How much do you use?
What Brand?
What is your supply voltage?
What is the impedence of the speakers you drive?
Thanks!
-Paul
What Brand?
What is your supply voltage?
What is the impedence of the speakers you drive?
Thanks!
-Paul
I don't see the need for more than 1000u per rail, per channel. That gives four caps rated 50V. They are allways from Black Gate, since using so little capacitance, allows to use better brand.
I tried both 18V & 24V on transformer and in both cases it sounded very similar, with 18V providing more bass. I used all kinds of speakers, never had a problem.😉
I tried both 18V & 24V on transformer and in both cases it sounded very similar, with 18V providing more bass. I used all kinds of speakers, never had a problem.😉
Thanks Peter.
I am using 1000uF per rail per channel, random cheap 50V electrolytics. I may have to upgrade them sometime.
I was thinking that is where some of my bass has gone.
I did try a second set of cables, but it didn't do anything.
-Paul
I am using 1000uF per rail per channel, random cheap 50V electrolytics. I may have to upgrade them sometime.
I was thinking that is where some of my bass has gone.
I did try a second set of cables, but it didn't do anything.
-Paul
I would try removing 15k resistor to ground after the pot. Maybe it makes your input impedance too low? I tried my gainclones with many different speakers, the bass was always fine. To get better bass, try lower voltage on transformer, 18V or less.
2000uf per sidePaulHilgeman said:
Rubicon ZL for the electros, Evox/Rifa for the input
24V ac
8 ohm
I have 1000uf mounted on the psu pcb and another 1000uf mounted at the chip.
Bridges are made with On Semi 100v 10a Schottkys.
Hi!
I use 1200 uF 50 V Panasonic FC for each polarity, and again 2 * 220 uF Panasonic FC in the transformer casing (for a little bit of smoothing before transferring the voltage through the cable to the clone)
Toroid voltage is 2*25 V, 225 VA (Nuvotem)
Speakers: None at all... Only built it for the fun of building, and for testing purposes... but I attached it to 8 ohms speakers (and blew my midrange woofer in initial faulty GC setup...)
Bye,
Cradle
I use 1200 uF 50 V Panasonic FC for each polarity, and again 2 * 220 uF Panasonic FC in the transformer casing (for a little bit of smoothing before transferring the voltage through the cable to the clone)
Toroid voltage is 2*25 V, 225 VA (Nuvotem)
Speakers: None at all... Only built it for the fun of building, and for testing purposes... but I attached it to 8 ohms speakers (and blew my midrange woofer in initial faulty GC setup...)
Bye,
Cradle
How much:
For supply:Started with 1000uf , 50V. Increased to as much as 35000uf, did not cause any audible improvement. Back to 1000uf.
For input: 2.2uf , 300V, WIMA MKS
Brand: Hitano (Very cheap)
Supply voltage: 15-0-15 Prim. 21-0-21 Sec. after standard 20A, 600V Bridge. Delivers 17Watt / channel, that is when it start to clipp. Run at about 30°C on smallish heatsinks.
Speaker Z: 8ohm
Get more bass then from my Aleph30/ XBoSoZ setup. Sounds extremely good for the price, but not the same as A30.
For supply:Started with 1000uf , 50V. Increased to as much as 35000uf, did not cause any audible improvement. Back to 1000uf.
For input: 2.2uf , 300V, WIMA MKS
Brand: Hitano (Very cheap)
Supply voltage: 15-0-15 Prim. 21-0-21 Sec. after standard 20A, 600V Bridge. Delivers 17Watt / channel, that is when it start to clipp. Run at about 30°C on smallish heatsinks.
Speaker Z: 8ohm
Get more bass then from my Aleph30/ XBoSoZ setup. Sounds extremely good for the price, but not the same as A30.
Hi!
If what I understood of this whole OPAmp stuff is right, then somewhat between 1000 and 2000 uF per rail is enough simply because the caps have a distance to the chip of about 0.5 cm to 1 cm, and the path in the chips is neglectible short...
And smaller caps are also faster. A lot of people mentioned here that increasing the capacity didn't improve sound at all, even if the caps were directly soldered to the IC pins, so why not stick to less, since I will also cost (considerably) less...
Bye,
Arndt
Bricolo said:
If what I understood of this whole OPAmp stuff is right, then somewhat between 1000 and 2000 uF per rail is enough simply because the caps have a distance to the chip of about 0.5 cm to 1 cm, and the path in the chips is neglectible short...
And smaller caps are also faster. A lot of people mentioned here that increasing the capacity didn't improve sound at all, even if the caps were directly soldered to the IC pins, so why not stick to less, since I will also cost (considerably) less...
Bye,
Arndt
I use no input caps
the powercaps are 1000 uF Philips ( BC now ) directly on the chip.
I normaly run tube amplifiers but this sounds perfecly clean.
No colloring at all !!
Dick
the powercaps are 1000 uF Philips ( BC now ) directly on the chip.
I normaly run tube amplifiers but this sounds perfecly clean.
No colloring at all !!
Dick
1000mF anonymus brand, must be Hitano but never heard before, on each rail directly in the legs IC and 2,2mF Evox-Rifa PHE at the input.
Even if I'm happy with them maybe I upgrade to BG or Panasonic FC just to see what happen.
Giorgio
Even if I'm happy with them maybe I upgrade to BG or Panasonic FC just to see what happen.
Giorgio
I am using Rubycon YXG 1000uF/63V per rail, an old Auricap 3.0uF input cap, Vishay-Dale 10k input resistor, Mills .16ohm output resistor and a cheap unknown brand MilSpec 220k feedback resistor. Pot is a cheap carbon audio but changing to a ALPS soon when I add a input selector switch. Supply voltage is 28VDC unloaded from dual torroids with dual BRs. Speakers are old B&W DM220s 8ohm.
I'm building an amp with LM3886. I'll be using 54,000 uF per side, with smaller supply decoupling caps at the chip. I've got a 320VA, dual 18V power transformer to supply two channels.
The 1000 uF close to the chip is a fine idea, but I hear a lot of people complaining about weakness in the bass. Think about it for a minute- the reason big amps use big caps is to prevent the voltage rails from sagging when the amp has to deliver large output power, typically at low frequencies. PSRR is a small-signal spec. It tells you how noise on the supply will be transferred to the output. It doesn't mean that the output power is almost completely independent of supply voltage. If you reduce the supply voltage, you reduce the output power. Look at the graphs of power output vs. supply voltage in the data sheet. A loud bass signal will try to dump lots of power into your speaker. If there is insufficient charge storage in the power supply, the rails will drop and the power to the speaker will drop with them.
If you want to use a 1000uF as a bypass cap for the chip, that's fine. But I wouldn't try to get any real power out of the amp without a more substantial supply reservoir. At these low voltages, large capacitors are very cheap.
MR
The 1000 uF close to the chip is a fine idea, but I hear a lot of people complaining about weakness in the bass. Think about it for a minute- the reason big amps use big caps is to prevent the voltage rails from sagging when the amp has to deliver large output power, typically at low frequencies. PSRR is a small-signal spec. It tells you how noise on the supply will be transferred to the output. It doesn't mean that the output power is almost completely independent of supply voltage. If you reduce the supply voltage, you reduce the output power. Look at the graphs of power output vs. supply voltage in the data sheet. A loud bass signal will try to dump lots of power into your speaker. If there is insufficient charge storage in the power supply, the rails will drop and the power to the speaker will drop with them.
If you want to use a 1000uF as a bypass cap for the chip, that's fine. But I wouldn't try to get any real power out of the amp without a more substantial supply reservoir. At these low voltages, large capacitors are very cheap.
MR
I thought long and hard in which thread to post it but now as I saw this thread and Mrehorst’s explanation it makes it easier to just post it here. Also I couldn’t find a good subject to this musing. I thought about headings like “The king is naked “ or Wolf in lamb’s clothing” perhaps. Sounds pretty dramatic but this is how I would describe the feeling I was left after the simple mod I did to my clone and am about to introduce.
Some of you maybe have read in this forum already, that few of us –the cloners - have been unhappy with the results of their experiment.
Namely the bass response and transient behavior of these puppies has not been adequate for reproducing the full natural character of the voices and instruments and hence has not been up to the task of recreating a true live performance illusion. The sound has been seductively pleasing and sweet overall but has failed to impress musical experience-wise in the long run.
Do yourself a favor….somebody once started this cloning saga in chip amp forum …..
…and when you have inverted and non-inverted and inverted again, reduced and increased the gain and/or supply voltages, have gone for log to linear to law fake with your pots, swapped no-name or big fame carbons, composites and films in and out like a exchange broker, have equipped your clone with capasitors from all ranks of Japanese clans, lined up the resistors and wires according to the moon’s direction, even spiked the damn creatures …
…add a normal power supply to your existing clone. By normal I mean 10 000 to 15 000 uF PS caps and 0.1uF bypass film caps of your favorite choice (with the existing 1000uF caps still remaining close to chip).
The following is my personal listening experience of so called inverted gainclones I have built, supported by the rest of equipment and room (and local power lines and station, if you mind) I am currently using, with resultant limiting and subjective factors and it does not endeavor to set the absolute truth about those chip amps, nor does it try to ruin someone else’s day and DIY-fun under the sun. 😉 😉 😉 😉 😉 😉 😉 (I need to put lot of those Smilies here in case anyone starts to think I am absolutely serious)
So I have added this normal power supply to the clone. What is immediately noticed, that all the shortcomings I complained about are gone. Completeness and equilibrium of tonal character of vocals and instruments is restored. Rhythm and punch is back. Transients and pitch sound real again. A good part of detail is added as supplement because of the mod.
