That is understandable even it is running fine now it's good to know what was smoking. A known problem is with SMPS PSU and the onboard 10 ohm PMI mention. They sit underneath PCB named R19/R20, and if it's them you could not have seen where smoke came from when PCB is mounted on a heatsink at startup. See you are aware of Post #2534, in first section point 5. deals R19/R20, and in second section point 2. deals how you can ommit R19/R20. What about chemicals from manufactures washing PCB, could little chemical rest do little smoke ?, and what type PSU is used ?
Atupi, that magic smoke happened to me when I was doing the humming test (touching the input of the amp). I also wondered when it came from. I thought one of the transistor burned but after checking all components, I didn't find anything wrong. Later on, I found out that R21 and R22 has some dark spots on top of the light blue coating of the resistor. I checked the resistance and still read 20 ohms like the other channel. It still working. So you might want to check that which might be the same case.
Thank you fredlock for the tip, i will check tonight.
I'm using a Cap Multiplier PSU (+/- 33.5V) from PMI and already shorted R19/R20 ,D5/D6, R23 even before first start. I find post#2543 very useful for latecomers like me.
I believe that smoke came out on the same conditions like fredlock, when touching input with screwdriver.
Adrian
I'm using a Cap Multiplier PSU (+/- 33.5V) from PMI and already shorted R19/R20 ,D5/D6, R23 even before first start. I find post#2543 very useful for latecomers like me.
I believe that smoke came out on the same conditions like fredlock, when touching input with screwdriver.
Adrian
@fredlock & Atupi: Don't take this the wrong way, b/c I am all in favor of experimenting and having fun, but I would say that if you want to see what it takes for the circuit to oscillate, or run high frequency tests, a) remove the Zobel resistors first, and b) have some spare parts ready.
Sort-of like revving a car engine to see how far above redline it will go without throwing a rod...
I have done this with my (well, Shaan's circuit) TH-version, and after trying a few different things, I have a few blown latfets in my collection. In industry, that is called UBNU "Undesirable, But Not Unexpected" test results.
The Zobel resistors are rated high enough, but they a physically still very small. In the normal audio range, the series cap blocks most of the output signal. During a high frequency test, or uncontrolled hf oscillation, the cap is essentially a short, so you have most of the output signal across a tiny 20R. When a small object has to dissipate a lot of power, it will get very hot.
Not to speculate what this is doing to the other semi's.
I think the 20 ohms resistor R21 and R22 should be replaced with 1 watt SMD rating, just my opinion. What I learned this, don't do the humming test while your meter is still connected on TP1 and TP2. 
I think it creates a loop that causes oscillation. That's why shorting the input to gnd is mandatory when performing the bias adjustment to avoid this kind of problem.
I think it creates a loop that causes oscillation. That's why shorting the input to gnd is mandatory when performing the bias adjustment to avoid this kind of problem.Fredlock was right, the two 20R (R21,R22) are a bit burned in the middle but the value is the same, 20ohms.
I will change them in the future.
Do i need to isolate the pcb from heatsink using plastic washers and connect after that the heatsink to the stargound ?
Pete you are right but sometimes enthusiasm to hear some sounds is so strong
I'm so eager so hear the sound of VSSA
Adrian
I will change them in the future.
Do i need to isolate the pcb from heatsink using plastic washers and connect after that the heatsink to the stargound ?
Pete you are right but sometimes enthusiasm to hear some sounds is so strong
I'm so eager so hear the sound of VSSA
Adrian
Last edited:
Nice it was only zobel - 
and a little fright for owner which he gets over, and no cost. Concerning isolation of two screws for VAS device, i would do isolation for counter ground loops. Then when modules sit in cabinet no need to make wire from heatsink to starground if physical metal connection in cabinet parts where ground is connected in one point. But if running test setup with loose heatsink/VSSA module on table, one have to do temperary connection from starground to heatsink. Pretty sure Andrej run his setups without isolation, so up to owner. Fredlock and Atupi will you report sound quality impression when time comes .
