Hi ttan98
Thanks for that, I was lucky my Dad was a carpenter so I picked up a few hints along the way.
Got any rain in Melbourne yet, I was down there a few weeks back for a break (my wife and I love Melbourne) and it looked pretty dry, the drought has well an truly broken in our little town.
Thanks for that, I was lucky my Dad was a carpenter so I picked up a few hints along the way.
Got any rain in Melbourne yet, I was down there a few weeks back for a break (my wife and I love Melbourne) and it looked pretty dry, the drought has well an truly broken in our little town.
Before this thread gets completely hijacked by open baffle speaker discussions (I'm also considering an open baffle midrange, so I'm eager to learn more about it, but not on my amp thread!) and Australian weather reports, let me respond to Daniel.
I was mistaken about Rin; I don't understand the purpose of the 1k series resistor. Okay, so now I know the 22k is Rin. Thanks for clearing that up. In that case, I probably won't mess with it. I intend to try the 330pf in parallel with Rin next.
However, the rest of your babbling about resistors leaves me totally confused. You really need to work on your communication skills. I mean that in the nicest way possible. I didn't understand a single thing you wrote, and I've noticed this tendency in some of your other posts, and similar complaints from other readers, so I know it's not just me. I do appreciate your contributions, but I admit I ignore certain parts of them.
Regarding the speaker ground: once you jumper around the original PCB ground trace, you must then break that trace somewhere between the speaker/zobel path and the large coupling cap or you will have dual paths to ground, and that's probably not good! Don't worry about permanently altering the PCB, because you'll never consider going back to the original layout. If you do, for some reason, you can simply jumper the break you made.
I'm almost tweaked out on this little amp, and pretty soon I'll have two identical amps to try in stereo. Glad you're all enjoying the journey along with me.
Peace,
Tom E
I was mistaken about Rin; I don't understand the purpose of the 1k series resistor. Okay, so now I know the 22k is Rin. Thanks for clearing that up. In that case, I probably won't mess with it. I intend to try the 330pf in parallel with Rin next.
However, the rest of your babbling about resistors leaves me totally confused. You really need to work on your communication skills. I mean that in the nicest way possible. I didn't understand a single thing you wrote, and I've noticed this tendency in some of your other posts, and similar complaints from other readers, so I know it's not just me. I do appreciate your contributions, but I admit I ignore certain parts of them.
Regarding the speaker ground: once you jumper around the original PCB ground trace, you must then break that trace somewhere between the speaker/zobel path and the large coupling cap or you will have dual paths to ground, and that's probably not good! Don't worry about permanently altering the PCB, because you'll never consider going back to the original layout. If you do, for some reason, you can simply jumper the break you made.
I'm almost tweaked out on this little amp, and pretty soon I'll have two identical amps to try in stereo. Glad you're all enjoying the journey along with me.
Peace,
Tom E
Hi Mad,
you should (must?) fit a pair of filters on the input.
The High Pass to block DC coming from the source and a Low Pass to attenuate the Radio Frequency interference that comes in through the cables.
The 1k0 series resistor and the 330pF capacitor from line in to signal ground are your RF (low pass) filter and are a single pole (-6db/octave) passive filter. They cut the treble by 3db at their turn over frequency.
If Rs (source impedance) is 100r then the total R in series with the input signal is 1k0+100r=1k1. The C is 330pF.
The filter frequency is (F-3db) 1/Pi/2/C/R=438kHz. Quite high. This could usefully be made a little lower, possibly as much as one octave lower, but at 200kHz you may notice a slight change in timbre of the highest harmonics.
The DC blocking capacitor can be fitted in the source equipment or in the next amplifying stage. Most builders put it in the power amp input and in the preamp input.
This is a series line input capacitor with a resistor to ground (Rin=22k).
If you use a 10uF series capacitor with 22k Rin the F-3db=0.7Hz
You should ensure that these two passive filters set the bandwidth of your amplifier.
If your amp cannot properly handle 0.7Hz, then you must move the high pass filter up until it becomes the bottleneck. i.e. make the capacitor lower in value, or add a smaller cap to the source equipment.
