Andrew, I have no idea
Bass is lush, warm, punchy, I love it. So, I probably misunderstood the "shouty" bit. To me it sounded like "the highs will shout". Sorry. I am a total nooob. In any event, the sound at normal listening levels is great. If I could tweak the treble at high volumes, I would do it but most of the time I listen to music at low levels.I don't think it is clipping because the "harshness" is halo-like, i.e. it is evenly spread around the soundstage when volume is around 60-70%
The speakers are KEF LS50's. I am not sure is this is to be considered a "reactive" speaker.
Internal cabling is silver plated copper Olflex 260. I read somewhere that silver can add harshness but I don't really think such short runs could have any influence at all.
The speakers are KEF LS50's. I am not sure is this is to be considered a "reactive" speaker.
Internal cabling is silver plated copper Olflex 260. I read somewhere that silver can add harshness but I don't really think such short runs could have any influence at all.
Hi,
Maybe wait few several hundred hours of listening..
LS50 is not so quiet for HF, silver plated copper also.
Choose naked silver or naked copper, wtih ptfe ( Polytetrafluoroethylene) insulation.
Phil.
If the power decoupling caps are too small, there's not enough coverage for signal, resulting in a harsh tone as the amp plays louder. The 100u amplifier board power decoupling caps provided with the chipamp.com kit, are too small. The value is not even plausible for full bandwidth operation with that chip.
So, if we were to crank that up as loud as it will go, the tone *may* be overly similar to an auctioneer directly after a breakup with her boyfriend (or, in other words, a bit off-putting).
This problem can be a lot like a car starved for gasoline, climbing a hill, and suddenly suffering pre-ignition overly advanced timing issues because shortfall of gasoline pressure has made the mix go wrong whenever the engine is heavily loaded. Likewise, an amplifier with a decoupling shortfall will sound like the stability has slipped off the mark whenever you crank it up.
Ci is not a power decoupling cap. Ci is a signal coupling cap (typically the inverting input coupler).
It performs as part of an RC, so the attached resistor value also has considerable importance when it comes to the getting the right size cap.
Ci's size will affect the bass harmonic thud/boom balance.
Sizing:
Far too small = bass blocker and/or loud midrange
Slightly too small = warm/boomy/blurry bass
Just right size = no distortion
Far too large = cold/thud-only bass
The cap included with the chipamp.com kit is Far too small.
An overspend on the wire adds no benefit nor problem whatsoever. I would suggest postponing the purchase of audiophile specialist market supplies. . . until after installing plausible/sensible component values.
Compare: We wouldn't want to try indirect/unrelated repairs as a substitute for addressing a problem directly.
If one wished an audiophile like expense to also result in fine tuned performance, then I suggest to purchase several different (but likely) pairs of Panasonic and Nichicon caps (within a sensible range of values, which did not come with the chipamp.com kit) to try for power decoupling caps, as well as some different options for the signal coupling caps (within a sensible range of values, which does not come with a kit).
Meanwhile, over at the hardware store, you can get Thermostat wire, and inside that brown jacket is a bundle of many different color 20ga solid copper insulated wires that can be re-purposed for really organized color-coded amplifier enclosure hookup cables. It is good to spiral twist pairs.
Hi, thanks for the detailed reply. I'll need some time to grasp all the details. The Olflex cable is quite cheap and has a PTFE insulation which I wanted. I think they use it mainly in high temperature industrial environments ovens and stuff like that). There is only about 50cm in the amp at a cost of about 1 or 2$.
But most definitely you are right to sort the values of the components before introducing overpriced "audio" spec merchandize into the build.
But most definitely you are right to sort the values of the components before introducing overpriced "audio" spec merchandize into the build.
Last edited:
Will do.. The amp has been playing for the last couple of days and is not fatiguing at all at low levels.
60 - 70 % of what?
If the vol pot was at maximum, would the maximum source signal cause the amplifier to voltage clip?
Is 60% feeding 60% of the Source voltage to the power amplifier?
Will this level cause voltage clipping of the amplifier when the Source is at maximum?
With a reactive speaker, even your KEFs, there could be current clipping/limiting at lower signal voltages.
Of the 360° of the vol button, i.e. around 14h (6h==0)... Way to loud for me anyways.
I think 100% would blow my speakers😕
Ps
I can get a cheap second hand oscilloscope if what you are describing could be measured with one
Last edited:
If the power decoupling caps nearest the chip are only 100 uF, try changing them to 1000 uF or so. And if there are also smaller decoupling caps, such as 0.1 uF, they could be increased to 1 uF or more, as long as the lead spacing is not larger than the original ones'.
