@Andrew: I missed the second part of your question from a few days ago, sorry:
Yes, the pass transistor will eventually "drop out", just like any similar regulator, if there is a momentary dip in voltage which is large enough. I plan to set mine so this will not happen during normal-to-moderate loud volume.
In my mind, there are two separate goals. One is to suppress the ripple and noise which originates "upstream", meaning in the transformer/rectifier portion of the power supply. Normally this includes the 100-120Hz ripple and the diode switching noise.
The second goal is to moderate and reduce the momentary dips caused by power peaks. That was what I was trying to show in the pics, and I just happened to have the negative rail on the screen when I captured a good example, so it looked like a peak and not a dip.
If the extra power demand is sustained for a period which is greater than the RC constant of the filter, the dropout voltage will start to dip gradually, compared to what it would do without the Cap Multiplier. The effectiveness of the Cap Multiplier will be reduced, but it will still do its job.
This type of circuit can actually be optimized or "tuned" by adjusting the size of the main filter cap, the RC constant, and Vdrop, so the likelihood of an actual dropout is almost eliminated. This shifts any power supply variation below the frequency of interest. In our case, that would be below hearing range. Unfortunately, this requires knowing the supply voltage, ripple at the input, quiescent and avg current demand, etc., so it would be different for every application.
It was set at 1V. In other words, the dropout voltage was set about 1V below the average (taking unfiltered ripple into account) DC voltage before the Cap Multiplier filter, as measured at the 10,000uF cap.
Yes, the pass transistor will eventually "drop out", just like any similar regulator, if there is a momentary dip in voltage which is large enough. I plan to set mine so this will not happen during normal-to-moderate loud volume.
In my mind, there are two separate goals. One is to suppress the ripple and noise which originates "upstream", meaning in the transformer/rectifier portion of the power supply. Normally this includes the 100-120Hz ripple and the diode switching noise.
The second goal is to moderate and reduce the momentary dips caused by power peaks. That was what I was trying to show in the pics, and I just happened to have the negative rail on the screen when I captured a good example, so it looked like a peak and not a dip.
If the extra power demand is sustained for a period which is greater than the RC constant of the filter, the dropout voltage will start to dip gradually, compared to what it would do without the Cap Multiplier. The effectiveness of the Cap Multiplier will be reduced, but it will still do its job.
This type of circuit can actually be optimized or "tuned" by adjusting the size of the main filter cap, the RC constant, and Vdrop, so the likelihood of an actual dropout is almost eliminated. This shifts any power supply variation below the frequency of interest. In our case, that would be below hearing range. Unfortunately, this requires knowing the supply voltage, ripple at the input, quiescent and avg current demand, etc., so it would be different for every application.
Last edited:
I am hoping that allowing the pass transistor to be mounted from below the board, will satisfy those who need a higher current and bigger heatsinks. For a moderate power demand like VSSA, the heatsink as shown is sufficient.
However, this is not a magic bullet with any size heatsink, for class A, or any other type of amp with a high idle current, you still need a good size capacitor after the rectifier, otherwise the ripple will exceed any reasonable dropout voltage.
Last edited:
Noted, thanks. As best I can tell, this is the list as of now.
IvanLukic(3)
Freeman(tbd)
Idefixes(4)
Touchdown(2)
vgeorge(2)
Jameshillj(2)
Jsixis(2)
Praudio(2)
Compressit(2)
Cambe(2)
Gannaji(1)
Potepuh(2)
Syklab(2)
quan(2)
tjencks(2)
BYRTT(4)
Ryssen(2)
rickmcinnis(3)
Joachim Gerhard(2)
still4given(4)
Allowing for a few extra for anyone who signs up later this week, around 50 boards, in addition to those shipped already, or about to ship today.
There are a couple of FT same size output capacitors also, you can have a slight glimpse of those at the picture's far right.
I have listened to VSSA too with that same multiplier. Better enough than with bridge and reservoirs alone.
It takes larger Vin-Vout setting for class AB to avoid ripple on peaks.
Regulated SMPS reigned supreme nonetheless.
Class A TSSA came easily on top versus VSSA AB with that multiplier as a PSU for both.
Nothing was as good as a homegrown SSA with regulated SMPS though.
That one was clearly better for general resolution and bass synchronicity.
No feedback capacitors you see. No cascode too, class AB MOSFET source follower, plus different bias scheme using IR LED.
@Salas: I only have different versions of VSSA to compare, not the other SSA and TSSA. I do have a regulated linear bench supply I am using as my base line. So far, listening results with the Cap Multiplier as it stands right now, with dropout voltage set 1V below Vin are as good as the bench supply. The improvement compared to cap bank alone, or even the CLCRC setup is small but noticeable.
At the same time, LC's latest VSSA seems to have a very good PSRR (tested, relative to ripple at least), so I am at a loss to explain or measure where exactly the improvement comes from.
At the same time, LC's latest VSSA seems to have a very good PSRR (tested, relative to ripple at least), so I am at a loss to explain or measure where exactly the improvement comes from.
I am talking LC's GB board I compared, not some homegrown VSSA. It was subjectively responding to all PSUs used. As the other CF simple amps did too.
This cap multiplier you make sports feedback so it may well be even closer to using a linear power reg. Pros and cons of feedback in a power multiplier included.
This cap multiplier you make sports feedback so it may well be even closer to using a linear power reg. Pros and cons of feedback in a power multiplier included.
Hi Salas
I read your experience with great interest and plan to test standard supply/PMI supply/Hypex/Connex when i get VSSA modules expected this month and will also do thru hole version with Sanken. This test shall then make base for my final build. Do i understand right you prefer regulated SMPS for the AB amps, and if is, is it connex ?.
BR Ricky
I read your experience with great interest and plan to test standard supply/PMI supply/Hypex/Connex when i get VSSA modules expected this month and will also do thru hole version with Sanken. This test shall then make base for my final build. Do i understand right you prefer regulated SMPS for the AB amps, and if is, is it connex ?.
BR Ricky
That would depend on why you want to do it.
So, for example, the bridge rectifier will be split in half. Also, if connected to a class A-B amp, you will no longer have a balanced circuit. Each half of the board will see current pulses corresponding to the demand from one rail. (The problem actually occurs not at maximum current, but at zero crossover, when the current stops abruptly). Filter caps not connected to the same point, etc...
Class AB amps responded more readily to regulated SMPS was the trend. Current swings are wilder in those than in class A, so it makes sense.
Was double mono connex 300. Although Mr.Evil/PMI multiplier may prove a wildcard especially in double mono. Due to feedback and lowered impedance plus film cap core element that is. Those CF amps as a family with either regulated or multiplier PSU sound as silent and informative as a good preamp on headphones rather than a typical power amp on speakers is the common ground. Then the old devil called curiosity grabs you and you want to configure them optimally since it pays back.
It should work. You may have some very sharp current transients in the connections. Where is the common point of the two large filter caps? I suppose you can just try it and see how well it works.
Last edited:
Thanks for answer and glad to hear it is an regulated SMPS that is available.
Still want to make trials of different PSU and amps setup before final build, but must say it is teaser against good SQ anyway with this amp as you descripe sound.
I am hobbyist and do not have skills for making serial valtage regulator of VSSA module, but i could not get over Andrej's manual claiming z out 5 mili ohm, hey that is perfect serial voltage regulator, so that is DEVIL again, but cant do my own, and it is not simple anymore but relative cheap (see picture).
PT on holiday danish coast, very nice wheater, but lack of dayly desktop makes slow following and updates.
BR Ricky
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.