If the source has a 2u2F cap then the effective capacitance is now 2u2+10uF=1u8F and f-3db=4Hz. This is noticeably bass light with most wideband speakers. Experiment with these two filters to find what suits your ears and your equipment. BUT, remember that bandlimiting rule.
you should (must?) fit a pair of filters on the input.
The High Pass to block DC coming from the source and a Low Pass to attenuate the Radio Frequency interference that comes in through the cables.
The 1k0 series resistor and the 330pF capacitor from line in to signal ground are your RF (low pass) filter and are a single pole (-6db/octave) passive filter. They cut the treble by 3db at their turn over frequency.
If Rs (source impedance) is 100r then the total R in series with the input signal is 1k0+100r=1k1. The C is 330pF.
The filter frequency is (F-3db) 1/Pi/2/C/R=438kHz. Quite high. This could usefully be made a little lower, possibly as much as one octave lower, but at 200kHz you may notice a slight change in timbre of the highest harmonics.
The DC blocking capacitor can be fitted in the source equipment or in the next amplifying stage. Most builders put it in the power amp input and in the preamp input.
This is a series line input capacitor with a resistor to ground (Rin=22k).
If you use a 10uF series capacitor with 22k Rin the F-3db=0.7Hz
You should ensure that these two passive filters set the bandwidth of your amplifier.
If your amp cannot properly handle 0.7Hz, then you must move the high pass filter up until it becomes the bottleneck. i.e. make the capacitor lower in value, or add a smaller cap to the source equipment.
If the source has a 2u2F cap then the effective capacitance is now 2u2+10uF=1u8F and f-3db=4Hz. This is noticeably bass light with most wideband speakers. Experiment with these two filters to find what suits your ears and your equipment. BUT, remember that bandlimiting rule.
Andrew:
Thanks very much for your response. I find your posts to be technically astute and yet mostly understandable by even a novice such as me.
There is a DC blocking cap present. Supplied in the kit was 1uf electrolytic which would yield a 7hz -3db. I found that a little bass shy even on my small speakers, so I replaced with a 2.2uf film cap. I like that sound a lot more.
Regarding RF filter, there is no provision for it in the QK50. Based on posts I've read in other LM1875 threads, I purchased 330pf ceramic caps to add a filter. I did not understand filters at that time, but I think I'm beginning to see it now. With the 1k series resistor supplied with the kit, the 330pf from input to ground would yield 438k -3db point. If I change the 1k to 1.5k (I have lots of extra resistors of different values, but no other caps on hand), I will get a low pass filter with a corner around 300k. That seems like a good compromise to me. Is all that correct?
One of the problems I have had in reading all these threads is the bewildering array of labels for various components in the simple circuit. The terms Cin and Ci and Rin and Ri and Rb mean nothing to me. I look at the NSC schematic, and all R's and C's are numerically labelled. On top of that, my kit also uses numbers, but they're slightly different. When I compare the two schematics, I can figure out which components are equivalent, but when someone says Rin or something similar, I have no idea what they mean. I'm beginning to understand, but it's been difficult.
Thanks to everyone for your patience in explaining these concepts. I promise I will put all your advice to good use and build an amplifier you would be proud of and pleased to hear.
Peace,
Tom E
Thanks very much for your response. I find your posts to be technically astute and yet mostly understandable by even a novice such as me.
There is a DC blocking cap present. Supplied in the kit was 1uf electrolytic which would yield a 7hz -3db. I found that a little bass shy even on my small speakers, so I replaced with a 2.2uf film cap. I like that sound a lot more.
Regarding RF filter, there is no provision for it in the QK50. Based on posts I've read in other LM1875 threads, I purchased 330pf ceramic caps to add a filter. I did not understand filters at that time, but I think I'm beginning to see it now. With the 1k series resistor supplied with the kit, the 330pf from input to ground would yield 438k -3db point. If I change the 1k to 1.5k (I have lots of extra resistors of different values, but no other caps on hand), I will get a low pass filter with a corner around 300k. That seems like a good compromise to me. Is all that correct?
