I think your surmise about the sound quality is correct, ginetto. Thick, solid connecting wire, silver plating, no fasteners......it all reads like 1980s pseudo-science. I would agree that the sound quality is really in the electronic circuit and its reasonably good PCB, connecting wire layout and a very basic electrical design.
Regarding nylon screws, we need to be careful how the term "melting point" is applied to such thermoplastics. There are about 4 different polymer groupings all called "Nylon" or "polyamide". These are not pure compounds with a defined molecular structure and physical properties. They are thermoplastics and really are a mixture of many molecular chain forms and lengths, so they don't have any true melting point, just a softening temperature range at which they can be injected into a metal die with extreme pressure.
The softening point is likely to be well below what might appear to be a melting point and it's very broad in temperature range. This is a problem because bolts and nuts not only stretch and distort as low as 75C but after some years, they may even become loose at normal ambient temperatures. That's the reason we don't use nylon thread fasteners in critical applications.
Regarding nylon screws, we need to be careful how the term "melting point" is applied to such thermoplastics. There are about 4 different polymer groupings all called "Nylon" or "polyamide". These are not pure compounds with a defined molecular structure and physical properties. They are thermoplastics and really are a mixture of many molecular chain forms and lengths, so they don't have any true melting point, just a softening temperature range at which they can be injected into a metal die with extreme pressure.
The softening point is likely to be well below what might appear to be a melting point and it's very broad in temperature range. This is a problem because bolts and nuts not only stretch and distort as low as 75C but after some years, they may even become loose at normal ambient temperatures. That's the reason we don't use nylon thread fasteners in critical applications.
Back on your topic, I had another look at #5 where DuncanF posted a schematic from simulation of a larger model but still essentially the same circuit. I noticed that that the feedback capacitor is quite small by hifi standards, at only 47μF. This does limit the low frequency response and hence the power requirements such that a small transformer would be less a problem because the current demand is less. I think 220 or even 470μF would be more normal there so try 220μ by bridging the existing cap (mind the polarity) in one channel only, to be certain the amplifier remains stable with audio output before going further.
I'm also fairly certain you can buy somewhat larger, low cost, general purpose transformers with multi-tap windings covering 18-20 or 24V because I can here in Oz, from 2 electronics retail chains or from on-line sources. They are all Asian imports so look around and look further than just your major trade distros like RS or Farnell.
I'm also fairly certain you can buy somewhat larger, low cost, general purpose transformers with multi-tap windings covering 18-20 or 24V because I can here in Oz, from 2 electronics retail chains or from on-line sources. They are all Asian imports so look around and look further than just your major trade distros like RS or Farnell.
Actually, putting a 47 or 100uF bipolar in there might improve the sound a touch.
Those o/p transistors seem to be good for 80V and 12A so the remaining question is whether they've got sufficient heatsinking. It might be an idea to put a slab of aluminium on the other side of the base, or even a small heatsink.
Personally I would build an external supply and split the power supply with each board getting its own pair of caps.
Those o/p transistors seem to be good for 80V and 12A so the remaining question is whether they've got sufficient heatsinking. It might be an idea to put a slab of aluminium on the other side of the base, or even a small heatsink.
Personally I would build an external supply and split the power supply with each board getting its own pair of caps.
This is quite, quite wrong. It probably comes from someone like Linn who, it seems to me, have got almost every facet of audio wrong. If the reservoir caps are close to the boards then so too are the current spikes. These are not to be underestimated - I have seen levels of 16A either side of the diode bridge in models that I have played with. The diodes are switched on only for a very short time at the top of the voltage peaks from the transformer and it has to charge the capacitor in that short time. One of the problems here is that if you double the value of the reservoir caps, so as to reduce supply ripple, then you also double the current.
What you would normally do is put a pair of capacitors next to the output stage and, though you probably can't measure this, match them up to the output impedance of your supply so that the transistors see a more resistive supply output impedance. It also gives you a chance, if your supply isn't split, to put a couple of diodes in so the other channel doesn't pinch charge from that side - and vice versa.
But by all means use a good thick set of cables to supply the DC after the res caps. The current in these will be a lot less than in the cables before the reservoir caps. An amp or two.
Hi ! thanks again for confirming the fact. I said nylon but i have to say that i have never seen them used in a unit. I said a silly thing.
The construction makes the maintenance a little tricky.
It seems tha a complete rework is needed and not easy to do
It will change completely the lay-out and i guess the concept of this amp.
