Agreed,
Besides other replacements with better parts I have ordered original Wima FKP2-1000 100P2 and FKP2-1000 47P, both with Content spacing 5mm, to replace ceramic capacitors. Also, judging from photos, the original doesn't use ceramic capacitors. Also, I will replace all other capacitors with better parts. I will replace the 330uF/35V as suggested by Ultima Legione with 470uf/100V (Panasonic FC-A 470U 100).
Perhaps, that wouldn't increase "sound quality" but I will know that I have done everything I could to do so.
I understand Chinese, they are driven by demand and economy. In general, people expect Chinese stuff to be cheap and they must make some compromises to sell at that price level. It is ridiculous that some people expect: quality, guarantee and technical support at that price level. No way! Taht is why we have to use bare boards and start from there. In the end, the main cost are heat sinks, large transformers and capacitors and fabrication of chassis which we purchase home at much higher prices.
EDIT: Now it is confirmed that both multiple pairs of devices as well as notorious "servo" deteriorate sound. Good! I don't need more than 10W for my listening habits on 97dB sensitivity loudspeakers. You have already explained me how to adjust the midpoint, so no problems for me at all.
Muito obrigado meu caro amigo
Last edited:
you can do this:
- cut a 2 mm thick rectangle of aluminum, abundant compared to the size of the PCB, let's say double between length and width
- on this rectangle, afterwards, you will fix both the driver and the final transistors, but do not bolt them: you get 2 perforated aluminum ELLE head and tail with which you crush the transistors against the aluminum framework
- for the heat sink you always get rectangles in aluminum or "U", base 40 mm and height 80 mm thickness 1 or 2 mm (they cost nothing)
- fixed, spaced at least 10 mm away, these home made heat sinks at the aluminum corner and the heat sink is ready, as Otala did in 76 with its Class A
Do not use hyper transformers, they are useless
Thanks Domenico for good advice but I really don't mind to spend some money on things that are real fun. Since September I have spend 3.000-4.000 EUR on this hobby and already have 3 pairs of large heat sinks and several smaller.
Yesterday I have finally found a shop for proffessionals where I have purchased a 2 sq. meters of 3mm aluminium sheet for chassis. These days it is impossible to buy anything except the food due to this Virus Madness. These days I will cut that sheet into smaller 500mmX500mm pieces for easier storage. I cut aluminium sheets on table saw and have also a 1,5 hp router for fine finishing of aluminium and wood. I also do some TIG/MIG welding in my garage. Those are several of my hobbies. Otherwise, I am clumsy chemist.
Similarly to your proposed technique I have recently cut an aluminium L-shaped holder for TO-3 output stage of mine KSA-50.
One question: I have ordered 200 alumina insulators for TO-220 and TO-264 devices to provide better heat transfer for my heat sinks. What I want is to provide conditions for safely increasing biass to higher levels to achieve better linearity. In my view that is the best way to have really good sound. I want to do this on NHB-108 by slightly changing values of resistor pairs R9/R10 and R11/R12 as suggested by Danny earlier in this thread. What is your opinion on that?
I wish you all a pleasant weekend and lots of fun.
Buon fine settimana 🙂
PS: What do you mean with "hyper transformers"? I use toroidal transformers.
Yesterday I have finally found a shop for proffessionals where I have purchased a 2 sq. meters of 3mm aluminium sheet for chassis. These days it is impossible to buy anything except the food due to this Virus Madness. These days I will cut that sheet into smaller 500mmX500mm pieces for easier storage. I cut aluminium sheets on table saw and have also a 1,5 hp router for fine finishing of aluminium and wood. I also do some TIG/MIG welding in my garage. Those are several of my hobbies. Otherwise, I am clumsy chemist.
Similarly to your proposed technique I have recently cut an aluminium L-shaped holder for TO-3 output stage of mine KSA-50.
One question: I have ordered 200 alumina insulators for TO-220 and TO-264 devices to provide better heat transfer for my heat sinks. What I want is to provide conditions for safely increasing biass to higher levels to achieve better linearity. In my view that is the best way to have really good sound. I want to do this on NHB-108 by slightly changing values of resistor pairs R9/R10 and R11/R12 as suggested by Danny earlier in this thread. What is your opinion on that?
I wish you all a pleasant weekend and lots of fun.
Buon fine settimana 🙂
PS: What do you mean with "hyper transformers"? I use toroidal transformers.
