cnx said:
Hi CNX,
Great documentation! I just wanted to pull the trigger on a board including filter when I noticed you have added the V2 version. What are the updates on this one? Thank you.
col said:
That's my idea. I rather support people who do research and design and spend time improving designs including writing concise and thorough documentation for the buyers, than wiring money to a trader in this case.
Sander said:
Hi cnx,
I would be interested in this development also, as the V1 version is winging it's way to me right now!
I'm preparing the next version of TA3020 Amplifier, so i would like to take in account your suggestions. What do you consider that should be added, what is more usefull, what is not and need to be removed, modified, improoved.
Hello CNX
I agree with Col
Post #111
I think if you make a balanced version of the TA3020 board then you have something that will really competes with the Hypex modules. This will be very interesting!
Balanced input is a must for me - the moment this is done I will
be buying 3 modules.
I would done so already but have been wishing /hoping that a balanced version would be made available.
A highly optimised version with premium parts would also be good
it would save me the trouble of having to do myself - not
necessarilly just capacitors, that's the easy part.
Similar to what 'Audiophool' does for his optimised Truepath at 41Hz.
To reiterate the moment you have a 'Balanced' version the first
three modules will be bought by me.
Yours Sincerely
John
I agree with Col
Post #111
I think if you make a balanced version of the TA3020 board then you have something that will really competes with the Hypex modules. This will be very interesting!
Balanced input is a must for me - the moment this is done I will
be buying 3 modules.
I would done so already but have been wishing /hoping that a balanced version would be made available.
A highly optimised version with premium parts would also be good
it would save me the trouble of having to do myself - not
necessarilly just capacitors, that's the easy part.
Similar to what 'Audiophool' does for his optimised Truepath at 41Hz.
To reiterate the moment you have a 'Balanced' version the first
three modules will be bought by me.
Yours Sincerely
John
To my opinion your previous design was flawed in many ways. I could point out a few if you like? These are based just on my visual evaluation of your board so don't pin me down on it. If there are things different in practise please kindly point them out.
I also have a few questions about your design.
Here's the list:
1. Close up caps are either far away/non existent or very large for this purpose. Or if they are bulk caps they are probably too small.
2. I have a hard time believing the rectifier is large enough for the whole amp, or is it just a VN10/5V rectifier?
3. Many board traces are just 5mm or so? I would render that VERY undersized!
4. Where are the rail to rail caps?
5. Far too small gate drive resistors. Of course depending on the fets, but for proper power they are too small.
6. Traces from the power supply to the hi/lo side fets are not symmetrical, so the inductances are not equal.
7. Are you feeding the 5V with a switchmode PS?
8.The distance from the LM2575HV SMPS to the input section and to the output FETs is a few millimeters?!?
9. Analogue feedback traces from the outputs pass under the main FETs?!?
10. Where are the FET current sense resistors?!? (The rather large 10mohm resistors...)
11. You use many cheap tantalum caps where other designs use nice low ESR high ripple and long life caps.
The boards itself raises most questions, big fat traces tend to have much lower inductance where your board has many very narrow traces where a lot of current has to flow through. Like the gate drive traces. IMO the parts and trace widths look more like suitable for a 50W amp. However, this may of course all work. I guess there is only one way to find out...
I would know what my choice is, and it's not to go with the cheapest available amp you can get....
I also have a few questions about your design.
Here's the list:
1. Close up caps are either far away/non existent or very large for this purpose. Or if they are bulk caps they are probably too small.
2. I have a hard time believing the rectifier is large enough for the whole amp, or is it just a VN10/5V rectifier?
3. Many board traces are just 5mm or so? I would render that VERY undersized!
4. Where are the rail to rail caps?
5. Far too small gate drive resistors. Of course depending on the fets, but for proper power they are too small.
6. Traces from the power supply to the hi/lo side fets are not symmetrical, so the inductances are not equal.
7. Are you feeding the 5V with a switchmode PS?
8.The distance from the LM2575HV SMPS to the input section and to the output FETs is a few millimeters?!?