The output transistors look robust ... but everything else must also be up to the task
Better not to mess with the design ?
However i think that this is not a very minimalist unit because i see many active parts ... the design is quite complex actually.
Recently i have become a fan of simpler circuits ... they usually sound good.
I like the KISS principle a lot.
Hi ! they are not alone i am afraid ... many manufacturers put the mains PS/reservoir caps directly on the amp boards and so very close to output devices especially for multichannel amps
https://www.audioadvisor.com/images/BY9BSST2M.jpg
Thank you very much again for all these very kind and valuable directions.
This has changed completed my wrong point of view. Thanks !
Actually i have seen fatter traces downstream the PS caps and often they put more solder lead on them in order to increase their section.
I could by-pass them with a good section wire i guess without touching them. Just solder a wire directly from the pd caps pins to the circuit. I have also seen this in some vintage amps. Not very elegant but i understand very beneficial
Thanks again and best regards, gino
That's not quite the same problem. The problem is separating the transformer and the PS caps with 30 or 40cm of cable. Then you get 16A going down that cable 100 times a second to charge the capacitors - usually complete with switching transients way above the audio bandwidth. These cables then radiate EMI all over the board which is added to the signal and can produce significant output.
Once you get past the PS caps the current carried is whatever is going to the speaker, or a rectified form of it on each side, and is much lower current - mostly well under 1/2 an amp. It won't (or shouldn't) have switching transients on it.
So the rule is to keep the transformer, rectifier and PS caps close together, with the leads connecting them as short as possible. The leads to the board don't have to be that thick, but why introduce a voltage drop? 1.5mm^2 (good mains lead) will be fine but in a more powerful amp you might go for 2.5mm^2 if you are feeling fussy.
No harm in screening these leads either.Incidentally, adding solder to the tracks is not really very effective, though you see it all the time. Lead and tin aren't as good conductors as copper by a huge distance, though you imagine they would be. It's a factor of around 10x worse, so you need a lot of solder to make a significant difference.
That's not quite the same problem. The problem is separating the transformer and the PS caps with 30 or 40cm of cable. Then you get 16A going down that cable 100 times a second to charge the capacitors - usually complete with switching transients way above the audio bandwidth.
These cables then radiate EMI all over the board which is added to the signal and can produce significant output.
Once you get past the PS caps the current carried is whatever is going to the speaker, or a rectified form of it on each side, and is much lower current - mostly well under 1/2 an amp. It won't (or shouldn't) have switching transients on it.
So the rule is to keep the transformer, rectifier and PS caps close together, with the leads connecting them as short as possible.
The leads to the board don't have to be that thick, but why introduce a voltage drop? 1.5mm^2 (good mains lead) will be fine but in a more powerful amp you might go for 2.5mm^2 if you are feeling fussy.
Incidentally, adding solder to the tracks is not really very effective, though you see it all the time. Lead and tin aren't as good conductors as copper by a huge distance, though you imagine they would be. It's a factor of around 10x worse, so you need a lot of solder to make a significant difference.
Thank you very much again. I have to check but i think that a 100VA same V out toroidal would fit in the same case without big problems. Not big cost and not much stress on the output stage. Something like this one here
100W R-Core Transformer for Audio Amplifier Power AMP - Selectable Input Outputs | eBay
with no other changes. Would that give any appreciable benefit ?
more or less 3 times the VA of the original i mean
R-Cores are a very good choice but would be best out of the box. I wouldn't worry too much about stressing the power transistors - they are well within their limits. That struck me as a touch expensive, but perhaps I'm out of date. India may be cheaper than China; Silchar is a brand I know.
I'm not sure why you don't follow my advice of putting everything outside the box and splitting the supply with 4 caps and 2 rectifiers. The boards are separate already so it would be made for the job. Separating channels is a really good idea and pays dividends in sound quality - especially if you are using vinyl where every dB of channel separation is needed.
But yes, that would be of appreciable benefit.
I'm not sure why you don't follow my advice of putting everything outside the box and splitting the supply with 4 caps and 2 rectifiers. The boards are separate already so it would be made for the job. Separating channels is a really good idea and pays dividends in sound quality - especially if you are using vinyl where every dB of channel separation is needed.
But yes, that would be of appreciable benefit.
Hi and thanks again I am not sure to have understood well. You suggest to build an external box with transformer + 2 rectifiers + 4 caps and then go with a DC umbilical to the amp board without nothing in between just wires ? and how many wires for the umbilical cord ? 6 (i.e. V+/V-/0 x 2) ?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.