2 kwts discreet class d need help

stewin

Member
hi guys i have constructed and have been using this ucd amp based for quite some time . i wanted to juice it up to use multiple output like 4 pairs of irfp250 / ifrp260 or 3pairs of irf540 like in iraud 9amp with 250and 260 i wanted to do +/-72vlts at 4 to 2 ohms i dont have ociloscope or don't know how to go about one. much participation will be appreciated. :):):) stewin
 

Attachments

  • ucd discreet pcb top.pdf
    51.1 KB · Views: 1,452
  • ucd discreet pcb bottom.pdf
    55.2 KB · Views: 995
  • ucd discreet components clear.pdf
    42.4 KB · Views: 993
  • ucd discreet pcb artwork stereo.jpg
    ucd discreet pcb artwork stereo.jpg
    17.7 KB · Views: 1,350
  • ucd discreet pcb  stereo.jpg
    ucd discreet pcb stereo.jpg
    20.3 KB · Views: 1,392
  • ucd discreet pcb artwork stereo.pdf
    530.3 KB · Views: 1,208
  • ucd discreet schema.pdf
    32 KB · Views: 2,552
  • ucd discreet components all.pdf
    44.6 KB · Views: 1,088
  • ucd single discreet .bmp
    900.1 KB · Views: 1,108

stewin

Member
i am using +/- 45vlts cause of irf540 voltage ratings.
i had written +/- 80vlts on the schema hoping to get my hands on a cool fet like irfb4227
 

norazmi

diysmps Senior Member
hmmm, to go with high power, this circuit is too complicated. u can use simple one with ir2110 and can go such as power by using right mosfet. or drive it as fullbridge. btw this circuit is discrete, did u build it stewin?
 

stewin

Member
yes i have built it it works. and has not been disturbing me with noise issues. i like discreet because when it fails or burns i only replace one or two 2n5404 near the outputs and the outputs them selves unlike using ir2110 . i have to replace the output everytime plus the expensive ir2110 ic every time something goes wrong.
 

norazmi

diysmps Senior Member
agreed with IR2110 is expensive, discrete just need more work to do but you can save $$, i have few schema with discrete component and i do simu with LTspice seems work good. I think i will build it when i have time here i leave schema maybe anyone interested.
 

Attachments

  • small_ucd_sch_2.pdf
    35.3 KB · Views: 2,084

norazmi

diysmps Senior Member
hi stewin,

here is the result with LTspice, i do modification from original schematic because original one have 1.3% thd and offset is 98mvdc. so this is result for the modified one and i`ve free time today i`ll make one board for it, i wanna test it how this amp working and how stable it is.

Tested mosfet IRFP250N, with +-/- 48 vdc, output voltage +/- 32 vac 8 ampere. 4 ohm load. Dc offset only 0.004vdc.
Starting source stepping with srcstepmethod=0
Source Step = 3.0303%
Source Step = 9.84848%
Source Step = 17.4242%
Source Step = 25%
Source Step = 32.5758%
Source Step = 40.1515%
vernier = 0.375
Source Step = 50.7576%
Source Step = 62.1212%
Source Step = 73.4848%
Source Step = 84.8485%
Source Step = 96.2121%
vernier = 0.5625
Source stepping succeeded in finding the operating point.

Fourier components of V(out)
DC component:0.004799

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+02 3.135e+01 1.000e+00 -123.45° 0.00°
2 2.000e+02 3.292e-03 1.050e-04 56.05° 179.50°
3 3.000e+02 3.236e-02 1.032e-03 -166.58° -43.13°
4 4.000e+02 2.101e-03 6.702e-05 148.97° 272.42°
5 5.000e+02 2.386e-02 7.612e-04 -61.54° 61.92°
6 6.000e+02 2.849e-04 9.087e-06 76.29° 199.74°
7 7.000e+02 1.236e-02 3.944e-04 47.00° 170.45°
8 8.000e+02 4.671e-04 1.490e-05 10.87° 134.33°
9 9.000e+02 8.908e-03 2.841e-04 157.27° 280.72°
Total Harmonic Distortion: 0.137746%


Fourier components of V(in)
DC component:-3.3813e-011

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+02 2.000e+00 1.000e+00 -0.00° 0.00°
2 2.000e+02 1.188e-10 5.939e-11 -171.43° -171.43°
3 3.000e+02 2.156e-10 1.078e-10 163.34° 163.34°
4 4.000e+02 6.074e-11 3.037e-11 78.61° 78.61°
5 5.000e+02 1.227e-10 6.136e-11 -2.60° -2.60°
6 6.000e+02 5.527e-11 2.763e-11 113.55° 113.55°
7 7.000e+02 1.044e-10 5.218e-11 51.52° 51.52°
8 8.000e+02 2.053e-10 1.026e-10 -33.31° -33.31°
9 9.000e+02 1.105e-10 5.525e-11 -161.60° -161.60°
Total Harmonic Distortion: 0.000000%



