The answer is yes. My favourite DIY amp, the AKSA 55, used a 470nF with Zin=47k IIRC and was more than satisfactory as was my reference / baseline. If my calculations are correct, a Zin=22k with a 1uF cap gives a corner frequency of 7.23Hz and this amp will be used with speakers with an F3 no lower than 60Hz.
I did spot this in a post (I think from lineup) which is handy:
My rule of the thumb is using ~50 uF kohm.
uF x kOhm = 50
----------------------
Input impedance is 10 kohm, I will go for 50/10 = 4.7 uF MKP
Input resistor is 22 kohm, I will go for 50/22k = 2.2 uF
Input resistor is 47 kohm, I will go for 50/47k = 1.0 uF
Input impedance is >47 kohm, I will still go for = 1.0 uF MKP
Ideally 2.2uF may be better but 1uF will suit the application. My other eBay PCB uses 2.2uF but had none available for this build.
Also mentioned in the same thread as the quote, if you use a bipolar electrolytic as an input cap, then multiply the film cap size by 10.
What is interesting is we have both gone down different paths. I've gone for small power supply decoupling and input caps where you have gone the other direction but on this PCB, we both have had results that we find satisfactory and put a smile on our face. Been a good kit whether built as is or modded.
Hi Rabbitz and others
I see at the LM3875 datasheet that this chip is more powerful but the current is still 4Amps max - so its comparable with the LM1875.
the THD and the PSRR is much better.
I know you are experts at chip amps + you (rabbitz) own a Arcam A19.
Can you "confirm" that a powerful chip sounds at lower power /volume level better then e.g. LM1875??
chris
Firstly, I am not an expert and know very little about amplifiers but get by with the experience I've gained over the years.
I can't say I've compared amps at low volumes as it is not how I listen. I will check over the next few days and put up a LM3886 P19, LM3875 against the LM1875.
Low volumes becomes an issue due to the Fletcher Munson Curve and the only amp I've had with a decent loudness circuit was that from Yamaha as it's variable. I do have an active BSC that I could try with it as well.
i am asking because in generally i planned to use a other chip in the next future to compare with the LM1875 - but i want to try composite + parallel LM1875 first...so a lot of work
chris
Parallel can be fun or a hassle as did try that with a LM4780 a decade ago. The DC offset went through the roof but I believe if you closely match the feedback network resistors so you get identical gain in both amps, the outcome is improved. I think there is some info on it in the LM4780 datasheet. The output current sharing resistors (Rout) I used were 0R15 3W and were supplied in the Audiosector kit.
I've attached the parallel info I got for the LM4780.
Peter
Thanks rabbitz
Sorry rabbitz...i mean with low level music...no boom boom... i measured it at home at my listening level "normal" = floating the room with enough energy = 1-3 Watt continuous and peaks about 15 - 30 Watt at 8R. depending which music you are listining
chris
Thank you rabbitz for sharing your experience -i just tried the UES caps in the input and maybe i am biased with this - now i have and my other amp (blue decoupling cap and wima 4,7µF at the input) +at my "best amp part" the foil cap. i have to try more with foils or others.
chris
Thanks rabbitz. yes i read about parallel amps - its not easy....DC offset is a killer.
chris
Hi Chris and rabbitz,
Parallel LM1875 designs with ready-made PCB layouts will be difficult to find "open source".
Could it be an idea if we start a common design in a thread here on the forum such that we discuss the design principles and do a common design which we first test in an ordinary SE configuration? Then, when the parallel LM1875 amplifier works well, we can try to include it in a composite amplifier loop and test that result. Finally, we can build a similar second composite amplifier and run the two in BTL.
We will have to make it on Vero-board for a start. For the real enthusiasts, a PCB layout can be made in the end.
I once made an estimate and concluded that three LM1875 in parallel would be suited for a BTL version. That should allow some 100W-120W in 8 Ohm.
I believe we can manage and all gain better understanding of such design principles.
Last edited:
...lets do this
thx
sorry but the last one - i can´t handle this
chris
here is the first test with stereo mode of my AVR reciver in stereo mode. source direct as the previous post- for sure.
technically the AVR390 is a power house! 20 Hz bis 20 kHz, THD <0,02 %80 WATT into 8R - thats enough. its not easy to "separate" a AVR with a stereo amp price wise - the AVR has a lot of extras ..
first impression is a very stable sound character - silent, no spectacular, good sound, lot of power!, it is very neutral, so no "preferred" area,sometime it sounds good but not really with sparkling - I miss here a bit rhythm and timing - you know when your legs are moving to the rhythm
I miss the micro details, e.g. the skin of a drum, saxophones "distorded sound" - hopefully you can follow me what i ant to tell you...
its is not soo dark i- the silents in the back stage a bit.
second test will follow.
