Re: Re: Re: Re: Silver Mica de-coupling caps
I'm confused.. can somebody attach a diagram with the latest values ?
I've attached my implementation.
Here's some quick review of the sound by the current owner:
- rather dry, without detail, sometimes a bit harsh
- no power, weak bass
my impression is, it sounds like a loud (not very loud) boombox.
the amp was left to burn in for up to 8hrs a couple of times but never for up to 24hrs unattended. In total the amp has been playing for a long time now since May 2003.
Unfortunately, due to Farnell erasing my parts list, I don't have the original pot but it's probably the cheapest from the catalog - probably a cermet type.
Currently using a single toroid for both channels, I guess 225-300VA.
I can't remember if it's 2 or 4 diode bridges but can check.
Can anyone guess what I've done wrong/how to improve ?
LS cables are single CAT5 strands, hardwired, which replaced some Mogami or Monster LS cables (can't remember which).
One last thing, dunno if its important, heatshrink and flame resistant sleeving were used alternately to cover leads and prevent shorts. The fr sleeving is very common type you often see for eg. inside cookers etc.
regards
Stelios
PS: are these the stacked film type caps that KYW recomends ?
I couldn't get the Wima ones. Would Wima MKS4 be an improvement ?
Kuei Yang Wang said:
I'm confused.. can somebody attach a diagram with the latest values ?
I've attached my implementation.
Here's some quick review of the sound by the current owner:
- rather dry, without detail, sometimes a bit harsh
- no power, weak bass
my impression is, it sounds like a loud (not very loud) boombox.
the amp was left to burn in for up to 8hrs a couple of times but never for up to 24hrs unattended. In total the amp has been playing for a long time now since May 2003.
Unfortunately, due to Farnell erasing my parts list, I don't have the original pot but it's probably the cheapest from the catalog - probably a cermet type.
Currently using a single toroid for both channels, I guess 225-300VA.
I can't remember if it's 2 or 4 diode bridges but can check.
Can anyone guess what I've done wrong/how to improve ?
LS cables are single CAT5 strands, hardwired, which replaced some Mogami or Monster LS cables (can't remember which).
One last thing, dunno if its important, heatshrink and flame resistant sleeving were used alternately to cover leads and prevent shorts. The fr sleeving is very common type you often see for eg. inside cookers etc.
regards
Stelios
PS: are these the stacked film type caps that KYW recomends ?
I couldn't get the Wima ones. Would Wima MKS4 be an improvement ?
Attachments
The sound that you describe from your Gainclone is not what most of us have experienced and I'm not quite sure why that should be.
I would recommend replacing the 18K/22K resistors to ground with a plain wire from non-inverting input to ground with no connection to ground from the inverting input. See what that sounds like.
You will find that you have more DC offset on the speaker terminals (do measure it before reconnecting the speakers) but many of us prefer the sound with that arrangement.
I would recommend replacing the 18K/22K resistors to ground with a plain wire from non-inverting input to ground with no connection to ground from the inverting input. See what that sounds like.
You will find that you have more DC offset on the speaker terminals (do measure it before reconnecting the speakers) but many of us prefer the sound with that arrangement.
Hi
a correction:
the 1uF cap across the +/-V is an EPCOS Stacked film polyester capacitor 1.0UF 63V.
There are only two diode bridges for both amps.
Would wiring the mains TX the wrong way significantly affect the sound ? I would suppose that since there 2 bridges, there's dc through the TX and if these two dc's are not opposing each other then that might lead to saturation of the core. What would that sound like ?
Nuuk, thanks for the suggestion : I can't try it right now though as the amp is not in my hands so any changes would have to be quite simple for a non-tec person to carry out safely... :|
It's definitely sth I'll try in the future though..
I'm thinking about removing the output resistor or perhaps replacing it with sth better.
Btw what are better resistors to use ?
Also is sticking to the exact values really important ?
Ie. mine didn't measure exactly 18k but closer to 17k.
I thought of using a vishay or sth in that league but they were far too expensive for me.. (£7 each) I'd rather get a separate transformer if I was to spend this kind of money.
regards
Stelios
PS: I'm about to build my own set to be able to experiment more with them. Do you know which resistors are the ones that Thorsten suggests changing to the new values ?
a correction:
the 1uF cap across the +/-V is an EPCOS Stacked film polyester capacitor 1.0UF 63V.
There are only two diode bridges for both amps.
Would wiring the mains TX the wrong way significantly affect the sound ? I would suppose that since there 2 bridges, there's dc through the TX and if these two dc's are not opposing each other then that might lead to saturation of the core. What would that sound like ?
Nuuk, thanks for the suggestion : I can't try it right now though as the amp is not in my hands so any changes would have to be quite simple for a non-tec person to carry out safely... :|
It's definitely sth I'll try in the future though..
I'm thinking about removing the output resistor or perhaps replacing it with sth better.
Btw what are better resistors to use ?
Also is sticking to the exact values really important ?
Ie. mine didn't measure exactly 18k but closer to 17k.
I thought of using a vishay or sth in that league but they were far too expensive for me.. (£7 each) I'd rather get a separate transformer if I was to spend this kind of money.
regards
Stelios
PS: I'm about to build my own set to be able to experiment more with them. Do you know which resistors are the ones that Thorsten suggests changing to the new values ?
I like the Welwyn RC55s which are not too expensive. The RC55Cs are even cheaper and I can't hear any difference. I prefer to double up for the feedback resistor.
Thorsten recommends changing the original 220K feedback resistor to 330K. The input resistor remains at 10K. I still prefer to use a wire on the non-inverting input to ground and no connection to ground from the inverting input.
