Folks:
I have finished four LM3875-based amps using Brian Bell's terrific PCBs and am thrilled with the result. But as I follow the various threads on this website and periodically review Brian's offerings on chipamp.com, I am left with the following question: what are the differences between the various chips, and how do they compare in terms of features, power handling and sonics?
And Brian: consistent with the foregoing, perhaps you could provide a little detail on your website distinguishing your various offerings from each other? Prospective customers (especially newbies such as myself) may want a little direction or advice on which option to choose.
Thanks to all for any insight you can provide.
Regards,
Scott
I have finished four LM3875-based amps using Brian Bell's terrific PCBs and am thrilled with the result. But as I follow the various threads on this website and periodically review Brian's offerings on chipamp.com, I am left with the following question: what are the differences between the various chips, and how do they compare in terms of features, power handling and sonics?
And Brian: consistent with the foregoing, perhaps you could provide a little detail on your website distinguishing your various offerings from each other? Prospective customers (especially newbies such as myself) may want a little direction or advice on which option to choose.
Thanks to all for any insight you can provide.
Regards,
Scott
SRMcGee said:
I will need to work on a FAQ (frequently asked questions) section for the website for questions like this. Here is what I see as for the difference:
LM3875 - simplest of the higher power chipamps out there by National, bigger brother to the LM1875. This amp has the fewest pins (5 pins used for +IN, -IN, OUT, V+ and V-). Sonically, it is said to be better due to it's simplicity. As for power, this amp has lower current handling, making it not as suitable for lower impedence loads.
LM3886 - newer chipamp based on the LM3875, with higher current handling and mute functionality. This chip is more capable with 4 impedence loads due to the higher current handling. There are more pins used also (8 pins, extra V+, GND and MUTE input pins).
LM4780 - newest chipamp, with 2 - LM3886 dies inside the larger package. The larger package is has higher thermal power dissipation properties. The pcb that I sell has the 2 LM3886 amps inside wired in parallel, see schematic. With the 2 amps in parallel, each amplifier sees double the output impedence, making this chipamp best for low impedence loads.
Datasheets will show power handling:
LM3875
LM3886
LM4780
As for a higher power solution, it seems like it would be the most sense to put a bridging chip with 2 LM4780 boards, for a setup similar to the BPA200 application note setup.
These are my observations. Sonically, the best amp might depend on how difficult your speakers are to drive. The LM4780 should be considered as 2 - LM3886 in 1 package. In terms of power supplies, the new larger power supply board with 10,000uF (will accept 22,000uF ) per rail and Carlo's snubber design, should also help with more difficult speakers.
--
Brian
G4ME said:
The LM3886 kits that I am putting together now will all include the new "snubberized" power supply with 10,000uF caps. The board will accept up to 22,000uF caps (largest 35mm diameter cap I could find in the catalog).
As for the LM3886 current handling, the datasheet goes over this:
http://www.national.com/ds.cgi/LM/LM3886.pdf
--
Brian
In my case of having small listening room ( ~3 x 4m) and have +88 db/w/m speakers, the bass and loundness of 25W from the LM3875 (even 20W from the LM1875) plenty more than enough for me.
More bass doesn't means better ,
Sound different doesn't means better,
It depends on your taste. your listening room, your music you listen, and your speakers.
To me, I likely stop DIY for a while to enjoy the MUSIC!
NAD 521iBBE + LM3875(25W) + 88 dB/w/m speakers + tons of CDs!
Cheer!
More bass doesn't means better ,
Sound different doesn't means better,
It depends on your taste. your listening room, your music you listen, and your speakers.
To me, I likely stop DIY for a while to enjoy the MUSIC!
NAD 521iBBE + LM3875(25W) + 88 dB/w/m speakers + tons of CDs!
Cheer!
G4ME said:
I could possibly offer these as an upgrade option. The main issue with the 22,000uF caps is the height. The diameter of both of them is 35mm, but the height of the 10,000uF is 30mm (1.2"), while the 22,000uF caps are 50mm tall (2"). If 3 more people express interest in the larger caps, I can modify the webpage to offer these as an upgrade option. (it would cost about $4 extra per cap given the price difference and the number of parts I am ordering)
--
Brian
2*10.000uF Vs 4*4.700uF
Brian and others
In stead of adding 2 * 10000uF on the rectifier board is it also possible to change the 1500uF caps for 4700uF caps. The 100uF electrolyte and the 0.1uF ceramic can be directly at the chip pins,
The snubber (1W, 1ohm + 0.1uF film) as wel as the 0.1uF film cap can be at the bottom of the board in the holes previously used for the panasonic FC caps. The leak resistor(2.2Kohm) can be soldered where the V+ and V- are connected to the board.
The 4700uF can be soldered where previously the Black gates could be placed. They are only 26mm dia and 30 mm high (BC components 40V)
From the schematic it looks like it should not be a problem, or am I overlooking something?
