Hi guys,
after reading about MyRef and MiniRef I stumbled upon this project, which should be similar to MiniRef I guess, but with bridged LM3886 and a remote feedback that is connected to the speaker terminals, thus taking cable impedance and inductance into consideration.
The guy that came up with this (Dragoljub Aleksijevic) wrote about all the advantages of a bridged approach as well as remote fb, but I presume that's all well known stuff, so no need to go into it.
My question is - is this similar to MiniRef design appart from the bridge and fb (I think so, but with different op-amps, chips and values I can't tell)? Also, do you think this approach could be an improvement?
scheme:
and the pcb:
after reading about MyRef and MiniRef I stumbled upon this project, which should be similar to MiniRef I guess, but with bridged LM3886 and a remote feedback that is connected to the speaker terminals, thus taking cable impedance and inductance into consideration.
The guy that came up with this (Dragoljub Aleksijevic) wrote about all the advantages of a bridged approach as well as remote fb, but I presume that's all well known stuff, so no need to go into it.
My question is - is this similar to MiniRef design appart from the bridge and fb (I think so, but with different op-amps, chips and values I can't tell)? Also, do you think this approach could be an improvement?
scheme:
An externally hosted image should be here but it was not working when we last tested it.
and the pcb:
An externally hosted image should be here but it was not working when we last tested it.
Take a look at the positive feedback paths, familiarize yourself with Howland charge pumps (if you're not already), and the topology differences from MyRef should be obvious. Which is better depends on the desired output impedance.
The schematic was originally posted on a serbian forum I think, but here I found someone copied everything, and it's in english.
geraldfryjr, did you try to build such an amp or had a chance to hear it somewhere?
twest820 - thanks for the link, it clarified the schematic a bit, though I can't say I get it all.
After several versions of my LM3886 amp I'm very itchy to try this, for it combines three concepts that seem very beneficial - Howland pump, bridging, and remote feedback. Though I would definitely optimize the layout and decoupling further, regulate the front-end power supply, and probably go for separate ps board.
geraldfryjr, did you try to build such an amp or had a chance to hear it somewhere?
twest820 - thanks for the link, it clarified the schematic a bit, though I can't say I get it all.
After several versions of my LM3886 amp I'm very itchy to try this, for it combines three concepts that seem very beneficial - Howland pump, bridging, and remote feedback. Though I would definitely optimize the layout and decoupling further, regulate the front-end power supply, and probably go for separate ps board.
What do you think of the power supply and the zener regulation? I think I would go for LM317/337 regulators and even make pads/pins on the pcb so they could be changed for something even better down the line.
The ps utilizes the benefits of the bridged configuration so there is only one big cap, and I would try that for the experience, but I would definitely go for a higher-quality ps like the concept danielwritesbac proposed for chipamps more than once. See here for example.
The ps utilizes the benefits of the bridged configuration so there is only one big cap, and I would try that for the experience, but I would definitely go for a higher-quality ps like the concept danielwritesbac proposed for chipamps more than once. See here for example.
The Zener setup is fine, Personally I prefer the 3 terminal regulators as well.
I have on the drawing board a BPA200 that can be expanded to a BPA400 or larger possibly using the LM4780 for the larger versions.
I was looking for this schematic in order to finish the design.
Since I want to use a Quad opamps I am thinking about changing the design to a PA150 that can be bridged as well.
I will be employing a OPA1664 in it providing the Vos of each opamp stage is below 1mv or at least equal between the stages.
http://www.ti.com/lit/ds/symlink/opa1664.pdf
The Vos of the LME49720's I have are not what they should be, although they seem to be in spec with the data sheet.
On the same chip one is about 1mv and the other one is about 9mv!
I prefer them that be as close to the same as the other as they can be 1mv or less.
All three of my sample LME49720 ( maybe it was the LME49860's I forget at this time) devices shows this trait.
I may have to switch to a single opamp for the DC servos if I can't find any multiple devices that match up, I am hoping the OPA1664's fit the bill.
I like the bridged configuration as well as the transformers I have produce 49.5v which is too high for a bipolar supply configuration.
My requirements aren't for maximum power but to be able to drive the low impedance loads of my DIYesl designs well being able to maintain a 20v to 30v peak signal.
Typically no more the 20v peak under normal conditions as my desktop ESL produces +105db at that level!!
This thread is a must read in it comes to layout design,
http://www.diyaudio.com/forums/chip-amps/252436-lm3886-pcb-vs-point-point-data.html#post3845470
I have been contemplating my design (up to 16 devices) since 2004 when the LM4780's First appeared.