chris
Hi - extended from post 284
id did again the compare Arcam AVR 390 - LM1875
if you use bi amping (rear speakers) for the different chassis - yes with my KEFQ100 and Sonus Faber Venere 1,5 is this possible - then the dynamic gets a small step better. For me it is not really clear that at my "requested" power this helps this big amp to sound a little better.
the dynamic:
the dynamic is behind the LM1875 amp. if you turn on and play the first seconds its clear a difference. i tried several times to give more volume at the arcam to compensate the difference, e.g. 75- 77 Volume level but this doesn´t help at all. good sounding amp. sound stage to closed together..
so once again 1:0 for the LM1875
chris
Hi Rabbitz
Yes the AVR do have another job
no. i dont try 2 LM1875 per channel - i haven´t a split (y-Cable) cable
Hi
i have to refer to post 380 and the comment by FF at 381 with the oscillation at higher output power and the possible avoiding to do´s.
i am sorry i do not post my measurements :
pic 1 is the output power at 8R with 500mVrms input with the small cap bank and new components. post 173
https://www.diyaudio.com/forums/chip-amps/341675-ebay-mono-lm1875-kit-18.html#post5932363
pic 2 is the "burn in" amp/ and new cap bank version 2 with the more input and no oscillation - for me its clear that a bigger cap bank is better.
post 385
https://www.diyaudio.com/forums/chip-amps/341675-ebay-mono-lm1875-kit-39.html#post6002877
i swear i did not change anything at the amp boards
the cap bank version 2 consist of:
8x 3900µF/35 chemicon europe caps
2x 47µ/50V Rubicon YKM
2x10µ/50V Nichicon audio blue
6x1µF/50V Rubicon YXF
4x MPC74 o,1 R resitors
output power is 10,8Vrms into 8R = 14,58Watt
output power is 11,5 Vrms into 8R = 16,5 Watt
chris
i have to refer to post 380 and the comment by FF at 381 with the oscillation at higher output power and the possible avoiding to do´s.
i am sorry i do not post my measurements :
pic 1 is the output power at 8R with 500mVrms input with the small cap bank and new components. post 173
https://www.diyaudio.com/forums/chip-amps/341675-ebay-mono-lm1875-kit-18.html#post5932363
pic 2 is the "burn in" amp/ and new cap bank version 2 with the more input and no oscillation - for me its clear that a bigger cap bank is better.
post 385
https://www.diyaudio.com/forums/chip-amps/341675-ebay-mono-lm1875-kit-39.html#post6002877
i swear i did not change anything at the amp boards
the cap bank version 2 consist of:
8x 3900µF/35 chemicon europe caps
2x 47µ/50V Rubicon YKM
2x10µ/50V Nichicon audio blue
6x1µF/50V Rubicon YXF
4x MPC74 o,1 R resitors
output power is 10,8Vrms into 8R = 14,58Watt
output power is 11,5 Vrms into 8R = 16,5 Watt
chris
Attachments
Last edited:
I had some listening tests with a few of these chipamps.
Source was Yamaha WXAD-10 streaming FLAC from NAS > pre amp with 2 x LM4562 with gain and buffer stages > power amp > modded Jeff Bagby Soprano (4 ohm) SB26STCN + SB12NRX. Levels at 2.5m from speaker was 80-85dB (C) with 92dB (C) peaks.
LM1875 original:
A larger than life presentation, bass over blown, can get congested, can get an edge on certain tunes. Still a very good listen especially if not comparing. Heatsinks (2 x 2.5°C/W) gets quite warm to hot.
LM1875 Brian Bell:
Foot tapping, clean and clear, tight bass, sound stage comes out of the mix. Heatsink 0.89°C/W gets quite warm to hot.
LM3886 P19:
Smoother presentation, very controlled, tight bass, sounds quite natural but not laid back. Heatsink 0.72°C/W runs luke warm.
LM3875 Audiosector / Chipamp.com
A bit tube like, controlled, smooths out rough edges, low sibilance, good bottom end but bass not as good as the 2 above. Slight pop and shutdown and takes a while for caps to drain. Some high frequency residual noise present. Heatsink 0.72°C/W runs luke warm. Really needs an input and feedback cap for stability etc but probably would ruin it's charm.
The difference between the higher power amps over the LM1875 is they seem to do it with a bit more ease. Hard to explain but the music seems to flow better... sort of presented to the listener and not played to the listener. The big difference is the was they dissipate heat as the LM1875 really pours out the heat when pushed (small package), especially with 4 ohm loads. Really my LM1875 would be better with 21VDC rails than 25VDC rails for 4 ohm speakers. Both my LM3886 and LM3875 have 25VDC rails which suits low impedance loads (checked with the Overture design guide spreadsheet).
The LM1875 Brian Bell can hold it's own against the LM3886 P19 as far as sonics go. All these are enjoyable amps and none are a dog, but we all have sonic preferences. I think the higher power amps would be better for longer term reliability.
I didn't put the Neurochome 3886DR into the test as I know it's a tad smoother than the P19 but that's a bit more engaging. I was going to throw a TPA3116D2 into the mix but I forgot.