and a little fright for owner which he gets over, and no cost. Concerning isolation of two screws for VAS device, i would do isolation for counter ground loops. Then when modules sit in cabinet no need to make wire from heatsink to starground if physical metal connection in cabinet parts where ground is connected in one point. But if running test setup with loose heatsink/VSSA module on table, one have to do temperary connection from starground to heatsink. Pretty sure Andrej run his setups without isolation, so up to owner. Fredlock and Atupi will you report sound quality impression when time comes .Fredlock and Atupi will you report sound quality impression when time comes
Hello again BYRTT,
allow me to report my findings. Short story I am in audio nirvana
, long story VSSA is in control, realistic, detailed and at the same time emotionally involving.Sad songs playing on VSSA make you cry and that was something I only experienced once before on a much, much more expensive system. What was once one note now is two or three, you can literally hear musician pausing finger on guitar string.
My amp was built some time ago so it is not euphoria of successfully building something speaking from me. My plan is to keep it for another 30 years, and I will probably buy extra ALF mosfets now while available should they ever require replacement in the future.
English is not my native language, and it is extra hard to describe something so subjective so I apologize if what I am saying is not making much sense.
Regards
Marko
Hello again BYRTT,
allow me to report my findings. Short story I am in audio nirvana
Sad songs playing on VSSA make you cry and that was something I only experienced once before on a much, much more expensive system. What was once one note now is two or three, you can literally hear musician pausing finger on guitar string.
My amp was built some time ago so it is not euphoria of successfully building something speaking from me. My plan is to keep it for another 30 years, and I will probably buy extra ALF mosfets now while available should they ever require replacement in the future.
English is not my native language, and it is extra hard to describe something so subjective so I apologize if what I am saying is not making much sense.
Regards
Marko
I agree. With a good source, you can feel the intention of the performer. Every little nuance is perceived, and it makes even uninvolving music quite interesting...I was listening to Brahm's violin and piano sonatas the other day and almost believed Baremboim was not boring at all, he, he.
Hi Marko
I like your explanation that these are not just wow impressions but rather long term observations.
You should definitely inform local DIY forum members that PMI's PCB-s are now available.
Regards, Lazy Cat
Hi Lazy Cat,
I am glad you replied because I forgot to thank you, the designer, sorry about that.
I consider your amp a bargain which you unselfishly provided for all of us as a gift to enjoy. I hope you will now profit from your work and I wish you success in your business plans, you certainly deserve that.
I will spread the word about PMI's PCB-s and your amp in Croatia for sure.
Big THANK YOU
Marko
Hi Marko thanks for sound reports. Myself have been too busy last months, must kick my a** too get my VSSA done so i also can listen and feel the reported SQ. Nice to hear from you again i remember sweet little story about hiding VSSA amp at job for your boss because if build at home you had small children who could possible hurt VSSA process, hope it went well with boss and job . Think your English is good, i understand and look forward to go further into soundstage as described and be hit by artists musicality expression. This desciption about sad song make you cry i for some years have had in headphones listenings, but after member "Esperado" had teach me about dampening my speakers by doing small circuits for speaker drivers to linearize the inpedance of driver, i tell you SQ from headphones gets out in room now, so people having time and curiousity hunting higher SQ should make trials with his "Motional impedance feedback" suggestions. At presence i have old Sansui widebander from first in 80's DC-300kHz, but i must find time so i can enjoy VSSA. Also maxlorenz thanks for posts with reports and build.