One of the problems I have had in reading all these threads is the bewildering array of labels for various components in the simple circuit. The terms Cin and Ci and Rin and Ri and Rb mean nothing to me. I look at the NSC schematic, and all R's and C's are numerically labelled. On top of that, my kit also uses numbers, but they're slightly different. When I compare the two schematics, I can figure out which components are equivalent, but when someone says Rin or something similar, I have no idea what they mean. I'm beginning to understand, but it's been difficult.
Thanks to everyone for your patience in explaining these concepts. I promise I will put all your advice to good use and build an amplifier you would be proud of and pleased to hear.
Peace,
Tom E
yes increasing the resistor value will lower the RF filter frequency.madisonears said:
What is the source impedance? If it's <<200r then changing the resistor to a higher value makes sense.
If Rs >>500r then retain the 1k0 or 1k5 and listen to the result. But it might be worth trying 2k0 and then paralleling an extra resistor to lower it to listen for any difference.
madisonears said:
My apologies. I really do try to be as "organized" as possible. Its just that I can't help but think of the many interrelationships, rather than thinking that one component does only one job.
EDIT: Thanks for the compliment!
An externally hosted image should be here but it was not working when we last tested it.
This is from LM1876, (the lesser power stereo LM1875 version with spike noise added). The picture shows the National Semiconductor resistor names corresponding to the Overture Design Guide spreadsheet.
You've probably already figured this out, but here's a handy chart anyway:
R1=Rb -input
R2= an optional noise-prevention
R3=Rin -input (the optional 330pF cap parallels this load)
R4=Ri -nfb
R5=Rf -nfb
R6= the output zobel resistor (values valid from 2.2R to 5R)
C1=Cin -input
C3=Ci -nfb
C4= the output zobel capacitor (values from 0.1uF to 0.22uF)
X1=Audio Input -Purpose of this "additional load" is noise-prevention.
Input bandwidth:
The input load with the most influence on sound is the load closest to the chip, "inboard" of the input filter cap. For purposes of compatibility with the design guide spreadsheet, this is still referenced as Rin. This is R3 on the K50 kit.
Relationship at the input:
Rin (k50's R3), representing a load, and Rb (k50's R1) in-series resistor, represent a relationship of an almost wide open potentiometer.
These two resistors are partners.
Simply put: 1k with 22k is a common figure, and so is 2.2k with 100k. Those are merely baselines (starting place).
Passive preamp:
The good news is that the K50 kit comes ready to use with a passive preamp or for use as pot+source. That's because it has a complete (also called "armored") input circuit.
So, your purchase of this kit was very smart.
On the passive preamp topic, lowering Ri (R4) down from 10k, will be louder. At some point, gain boost will also boost the sonic signature, badly. Perhaps the "almost unnoticable" borderline is 4.7k (when partnered with 150k).
Active preamp:
Past a certain amount of gain, then hiding the amplifier's sonic signature (by voicing) becomes mighty difficult. This is when you choose to add another amplifier--an active preamp, and the point is that neither one has much gain, and so neither highlights its own sonic signature to excess.
If using an active preamp, then C1, X1, and R2 (passive pre input circuit components onboard K50) aren't required if those safety features are, as expected (not guaranteed), already provided by the active preamp.
Bridged:
A potential way around the problem is using bridged mode. As seen in the vastly award-winning miniature Tripath amplifiers, Bridged amplifiers apply equal and opposite force to both speaker output poles. This also happens to apply equal and opposite force to the amplifier's sonic signature. Other potential benefits include increased soundstage and power handling/output, depending on application. I think that this is either unexplored for LM1875 or a well-kept "shop secret" in case this potential is realized.
Relationship between the input and the NFB:
There is a relationship between Rin (k50's R3) and Rf (k50's R5).
Perhaps those wiser than myself can explain it.
madisonears said:
Problem. The stock K50's Zobel has invalid values of 1R + 0.22uF. That zobel is audible. Any change will be heard inside the audio band. Valid resistor values start at 2.2R for partnering lousy caps, and at 3R for partnering quality caps.