People having boards already, having space limitations or want ease of upgrade might be interested
Brian and others
In stead of adding 2 * 10000uF on the rectifier board is it also possible to change the 1500uF caps for 4700uF caps. The 100uF electrolyte and the 0.1uF ceramic can be directly at the chip pins,
The snubber (1W, 1ohm + 0.1uF film) as wel as the 0.1uF film cap can be at the bottom of the board in the holes previously used for the panasonic FC caps. The leak resistor(2.2Kohm) can be soldered where the V+ and V- are connected to the board.
The 4700uF can be soldered where previously the Black gates could be placed. They are only 26mm dia and 30 mm high (BC components 40V)
From the schematic it looks like it should not be a problem, or am I overlooking something?
People having boards already, having space limitations or want ease of upgrade might be interested
LM3886 and LM4780 with 8 ohm speakers
Brian's comment's earlier in this thread, and from reading the datasheets, I am lead to believe that if using 8 ohm speakers, one should stick with LM3875, and stay away from LM3886 and LM4780. Even at higher power levels, the 3875 offers best sound quality. Is that correct?
Brian's previous comments:
"LM3875 - simplest of the higher power chipamps out there by National, bigger brother to the LM1875. This amp has the fewest pins (5 pins used for +IN, -IN, OUT, V+ and V-). Sonically, it is said to be better due to it's simplicity. As for power, this amp has lower current handling, making it not as suitable for lower impedence loads.
LM3886 - newer chipamp based on the LM3875, with higher current handling and mute functionality. This chip is more capable with 4 impedence loads due to the higher current handling. There are more pins used also (8 pins, extra V+, GND and MUTE input pins).
LM4780 - newest chipamp, with 2 - LM3886 dies inside the larger package. The larger package is has higher thermal power dissipation properties. The pcb that I sell has the 2 LM3886 amps inside wired in parallel, see schematic. With the 2 amps in parallel, each amplifier sees double the output impedence, making this chipamp best for low impedence loads.
Datasheets will show power handling:
LM3875
LM3886
LM4780
As for a higher power solution, it seems like it would be the most sense to put a bridging chip with 2 LM4780 boards, for a setup similar to the BPA200 application note setup.
These are my observations. Sonically, the best amp might depend on how difficult your speakers are to drive. The LM4780 should be considered as 2 - LM3886 in 1 package. In terms of power supplies, the new larger power supply board with 10,000uF (will accept 22,000uF ) per rail and Carlo's snubber design, should also help with more difficult speakers."
Brian's comment's earlier in this thread, and from reading the datasheets, I am lead to believe that if using 8 ohm speakers, one should stick with LM3875, and stay away from LM3886 and LM4780. Even at higher power levels, the 3875 offers best sound quality. Is that correct?
Brian's previous comments:
"LM3875 - simplest of the higher power chipamps out there by National, bigger brother to the LM1875. This amp has the fewest pins (5 pins used for +IN, -IN, OUT, V+ and V-). Sonically, it is said to be better due to it's simplicity. As for power, this amp has lower current handling, making it not as suitable for lower impedence loads.
LM3886 - newer chipamp based on the LM3875, with higher current handling and mute functionality. This chip is more capable with 4 impedence loads due to the higher current handling. There are more pins used also (8 pins, extra V+, GND and MUTE input pins).
LM4780 - newest chipamp, with 2 - LM3886 dies inside the larger package. The larger package is has higher thermal power dissipation properties. The pcb that I sell has the 2 LM3886 amps inside wired in parallel, see schematic. With the 2 amps in parallel, each amplifier sees double the output impedence, making this chipamp best for low impedence loads.
Datasheets will show power handling:
LM3875
LM3886
LM4780
As for a higher power solution, it seems like it would be the most sense to put a bridging chip with 2 LM4780 boards, for a setup similar to the BPA200 application note setup.
These are my observations. Sonically, the best amp might depend on how difficult your speakers are to drive. The LM4780 should be considered as 2 - LM3886 in 1 package. In terms of power supplies, the new larger power supply board with 10,000uF (will accept 22,000uF ) per rail and Carlo's snubber design, should also help with more difficult speakers."
I am currently comparing a BrianGT 3875 premium to a CarlosFM Regulated 3886 IGC built on modified Digi01 boards.
The Regulated 3886 has the snubberized values between the diodes and the regulator, so it is not a pure CarlosFM regulated IGC. This may or may not be a good idea. Power supply caps are the apexjr.com 12,000uF Nichicon. No buffering on either amp but both have no pot on the input. Both are driven by the same preamp.
I am driving speakers crossed over to a sub 12db at 90Hz
As has been observed, the regulated 3886 is much more accurate. This is clear when playing loud passages. An example is cranking up Dire Straits, Money for Nothing from the Brothers in Arms CD. (Warning, some may not find this song PC unless you realize it is quoting someone they are poking fun at)
This track has strong kick drum and bass in it so is probably pretty demanding on the smallish power supply cap. At the same time though there are many overlayed sounds that come in and out, some of them quite detailed.