Last edited:
The topology of the big 458 looks like famous project in our forum, the VSSA and its variants.
However, it feedback from output of VAS transistors. This thing will be hard to control DC offset as output.
However, it feedback from output of VAS transistors. This thing will be hard to control DC offset as output.
the resistors from 390R and 7.5 Kohm?
I don't know, I left them as they are, as values.
I'm on the signal path and if you really want to feel an objective difference, listen to it with these resistances and then replace them with Texas 2575
it seems that the control of the offset is entirely borne by the 2 x 15 millifarad capacitors each
do you know if there has been talk of the Viola Stereo scheme?
@Domenico,
This post below is what I was referring to. It appears that Danny used another, black PBC with different values of resistors. This means that our newer version already has new suggested values. Obviously, these clones evolve with time but I beleive there is thermally still some room to further increase biassing because my amplifier with large heat sink is hot after using it overnight but I can touch and hold fingers on the heatsink and output devices which are hot but I believe below 60°C. To test this I plan to solder a wire in place of these resistors and try to connect different resistors. I hope to improve sound with further increasing bias. Ofcourse, I will use thermostat switches on both heat sinks for safety.
This post below is what I was referring to. It appears that Danny used another, black PBC with different values of resistors. This means that our newer version already has new suggested values. Obviously, these clones evolve with time but I beleive there is thermally still some room to further increase biassing because my amplifier with large heat sink is hot after using it overnight but I can touch and hold fingers on the heatsink and output devices which are hot but I believe below 60°C. To test this I plan to solder a wire in place of these resistors and try to connect different resistors. I hope to improve sound with further increasing bias. Ofcourse, I will use thermostat switches on both heat sinks for safety.
Last edited:
It is often a good choice to polarize properly drivers and pre drivers and having TO 220s like MJE 15033 and & you can do it.
Important not to load the pair of endings that does what it can and that I always believe, generally, not to warm up too much to avoid the pop corn effect and not to stress them that is useless
Important not to load the pair of endings that does what it can and that I always believe, generally, not to warm up too much to avoid the pop corn effect and not to stress them that is useless
The max dissipation for the 7.5k may be getting close to the red line at full power. Apart from that these are the best sounding resistors money can buy, too bad 8 are required for stereo 🙂
if put in the right places they can change an electronics.
Yet many laugh at it
I don't even understand why to insist with the filter cells for the finals, much better stabilized power supplies, they cost almost nothing and much less than the filter electrolytics and they work much better , bah
Yet many laugh at it
I don't even understand why to insist with the filter cells for the finals, much better stabilized power supplies, they cost almost nothing and much less than the filter electrolytics and they work much better , bah
Last edited:
But how is it possible to rise the bias? I tried some suggestions I got here but it did not work. Seems like it oscillated instead.
Rails 2 x 56 Vdc possible? And 50 ma pr transistor set? 2 sets output transistors.
I though you wanted to raise the bias, not reduce it, right?
@56v rails the drivers run at 110mA and the outputs at 262mA...This is for single outputs. 12W dissipation per output transistor, 6W per driver and 5W per CCS. 30W overall dissipation per rail.
For 2 sets the bias can be raised within reason.
Of course these are only predictions and before making any adjustments it is advisable to take measurements of a working board.
@56v rails the drivers run at 110mA and the outputs at 262mA...This is for single outputs. 12W dissipation per output transistor, 6W per driver and 5W per CCS. 30W overall dissipation per rail.
For 2 sets the bias can be raised within reason.
Of course these are only predictions and before making any adjustments it is advisable to take measurements of a working board.
I will measure this night then. I am sure bias is very low. It takes hours to get just more then hand warm. I use a two set version with 50 V rails. Planning to change one 1500 Va transformer to 500 Va for each channel 40 vac.
I have only 0.2 MV across 0.22 ohm emitter resistor. Maybe I have caused confusion since this is 108?
Last edited:
Wow. That is low. Do you think that apart from the added output pair the rest of the circuit follows the "official" one?
Btw 0.22ohm is an excessively high value for the emitter resistors. It seems with two pairs the damping factor is now less than the original with a single pair. Perhaps because they didn't bother matching the pairs.
What voltage drop do you have across the large 39ohm resistor?
Btw 0.22ohm is an excessively high value for the emitter resistors. It seems with two pairs the damping factor is now less than the original with a single pair. Perhaps because they didn't bother matching the pairs.
What voltage drop do you have across the large 39ohm resistor?