9. Analogue feedback traces from the outputs pass under the main FETs?!?
10. Where are the FET current sense resistors?!? (The rather large 10mohm resistors...)
11. You use many cheap tantalum caps where other designs use nice low ESR high ripple and long life caps.
The boards itself raises most questions, big fat traces tend to have much lower inductance where your board has many very narrow traces where a lot of current has to flow through. Like the gate drive traces. IMO the parts and trace widths look more like suitable for a 50W amp. However, this may of course all work. I guess there is only one way to find out...
I would know what my choice is, and it's not to go with the cheapest available amp you can get....
Thx johnYks
i write down your suggestions, but i have few more proposals. Speaker protection with relays mounted on board, thermal monitor for the power stage, which will shut-down when the temperature rise too much, possible low-pass band-pass or hi-pass filters in case the board is used for a specific frequency range.
one question remain, the desired size of the board and parts arrangement. now i have 2 versions, one is 152x104x40mm and the other one is small, just 100x100x40mm. i think the next one which will include more options will be a little bigger. there is some space and size constraints on usual applications ? i saw that most of the peoples which bought this boards used them in cases which have plenty of space arround.
i write down your suggestions, but i have few more proposals. Speaker protection with relays mounted on board, thermal monitor for the power stage, which will shut-down when the temperature rise too much, possible low-pass band-pass or hi-pass filters in case the board is used for a specific frequency range.
one question remain, the desired size of the board and parts arrangement. now i have 2 versions, one is 152x104x40mm and the other one is small, just 100x100x40mm. i think the next one which will include more options will be a little bigger. there is some space and size constraints on usual applications ? i saw that most of the peoples which bought this boards used them in cases which have plenty of space arround.
Hello CNX
The bigger board is fine with me!
A premium/optimised board idealy 'Balanced' is what I would prefer.
Yours Sincerely
John
The bigger board is fine with me!
A premium/optimised board idealy 'Balanced' is what I would prefer.
Yours Sincerely
John
v-bro said:
seems that i start to iritate some "friends" which start spitting
Answers:
1. Untill u make-up your mind if the capacitors exist or not, are too big or too small, i will let you know that they EXIST and they are proper sized, good quality, Low ESR and in addition to this trere are X7R caps placed as close as possible to the MOS-FET pins. of course you can't see them because they are on the other side.
2. 25A bridge is not enoug for this amplifier ??? please take a look at the schematic and maybe you will understand where is connected.
3. some of the board traces have more some less than 5 mm. try to calculate their resistance and temperature rising at the maximum current considering 2 oz copper.
4. on the board.
5. take a look into the datasheet, into application notes and reference design.
6. can you proove me mathematical this aspect. and the result also. are the wires which come from the PSU to the amplifier perfectly equal and with same impedance ???
7. try to guess
8. yes, few mm, (42 mm from the pins of the LM to the pins of the TA and surrounded with GND copper pour)
9. i tried to pass over them but it was much more difficult. so i choose under. i would like to remind you that two intersecting tracks on two different layers, one which carry large current and another is signal feedback, but close connected to the previous one and they are perpendicular one to another, will not influence eachother (high-school phisycs)
10. i forgot to put them. or maybe they are too small to bee seen. anyway i didn't use the thru-hole with 20 mm legs on each side which have at least 20nH on each side. (how aout this ????)
11. Epcos and AVX are not quite cheap, maybe you have some "supplier" which can provide you cheap this parts. read again some specs about this parts. and those "many" are just 4.
"I would know what my choice is, and it's not to go with the cheapest available amp you can get...."
Price is not the only thing which can characterize a product. just because your product is more expensive doesn't mean that all the others are $hit. and btw, i just released few days ago another board even smaller than this, complete and much cheaper than your lovely board.
Why don't you try to design a better one and more complete if you have so good advices. i had 41's board, i have the Tripath ref board, and i compared all of them. if you search on another topic you will find some test results of this board.
You may blame me for being over critical, but not irritated. I'm sorry if I irritate you.
My being over critical is the reason why some designers like me to test their boards and appreciate my humble advise. I'm not a designer myself, I have a busy day time job in the broadcasting field so I'm not up to my neck into this. I do have some test equipment and ideas about how I like a design to be laid out. I'm not in any way affiliated with any business, nor am I making money out of this myself. The only way I get compensated for my hours and trouble is fun and hardware.