On the regulated 3886 the sounds like the distorted synth sounds that come in once in a while or the growly guitar are much more convincing. The whole sound and vocal is much more solid. The subtle cymbal in one part was identifiable on the Reg 3886. It was a tad fuzzy on the 3875 Gainclone and not as convincing.
Same impression with Green Day - American Idiot. This recording has a lot of noise and effects in parts of it, even on the vocals. On a muddy system, it sounds..well..like a muddy clump of sound. On the Regulated 3886 things are pretty distinct and you can hear the reverb and modulation effects on the vocals apart from the overall noise. Still a "wall of sound" type recording on some passages but gives a nice sound to this recording.
So if you like your rock cranked you will likely prefer this over the BrianGT 3875 implementation.
But the BrianGT had advantages. The sound was a bit more ethereal (?). Perhaps this is what people mean by "air". Vocals were less specifically placed which was a nice effect at times. I will come back to this later.
What surprised me was when I played quieter passages.
For example I played a 1964 vinyl recording of Dave Brubeck - Impressions of Japan. (Highly recommended!). On Fujiyama the
piano is much clearer after the note is struck on the Reg 3886. (What a hauntingly beautiful thing to listen to.) On the 3875 Gainclone it is a bit fuzzy in comparison. Same for the delicate cymbal on "The City is Crying". Definitely more delicate and distinct on the Reg 3886.
So I can see wanting to use the Reg 3886 for this type of music when I am really sitting down and focusing on it.
But there is something about the BrianGT 3875 premium that is very appealing, particularly when walking around the room. It may be that that slightly less detail and less distinct positioning is a benefit at times. It also gives an impression of a larger taller soundstage.
So I am not sure that one always wins out over the other. I can see using them both of them at times depending on my mood.
Clearly I need both
The regulated 3886 will probably get the majority of play. Maybe 70/30. But I think there is a place for each.
The Regulated 3886 has the snubberized values between the diodes and the regulator, so it is not a pure CarlosFM regulated IGC. This may or may not be a good idea. Power supply caps are the apexjr.com 12,000uF Nichicon. No buffering on either amp but both have no pot on the input. Both are driven by the same preamp.
I am driving speakers crossed over to a sub 12db at 90Hz
As has been observed, the regulated 3886 is much more accurate. This is clear when playing loud passages. An example is cranking up Dire Straits, Money for Nothing from the Brothers in Arms CD. (Warning, some may not find this song PC unless you realize it is quoting someone they are poking fun at)
This track has strong kick drum and bass in it so is probably pretty demanding on the smallish power supply cap. At the same time though there are many overlayed sounds that come in and out, some of them quite detailed.
On the regulated 3886 the sounds like the distorted synth sounds that come in once in a while or the growly guitar are much more convincing. The whole sound and vocal is much more solid. The subtle cymbal in one part was identifiable on the Reg 3886. It was a tad fuzzy on the 3875 Gainclone and not as convincing.
Same impression with Green Day - American Idiot. This recording has a lot of noise and effects in parts of it, even on the vocals. On a muddy system, it sounds..well..like a muddy clump of sound. On the Regulated 3886 things are pretty distinct and you can hear the reverb and modulation effects on the vocals apart from the overall noise. Still a "wall of sound" type recording on some passages but gives a nice sound to this recording.
So if you like your rock cranked you will likely prefer this over the BrianGT 3875 implementation.
But the BrianGT had advantages. The sound was a bit more ethereal (?). Perhaps this is what people mean by "air". Vocals were less specifically placed which was a nice effect at times. I will come back to this later.
What surprised me was when I played quieter passages.
For example I played a 1964 vinyl recording of Dave Brubeck - Impressions of Japan. (Highly recommended!). On Fujiyama the
piano is much clearer after the note is struck on the Reg 3886. (What a hauntingly beautiful thing to listen to.) On the 3875 Gainclone it is a bit fuzzy in comparison. Same for the delicate cymbal on "The City is Crying". Definitely more delicate and distinct on the Reg 3886.
So I can see wanting to use the Reg 3886 for this type of music when I am really sitting down and focusing on it.
But there is something about the BrianGT 3875 premium that is very appealing, particularly when walking around the room. It may be that that slightly less detail and less distinct positioning is a benefit at times. It also gives an impression of a larger taller soundstage.
So I am not sure that one always wins out over the other. I can see using them both of them at times depending on my mood.
Clearly I need both
The regulated 3886 will probably get the majority of play. Maybe 70/30. But I think there is a place for each.
I have just finished my first GC on brianGT's LM3875 boards and added carlos' snubber with 7200uF on each row, the sound is amazing, I don't think that anybody will be dissappointed over the result of building this.
You will get a lot of good sound, and it is cheap to build.
I have only used cheap components, and i'm very satisfied with the result.
Just go for it !
My personal judgement :
fore the sound.
You will get a lot of good sound, and it is cheap to build.
I have only used cheap components, and i'm very satisfied with the result.
Just go for it !
My personal judgement :
fore the sound.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.