Source was Yamaha WXAD-10 streaming FLAC from NAS > pre amp with 2 x LM4562 with gain and buffer stages > power amp > modded Jeff Bagby Soprano (4 ohm) SB26STCN + SB12NRX. Levels at 2.5m from speaker was 80-85dB (C) with 92dB (C) peaks.
LM1875 original:
A larger than life presentation, bass over blown, can get congested, can get an edge on certain tunes. Still a very good listen especially if not comparing. Heatsinks (2 x 2.5°C/W) gets quite warm to hot.
LM1875 Brian Bell:
Foot tapping, clean and clear, tight bass, sound stage comes out of the mix. Heatsink 0.89°C/W gets quite warm to hot.
LM3886 P19:
Smoother presentation, very controlled, tight bass, sounds quite natural but not laid back. Heatsink 0.72°C/W runs luke warm.
LM3875 Audiosector / Chipamp.com
A bit tube like, controlled, smooths out rough edges, low sibilance, good bottom end but bass not as good as the 2 above. Slight pop and shutdown and takes a while for caps to drain. Some high frequency residual noise present. Heatsink 0.72°C/W runs luke warm. Really needs an input and feedback cap for stability etc but probably would ruin it's charm.
The difference between the higher power amps over the LM1875 is they seem to do it with a bit more ease. Hard to explain but the music seems to flow better... sort of presented to the listener and not played to the listener. The big difference is the was they dissipate heat as the LM1875 really pours out the heat when pushed (small package), especially with 4 ohm loads. Really my LM1875 would be better with 21VDC rails than 25VDC rails for 4 ohm speakers. Both my LM3886 and LM3875 have 25VDC rails which suits low impedance loads (checked with the Overture design guide spreadsheet).
The LM1875 Brian Bell can hold it's own against the LM3886 P19 as far as sonics go. All these are enjoyable amps and none are a dog, but we all have sonic preferences. I think the higher power amps would be better for longer term reliability.
I didn't put the Neurochome 3886DR into the test as I know it's a tad smoother than the P19 but that's a bit more engaging. I was going to throw a TPA3116D2 into the mix but I forgot.
Last edited:
Chris
I know you been comparing commercial amps to the chip amp. Besides HT amps, I've tried Yamaha A-S1000, Cambridge 840A, Yamaha R-N500,Yamaha R-S700, NAD 316BEE, over the last decade and wasn't impressed as none were better than any of my SS DIY discrete amps at the time. The first three were boring (2 died within a day), the NAD made my ears bleed (returned) but the R-S700 was listenable.
I also have a Denon CEOL Piccolo which is fun but any of the chipamps would better it.
I know you been comparing commercial amps to the chip amp. Besides HT amps, I've tried Yamaha A-S1000, Cambridge 840A, Yamaha R-N500,Yamaha R-S700, NAD 316BEE, over the last decade and wasn't impressed as none were better than any of my SS DIY discrete amps at the time. The first three were boring (2 died within a day), the NAD made my ears bleed (returned) but the R-S700 was listenable.
I also have a Denon CEOL Piccolo which is fun but any of the chipamps would better it.
Lm1875 is a very simple toy.
If I have time. Will design a circuit board to upload on the network.
In fact, the business technology of designing lm1875 is not very good.
Because I don't think professionals will have much interest in this.
I think we can add high-performance op amp.
Input capacitance is not used. And negative feedback capacitance.
All DC circuits are used.
If I have time. Will design a circuit board to upload on the network.
In fact, the business technology of designing lm1875 is not very good.
Because I don't think professionals will have much interest in this.
I think we can add high-performance op amp.
Input capacitance is not used. And negative feedback capacitance.
All DC circuits are used.
Hi Rabbitz
Thanks you for your time to explain your experience and competence .
actually I am nearly finished with a compare test of a Marantz PM6006....give me 1 or 2 days.....
yes the more powerful amps should do it more easy with 4R load. the LM1875 is build for 8R and not heavy current - clear for me from the beginning - but I want to squeeze this chip out......look at the other thread.... project is running
chris
Thanks you for your time to explain your experience and competence .
actually I am nearly finished with a compare test of a Marantz PM6006....give me 1 or 2 days.....
yes the more powerful amps should do it more easy with 4R load. the LM1875 is build for 8R and not heavy current - clear for me from the beginning - but I want to squeeze this chip out...
...look at the other thread.... project is runningchris
Hi LJM
i do not really agree completely. look at the test of rabbitz and me to compare a LM1875 to other "professional/commercial amps".
if i would sit as an EE at a amp company like NAD, Yamaha, Denon, Marantz , cambridge...etc i get a target by the managment - i want to say the budget is low for EE and the target and the margin should be high if you want to sell a amp/product. - so as a customer you maybe have to pick up i higher level amp to get something good in SQ as a simple 1875....
DIY goes a different way - but you know that
chris