If you increased the quality of that cap (decreased the ESR), then expect to see brown spots, or smoke/fire, forming on R6, the zobel resistor. Have I done this? Make a guess--just one.
There's no way to tell which of the changes quoted above resulted in the improvement. But, I'd sure like to know.
P.S. Its good to remove the zobel before voicing the amplifier.
A higher quality, lower ESR cap at the zobel with only 1 ohm resistance, will have depressed the treble. That's abusive to the amplifier (too much load) and its a masking practice. Instead. . . Decreasing the upper treble may also be done by "sliding" the NFB values down (smaller numbers both resistors) by 16% to 20%. Decreasing the "middle ground treble"/"upper midrange" can also be done by increasing the onboard capacitance from 220uF to some larger figure. As a last step, after voicing, add the zobel back on, but not within the audio band.
Daniel:
Thanks for the excellent explanation of components within this amplifier. I knew the LM1876 was a related chip, but I never would have thought of looking at that to assist in identifying the components in the 1875 schematic.
I don't have any need or desire to obtain an active preamp. I've used quite few of them over the years, from Dynaco to BAT VK30 to Rowland Consummate, and I can tell you that they all color the sound. Not unpleasantly, but they do change it. My passive Sonic Euphoria is the least colored of anything I've ever heard. My SACD player puts out 5 volts, and most active pre's provide too much gain.
I am quite satisifed with the amplifier gain at 16. It sounds better than the higher gain from the stock kit.
I am not trying to "voice" anything. I want this amp to be accurate. I want the music to sound like music, not like it's been filtered by a bunch of arbitrarily chosen components.
I am puzzled by your evaluation of the zobel circuit. Of course it's audible. When wouldn't it be? I guess I'm not astute enough to understand your suggested values. I can tell you, without a doubt, that replacing the stock kit polyester cap with a better quality MKP cap has not resulted in anything burning or smoking. (I don't think it "depressed" the treble, either, unless you consider reducing painful shrillness to average SS brightness depressing.) In fact, these amps are running very coolly, and I'm not even using any thermal compound yet between chip and heatsink--they're just bolted on right now, and they're not even getting warm. Based on this claim and some of the responses you've received from others, sometimes I doubt your technical capabilites and that calls your other statements and judgments into question. If other novices reading this follow some of your advice, they'll be quite disappointed in the results.
I am attempting to make a controlled, systematic, incremental evaluation of these amazing little amp kits by changing only one or two variables at a time, according to accepted engineering and design principles. It appears that sometimes you try things willy-nilly and jump to some irrational conclusions based on imperfect experiments and without knowing exactly what happened or why. I appreciate your experience and much of your input, but some of it doesn't make any sense.
Your proclamation that carbon resistors in this circuit sound so much better than metal film resistors is pretty flakey. I had one amp built with the stock carbon, and the other amp built with all but one metal film resistors, and I must tell you that the metal film parts make a much better sound. This is not a subjective evaluation; this is based on careful observation of two amps with only that variation between them.
Thanks again for all your input. I just need the proper capacitor in my brain's circuit to block some of it.
Peace,
Tom E
Thanks for the excellent explanation of components within this amplifier. I knew the LM1876 was a related chip, but I never would have thought of looking at that to assist in identifying the components in the 1875 schematic.
I don't have any need or desire to obtain an active preamp. I've used quite few of them over the years, from Dynaco to BAT VK30 to Rowland Consummate, and I can tell you that they all color the sound. Not unpleasantly, but they do change it. My passive Sonic Euphoria is the least colored of anything I've ever heard. My SACD player puts out 5 volts, and most active pre's provide too much gain.
I am quite satisifed with the amplifier gain at 16. It sounds better than the higher gain from the stock kit.
I am not trying to "voice" anything. I want this amp to be accurate. I want the music to sound like music, not like it's been filtered by a bunch of arbitrarily chosen components.