All this and more information on the Decibel Dungeon Gainclone pages .
Just another thought Stelios. One strand of Cat 5 may not be enough. Have you tried other speaker cables? Or even two strands of Cat 5?
I am using a speaker cable that I 'wrote off' years ago but which seems to work very well with the IGC. It is a flat type braided cable but there is no name on it to identify exactly what it is!
I am using a speaker cable that I 'wrote off' years ago but which seems to work very well with the IGC. It is a flat type braided cable but there is no name on it to identify exactly what it is!
My description of "fairly little usable" is that the volume changes very little in the last 5 positions and goes out when in the first position
I just measured the Nobel pot and in center position the impedance split is as follows:
10k/40k
which indicates that to have any usable control in the first half of control range the curve has to be logarythmic and not linear. Although I had only 6 position on my linear pot the gradation of sound level between positions wasn't proportional (by ear) which might indicate that a linear pot is not the best choice here, and logarythmic would be better. Unless, I'm missing something in my reasoning.
I just measured the Nobel pot and in center position the impedance split is as follows:
10k/40k
which indicates that to have any usable control in the first half of control range the curve has to be logarythmic and not linear. Although I had only 6 position on my linear pot the gradation of sound level between positions wasn't proportional (by ear) which might indicate that a linear pot is not the best choice here, and logarythmic would be better. Unless, I'm missing something in my reasoning.
Re: Re: Silver Mica de-coupling caps
How would the sound of plastic track pots different from that of carbon track ones?
I cannot get this. Is it either a 15k or 18k resistor connected input to ground? Or 15k paralleled with 18k and connected from input to ground? Or...
Thanks,
Yeo
Kuei Yang Wang said:
How would the sound of plastic track pots different from that of carbon track ones?
Kuei Yang Wang said:
I cannot get this. Is it either a 15k or 18k resistor connected input to ground? Or 15k paralleled with 18k and connected from input to ground? Or...
Thanks,
Yeo
Konnichiwa,
I don't disagree, BUT, as said, 6 positions are inherently completely unusable and secondly you made a 50k Lin pot. Not a 100k lin pot....
I am more used to se a 1/9 split at this position, so the pot you are using is somewhere between the usual log pots and a linear one.
AND THE WRONG IMPEDANCE.
Of course not. You implement things as you like, not as told and they don't work. Big deal. What did you expect?
Certainly not an extremely coarse one of the wrong impedance.
First, when I write "POT" I mean "POT", not stepped attenuator. Secondly, when i write 100k I mean 100k, not 50k.
In the same sense in which Winnie the Poh concludes "Honey is a good thing"?
Have you ever compared "identical construction" (as close as possible anyway) lin & log pot's. Noticed any difference in sound?
Ahhm, lets see....
1) The way a linear Pot is "lawfaked" - which implied that 100k - 220k Pot's are required and that 50K is way marginal to work, ignoring the extremely coarse steps.
2) The way making a log track (changing composition across the length of the track) makes any log pot sound much worse than it's linear equivalent.
3) The way most log pots at low volumes lackk tracking and shift the soundstage unpredictably all over the place.
Enough for starters?
Sayonara
Peter Daniel said:My description of "fairly little usable" is that the volume changes very little in the last 5 positions and goes out when in the first position
I don't disagree, BUT, as said, 6 positions are inherently completely unusable and secondly you made a 50k Lin pot. Not a 100k lin pot....
Peter Daniel said:
I am more used to se a 1/9 split at this position, so the pot you are using is somewhere between the usual log pots and a linear one.
Peter Daniel said:
AND THE WRONG IMPEDANCE.
Peter Daniel said:
Of course not. You implement things as you like, not as told and they don't work. Big deal. What did you expect?
Peter Daniel said:
Certainly not an extremely coarse one of the wrong impedance.
First, when I write "POT" I mean "POT", not stepped attenuator. Secondly, when i write 100k I mean 100k, not 50k.
Peter Daniel said:
In the same sense in which Winnie the Poh concludes "Honey is a good thing"?
Have you ever compared "identical construction" (as close as possible anyway) lin & log pot's. Noticed any difference in sound?
Peter Daniel said:
Ahhm, lets see....
1) The way a linear Pot is "lawfaked" - which implied that 100k - 220k Pot's are required and that 50K is way marginal to work, ignoring the extremely coarse steps.
2) The way making a log track (changing composition across the length of the track) makes any log pot sound much worse than it's linear equivalent.
3) The way most log pots at low volumes lackk tracking and shift the soundstage unpredictably all over the place.
Enough for starters?
Sayonara
Kuei Yang Wang said:
I just checked 3 pots at the input of minimized IGC. 100k cermet with 10k/90k split (Spectrol) : it's definitely no good and without any usable range. 100K linear pot: it is better, but at lower volume (1/3 of the initial turn) the control range doesn't have enough resolution, which with efficient speakers might be problematic to adjust volume properly. 50K log pot with 10k/40k split: this pot provides the best overall volume control range.
My point here is not to argue which pot is the best or if a pot has to be used at all. I'm just trying to say that for proper control range the linear curve on whatever volume device one might use is not the best choice. Also using 100k pot increases tendency to oscillate, not to mention the inconvenience it presents when using passive LPF arrangement.
I don't question (or doubt) whether cermet is better than the usual Japanese conductive plastic pots. Enough people said it is.However when you say:
it may be misleading, because not all pots have 10/90 split (at mid position) and anything with more like 10/40 or 20/80 split (which is some Nobels and Alps) works perfectly well. I will not comment as to sonic qualities of those pots, as this is not my intention here.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.