The points I stated above are by all means meant as constructive criticism.
Your response however makes me doubt I'll continue with it.
I will dig into your design some more and will try to be more specific and more profound IF I even decide to respond again.
My being over critical is the reason why some designers like me to test their boards and appreciate my humble advise. I'm not a designer myself, I have a busy day time job in the broadcasting field so I'm not up to my neck into this. I do have some test equipment and ideas about how I like a design to be laid out. I'm not in any way affiliated with any business, nor am I making money out of this myself. The only way I get compensated for my hours and trouble is fun and hardware.
The points I stated above are by all means meant as constructive criticism.
Your response however makes me doubt I'll continue with it.
I will dig into your design some more and will try to be more specific and more profound IF I even decide to respond again.
v-bro said:
you don't have to sorry, this was your purpose and you've achieved.
i have doubts that you are not affiliated with any bussiness since you play that song which resonate at 41Hz over and over again. your observations are ment to blame my product and rise-up other. but at least if you know exactly the details not what can be seen from photos. you start a similar argue with somebody else a while ago, on the same subject. i didn't went in at that time. i wanted to see what will be next. but he gave up, and everything finish before even began.
so, if you want to make observations, take a board, test-it as good as you know and tell me the results. i did this already and i know what to expect.
cnx said:
Hey man, it's just my opinion and there are enough people that know I'm being objective. The reason I rave about 41hz products is merely because I find they are very well designed.
I think I know what I'm talking about since I've had a lot of Tripath reference designs on my desk and recently tried out an IRAUDAMP7 implementation (which I also found pretty poor as well) I've had some TI amps pass the revue. Tried Audiodigit amps. Seen a lot of amps go up in smoke or behave nasty in other ways. 41hz amps aren't perfect either, but at least they are open for criticism and adapt their designs accordingly.
I have published many of my objective tests of different amps on 41hz forum, so if you're interested you can have a read there...
v-bro said:
off course that they are very well designed, since some of them are designed by you. (Read just post#4 on this thread).
i deduct from what you say that even IR and other big monsters in the field have poor designs. you're like those guys who claim that can tune a mini in their garage to beat veyron.
why you don't say nothing about Hypex ? those are really good modules, compared with your favorites. just to be objective....
Some of my friends have Hypex modules and they are GREAT! They are some of the most clean sounding amplifiers BUT.... my preference goes out to the somewhat more magical sound of Tripath amps. That's just my opinion... On top of that they are modules, so not much to DIY about, not much challenge. For those who are just interested in a working amplifier they are perhaps the no1 choice if you ask me. I have had some email conversations with Bruno (the inventor/developer of UCD amps) in the past, must admit that much of what he said goes waaaay above my head. Perhaps in time I will be able to have such knowledge on the matter...
My knowledge is mostly hardware based, I do have some basic knowledge of electronics but not yet at the level of a designer like Bruno. I have to admit that it's easier to look upon a design from the side lines and give criticism than to design myself. For that I must praise your endeavours with your design and certainly didn't mean to burn it down completely. Like I said I'm generally over critical and often based upon hunches and don't mean to hurt your business, being a hobbyist myself I sometimes overlook the fact that there are people posting here with commercial interest.
My knowledge is mostly hardware based, I do have some basic knowledge of electronics but not yet at the level of a designer like Bruno. I have to admit that it's easier to look upon a design from the side lines and give criticism than to design myself. For that I must praise your endeavours with your design and certainly didn't mean to burn it down completely. Like I said I'm generally over critical and often based upon hunches and don't mean to hurt your business, being a hobbyist myself I sometimes overlook the fact that there are people posting here with commercial interest.
Oh and by the way, I said in post #4 that I was working on the design, I never said it was my own design...
The design is a collectively evaluated design that is updated with all ideas merged by 41hz.