Last edited:
Hi, Just to resume the sound of my current build:
At low to mid volume the sound is fantastic. Bass is perfect, mids and highs very detailed and clear. At high volumes (just a little difficult to understand someone talking next to you) the mids and highs seem to fuse and become a tad harsh especially with denser types of music. Acoustic stuff still plays nicely but there is a little bit of an edge
I did try and find the replacement capacitors at Digikey but there are so many different types that I got lost within a couple of minutes. Could someone knowledgeable point me in the right direction and tell me which capacitors should to be swapped ? I gather, that they are the ones around the chip. But as I said, for instance, I couldn't find one with 1000uF/50v. The only ones that came up had 500v and cost more than the entire build
thank's a lot
This is the original B-GT BOM from which my amps are built
Amplifier boards
- 22k ohm – Phoenix SFR16S resistors (R2 and RF) 4
- 1k ohm – Phoenix SFR16S resistors (R1) – band colored black 2
- 680 ohm – Phoenix SFR16S resistors (R3) – band colored blue 2
- 10k ohm – Phoenix SFR16S resistors (RM) – band colored red 2
- 2.7 ohm 2w – Phoenix PR02 resistors (RZ) 2
- 100uF 50v – Panasonic FC capacitors (C1, C2, CM) 6 [change to 1000uF]
- 47uF 50v - Panasonic FC capacitors (Ci) [optional component] 2 [change ?]
- 0.1uf 63v – BC polypropylene capacitors (C3, C4, Cz) – small blue caps 6 [change to 1uF]
Power Supply boards
- 10k ohm – Phoenix SFR16S resistors (R5) – band colored red 2
- 1 ohm 2w – Phoenix PR02 resistors (R1, R2) 4
- 2.2k ohm 2w – Phoenix PR02 resistors (R3, R4)4
- MUR860 – On Semiconductor fast diodes (D1 – D8) 16
- 1N4004 diodes – On Semiconductor standard diodes (D9) 2
- 0.1uf 63v – BC polypropylene capacitors (C1 - C4) – small blue caps 8 [change ?]
- 10uF 50v - Panasonic FC capacitor (C7) 2 [change ?]
- Blue LED – Lite-on LEDs (Power LED) 2
- Terminal blocks – 2 pin (for AC input) 4
- 10,000uF 50v – Panasonic TS-U 35mm x 30mm capacitors (C5, C6) [change ?]
The sound you are discipling could also be caused by the thermal protection build into the LM3886. Did you use thermal past?
Did you twist together the leads between the PSU and the amp boards?
Here is a 470 uF Panasonic FC with the same lead spacing and case diameter, it is the highest value FC capacitor of the same diameter.
EEU-FC1H471L Panasonic | Mouser
Did you twist together the leads between the PSU and the amp boards?
Here is a 470 uF Panasonic FC with the same lead spacing and case diameter, it is the highest value FC capacitor of the same diameter.
EEU-FC1H471L Panasonic | Mouser
Hi r100,
PSU 10uF 50v - Panasonic FC capacitor (C7), it's for led, don't change..
Maybe, a little help :
Cf is optional, RF filter..
Caddock is optional too, you can use Phoenix SFR16S or something like this.
Currently I built it.
Phil.
PSU 10uF 50v - Panasonic FC capacitor (C7), it's for led, don't change..
Maybe, a little help :
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Cf is optional, RF filter..
Caddock is optional too, you can use Phoenix SFR16S or something like this.
Currently I built it.
Phil.
The LM3886TF chips have no thermal paste between them and the large copper bar they are screwed to. I was under the impressions that the "TF" version didn't need paste.
The leads between the power boards and the amps are in parallel, i.e. not twisted as you suggest. They pass below the copper bar.
Maybe I should start by applying some paste an twisting the wires before looking any further.
Thank's for the reference to the capacitor. I would never of found it
Attachments
All hard surface Thermal Interfaces need some compound to exclude air from the interface.
Only the soft Thermal washers allow for a "no Thermal Compound" interface.
The TF is even worse because it has the extra thermal resistance of the plastic package on the back plate and air in the interface.
Twisted pairs is an interference attenuation tweak. I recommend this tweak for EVERY Flow and Return Pair. No exceptions.
Only the soft Thermal washers allow for a "no Thermal Compound" interface.
The TF is even worse because it has the extra thermal resistance of the plastic package on the back plate and air in the interface.
Twisted pairs is an interference attenuation tweak. I recommend this tweak for EVERY Flow and Return Pair. No exceptions.
Hi,
I agree, here is my way of doing..
But I don't said it's the best, but it's works well.
First I observe CAT5 cable and their twists, all are differents, but that is because they are all in the same sheath.
With only two wires, lenght twist is not a big problem..
An externally hosted image should be here but it was not working when we last tested it.
Phil.
That impression is not factual.
It does require thermal paste.
Plastic to metal interface is optimal with Arctic Silver; however, that is conductive (has to be applied with care to avoid shorts) so I would prefer to use Arctic Ceramique or GC Type 44 thermal paste, which works extremely well and is non-conductive.