. Think your English is good, i understand and look forward to go further into soundstage as described and be hit by artists musicality expression. This desciption about sad song make you cry i for some years have had in headphones listenings, but after member "Esperado" had teach me about dampening my speakers by doing small circuits for speaker drivers to linearize the inpedance of driver, i tell you SQ from headphones gets out in room now, so people having time and curiousity hunting higher SQ should make trials with his "Motional impedance feedback" suggestions. At presence i have old Sansui widebander from first in 80's DC-300kHz, but i must find time so i can enjoy VSSA. Also maxlorenz thanks for posts with reports and build.Hi Marko thanks for sound reports. Myself have been too busy last months, must kick my a** too get my VSSA done so i also can listen and feel the reported SQ. Nice to hear from you again i remember sweet little story about hiding VSSA amp at job for your boss because if build at home you had small children who could possible hurt VSSA process, hope it went well with boss and job
Hi Brytt,
you remember well. I am still employed fortunately
, rare thing over here these days.For me casing is always the hardest and longest part. This time I invested in Pessante Dissipante 2U from modushop, perfectly tight fit for VSSA, worth every cent, you wont be sorry. I am driving 6 ohm speaker and heatsinks are more than adequate. It will drastically speed up your build and look pretty at the same time.
I also listen to headphones daily but they never gave me such emotional "hit".
Perhaps we can discuss our systems via PM.
Regards
Marko
I cannot power both channels with one trafo; 500VA for 4 modules may be on the lower limit according to my taste, and probably the voltage will fell too much in peaks.
The trafo is built with 4 coils of 28Vac to power 4 diode bridges each...
Tough decisions
sounds like a great trafo
Thanks for cabinet report tip, i will also use modushop heatsink for trial setup, a huge one 200x300x40mm. And as starting PSU the recommended Hypex as i think i remember you use too. Earlier Andrej allowed off topic in thread, as he expressed he like to hear/learn what members are messing with. Else PM.
Regards Ricky
Hi All,
Just to give VSSA news...i just place an order for 2 SMPS300RE +/-45Vdc to give a try. The Goal is to install the 2 VSSA with 280x75x40mm heatsink, the 2 SMPS300RE in a Hifi2000 330x280x80mm galaxy case. I will use a second same case to house my ES9023/preamp/motorized pot. So a neat compact but performer system.
Marc
Just to give VSSA news...i just place an order for 2 SMPS300RE +/-45Vdc to give a try. The Goal is to install the 2 VSSA with 280x75x40mm heatsink, the 2 SMPS300RE in a Hifi2000 330x280x80mm galaxy case. I will use a second same case to house my ES9023/preamp/motorized pot. So a neat compact but performer system.
Marc
Nice news update, think we all ear to hear result later, +/-45V = VSSA max power
Guys, is it great that everybody is having fun but I am still worried because I confess that I don't fully understand the philosophy behind the diagrams proprosed for the bridged VSSA connection particularly the sharing of the feedback caps between the two modules and the bipolar supply with virtual ground: that one will need high voltage main supply caps (which are more expensive) and I suppose that virtual ground, if not connected to central or star ground will make the whole amp floating...please some kind soul explain what are the best decisions, as my first inclination was to connect everything from a +/-25DVC regulated supply (and use my cheap and good Rubycon ZL) in the classic way.
Hi i try explain, but if Andrej, metallicus69, Esparado pop up trust them more.
Classic:
You can certainly do all in classic way with +/- supply rails, and feedback caps as on standalone module.
Proposed:
When everything is floating you don't have potential difference to other sounddevices in audio setup keeping noice and amp perfect invisible. You right cost with higher voltage PSU caps will be downside, but you don't need 90V as in standard setup. The feedback sharing (floating) brings better groupdelay in lows which are VSSA weakness compared to widebanders going from DC-high Kz. With 2,2mF as standard value cap, signal delay (error) at 20Hz is ~0,5mS, this being half ~0,25mS with X'ed sharing caps (see Picture in link: http://www.diyaudio.com/forums/vend...lateral-mosfet-amplifier-287.html#post3701747), and in same picture see member metallicus69 gain by using 6.8mF (12,5x35,5) 6,3V. Think it's clever ways to get even better setups, i have heard several different older widebanders from Sansui using these schematics and they all sound better than their smaller models without these circuits. If in doubt one can make trials and do listening and measuring if it's worth, but i think it pays back in SQ.