The design is a collectively evaluated design that is updated with all ideas merged by 41hz.
so, eventually you admit that your comments are made without having enough knowledges in the field and without understandig all the phenomenons involved here, like the question with signal feedback or the copper tracks width question, or the gate resistors (what's the average dissipated power on them ? ), or the tantalum capacitors ESR which, btw are connected in series with 240 ohms resistors, try to guess how much does it count their ESR.
i think you're not that busy as you said before if you still have time for this useless questions. i am, but i still make some time to answer.
i think you're not that busy as you said before if you still have time for this useless questions. i am, but i still make some time to answer.
I have enough knowledge to see the difference between designs and I have enough knowledge to use my measurement equipment rather than calculate stuff hypothetically. That's enough for me to make my choices.
Right now I'm at work and everything is quiet, in a few hours hell will burst loose here. That's how it sometimes goes in my line of work...
The gate drive resistors I've measured can get pretty hot, the average dissipated power on them depends on many things, gain setting, the fets that you're driving etc. etc. This I would have to measure...
If you like me to make some measurements of your design just send me a sample and I'll let it break out some sweat 🙂
Right now I'm at work and everything is quiet, in a few hours hell will burst loose here. That's how it sometimes goes in my line of work...
The gate drive resistors I've measured can get pretty hot, the average dissipated power on them depends on many things, gain setting, the fets that you're driving etc. etc. This I would have to measure...
If you like me to make some measurements of your design just send me a sample and I'll let it break out some sweat 🙂
v-bro said:
you haven't proved me that yet
v-bro said:
this is very interesting answer. i didn't know that their dissipated power depends on amplifier's gain. maybe you can explain me this. i know only that the gain depends on the transistors gate charge, GS capacity, Drive voltage, frequency, DT. if the gain is 10 or 100 none of the parameters which i mentioned before will change, nor the sw frequency, the drive voltage (10V for TA3020) or transistors parameters other than what is changing due temperature rise.
try to make an experiment and tell me the result. use a TA3020 board and play 1Khz sine with let's say gain=20 and after that play with same sine and gain 50. make sure that you will have the same input signal amplitude, but not too much that the amp will clip at gain = 50.
I was under the impression that modulator gain did have an influence on the power going through the gates, but a nearer look teaches me they probably don't.
From the chip datasheet I read:
"Gate resistor, which is used to control the MOSFET rise/ fall times. This resistor
serves to dampen the parasitics at the MOSFET gates, which, in turn, minimizes
ringing and output overshoots. The typical power rating is 1 watt".
The gate drive resistors you are using are 0805 size? Where did you get them to be 1watt rated? Or did you feel safe to use lighter resistors in your design? Probably the IRFP fets you used allow this, but then again someone may want to change the fets one day...
By the way, for the TA3020 capabilities I do find the rectifier bridge a bit light. With the right fets the current/ch can reach up to 20 to 30A at least.
You also seem to connect the 5V ground to power/virtual ground. And indeed drive it directly from the SMPS. How's the noise level? I mean measured....
The data you publish with your amp seems to be a 100% copy from the TA3020 datasheet, but looking at such design features makes me doubt it'll be able to perform that well....
I would be thrilled if you can prove it does by some measurement data.
From the chip datasheet I read:
"Gate resistor, which is used to control the MOSFET rise/ fall times. This resistor
serves to dampen the parasitics at the MOSFET gates, which, in turn, minimizes
ringing and output overshoots. The typical power rating is 1 watt".
The gate drive resistors you are using are 0805 size? Where did you get them to be 1watt rated? Or did you feel safe to use lighter resistors in your design? Probably the IRFP fets you used allow this, but then again someone may want to change the fets one day...
By the way, for the TA3020 capabilities I do find the rectifier bridge a bit light. With the right fets the current/ch can reach up to 20 to 30A at least.
You also seem to connect the 5V ground to power/virtual ground. And indeed drive it directly from the SMPS. How's the noise level? I mean measured....
The data you publish with your amp seems to be a 100% copy from the TA3020 datasheet, but looking at such design features makes me doubt it'll be able to perform that well....
I would be thrilled if you can prove it does by some measurement data.
seems more and more that i'm waisting my time replying to you. it's obvious that your lack of knowledges push you to read between the lines and make assumptions which are favorable to you. you don't even know which are those resistors, and their size (sure not 0805). "typical" is not "minimal"
not to mention that in your perfect design were used resistors which have inductance of few tents of nH
show me calculation, theory and practice that the 25 A bridge is not enough. let me remind you that at 25A, +-50V will be 2.5KW !!! maybe you miss a zero. the current limit is set below the bridge nominal current. when the amp will reach that curent, that mean that something went wrong so, is not a normal working situation. i tested this, and it passed successfully.
open your eyes and you'll see that there are 2 GND's one is called PGND and one SGND.
maybe in your design try to connect the gnd to V- and see what you will get. the noise is very warm and pleasent to hear.
it is so surprisingly that the measured data corespond to the datasheet ? of course, your, yours, them, or who else made that design which is TE BEST EVER the measured values are far away form the datasheet and this made you so susceptive.
i already posted results of some measurements, make some time and search for them.
so what's next ? why the pcb is green and not red ? this will improove the soud quality....
i'm wondering who's behind you and push you to write this.
so, if you have to say something at least close to reality, you can post, otherwise i will ignore what you say.
not to mention that in your perfect design were used resistors which have inductance of few tents of nH
show me calculation, theory and practice that the 25 A bridge is not enough. let me remind you that at 25A, +-50V will be 2.5KW !!! maybe you miss a zero. the current limit is set below the bridge nominal current. when the amp will reach that curent, that mean that something went wrong so, is not a normal working situation. i tested this, and it passed successfully.
open your eyes and you'll see that there are 2 GND's one is called PGND and one SGND.
maybe in your design try to connect the gnd to V- and see what you will get. the noise is very warm and pleasent to hear.
it is so surprisingly that the measured data corespond to the datasheet ? of course, your, yours, them, or who else made that design which is TE BEST EVER the measured values are far away form the datasheet and this made you so susceptive.
i already posted results of some measurements, make some time and search for them.
so what's next ? why the pcb is green and not red ? this will improove the soud quality....
i'm wondering who's behind you and push you to write this.
so, if you have to say something at least close to reality, you can post, otherwise i will ignore what you say.
I can follow your calculations perfectly, they are simple ohm's law.
U=IxR
P=UxI
In reality the current through a rectifier can peak to more than 100A in a good power supply design for a several hundred watt amp. In my experience (and many commercial amplifiers show proof of this) it is a good idea to use a BIG bolt down 35A type rectifier on each rail to ensure minimal losses and to be certain it can hold up to current peaks such as from inrush currents from normally used BIG bulk caps. I'm talking about some 80000uF capacity which is nice to have to stiffen up the voltage a bit.
I don't think I need to give elaborate calculations of this as that's just theory, nice to know the aproximate boundaries and then double the rating
Current peaks are nothing to calculate, they just happen in practise.
Like I said, I don't mean to attack you. For me this is nothing to become hostile about. It's just that I have another philisophy on certain parts and am curious to why you did it like you did.
But I just don't seem to be allowed to criticize, so I'll stop then.
I could just as well ask you for proof of your statements, but so far I've mainly seen you scream and shout some theory.
U=IxR
P=UxI
In reality the current through a rectifier can peak to more than 100A in a good power supply design for a several hundred watt amp. In my experience (and many commercial amplifiers show proof of this) it is a good idea to use a BIG bolt down 35A type rectifier on each rail to ensure minimal losses and to be certain it can hold up to current peaks such as from inrush currents from normally used BIG bulk caps. I'm talking about some 80000uF capacity which is nice to have to stiffen up the voltage a bit.
I don't think I need to give elaborate calculations of this as that's just theory, nice to know the aproximate boundaries and then double the rating
Current peaks are nothing to calculate, they just happen in practise.Like I said, I don't mean to attack you. For me this is nothing to become hostile about. It's just that I have another philisophy on certain parts and am curious to why you did it like you did.
But I just don't seem to be allowed to criticize, so I'll stop then.
I could just as well ask you for proof of your statements, but so far I've mainly seen you scream and shout some theory.
- Status
- Not open for further replies.