Thank you for your answers, very interesting. But without inductance, I have a bit of fear for my dear tweeters.
BJT output? if so does that mean the current ability into low impedance should be better than all the FET output class-D's??
Cheers George
Of course not at the output. The so-called BCD process includes Bipolar, CMOS and DMOS technology in one process. No one wrote bipolar outputs.
Actually i am not agree with you. The issue is the ferrites used on Sabaj is very poor properties. I tested them. I bought for benchmarking. The device is also have some pop noise. So i sold it. Now i developed my own board with 4 ferrite on each channel. I tryed lots of ferrites BLE32PN300SN is the one i suggest. Works like a charm for long cables.
No pop noise for me with the Sabaj A20a, you should know that there was a second revision of this model ....
The question was not about the Sabaj, which suits him well ...
But on my future infineon Ma12070 evaluation boards, without filter or output choke.
The question was not about the Sabaj, which suits him well ...
But on my future infineon Ma12070 evaluation boards, without filter or output choke.
ok, thanks i just found ..... https://www.futureelectronics.com/f...--ferrite-beads/ble32pn300sn1l-murata-5087662
Should I study the way to implement them.
Should I study the way to implement them.
Hi mamocel,
my Sabaj has no pops, I have version 2.
Here is a link to the document for Infineon's recommendations on output filters. You will find everything in this document ... at least I think so.
EMC output filter recommendations for MA120XX
That's why I asked with a ? , it wasn't specified where the bi-polars were used as posted by potstip.
Cheers George
Could you please unplug mains power cord during music play. Is there any pop noise?
Hey everyone,
I'm making progress with my ma12070/p amp PCB and thought I share it with you. It is more or less using the reference design provided by Infineon using the filter design they recommended. The schematic and board is made with Kicad 5.99 (nightly) (kind of by accident, but now I have to stick with it ). Therefore you cant open it with the stable version of Kicad.
I used @matt_garman `s design for reference but build it up from scratch separately, also the Component refs have changed, they have a standard format with prefix and number.
The dc/dc converter also changed to a simpler but more expensive IC for easier soldering and simplified design.
All components (excluding the LC filter) are SMD 1206 with some exceptions using hand solder footprints.
For the output LC filter, I went with TH components with reasonable packaging and pricing.
The overall component price is about 50 euros.
https://github.com/fabianmuehlberger/ma12070p_amp_kicad
I will work on the documentation in the next couple of days,
I would very much appreciate it if someone is willing to look over and review my PCB, I never have done a PCB design before and have the feeling that I made mistakes choosing the critical parts, especially the filter caps in the input section and dc/dc converter.
stay healthy and have happy holidays.
Best Fabian
I'm making progress with my ma12070/p amp PCB and thought I share it with you. It is more or less using the reference design provided by Infineon using the filter design they recommended. The schematic and board is made with Kicad 5.99 (nightly) (kind of by accident, but now I have to stick with it ). Therefore you cant open it with the stable version of Kicad.
I used @matt_garman `s design for reference but build it up from scratch separately, also the Component refs have changed, they have a standard format with prefix and number.
The dc/dc converter also changed to a simpler but more expensive IC for easier soldering and simplified design.
All components (excluding the LC filter) are SMD 1206 with some exceptions using hand solder footprints.
For the output LC filter, I went with TH components with reasonable packaging and pricing.
The overall component price is about 50 euros.
https://github.com/fabianmuehlberger/ma12070p_amp_kicad
I will work on the documentation in the next couple of days,
I would very much appreciate it if someone is willing to look over and review my PCB, I never have done a PCB design before and have the feeling that I made mistakes choosing the critical parts, especially the filter caps in the input section and dc/dc converter.
stay healthy and have happy holidays.
Best Fabian
Last edited:
@Fabianmu - can you generate a PDF version of your schematic and post it? Similarly, you can post a couple screenshots of the PCB layout with different traces shown? That would make it easier for everyone to review (assuming most are like me and don't want to go installing the nightly build of KiCad!).
