
 
Join Date: Dec 2006
i have one and just replace all the diodes and capacitors of the amp section.
and remove the input capacitor. . my only concern is the mesure at the bias point . i have 75mv as needed but it is not very stable during the mesure . is there a malfunction somewhere even if it sounds very good
__________________
French is not a crime.
  Reply With Quote
Old 16th January 2018, 09:43 PM   #16
maxlorenz is offline maxlorenz  Chile
diyAudio Member
 
maxlorenz's Avatar
 
Join Date: Oct 2003
Location: osorno , Chile
Quote:
Originally Posted by juanitox View Post
i have one and just replace all the diodes and capacitors of the amp section.
and remove the input capacitor. . my only concern is the mesure at the bias point . i have 75mv as needed but it is not very stable during the mesure . is there a malfunction somewhere even if it sounds very good
Congratulations, colleague.

Bias shall be measured after a while when amp is "warm", say 30min or so...then it should not move much, safe if you have signal passing

If I remember correctly, the last recommendation was 50mV.

Enjoy this excellent amp from the 70īs.
M.
  Reply With Quote
Old 27th January 2018, 12:57 PM   #17
maxlorenz is offline maxlorenz  Chile
diyAudio Member
 
maxlorenz's Avatar
 
Join Date: Oct 2003
Location: osorno , Chile
OK, I finally had the time to do some tests.
It happens that my low bias amp doesn't show crossover distortion whatsoever. I tested sine at 4 Watts (the more that my ears bear; I did not found dummy load) and there was a perfect sine on both channels. I imagine that VFETs in this arrangement never go off...
(see example picture; lower power)

I also reduced the voltage consumption of the common-base cascode BJT to only 4V.
Vgs is now around 16.5V for the VFETs. Bias is 6ma!
Now I will begin listening tests.

I only tested music until around 25Watts and it did not clip.

I attach a concept schematic (not a finished one; probably contains a lot of errors) of the modded 5650, based on the several discussions about "thermal memory distortion" going around. I plan to test some of these on my amp. Who nows, perhaps I end with a "memoryless" amp...I particularly like the cascoded drivers which would allow tight control of the bias voltage for the VFETs, which in my case should be lower, and even allow the possibility of using another different power supply, other VFETs or make a clone amp...apart the known effects of this strategy.

Cheers,
M.
Attached Images
File Type: jpg Sine.jpg (744.6 KB, 157 views)
File Type: jpg SONY TA5650 AMPLI CASCODO-CASCODO2 CFP SCHEMA.jpg (278.2 KB, 160 views)
  Reply With Quote
Old 27th February 2018, 12:05 PM   #18
maxlorenz is offline maxlorenz  Chile
diyAudio Member
 
maxlorenz's Avatar
 
Join Date: Oct 2003
Location: osorno , Chile
To keep this thread updated, I copy the mod posted on my other thread about "thermal memory":

"OK. I know have the input section of my Sony TA-5650 properly modified with Peufeu's Cascoded CFP trick (see diagram on previous post). The effect was easily detected as being the same caused on the Blame. Less aggressive highs, more midrange presence (the bodies of the instruments are very pronounced), better delineation of melodic lines, in all, if I knew how to speak Latin, I would say the sound is now rotundum

Fortunately, the DC offset was low and I had not to mod the circuit: 10mV and 1,8mV."

This is the thread:

The AMNESIS amp: a good amplifier, like a gentleman, has no memory.

I had a little difficulty in fitting the units in the existing space for the differential pair. I had to desolder some film caps and the re-solder them.

Very worthwhile mod.
Cheer,
M.
Attached Images
File Type: jpg CASCODED CFP unit.jpg (633.6 KB, 141 views)
  Reply With Quote
Old 4th May 2018, 02:55 PM   #19
maxlorenz is offline maxlorenz  Chile
diyAudio Member
 
maxlorenz's Avatar
 
Join Date: Oct 2003
Location: osorno , Chile
I see that my vertical cascade of the output elements, while providing me a stable amp to enjoy, is technically wrong or far from optimal.

I plan to mod it as follows (see attachment), as time and my Amnesis amp project allows. That would be, IMHO, a proper cascoded common-source output for the VFETs, maintaining the original bias, hopefully...
(please ignore the driver's mod)

Cheers,
M.
  Reply With Quote
Old 12th March 2019, 11:14 PM   #20
maxlorenz is offline maxlorenz  Chile
diyAudio Member
 
maxlorenz's Avatar
 
Join Date: Oct 2003
Location: osorno , Chile
Now that my Amnesis experimental amp project is on its final state, I can update.

Here, some pics about the present state of the SONY TA-5650, while I rest and plan the next move--> use big power devices (BJT or MOSFETs) instead of that little BJTs, cascodying the VFETs at higher V reference point, "closer" to the VFETs.

With the present arrangement, normal current biasing is possible. No oscillation whatsoever...

Cheers,
M.
Attached Images
File Type: jpg SONY TA5650 -CASCODO2 CFP SCHEMA.jpg (294.1 KB, 61 views)
File Type: jpg SONY TA5650 CFP input.jpg (950.0 KB, 32 views)
File Type: jpg SONY TA 5650 HS.jpg (580.7 KB, 23 views)
File Type: jpg SONY TA 5650 speakerBP.jpg (498.2 KB, 27 views)

Last edited by maxlorenz; 12th March 2019 at 11:17 PM.
  Reply With Quote

Reply


My Sony TA-5650 VFET amp adventure.Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
VFET-X (or my 1/3 40th Anniversary Sony VFET Clone) WalterW Pass Labs 177 27th April 2018 11:14 PM
Sony TA-5650 vfet rebuild- Still cant get it stable! sd6 Solid State 6 2nd January 2018 05:08 PM
Sony TA-5650 V-Fet HermanKarton Solid State 0 26th July 2014 10:46 PM
Sony TA-5650 - coupling caps jdelange Solid State 10 5th January 2013 10:10 AM
VFETs replacement, Sony TA-5650 pietres Solid State 2 7th October 2008 02:07 PM


New To Site? Need Help?

All times are GMT. The time now is 01:43 AM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2019 DragonByte Technologies Ltd.
Copyright ©1999-2019 diyAudio
Wiki