Can I be satisfied? Is it all perfect now? Huh –far from that.
What this added PS capacitance has pulled of, is like opening the Pandora’s box.
All those nice, must have features, I have described, were missing there before but while now emerged, are overshadowed by the ugly, flashy, solid-state nature (in the bad sense of it) of the chip. So much for the slightly romantic, warm and soft tone, with good clarity.
You may say -those big elcos , because of their quality, are causing the degradation of sound.
It may be - I wont argue that those big Russian military grade, skinned cans, I used, can add their own character. But more likely that the grain and coarse sound was already there in the chip and was only held back by the inability of the PS to provide the sufficient juice. By dumping up the capacitance, the amp can now effortlessly reproduce all the transients and musical peaks and all the pitches can go to their full scale but at the same time they bring that irritating “chip sound” (or whatever it is called) with them.
Kimura San was apparently aware of the beast nature hidden in the chip, so he stick to the 1000uF. By doing this he killed two rabbits with one blow of the stick – smaller cap is way cheaper and it covers up the chip’s uglier parts amazingly well, only letting you to hear it’s nicer features
Argo
Some of you maybe have read in this forum already, that few of us –the cloners - have been unhappy with the results of their experiment.
Namely the bass response and transient behavior of these puppies has not been adequate for reproducing the full natural character of the voices and instruments and hence has not been up to the task of recreating a true live performance illusion. The sound has been seductively pleasing and sweet overall but has failed to impress musical experience-wise in the long run.
Do yourself a favor….somebody once started this cloning saga in chip amp forum …..
…and when you have inverted and non-inverted and inverted again, reduced and increased the gain and/or supply voltages, have gone for log to linear to law fake with your pots, swapped no-name or big fame carbons, composites and films in and out like a exchange broker, have equipped your clone with capasitors from all ranks of Japanese clans, lined up the resistors and wires according to the moon’s direction, even spiked the damn creatures …
…add a normal power supply to your existing clone. By normal I mean 10 000 to 15 000 uF PS caps and 0.1uF bypass film caps of your favorite choice (with the existing 1000uF caps still remaining close to chip).
The following is my personal listening experience of so called inverted gainclones I have built, supported by the rest of equipment and room (and local power lines and station, if you mind) I am currently using, with resultant limiting and subjective factors and it does not endeavor to set the absolute truth about those chip amps, nor does it try to ruin someone else’s day and DIY-fun under the sun. 😉 😉 😉 😉 😉 😉 😉 (I need to put lot of those Smilies here in case anyone starts to think I am absolutely serious)
So I have added this normal power supply to the clone. What is immediately noticed, that all the shortcomings I complained about are gone. Completeness and equilibrium of tonal character of vocals and instruments is restored. Rhythm and punch is back. Transients and pitch sound real again. A good part of detail is added as supplement because of the mod.
Can I be satisfied? Is it all perfect now? Huh –far from that.
What this added PS capacitance has pulled of, is like opening the Pandora’s box.
All those nice, must have features, I have described, were missing there before but while now emerged, are overshadowed by the ugly, flashy, solid-state nature (in the bad sense of it) of the chip. So much for the slightly romantic, warm and soft tone, with good clarity.
You may say -those big elcos , because of their quality, are causing the degradation of sound.
It may be - I wont argue that those big Russian military grade, skinned cans, I used, can add their own character. But more likely that the grain and coarse sound was already there in the chip and was only held back by the inability of the PS to provide the sufficient juice. By dumping up the capacitance, the amp can now effortlessly reproduce all the transients and musical peaks and all the pitches can go to their full scale but at the same time they bring that irritating “chip sound” (or whatever it is called) with them.
Kimura San was apparently aware of the beast nature hidden in the chip, so he stick to the 1000uF. By doing this he killed two rabbits with one blow of the stick – smaller cap is way cheaper and it covers up the chip’s uglier parts amazingly well, only letting you to hear it’s nicer features
Argo
For the cost of a complete good standard power supply you can build other standard clone, even with of the shelf component, and use them in a multiamp system I can assure you won't miss punch and transient attack.
Giorgio
Giorgio
argo said:
How many capacitance to use on the gainclone, as you very well put and found out on your tests, is a tricky question.
In my opinion you should add capacitance in steps, like say more 1000uF, then more 2200uF, then 4700uF, etc. and listen after every step.
If I am not wrong, the original 47 Labs Gaincard gave you the option of an external supply: the Humpty.
The other respected design using these chips was by Jeff Rowland, who used bridge/parallel arrangements and battery power.
Maybe the larger caps versions need some more time to settle down and come through.