I am puzzled by your evaluation of the zobel circuit. Of course it's audible. When wouldn't it be? I guess I'm not astute enough to understand your suggested values. I can tell you, without a doubt, that replacing the stock kit polyester cap with a better quality MKP cap has not resulted in anything burning or smoking. (I don't think it "depressed" the treble, either, unless you consider reducing painful shrillness to average SS brightness depressing.) In fact, these amps are running very coolly, and I'm not even using any thermal compound yet between chip and heatsink--they're just bolted on right now, and they're not even getting warm. Based on this claim and some of the responses you've received from others, sometimes I doubt your technical capabilites and that calls your other statements and judgments into question. If other novices reading this follow some of your advice, they'll be quite disappointed in the results.
I am attempting to make a controlled, systematic, incremental evaluation of these amazing little amp kits by changing only one or two variables at a time, according to accepted engineering and design principles. It appears that sometimes you try things willy-nilly and jump to some irrational conclusions based on imperfect experiments and without knowing exactly what happened or why. I appreciate your experience and much of your input, but some of it doesn't make any sense.
Your proclamation that carbon resistors in this circuit sound so much better than metal film resistors is pretty flakey. I had one amp built with the stock carbon, and the other amp built with all but one metal film resistors, and I must tell you that the metal film parts make a much better sound. This is not a subjective evaluation; this is based on careful observation of two amps with only that variation between them.
Thanks again for all your input. I just need the proper capacitor in my brain's circuit to block some of it.
Peace,
Tom E
danielwritesbac said:
I'm honestly at a loss for words... I would ask you to explain all of that, but I'm afraid of what you might come up with!
The Tripath amps are bridged to get higher power output from a single 12V supply, simple as that. Bridged amplifiers do have outputs that are out of phase but that doesn't mean they "apply equal and opposite force to the amplifier's sonic signature". Some even order harmonics may be cancelled out due to the nature of the circuit.
Increased soundstage? How would bridging affect that?
Bridging is no "well-kept shop secret". Many amplifiers use it - just search this forum! You may want to look into using chip amps in parallel since that would probably be more beneficial than bridging, but the simple fact of the matter is that most choose to go with the LM1875's more capable brethren (LM3875, LM3886) for higher output power. Properly implemented, the LM3875 sounds superb. I'm sure the LM3886 does too, but I haven't heard an amp using that chip yet.
hi Tom
A couple of things to ponder, though you may well have considered all this already.
The LM1875 is a fairly delicate type of amp, in that it really is meant to drive easy loads, it works brilliantly with things like full range drivers in horn loaded cabinets, and efficient open baffles. it is in a way not dissimilar to a low powered tube amp producing a really smooth grain free sound if well implemented (which is what I have aimed for).
On many forums I have advocated the concept of system synergy, in other words designing the system as a whole so that all parts complement each other.
Now the problem with the LM1875 to my ears is that it lacks in any std form the ability to produce real bass with low efficiency speakers, it is not a fault just a limitation. It can do so but this means you need more capacitance near the chip and within the power supply, but.....
doing so really stifles the mid range and upper end sweetness for some reason, which though noted by many has never been definitively explained to the best of my knowledge though CarlosFM has a good take on this in his threads on snubbered supplies and regulated supplies for gainclones.
Often I think folks confuse this lack of bass drive with problems created by the input cap, and certainly going capless may boost the lower bass a bit but only if the amp has adequate capacitance at the chip and in the supply. I have tried capless and all the regular variations, in the end a good 2.2uf poly with small poly bypass seemed to sound pretty much transparent (ie not better or worse than no cap).
The LM1875 like all Nat Semi chip amps is a precision piece that works brilliantly for its intended purpose, that is the beauty of the things, but really once you have replaced the std bits with better quality bits then no amount of changing of the circuit etc is going to reap big dividends, in fact as you get away from the suggested Nat Semi circuit chances are you will just get a more coloured sound, which of course could work if it complements your speakers and room etc but......
What does make the biggest changes, is the power supply and a layout that removes the likelyhood of RF pickup and instability and indeed here little things can make big differences.