But I never pursued it because of the mixed issues that some have reported about their chipamp designs through the years.
Even though many are using the exact same schematic many have had mixed results mostly due to layout design.
I still have the very same samples I obtained back in April of 2004 unused.
I do have plenty of both the LM3886's and Lm4780's to work with.
The thread I posted explains those issues in detail and I am now ready to start experimenting with the LM3886's and LM4780's now.
In fact the only thing that is holding me up is that I had misplaced my OPA1664's and I can't seem to find them right now !!
I have also had the thought of using some of the newer HV opamp's in order to not have to have the extra regulation circuits but I have not tested them yet and it may not be a cost effective solution anyhow.
jer
I have on the drawing board a BPA200 that can be expanded to a BPA400 or larger possibly using the LM4780 for the larger versions.
I was looking for this schematic in order to finish the design.
Since I want to use a Quad opamps I am thinking about changing the design to a PA150 that can be bridged as well.
I will be employing a OPA1664 in it providing the Vos of each opamp stage is below 1mv or at least equal between the stages.
http://www.ti.com/lit/ds/symlink/opa1664.pdf
The Vos of the LME49720's I have are not what they should be, although they seem to be in spec with the data sheet.
On the same chip one is about 1mv and the other one is about 9mv!
I prefer them that be as close to the same as the other as they can be 1mv or less.
All three of my sample LME49720 ( maybe it was the LME49860's I forget at this time) devices shows this trait.
I may have to switch to a single opamp for the DC servos if I can't find any multiple devices that match up, I am hoping the OPA1664's fit the bill.
I like the bridged configuration as well as the transformers I have produce 49.5v which is too high for a bipolar supply configuration.
My requirements aren't for maximum power but to be able to drive the low impedance loads of my DIYesl designs well being able to maintain a 20v to 30v peak signal.
Typically no more the 20v peak under normal conditions as my desktop ESL produces +105db at that level!!
This thread is a must read in it comes to layout design,
http://www.diyaudio.com/forums/chip-amps/252436-lm3886-pcb-vs-point-point-data.html#post3845470
I have been contemplating my design (up to 16 devices) since 2004 when the LM4780's First appeared.
But I never pursued it because of the mixed issues that some have reported about their chipamp designs through the years.
Even though many are using the exact same schematic many have had mixed results mostly due to layout design.
I still have the very same samples I obtained back in April of 2004 unused.
I do have plenty of both the LM3886's and Lm4780's to work with.
The thread I posted explains those issues in detail and I am now ready to start experimenting with the LM3886's and LM4780's now.
In fact the only thing that is holding me up is that I had misplaced my OPA1664's and I can't seem to find them right now !!
I have also had the thought of using some of the newer HV opamp's in order to not have to have the extra regulation circuits but I have not tested them yet and it may not be a cost effective solution anyhow.
jer
Last edited:
Yes, I follow that thread and I can witness that solving layout and decoupling issues takes the sound to a new level, as it did for my LM3886 amp.
My requirements are also not for the maximum power, but for best possible quality, as the chip is very dynamic as is, even on lower voltage, but with a good CRC supply.
Now this Macola guy (that's his nick) proved this amp to be extremely stable with capacitive loads, hooking up a 8uf capacitor while playing music, and the amp got warmer but continued to work flawlessly.
I can even find a photo of that, but I'm a bit tired right now. I could also translate what he wrote about bridging chips like in BPA amps and why he finds it's not as good as this approach.
I think I'll try to design another pcb for this amp, with regs and power supply on separate boards, so we can pay attention to layout and decoupling easier.
My requirements are also not for the maximum power, but for best possible quality, as the chip is very dynamic as is, even on lower voltage, but with a good CRC supply.
Now this Macola guy (that's his nick) proved this amp to be extremely stable with capacitive loads, hooking up a 8uf capacitor while playing music, and the amp got warmer but continued to work flawlessly.
I can even find a photo of that, but I'm a bit tired right now. I could also translate what he wrote about bridging chips like in BPA amps and why he finds it's not as good as this approach.
I think I'll try to design another pcb for this amp, with regs and power supply on separate boards, so we can pay attention to layout and decoupling easier.
Well now, I got a bit ahead of myself here (and I love it - there is a lengthy forum thread on another forum on this topic (in croatian though) so now I have to read all 79 pages to find out what's new
geraldfryjr, I'll write updates as I can, this has become way more interesting than I thought
- there is a lengthy forum thread on another forum on this topic (in croatian though) so now I have to read all 79 pages to find out what's new geraldfryjr, I'll write updates as I can, this has become way more interesting than I thought
So, it turns out that there are already newer through-hole and smd based pcb designs available. Several amps have already been built, and I have info that there is a sort of get-together going on these days to compare different versions.