That said, if you followed the reference design from the datasheet, you should be fine. IIRC, the datasheet gave pretty explicit specs for all the parts used in the reference design. So as long as you pick parts with same or better specs, it should be OK. The reference design generally uses smaller-than-1206 size components, which makes it easier for you to meet or exceed specs in component selection. For resistors, I generally look for 1% or better, thin film, highest power rating for the given resistance. For SMD capacitors, I get C0G/NPO when possible, and X7R for the rest. I shoot for the highest precision and voltage rating possible (particularly with X7R, I shoot for 100v or more). These are just my personal general rules of thumb.
I also remember the reference design has a 1R or maybe 10R resistor right at the power input... earlier in this thread, I did the math, this "theoretically" needs to be fairly big in terms of power rating... I say "theoretically" because the power rating is only likely to come into play if you are running at max voltage, max volume and pushing a sine wave through the amp. In terms of more realistic usage (normal volumes, actual music), it's unlikely you'll get anywhere near that kind of current through that resistor, but it's something to think about.
Definitely use KiCad's bug checker, both on schematic and layout. It's not perfect, but it's like another set of eyes.
How you plan to solder the ma12070p chip? I actually made two PCBs for this amp, my later one was the "premium" version with crazy overkill components. It took me several tries to get a working build with that board; and even my working one stopped working after I handled it too much. I was never able to debug it. My suspicion is that the ma12070p chip does not have a good solder, e.g. maybe a solder bridge somewhere that I can't see (such as a solder bridge to the power pad). I have invested in some better SMD soldering tools, but haven't attempted another go with this chip yet. I have been using JLCPCB for my PCBs, and they have an assembly service that looks to be fairly reasonable in terms of cost. I"ve been thinking about just having them solder the ma12070p chip for me (as the rest of the components are easy to solder).
Also, what 5v reg are you using? I mated my PCB to a Raspberry Pi, so I just use the RPI's 5v supply for the ma12070p. My board has pads for a 5v regulator (just like the datasheet/reference implementation), but I never actually populated it. I have been thinking about revisiting this design, and using a bigger 5v regulator, one with enough current capability to drive the Raspberry Pi.
Good luck, let us know how it goes!
That said, if you followed the reference design from the datasheet, you should be fine. IIRC, the datasheet gave pretty explicit specs for all the parts used in the reference design. So as long as you pick parts with same or better specs, it should be OK. The reference design generally uses smaller-than-1206 size components, which makes it easier for you to meet or exceed specs in component selection. For resistors, I generally look for 1% or better, thin film, highest power rating for the given resistance. For SMD capacitors, I get C0G/NPO when possible, and X7R for the rest. I shoot for the highest precision and voltage rating possible (particularly with X7R, I shoot for 100v or more). These are just my personal general rules of thumb.
I also remember the reference design has a 1R or maybe 10R resistor right at the power input... earlier in this thread, I did the math, this "theoretically" needs to be fairly big in terms of power rating... I say "theoretically" because the power rating is only likely to come into play if you are running at max voltage, max volume and pushing a sine wave through the amp. In terms of more realistic usage (normal volumes, actual music), it's unlikely you'll get anywhere near that kind of current through that resistor, but it's something to think about.
Definitely use KiCad's bug checker, both on schematic and layout. It's not perfect, but it's like another set of eyes.
How you plan to solder the ma12070p chip? I actually made two PCBs for this amp, my later one was the "premium" version with crazy overkill components. It took me several tries to get a working build with that board; and even my working one stopped working after I handled it too much. I was never able to debug it. My suspicion is that the ma12070p chip does not have a good solder, e.g. maybe a solder bridge somewhere that I can't see (such as a solder bridge to the power pad). I have invested in some better SMD soldering tools, but haven't attempted another go with this chip yet. I have been using JLCPCB for my PCBs, and they have an assembly service that looks to be fairly reasonable in terms of cost. I"ve been thinking about just having them solder the ma12070p chip for me (as the rest of the components are easy to solder).