Consider that many of the adjustment folks talk about to filter caps etc are really just fixes for problems that should have been dealt with elswhere. Like RF, poor input cable shielding, input cables being run too close to power supply lines etc, lack of smoothness in the power supply or noisy lines.
As an example here is a tip if you can get the shielding uttely perfect (well close anyway) then you can use a lower value cap for the rf filter, doing so in my system anyway made for a better top end, but if you change the cap without attending to the shielding you will get worse sound due to instability etc.
To my mind the way I can judge the success of such changes is a) the chip will run cool even under reasonable load b) the lack of grain c) a sense of really great detail without fatigue.
If you want to get to nirvana with this amp, and I believe you can I feel you need to really think about that power supply, I would say it is the limiting factor, also consider system synergy.
Finally do not discount the pre-amp or some alternative. I went down the same path as you, but once I tried both a JFet Buffer (have a look on the decibel dungeon website) and later a really good preamp the difference was utterly obvious and I was using a CD player directly like yourself before that.
If you don't need lots of power but don't want to mess with batteries and chargers, a pair of 12V SMPS will do the job and deliver a sound quality that I could never get with any transformer based supply, I used a pair of computer supplies (older ones with high output 12v rails, but you can buy high output suppiles pretty cheap these days and in fact you can even get 15V versions (once again have a look at decibel dungeon for a bit more info.
Hope it helps.
A couple of things to ponder, though you may well have considered all this already.
The LM1875 is a fairly delicate type of amp, in that it really is meant to drive easy loads, it works brilliantly with things like full range drivers in horn loaded cabinets, and efficient open baffles. it is in a way not dissimilar to a low powered tube amp producing a really smooth grain free sound if well implemented (which is what I have aimed for).
On many forums I have advocated the concept of system synergy, in other words designing the system as a whole so that all parts complement each other.
Now the problem with the LM1875 to my ears is that it lacks in any std form the ability to produce real bass with low efficiency speakers, it is not a fault just a limitation. It can do so but this means you need more capacitance near the chip and within the power supply, but.....
doing so really stifles the mid range and upper end sweetness for some reason, which though noted by many has never been definitively explained to the best of my knowledge though CarlosFM has a good take on this in his threads on snubbered supplies and regulated supplies for gainclones.
Often I think folks confuse this lack of bass drive with problems created by the input cap, and certainly going capless may boost the lower bass a bit but only if the amp has adequate capacitance at the chip and in the supply. I have tried capless and all the regular variations, in the end a good 2.2uf poly with small poly bypass seemed to sound pretty much transparent (ie not better or worse than no cap).
The LM1875 like all Nat Semi chip amps is a precision piece that works brilliantly for its intended purpose, that is the beauty of the things, but really once you have replaced the std bits with better quality bits then no amount of changing of the circuit etc is going to reap big dividends, in fact as you get away from the suggested Nat Semi circuit chances are you will just get a more coloured sound, which of course could work if it complements your speakers and room etc but......
What does make the biggest changes, is the power supply and a layout that removes the likelyhood of RF pickup and instability and indeed here little things can make big differences.
Consider that many of the adjustment folks talk about to filter caps etc are really just fixes for problems that should have been dealt with elswhere. Like RF, poor input cable shielding, input cables being run too close to power supply lines etc, lack of smoothness in the power supply or noisy lines.
As an example here is a tip if you can get the shielding uttely perfect (well close anyway) then you can use a lower value cap for the rf filter, doing so in my system anyway made for a better top end, but if you change the cap without attending to the shielding you will get worse sound due to instability etc.
To my mind the way I can judge the success of such changes is a) the chip will run cool even under reasonable load b) the lack of grain c) a sense of really great detail without fatigue.
If you want to get to nirvana with this amp, and I believe you can I feel you need to really think about that power supply, I would say it is the limiting factor, also consider system synergy.
Finally do not discount the pre-amp or some alternative. I went down the same path as you, but once I tried both a JFet Buffer (have a look on the decibel dungeon website) and later a really good preamp the difference was utterly obvious and I was using a CD player directly like yourself before that.
If you don't need lots of power but don't want to mess with batteries and chargers, a pair of 12V SMPS will do the job and deliver a sound quality that I could never get with any transformer based supply, I used a pair of computer supplies (older ones with high output 12v rails, but you can buy high output suppiles pretty cheap these days and in fact you can even get 15V versions (once again have a look at decibel dungeon for a bit more info.
Hope it helps.
madisonears said:
The primary definition of voicing is: "Hide the sonic signature of the amplifier so that it doesn't color the sound."
Unsafe Zobel: A 1 ohm load is inappropriate because it is below the safety limit for maximum reference load, and this limit is documented by National Semiconductor.
Usual zobels have inexpensive caps, and that's when National Semiconductor specifies 2.7 ohms resistance for output zobel.
Advice: Hopefully, anything that I have advised will provide a difference. A difference is a tool. One can hardly be disappointed in having as many tools as possible. Use whatever is most pleasant to your application.
BWRX said:
Hi Brian!
It is the frequently observed benefits of bridged amps that require more exploration. I don't have a bridged LM1875. That's all there was to it.
EDIT: Shop secret question had to do specifically with bridged LM1875 and not any wider definition (not any other amplifier). Its because the LM1875 is already so nice, then that is the source of my curiosity about bridging it.
Indeed, it can be improved with premium parts, but circuit design is also important. I have evaluated the following progression of improvement in the Quality Kits QK50 circuit, in order of significance. All original values were maintained unless noted otherwise.
Input DC blocking capacitor is most important, of course. I moved from 1.0uf cheap electrolytic to a 1.0uf stacked film polypro (immense improvement in high frequency detail and smoothness), then to 2.2 stacked film polypro (much better bass), then to 2.2uf Bennic wound XPP metallized polypro (another step up in high frequency air and smoothness). I intend to upgrade this cap further, and will report the result.
Zobel cap was improved from cheap polyester to nice stacked film polypro. At the same time, I moved the speaker/Zobel ground away from a path shared with PSU decoupling caps. Sound improved significantly. I can't be certain which change was more audible, but I suspect the ground path. I intend to investigate additional grounding alternatives.
Power supply decoupling caps and feedback loop cap are also important. I improved cheap electros to Panny electros, which led to smoother, more refined highs with more air. I believe the feedback loop cap is more important, but the two are almost equal. Maybe a Black gate there would be the best, and Panny's to decouple.
Changed all carbon film resistors to Vishay metal film. This was not a significant change, but it did contribute to overall smoothness of the highs. I'm not sure Caddocks or Rikens would be justified, but I could be persuaded.
My amps sound very smooth yet detailed in the highs, with plenty of air and dimensionality, and a very believable midrange, totally different from what I started with. The parts didn't cost more than a few bucks per amp. I'm not sure there is much more to be gained by spending a lot more money on components.
Peace,
Tom E
Input DC blocking capacitor is most important, of course. I moved from 1.0uf cheap electrolytic to a 1.0uf stacked film polypro (immense improvement in high frequency detail and smoothness), then to 2.2 stacked film polypro (much better bass), then to 2.2uf Bennic wound XPP metallized polypro (another step up in high frequency air and smoothness). I intend to upgrade this cap further, and will report the result.
Zobel cap was improved from cheap polyester to nice stacked film polypro. At the same time, I moved the speaker/Zobel ground away from a path shared with PSU decoupling caps. Sound improved significantly. I can't be certain which change was more audible, but I suspect the ground path. I intend to investigate additional grounding alternatives.
Power supply decoupling caps and feedback loop cap are also important. I improved cheap electros to Panny electros, which led to smoother, more refined highs with more air. I believe the feedback loop cap is more important, but the two are almost equal. Maybe a Black gate there would be the best, and Panny's to decouple.
Changed all carbon film resistors to Vishay metal film. This was not a significant change, but it did contribute to overall smoothness of the highs. I'm not sure Caddocks or Rikens would be justified, but I could be persuaded.
My amps sound very smooth yet detailed in the highs, with plenty of air and dimensionality, and a very believable midrange, totally different from what I started with. The parts didn't cost more than a few bucks per amp. I'm not sure there is much more to be gained by spending a lot more money on components.
Peace,
Tom E
madisonears said:
Hi Tom! I have a question for you.
Did you change the 1R in the zobel to some safer value like 2.2R, 2.7R, 3R, etc. . . ?
I ask because I like to push my LM1875's with some voltage and 4 ohm loudspeakers. That 1 ohm zobel (factory standard K50 zobel) would present my amp with a nearly dead short load at the treble, set off the limiter, and not let my dynamics come out at 80 watts (which it can do). I'd like to try out some of your excellent ideas, but I can't do it with that much load at the high frequencies.
ttan98:
Yes, I bought the Vishay's from Digi-key. All are Phoenix style 0.4 watt with the exception of one Vishay/Dale which is 1/4 watt, and the Zobel resistor which is 1 watt.
I must retract part of one statement: after replacing all the carbon resistors with metal film, I then lowered the gain from 19 to 16 by replacing the 180k resistor with 150k. I did not have a metal film 150k on hand, so that one resistor is carbon. When I changed the gain, I noticed a substantial reduction in high frequency harshness and hash. I wonder if, as you suggest (and Daniel claims), the carbon resistor also contributed to the increased smoothness of the highs. I have some 150k metal film on order, and I'll try those and let you know the effect.
Daniel:
I did not change the value of the Zobel resistor from 1 ohm. Both you and Andrew have remarked that value seems too small. I don't know what effect that might be having on the sound, but I sure like the way these amps sound right now. They are running very cool--the heatsink hardly gets warm, and I can easily hold my finger directly on the chip and feel no heat whatsoever. I believe that is the best indication that the chip is running well within performance limits. Maybe I'm just not pushing them hard enough, but the present volume is adequate if not overwhelming.
Please tell me, in the best way you can, what effect a different value Zobel resistor would have on the sound. I don't care AT ALL about power. It's a little hard to believe you can get 80 watts from a chip rated at 30 max! If you really need that much power, why don't you just use a chip designed for it instead of pushing the 1875 so far outside its optimium parameters? I mean: doesn't that defeat the purpose of a chip designed to provide low levels of quality sound? I am interested ONLY in the quality of the sound. Both NSC and QK50 specify 1 ohm, and I'm not willing to bet against both of them.
Peace,
Tom E
Yes, I bought the Vishay's from Digi-key. All are Phoenix style 0.4 watt with the exception of one Vishay/Dale which is 1/4 watt, and the Zobel resistor which is 1 watt.
I must retract part of one statement: after replacing all the carbon resistors with metal film, I then lowered the gain from 19 to 16 by replacing the 180k resistor with 150k. I did not have a metal film 150k on hand, so that one resistor is carbon. When I changed the gain, I noticed a substantial reduction in high frequency harshness and hash. I wonder if, as you suggest (and Daniel claims), the carbon resistor also contributed to the increased smoothness of the highs. I have some 150k metal film on order, and I'll try those and let you know the effect.
Daniel:
I did not change the value of the Zobel resistor from 1 ohm. Both you and Andrew have remarked that value seems too small. I don't know what effect that might be having on the sound, but I sure like the way these amps sound right now. They are running very cool--the heatsink hardly gets warm, and I can easily hold my finger directly on the chip and feel no heat whatsoever. I believe that is the best indication that the chip is running well within performance limits. Maybe I'm just not pushing them hard enough, but the present volume is adequate if not overwhelming.
Please tell me, in the best way you can, what effect a different value Zobel resistor would have on the sound. I don't care AT ALL about power. It's a little hard to believe you can get 80 watts from a chip rated at 30 max! If you really need that much power, why don't you just use a chip designed for it instead of pushing the 1875 so far outside its optimium parameters? I mean: doesn't that defeat the purpose of a chip designed to provide low levels of quality sound? I am interested ONLY in the quality of the sound. Both NSC and QK50 specify 1 ohm, and I'm not willing to bet against both of them.
Peace,
Tom E
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.