But I do agree on using a small capacitor close to the pins and larger caps on a separate place.
Carlos
I am using now 225 VA ,20A schottky bridges and 4700uf caps on the chip , for each channel separately. It is better than the 120VA , normal bridges,1000uf caps for both channels in my other gainclone. It is hard to say which of the above are contributing and by what amount to the better sound unless I start substituting each part separately etc.I wouldn't say it is a dramatic difference and the tonality hasn't changed, but the separation and solidity are increased.
I have no problem at all with lack of bass or bright "solid state" sound as you mention. I find that a well implemented gc sounds very smooth. This is in comparison to a Naim 180(production not DIY) and Aleph 5 monblocks (DIY). What I find with the gc is that being very transparent to the source it will show up problems with bad recordings or other mediocre components. Of course each persons mileage may vary but I am surprised you find the gc problematic. I think there may be differing views as always but the majority of the forum members that have actually built gc's and compared them to good amps have very positive things to say. Perhaps we should hold a forum opinion poll where we could rate our gc's and that could clear up it's " satisfaction" rating for people confused about whether it's worth all the commotion.
I have no problem at all with lack of bass or bright "solid state" sound as you mention. I find that a well implemented gc sounds very smooth. This is in comparison to a Naim 180(production not DIY) and Aleph 5 monblocks (DIY). What I find with the gc is that being very transparent to the source it will show up problems with bad recordings or other mediocre components. Of course each persons mileage may vary but I am surprised you find the gc problematic. I think there may be differing views as always but the majority of the forum members that have actually built gc's and compared them to good amps have very positive things to say. Perhaps we should hold a forum opinion poll where we could rate our gc's and that could clear up it's " satisfaction" rating for people confused about whether it's worth all the commotion.
Sorry to revive an old post, but I was searching for some info on why everyone doing the gainclone thing is using so little power supply capacitance.
So, from what I understand, more is less when it comes to the GC? I can't imagine how that would be! Of course, I haven't built one yet, so I have no valid opinion. I was just researching.
Was there ever a consensus reached about what the optimum capacitance is for the GC?
I have 4 ohm speakers, and was wondering if it would be better to double up on the capacitance.
I still can't buy into the "faster" theory- it sounds like a lot of nonsense to me. I can understand that there may be a certain euphonic character produced by the inability of the supply to not sag under load. This has been used in guitar amps for decades now- producing a musical sustaining compression. I'm just not sure that is what I want in my audio system.
So, how much capacitance?
So, from what I understand, more is less when it comes to the GC? I can't imagine how that would be! Of course, I haven't built one yet, so I have no valid opinion. I was just researching.
Was there ever a consensus reached about what the optimum capacitance is for the GC?
I have 4 ohm speakers, and was wondering if it would be better to double up on the capacitance.
I still can't buy into the "faster" theory- it sounds like a lot of nonsense to me. I can understand that there may be a certain euphonic character produced by the inability of the supply to not sag under load. This has been used in guitar amps for decades now- producing a musical sustaining compression. I'm just not sure that is what I want in my audio system.
So, how much capacitance?
Caps
On page 12 of AN-1192 application note:
"7.2.3 Power Supply Circuit
The power supply portion of the amplifier is made up of a
typical unregulated bipolar power supply. The supply is comprised
of an input AC line filter, surge protecting MOVs, a
separate 385VA toroidal transformer for each channel, and
40,000µF of supply reservoir capacitance for each supply
voltage rail."
Are they joking?
I don't think so...
This is for a bridge-parallel configuration.
For a normal implementation I would say 10.000uf per rail.
But there's nothing like testing.
The results can be very different depending on transformer capacity.
Of course an under-dimensioned transformer drops in voltage more easily.
These 1000uf caps alone really make me scratch my head.
Oh, and the original 47 Labs Gaincard, with Power Humpty, for me is an ultrage, a robbery!
So much money...
I would buy a Krell.
But I'm making a Gainclone.
On page 12 of AN-1192 application note:
"7.2.3 Power Supply Circuit
The power supply portion of the amplifier is made up of a
typical unregulated bipolar power supply. The supply is comprised
of an input AC line filter, surge protecting MOVs, a
separate 385VA toroidal transformer for each channel, and
40,000µF of supply reservoir capacitance for each supply
voltage rail."
Are they joking?
I don't think so...
This is for a bridge-parallel configuration.
For a normal implementation I would say 10.000uf per rail.
But there's nothing like testing.
The results can be very different depending on transformer capacity.
Of course an under-dimensioned transformer drops in voltage more easily.
These 1000uf caps alone really make me scratch my head.
Oh, and the original 47 Labs Gaincard, with Power Humpty, for me is an ultrage, a robbery!
So much money...
I would buy a Krell.
But I'm making a Gainclone.
- Status
- Not open for further replies.