Also, what 5v reg are you using? I mated my PCB to a Raspberry Pi, so I just use the RPI's 5v supply for the ma12070p. My board has pads for a 5v regulator (just like the datasheet/reference implementation), but I never actually populated it. I have been thinking about revisiting this design, and using a bigger 5v regulator, one with enough current capability to drive the Raspberry Pi.
Good luck, let us know how it goes!
Hi, I kinda messed up using the nightly build. The schematics created with it can`t be opened with kicad 5.1 ... I just tried it because there is a jlcpcb plugin that lets you choose components from their library. this could be interesting for you to try out if you want the PCB assembled by jlcpcb.
But it looks nice 🙂
Here are some pictures of the board
If I find a more adequate solution for sharing the design is update
For the Rin, I used a 3 W 0.1 Ohm current sensing resistor.
A friend recommended the pan and hot air method, putting the PCB with some distance blocks in a pen that has some sand in it (to keep the heat more stable ) and using a hot air station to reflow the IC. He told me that this method is reliable when working with a stencil and low heat solder paste.
I never have done a QFN package myself, so I don't know how good this method is.
Best Fabian
But it looks nice 🙂
Here are some pictures of the board
If I find a more adequate solution for sharing the design is update
I used the same technique for choosing parts as you described, I hope I'm fine just going overkill 😉
For the Rin, I used a 3 W 0.1 Ohm current sensing resistor.
I think I buy a reflow oven or build one myself using a cheap Infrared mini-oven, there are some controllers made for that.
A friend recommended the pan and hot air method, putting the PCB with some distance blocks in a pen that has some sand in it (to keep the heat more stable ) and using a hot air station to reflow the IC. He told me that this method is reliable when working with a stencil and low heat solder paste.
I never have done a QFN package myself, so I don't know how good this method is.
I use a MAX16956AUBD. mainly because of low noise, high efficiency and the design can be slightly simplified. it is not delivering enough power for the raspy, but I could look into getting one that works.
Best Fabian
Why ... do that? .. What country do you live in?
For info ... My whole system (Hifi is on a 1500kva inverter) ...
no voltage flow problem
This sabaj at 230 euros is the current best deal ....
I only listen to flac or high resolution files .. on Dac Akm 4497 in usb.. It is amp are revealing the source, I say no more you understand ....
Unplug the amplifier on.. It's a bit crooked ... To accompany my Sabaj A20a
I have 2 Merus 12070 evaluation cards to integrate into my system to have 5.1 .... tiny they are ..
1 in BTL 2.0
1 in PBTL 1.0.
I need to look at the output inductors.
It's not the joy ... either at the level of dac 5.1 ... there is not too much choice, I went on a Soundblaster X3 sound card with akm chip ...
Do not type ... It's for HC. 🙂🙂🙂
Last edited:
I think you are on the right track ...
Sharing is wonderful ....
These chips are very interesting and will break some mythology.😉👍
Looks interesting to me. Which output inductors did you choose? They look like 5mm pitch and may be a bit too small. Undersized output inductors degrade THD numbers significantly, a lesson I learned with TPA3251 designs.
You may consider some sort of post-filter feedback as well to avoid ringing of output LC-Filter.
Well i live in Turkey, and i am used to sudden mains lose. I sold mine because pop noise occurs when mains gone sudden which means there is no sense circuit in design. So i wonder if v2 has this improvemetn or not.
No need to during play music just unplug from mains before close it and see if there is pop noise or not.
Ok, I understand ... the micro cuts ... it can destroy everything ... My inverter comes from the dumpster, I just had to change the batteries ...
The frequency and voltage is fixed.
This reassures me, the source being a media center pc, dac connected via usb.
Happy New Year !!!! 😉😉😉
The frequency and voltage is fixed.
This reassures me, the source being a media center pc, dac connected via usb.
Happy New Year !!!! 😉😉😉
Hi,
I had to change the Inductor, due to availability. They are Murata´s 15682C, I thought I had some space left on the Board, now it's getting packed. What is a post-filter feedback?
I noticed that that kicad released v 6.0, my nightly build is stable now, therefore, there isn't an issue anymore with my filetype 🙂
I attached a pdf with the board layout.
Last edited: