another boring story
Hello again guys,
been a long time.....
here just showing what I have been busy about for the past few weeks.....
the torroid core
smps top view
smps near the CD player
here is what I have made... an SMPS 12V -> +/-48VDC.... used TL494 running @ 100KHz..no output voltage regulation to maximize efficiency... primary of torroid is bifilar wound (3 strands in parallel) 6T+6T.... secondary is bifilar wound (3 strands in parallel) 24T+24T.. man, takes a lot of time and effort winding paralleled magnet wires... 😀 anyway, its done.. torroid ferrite has 1.4inch OUTER diameter.. got it from a friend... part swap from a chip he doesnt have.. .nice deal!!
estimated output was/should have been +-48VDC.. but I only measure +/-45VDC... measured it btw, w/o filter caps...just after the filter coil/s.... when I soldered 4700uF and 1000uF to the smps board, I measure +/-48VDC.... output current of SMPS? maybe 4-5A?? im using BYV 28-200 for the output rectifiers... from surplus..these diodes ARE low loss type with 3.5A contintinous/200V/<50ns recovery time..... I cant find diodes with HIGH voltage ratings in our area...dual diodes in TO220 package, highest I could find is 35V 10A... too low voltage rating...so they are off the list..... switching mosfets used? 2 paralleled IRFZ44 (.5ohm 55A 55V) per side... 4 pcs total.... As you can SEE, I -just- mounted them mosfets in small heatsink....
heat issues??? hhhmmmmm MOSFETS, even with SMALL heatsinks?? no HEAT... EVEN WITH DIY UCD AT FULL BLAST... the rectifier diodes (BYV).. nop.. no heat either... punny diodes?? no!! THEY are good !! good in the sense, LOW LOSS 😀 THE ONLY PART HEATING IS THE TORROID.... and NO!!! I have not made ANY error on my smps... IT only HEATS up when I connect it to the D-aMP....otherwise, its cold with NO LOAD.... MORE in THIS =issue=later....
the AMP.. running
the amp(running at +/-48VDC @5A)... see the BLUE LED?? yeah... -that- is functional indicator... not just ANY power ind.... I thank Jan-Peter for posting his over_voltage protection and power_on_delay for the current source... and the OP-amp input buffer... I have added them all...as can be seen.. board IS Now longer than before.... just like the real deal, op-amp(5532) input buffer, 1 sec -ON- delay, over_voltage "power supply pumping " protection .... AND how do I know IF the over_voltage protection works?? simple... remove THOSE 470uF decoupling caps in amp.. and lets see(hear) if you can drive your amp to max volume.. 😀 I bet youl only get about 30% before the over_voltage prot starts turning -OFF- the current source, and hence the AMP... 😀 it has also the option like when you want to run AMP in standby mode.... made a jumper of w/c I could simply turn amp to standby/ON..... over_current protection?? hhmmmmmmm im roughly 30% from the truth.. hahaha I know.. im not a circuit designer/engineer...Im trying to apply what I have learned from my past electronic "adventures" .. that dont bother me, JUST DONT short speaker terminals 😀 Amp RMS output is MORE than 100W rms... I am planning to supply it in the near future, with +/-65VDC... Of course I have to change a higher voltage version in the IRF540(but I think it still can handle that supply).. more like IRFBN23N15(If I could get hold of one)... and would 'upper' the over_voltage sensing part....
and about amp, 1 part ONLY that heats up.. and its the voltage regulator for the high-side driver.. I guess thats normal....thats what the heatsink IS for... 🙂 its a =SAFE= heat... not HEAT that
would get burned in just a few minutes... all doides used in amp: 1N4148.. no BYV for the high side...yeah!! AND no more diode BURNING....
I have been toying with this diy amp long enough to learn the tricks/basics.... DC offset at speaker terminal??? 1.9mV 😀 yup!!! as compared to the single transistor current source, 2 transistor
current source IS WAY much better(see orig hypex patent)....
recently, I had to make side-by-side comparison with my diy UcD and my 100WRMS class AB amp......
results: these ARE based on what I have observed
>> I am using my SMPS as power supply... and use PC PS (12V 10A) to supply the SMPS
Class AB (OCL type)= lack off RMS power.. hehe I know I know its just 100W... (but I make it a point that I am driving the ucd to more/less 100W too)..... lacks bass..... some "timbre" (right term?) missing??? in laymans term, I can hear more "tonal" quality in UcD than my AB... MORE heat of AB amp(as always).....long story made short, THERE IS a DIFFERENCE between AB amp and D-amp.....
diy UCD = as I would always say.. ONE HELL OF AN AMP!!!!! I have done amp testing using my PC... been playing with the mediaplayer graphic EQ, UCD -was- VERY responsive..... having this D-amp, its like having "multiple" amps in a single AMP.. thats how I feel it.... its like, It has its built-in highpass and low pass... BASS sounds THUMP!!! while tweeter squeeks.. heheh as If I dont need a low pass filter+amp and highpass filter+amp.... Thats IN MY opinion ok.... 🙂
now back to SMPS heating
>>>>>>>>>>> SMPS used in AB amp, amp run full blast, HEAT from AB amp heatsinks && minimal heating of smps torroid winding....
>>>>>>>>>>>SMPS used in D-amp, amp normal heat(full blast), HEAT from smps torroid winding is HOT.... needs fan!!
that happens too when I use linear power supply (+/-30@8A 6800uF p/r rail), transformer tend to get HOT after an hour of playing, even with D-amp at 20% volume setting..... whereas, when I use supply with AB amp, It dont heat up UNLESS I really crank UP the volume ....
IS THIS NORMAL???? you guys out there with UCDs... please , same experience as mine??
if this is so, then I could conclude that AB amps are like CMOS chips....when in low power mode(or low volume), they dont heat heat up much(power supply and amp).. whereas, switching AMPs are like TTL.... even in standby (or D-amp in low volume settings) current consumption per chip is a high 50mA....... as compared to CMOS w/c is in the uA range... THIS Is just my opinion... maybe because I have only a DIY UCD.. that is understandable....
BUT!!!! that dont bother me much..... just add a fan FOR SMPS and diy UCD...and all is COOL!!!
once AGAIN UCD amp ROCKS!!!! no one is paying me to "SAY" this.. I just feel the need to say it!!
If you wanna go MY way, read and understand my past posts.... if not, BUY from hypex.... Im sure you wont regret it 😀
till then, im off making a stereo version and a nice amp case... and maybe I can get the over_current protection to work my way..
-xiao
Hello again guys,
been a long time.....
here just showing what I have been busy about for the past few weeks.....
An externally hosted image should be here but it was not working when we last tested it.
the torroid core
An externally hosted image should be here but it was not working when we last tested it.
smps top view
An externally hosted image should be here but it was not working when we last tested it.
smps near the CD player
here is what I have made... an SMPS 12V -> +/-48VDC.... used TL494 running @ 100KHz..no output voltage regulation to maximize efficiency... primary of torroid is bifilar wound (3 strands in parallel) 6T+6T.... secondary is bifilar wound (3 strands in parallel) 24T+24T.. man, takes a lot of time and effort winding paralleled magnet wires... 😀 anyway, its done.. torroid ferrite has 1.4inch OUTER diameter.. got it from a friend... part swap from a chip he doesnt have.. .nice deal!!
estimated output was/should have been +-48VDC.. but I only measure +/-45VDC... measured it btw, w/o filter caps...just after the filter coil/s.... when I soldered 4700uF and 1000uF to the smps board, I measure +/-48VDC.... output current of SMPS? maybe 4-5A?? im using BYV 28-200 for the output rectifiers... from surplus..these diodes ARE low loss type with 3.5A contintinous/200V/<50ns recovery time..... I cant find diodes with HIGH voltage ratings in our area...dual diodes in TO220 package, highest I could find is 35V 10A... too low voltage rating...so they are off the list..... switching mosfets used? 2 paralleled IRFZ44 (.5ohm 55A 55V) per side... 4 pcs total.... As you can SEE, I -just- mounted them mosfets in small heatsink....
heat issues??? hhhmmmmm MOSFETS, even with SMALL heatsinks?? no HEAT... EVEN WITH DIY UCD AT FULL BLAST... the rectifier diodes (BYV).. nop.. no heat either... punny diodes?? no!! THEY are good !! good in the sense, LOW LOSS 😀 THE ONLY PART HEATING IS THE TORROID.... and NO!!! I have not made ANY error on my smps... IT only HEATS up when I connect it to the D-aMP....otherwise, its cold with NO LOAD.... MORE in THIS =issue=later....
An externally hosted image should be here but it was not working when we last tested it.
the AMP.. running
the amp(running at +/-48VDC @5A)... see the BLUE LED?? yeah... -that- is functional indicator... not just ANY power ind.... I thank Jan-Peter for posting his over_voltage protection and power_on_delay for the current source... and the OP-amp input buffer... I have added them all...as can be seen.. board IS Now longer than before.... just like the real deal, op-amp(5532) input buffer, 1 sec -ON- delay, over_voltage "power supply pumping " protection .... AND how do I know IF the over_voltage protection works?? simple... remove THOSE 470uF decoupling caps in amp.. and lets see(hear) if you can drive your amp to max volume.. 😀 I bet youl only get about 30% before the over_voltage prot starts turning -OFF- the current source, and hence the AMP... 😀 it has also the option like when you want to run AMP in standby mode.... made a jumper of w/c I could simply turn amp to standby/ON..... over_current protection?? hhmmmmmmm im roughly 30% from the truth.. hahaha I know.. im not a circuit designer/engineer...Im trying to apply what I have learned from my past electronic "adventures" .. that dont bother me, JUST DONT short speaker terminals 😀 Amp RMS output is MORE than 100W rms... I am planning to supply it in the near future, with +/-65VDC... Of course I have to change a higher voltage version in the IRF540(but I think it still can handle that supply).. more like IRFBN23N15(If I could get hold of one)... and would 'upper' the over_voltage sensing part....
and about amp, 1 part ONLY that heats up.. and its the voltage regulator for the high-side driver.. I guess thats normal....thats what the heatsink IS for... 🙂 its a =SAFE= heat... not HEAT that
would get burned in just a few minutes... all doides used in amp: 1N4148.. no BYV for the high side...yeah!! AND no more diode BURNING....
I have been toying with this diy amp long enough to learn the tricks/basics.... DC offset at speaker terminal??? 1.9mV 😀 yup!!! as compared to the single transistor current source, 2 transistor
current source IS WAY much better(see orig hypex patent)....
recently, I had to make side-by-side comparison with my diy UcD and my 100WRMS class AB amp......
results: these ARE based on what I have observed
>> I am using my SMPS as power supply... and use PC PS (12V 10A) to supply the SMPS
Class AB (OCL type)= lack off RMS power.. hehe I know I know its just 100W... (but I make it a point that I am driving the ucd to more/less 100W too)..... lacks bass..... some "timbre" (right term?) missing??? in laymans term, I can hear more "tonal" quality in UcD than my AB... MORE heat of AB amp(as always).....long story made short, THERE IS a DIFFERENCE between AB amp and D-amp.....
diy UCD = as I would always say.. ONE HELL OF AN AMP!!!!! I have done amp testing using my PC... been playing with the mediaplayer graphic EQ, UCD -was- VERY responsive..... having this D-amp, its like having "multiple" amps in a single AMP.. thats how I feel it.... its like, It has its built-in highpass and low pass... BASS sounds THUMP!!! while tweeter squeeks.. heheh as If I dont need a low pass filter+amp and highpass filter+amp.... Thats IN MY opinion ok.... 🙂
now back to SMPS heating
>>>>>>>>>>> SMPS used in AB amp, amp run full blast, HEAT from AB amp heatsinks && minimal heating of smps torroid winding....
>>>>>>>>>>>SMPS used in D-amp, amp normal heat(full blast), HEAT from smps torroid winding is HOT.... needs fan!!
that happens too when I use linear power supply (+/-30@8A 6800uF p/r rail), transformer tend to get HOT after an hour of playing, even with D-amp at 20% volume setting..... whereas, when I use supply with AB amp, It dont heat up UNLESS I really crank UP the volume ....
IS THIS NORMAL???? you guys out there with UCDs... please , same experience as mine??
if this is so, then I could conclude that AB amps are like CMOS chips....when in low power mode(or low volume), they dont heat heat up much(power supply and amp).. whereas, switching AMPs are like TTL.... even in standby (or D-amp in low volume settings) current consumption per chip is a high 50mA....... as compared to CMOS w/c is in the uA range... THIS Is just my opinion... maybe because I have only a DIY UCD.. that is understandable....
BUT!!!! that dont bother me much..... just add a fan FOR SMPS and diy UCD...and all is COOL!!!
once AGAIN UCD amp ROCKS!!!! no one is paying me to "SAY" this.. I just feel the need to say it!!
If you wanna go MY way, read and understand my past posts.... if not, BUY from hypex.... Im sure you wont regret it 😀
till then, im off making a stereo version and a nice amp case... and maybe I can get the over_current protection to work my way..
-xiao
Nice work!
Did you base your 12V SMPS on ESP Project 89?
BTW: nice diodes for the secondary are, for example, MUR1520, the work great for the job. You can get them at many places, for example Digikey.
Best regards,
Pierre
Did you base your 12V SMPS on ESP Project 89?
BTW: nice diodes for the secondary are, for example, MUR1520, the work great for the job. You can get them at many places, for example Digikey.
Best regards,
Pierre
Pierre said:
SMPS??? partly... I have DONE that SMPS long ago.... before this UCD.... I had 3 sources.... ESP site, valveaudio(forum member), and djquan(forum member).......
its a redone circuit....my 1st board was BIG and I used U-U core from TV flybacks... hehehehe but it did work... although BIG... it was I think? last year?? I forgot.. back before I even know what a D-amp was... hehe
yes MUR are nice.. but im "allergic" to ordering parts.. 🙂 I use parts that are locally available...
Sniff.... that's the coolest thing I've seen in a long time 🙂
Just curiouse if you tried loading down your SMPS with a dumby load... just to see how it heats with that?
Three wires is trifilar winding isn't it?
You really did your digging to have come up with all those circuits🙂 Really one hell of an effort. Well done excellent example 🙂
Just curiouse if you tried loading down your SMPS with a dumby load... just to see how it heats with that?
Three wires is trifilar winding isn't it?
You really did your digging to have come up with all those circuits🙂 Really one hell of an effort. Well done excellent example 🙂
Hi there. Excelent job!
Nice to hear that my article was useful to you (ESP one). At those times it was haaaard for me to find information on push-pull SMPS, that was what motivated me to write it.
We are also preparing a 12V supply at coldamp to complete our series of switching supplies.
MUR diodes can be ordered as samples at ON Semiconductor (they will charge you for the delivery, 12 euro or so).
Best regards,
Sergio
Nice to hear that my article was useful to you (ESP one). At those times it was haaaard for me to find information on push-pull SMPS, that was what motivated me to write it.
We are also preparing a 12V supply at coldamp to complete our series of switching supplies.
MUR diodes can be ordered as samples at ON Semiconductor (they will charge you for the delivery, 12 euro or so).
Best regards,
Sergio
Build me a 12v --> -5v, +5v and 42v smps bitte, so i can use my D-Amp in a car without having to lug around with a inverter and heavy transformers.
Hello Chris,
Geeee, id love too!😀 but somethings more important right now than loading up that SMPS , namely overcurrent/short protection.... if I could ever...
was it??? hmmmmmm i forgot? hahahah 😀
you had me started with it.... thanks for the journey.. but IT still aint over for me... one more problem still....
hello Sergio,
yah it was... your ESP and Valveaudio.. where I think the only ones available YET at that time....
at 1st, it was in this VERY forum where it started.. down in the power supply section.... I was kinda "waiting" for a TL494 flavor.... TL494 I can get cheap and easy.. as compared to SG3525 w/c had to be ordered and WAS more costly than TL494.. my ONLY problem with the 494 was it kept mt MOSFETs overheating WITHOUT any load.... so I "junked" the 494 scheme and HAD to use SG3525.. I had to let my frend order a couple of those ... and I had to pay them when they arrive.... well, it WAS my 1st working SMPS with 3525... I used U-U core from TV flyback then.. But it worked.. just a bit ugly and big though... up until I really wanted to TRY use 494... I had many reference circuits... 494 with 4050 buffer outputs to drive mosfets... did not work... I then tried totem-pole output transistors.. It did work and no heating... and up till I saw djquan link his site to a 494 SMPS with a much simpler output buffer(looks exactly like ucd driver, lower side.... a diode/resistor/pnp)... thats WHAT I use now .. that SMPS IS equipd also with shutdwn feature.. I just enable it always by tying to +12V
what I really wanted to see/build is an SMPS from 220VAC, down to +/-48V..... 🙂 I =CAN= modify an AT PC PS to DO this, But maybe -too much- of a hasle in breaking the CORE... might "BREAK" it.. hahah and render It useless... I finally figured out HOW to re rewind the 5V/12V and changing them to thinner AWG.. so as to accomodate another thicker wire for the new +/- windings....
Hello Tekko,
you =finally= have a D-amp now?? thats great 😀 w/c one btw?
again, thanks for the replies guys.... You are all welcome..
-Raff
Hi,
You make the party very well.
I glad too. It can motivate other peoples to do and learn. Most peoples think very complicated for not more than 10% from part of issue, while you take the other 90% simple part of how to make your own amp, that is why you are successfull.
well done indeed.
You make the party very well.
I glad too. It can motivate other peoples to do and learn. Most peoples think very complicated for not more than 10% from part of issue, while you take the other 90% simple part of how to make your own amp, that is why you are successfull.
well done indeed.
kartino said:
thanks 🙂 and well said.....
bUt still... I want it to be perfect.... I would never do THOSE "extras" if amp sounded bad... BUt AMP sounded WAY GOOD
worth the while adding those up ....
meanwhile, Im a bit "learning" how to use autocad...to edit some tracks you have done for the board.. 😀
hhhmmmm since you guys fail to answer my question regarding heating power supply, I re-read the "referrence for class D" thread.. to refresh my mind.....I found the answer....
shoothrough...caused by IRF540... Am I correct?
if so... need some FREE samples then....
🙄
shoothrough...caused by IRF540... Am I correct?
if so... need some FREE samples then....
🙄
shoothrough from the smps mosfets? Yes possibly. Try messing with deadtime or something before just reverting to swapping to other mosfets.
hello again,
I think there are 2 factors leading to my supply "heating"
#1. because of shootthrough..... of "improper" MOSFETs selection....I should get a fairchild MOSFET (lower gate-charge) faster reverse diode recovery time....
#2. I remember one of Jan-Peter's post regarding the trimmer (on the UCD).. used to adjust the quiscent supply?? was it??? so this trimmer adjusts deadtime.....next time I make a board, ill include a trimmer adjustment for deadtime 🙂
yes Chris, changing the deadtime setting did lower the "heat".... but not totally.. If I make it much
higher, the amp has a bit of "whining"... due to unshielded SMPS... maybe thats why ALMOST if not all UCD amps USE linear power supply .... Ive never seen(or heard) a commercial unit UCD
powered from SMPS...or am I wrong??
to those =masters= of SMPS.... what max switching frequency can I safely go with??? im currently running mine @ 100KHz.... maybe if I upper the switching freq of SMPS, UCD wont be affected anymore...
I think there are 2 factors leading to my supply "heating"
#1. because of shootthrough..... of "improper" MOSFETs selection....I should get a fairchild MOSFET (lower gate-charge) faster reverse diode recovery time....
#2. I remember one of Jan-Peter's post regarding the trimmer (on the UCD).. used to adjust the quiscent supply?? was it??? so this trimmer adjusts deadtime.....next time I make a board, ill include a trimmer adjustment for deadtime 🙂
yes Chris, changing the deadtime setting did lower the "heat".... but not totally.. If I make it much
higher, the amp has a bit of "whining"... due to unshielded SMPS... maybe thats why ALMOST if not all UCD amps USE linear power supply .... Ive never seen(or heard) a commercial unit UCD
powered from SMPS...or am I wrong??
to those =masters= of SMPS.... what max switching frequency can I safely go with??? im currently running mine @ 100KHz.... maybe if I upper the switching freq of SMPS, UCD wont be affected anymore...
RX5,
UCD has been powered by at least an SMPS from Coldamp SPS80. I have read some examples, and the users who did that found no noises at all. Have a look at:
http://www.diyaudio.com/forums/showthread.php?threadid=83670
http://www.diyaudio.com/forums/showthread.php?threadid=80335
So I think that your PSU has too much switching noise at the output. If I am not wrong, SPS80 runs at around 90KHz, very similar to yours. Perhaps ssanmor can give more exact data about this.
UCD has been powered by at least an SMPS from Coldamp SPS80. I have read some examples, and the users who did that found no noises at all. Have a look at:
http://www.diyaudio.com/forums/showthread.php?threadid=83670
http://www.diyaudio.com/forums/showthread.php?threadid=80335
So I think that your PSU has too much switching noise at the output. If I am not wrong, SPS80 runs at around 90KHz, very similar to yours. Perhaps ssanmor can give more exact data about this.
Hello Pierre,
too much you say?? I think too little I have?? TL494 is 100KHz but IS in push-pull mode.... meaning ( I think?) is roughly 100KHz / 2 = 50 KHz...... I dont know if at all possible to increase switching freq to 200KHz maybe????
Tekko
can you post a pic of your Damp? diy or commercial? 😀
It would be better to use IR2153D (very impressive IC) 🙂 in push-pull config, i.e. shorted bootstrap diode.
It automaticaly would keep deadtime at reasonable value, so your IRF540 won`t do crossconduct.
Btw, it`s really cheap and easy to buy. OR you can get it from dead chinese daylight lamp ballast.
It automaticaly would keep deadtime at reasonable value, so your IRF540 won`t do crossconduct.
Btw, it`s really cheap and easy to buy. OR you can get it from dead chinese daylight lamp ballast.
Hello, RX5.
I said that you must have too much noise, not too high frequency. I don't think it is a matter of frequency, as if your conducted or radiated noise is low enough it shouldn't beat with the amplifier.
Did you use snubbers?
Just an opinion
I said that you must have too much noise, not too high frequency. I don't think it is a matter of frequency, as if your conducted or radiated noise is low enough it shouldn't beat with the amplifier.
Did you use snubbers?
Just an opinion
updated pics of SMPS....
after seeing the coldamp sps80 output filters, I thought, might give it a try.. and YES.. noise has lowered significantly.... coil near the rectifiers are 22uH and coils between the filtercaps, 2uH.... ferrite material.... from dead ATX power supplies... and YES....much lower heat from torroid core....iron powder core just wont do it....nice to have some junk parts in hand.... 😀
Kuzmenko,
I like this kind of reply.. 😀 ok ill try to look for those ballast...and maybe salvage some of those IR chips..
Pierre,
on SMPS? yes.. in the primary side of torroid.... 50R + 10nF... per side... are these values OK with 100KHz? man I misplaced the equation for this somewhere.... anyone still got equation for one???
Tekko,
nice...😀 how many watts RMS??
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
after seeing the coldamp sps80 output filters, I thought, might give it a try.. and YES.. noise has lowered significantly.... coil near the rectifiers are 22uH and coils between the filtercaps, 2uH.... ferrite material.... from dead ATX power supplies... and YES....much lower heat from torroid core....iron powder core just wont do it....nice to have some junk parts in hand.... 😀
Kuzmenko,
I like this kind of reply.. 😀 ok ill try to look for those ballast...and maybe salvage some of those IR chips..
Pierre,
on SMPS? yes.. in the primary side of torroid.... 50R + 10nF... per side... are these values OK with 100KHz? man I misplaced the equation for this somewhere.... anyone still got equation for one???
Tekko,
nice...😀 how many watts RMS??
- Status
- Not open for further replies.