Heatsink and 2 necessary vents:
Convection thermal exchange requires a vent below the radiator (typically a vent directly under the heatsink--that enclosure will also need feet) for cool air intake and a vent above the radiator for hot air exit (a vent at or near the top of the enclosure or near the top of the rear panel is also typical). Have a look at an ordinary home baseboard heater (or photos) to see that cool air enters the bottom vent, runs over the radiator and hot air exits the top. An amplifier heatsink's needs of vents, are not different from that of a convector heater, save for one difference--providing good airflow (both intake and output vents) for the audio amplifier is vitally necessary for appropriate lifespan of the capacitors.
So, do you have a vent under your heatsink, AND does the enclosure have a vent at or near the top?
If an LM3886 is used in non-inverting mode, stability, layout and cable routing is of great concern as it applies to heat output.
Speaker load is also of great concern and this matter is affected at HF by adding the full Thiele Small output parts, such as the often forgotten inductor parallel with resistor output parts, which could possibly result in a happier amplifier that runs slightly cooler.
RF control parts can also reduce extraneous workout thereby reducing heat.
If your various stability measures have indeed worked, the amplifier will run cooler and possibly/probably output a more pleasant and useful tone to the speaker.
OR, if the stability has gone badly, the possible ugly tone consequences could be of at least as much concern as the extraneous waste heat output. Various language to describe the unfortunate tonal consequences of instability includes: "Harsh" "Edge" "Shouty" and "Forwards" all of which are most likely to indicate a problem with the insufficient stability and/or insufficient decoupling, usually both.
Ear tomfoolery: Unstable amplifiers can image amazingly well, but with such an impractically poor tone that cranking it up can be unpleasant. Those functions typically conflict; so, try for a useful balance between those two audio functions.
Adjustments that improve the tone without reducing the imaging performance do happen to fascinate me, but not for long, since it is a very very short list. Not surprisingly, the locales for that sort are input impedance, output impedance and better quality decoupling.
After the layout and cable routing and RF-reduction work has been completed, and after assuring that the power decoupling is in working condition, then remaining stability control for a chip amplifier is the gain setting, as in gain comp, which can be done simply by changing the gain for the entire band (very easily done, but may accrue the cost of needing cleaner power, which is most easily done by better working decoupling that may, at worst, inconvenience the layout) or a more complex measure can be done by changing the gain for only non-audio pitches. Compensation is not a spot to begin, since the eventual optimal setting is affected by everything else. The gain setting is not there for your convenience with an arbitrary figure; so, for best quality results you'll need to dial in whatever gain setting happens to result in optimized output quality, done fairly close to the finish of the build.
Compensation, especially gain comp, can reduce imaging performance in trade for a more practical tone, and that seems almost a crime until we remember that practical output is mandatory. This is a good enough reason to do everything you can to reduce the need of compensation prior to actually doing the compensation work. For example, be sure to use AndrewT's ideas on filtering, decoupling and cabling before you adjust the amp compensation/gain directly.
Although there are many ways to combat a stability difficulty, the fastest and easiest way with a chip amp is to use inverting mode; or, in other words, you're far more likely to have extra labor for stabilizing when a chip amp is used in non-inverting mode (unfortunately that is the case with the majority of kits). If it so happens that your stability efforts have not resolved stability/heat/tone difficulty, leaving you with the need of a more practical amplifier see Decibel Dungon for an inverting solution: The Decibel Dungeon Gainclone index page with links to all chip amp articles. That will be faster and easier than a long fight with a kit.
I thank you all, especially Daniel for his extensive output of experience and knowledge which will, I am sure, help many builders, including myself.
I have applied the paste to the chips and yes, they run much, much, much, much,... cooler
I also swapped the pre-amp. I reintroduced the preamp of a Yulong DA8 DAC and now I can push the volume as hard as I like without getting the feeling that something is wrong. The passive LDR pre I was using was superbe with my Quad 306 but I think it is not a good match for the LM3886 which seems to be less forgiving than the Quad 306.
So, these first steps have made an enormous positive change. The edginess has disappeared and I can push the volume much harder. The soundstage is almost scary !
I will implement the other changes like the bigger capacitors, winding the cables , etc. in the near future. But I must say that I am extremely happy with the sound at the moment.
Thank's all. I love you.
I have applied the paste to the chips and yes, they run much, much, much, much,... cooler
I also swapped the pre-amp. I reintroduced the preamp of a Yulong DA8 DAC and now I can push the volume as hard as I like without getting the feeling that something is wrong. The passive LDR pre I was using was superbe with my Quad 306 but I think it is not a good match for the LM3886 which seems to be less forgiving than the Quad 306.
So, these first steps have made an enormous positive change. The edginess has disappeared and I can push the volume much harder. The soundstage is almost scary !
I will implement the other changes like the bigger capacitors, winding the cables , etc. in the near future. But I must say that I am extremely happy with the sound at the moment.
Thank's all. I love you.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.