Unfortunately I can't register on that forum yet so I can't download all the files and images to share here.
Possibly after the comparisons some of the original builders will post all available info and first-hand experience here.
Unfortunately I can't register on that forum yet so I can't download all the files and images to share here.
Possibly after the comparisons some of the original builders will post all available info and first-hand experience here.
Hi, there's another thread here about it : http://www.diyaudio.com/forums/chip-amps/233411-batomm-amp-lm-bridge-consider-try.html
(waiting for a order of 10 smd pcb )
(waiting for a order of 10 smd pcb )
And no more explanation ? Have you really looked to this design ? It have been built several times on in serbia, it seems popular. It drive perfectly a pure 8uf load if i remeber well, and it's 200hkz bandwith. From what i have readen, even thought it seems simple, this design is well compensated to the manage to LME4920 correct errors very well. Look at smd pcb and ground path not sharing inductance, a lot of work. It's far from what i can understand, but it seems not as simple as it. I don't remember exactly, but i seems me that distortion was measured really low. It's hard to read fully 87 page translated from serbian to english (not even talking about translation from serbian to french, it can be epic ^^), and even more hard to try to re-find those informations.
So i would really need more explanation than those classic recommendation as i have ever seen here for basics lm3886 design,
thanks
(With the V0.8 pcb) Is that good enough :
Google traduct it: http://yu3ma.net/wp/?tag=batomm
So i would really need more explanation than those classic recommendation as i have ever seen here for basics lm3886 design,
thanks
(With the V0.8 pcb) Is that good enough :
Google traduct it: http://yu3ma.net/wp/?tag=batomm
Last edited:
This is certainly interesting, and differs from both the MyRef and MiniRef (both of which have a transconductance power stage nested within an outer voltage-series feedback loop). This one is a voltage-voltage nested-feedback topology with an inverting stage to create a push-pull bridge. The compensation also differs - it's basically an integrator (dominant-pole) across the outer feedback opamp.
It's simple, and will most probably work well. The only downside that I can see is that it doesn't have a speaker protection circuit, for which an external board can probably be plugged in series if required.
It's simple, and will most probably work well. The only downside that I can see is that it doesn't have a speaker protection circuit, for which an external board can probably be plugged in series if required.
Thanks a lot lunixguru for giving so simple explanation and right names on what it is. It is really interesting, and give the puzzle piece i was searching, even if it there is a infinity of pieces in this puzzle ^^
Now, i have to seach to be able to learn what i need to someday understand what it means
Now, i have to seach to be able to learn what i need to someday understand what it means
Hello to everybody, I'm first time here
I'm glad that You are interested in BatoMM amplifier.
If You have any question regarding it, fell free to ask. I will try to answer whatever I know.
BatoMM is designed by my colleague Macola and I have done PCB for version 0.8 and 1.0.
Anyway, try this amplifier. It is so simple but works very very nice. You will be surprised how good it sounds
I'm glad that You are interested in BatoMM amplifier.
If You have any question regarding it, fell free to ask. I will try to answer whatever I know.
BatoMM is designed by my colleague Macola and I have done PCB for version 0.8 and 1.0.
Anyway, try this amplifier. It is so simple but works very very nice. You will be surprised how good it sounds
Really happy to see you here ! I've gone to the forum many time. Hard to translate all, but it was a impressive work. I didn't manage to understand all, from what i remember, i'm not sure but, it seemed me that there were once a problem using batomm with a smps, a ground problem maybe, linked to current path...don't know. Google translate is not always really practical. I know there's some LLC project, but it's too hard now for me. So i would go for a hard switching unregulated psu with ir2153/atx transformer/LC filter and pfc front end(got boost inductor sample from coilcraft ^^). I would power a 3 way batomm amp. So ground star from psu to each way. Does i plan maybe resistors (let say 0,05ohm) to separate power and ground path of the three amp ? (i've seen this once on a french forum)
How much max current (for transient) does each amp needs ? I will maybe separate the three way psus with regulators as it adds security features...(ld1084, 5A ldo are cheap and fast...can be parralleled with front balancing op amp).
How much max current (for transient) does each amp needs ? I will maybe separate the three way psus with regulators as it adds security features...(ld1084, 5A ldo are cheap and fast...can be parralleled with front balancing op amp).