Date: Fri Feb 24 12:19:12 2012
Total elapsed time: 1037.290 seconds.

tnom = 27
temp = 27
method = modified trap
totiter = 17672589
traniter = 17670226
tranpoints = 4732823
accept = 3348150
rejected = 1384673
matrix size = 96
fillins = 202
solver = Normal
Matrix Compiler1: 23.3 KB object code size 9.0/5.9/[2.5]
Matrix Compiler2: off [3.0]/3.6/16.1
 

stewin

Member
wow thanks nozrami for the help . are you sure your circuit can drive irfp250 without heating . with my experience with discreet ucd R1 in your schematic needs to be of a high value or the output will heat like crazy. why haven't you include the op-amp preamp section is the gain sufficient enough? thanks for the simulation man GOD almighty bless you big time.
 

stewin

Member
BTW nozrami does driving the irfp250 with small bjts like in my schema above q22 ,q14 ,q6 , q3 help in any way in terms of power output and t.h.d?
 

norazmi

diysmps Senior Member
yes R1 need to set high, its original schema, btw Stewin, im trying to build this one but not finish yet. For your schema, i`ll do some research let me see what we can do.
 

norazmi

diysmps Senior Member
Oo, as i can see ur schema is more better and include op -amp, i`ll try simulate it, and tell u the result
 

stewin

Member
thanks nozrami. my main aim was to get it up to 2kw at 4ohms using irfp250 or irfp260 even though i will use five pairs . but if it reaches up to 1kwts or 1.2kwts it will be o.k but i will be forced to bridge which is no problem because here irfp250 or irfp260 are very available even irf540, irf640 are also avail but those can go up to 400wts . i will post picts soon thanks again nozrami for simulating. btw what is the t.h.d of your 200 watts amp i love it because it is simple will try it and post picts too in that forum?

http://www.diysmps.com/forums/showthread.php?237-Class-D-200-Wrms-with-2-mosfet-cheap
 

norazmi

diysmps Senior Member
hihi,

Stewin im finish build that amp, schema i posted, lol, its worked nice... Btw i think R1 really need to be trimpot lol. because that value i think depends on supply voltage. But i did not test yet with variable voltage, just tested today with 24 0 24 vac and it work. What dimension of your pcb? mine i think little big lol, i dunno its about 8cm x 7 cm. cant make it more compact because too many discrete parts.

haha its funny lol. i like this topology because its Discrete with N-channel only mosfet in class D. my topology is little different with yours. first time im power it up i need to play with R1, this value really killer hahahaha. too low value then u will heat the mosfet alot, too high will be no sound at all lol :D . I think i`ll put trimmer with 500 ohm or maybe 1k. I dont play yet with duty cycle and gate resistor. maybe tomorrow i`ll do. running this amp hard mosfet stay cool all the time. and i`m not yet burn anything from this amp hahahaha.

Stewin i dont know with your topology what the advantage with 2 bjts at the mosfet gate. but i`ll do simulate it for ya, gimme some time okay. i`ll take some picture too and maybe video when playing this amp. Okay for your question regarding KW power output, hmmm.. maybe can with IRFP250 but i think it will distort much at that power i`m not sure i need to simulate it, and remember on simulation we dont know exactly if the mosfet overheat so u need to do it at manually in real live. But still we can give it a try, why not? :p.

Regarding simple 200wrms the thd is below 1%, trust me, u just need proper power supply with enough current to drive with it.

regards,
 

stewin

Member
thanks again man. the board dimensions are exactly like in the artwork pdf , when you are printing with your laser jet do not select the fit to page option . just select none option,
and print. it will give you the exact measurement of board . just measure the printed paper.
the circuit works perfectly without the two bjts gate drivers, but i have used the 2 bjts gate drivers because i wanted to drive like four pairs of irfp250 or 260 due to lack of affordable good fets like irfb4227 or irfp4227 i saw the topology in iraud amp 9 and here http://www.diyaudio.com/forums/class-d/184954-gzpa-class-d-4000w-irfp264.html
 

norazmi

diysmps Senior Member
Hi stewin,

tell me about your amp sound. I`ve simulate with LTspice and its weird regarding the sound at output, distort at high volume ? did it produce noise? I`m asking this because the value resistor R6, R11 and R35, R28, R19 and R12. If this resistor that i mention is not have proper value, than this amp ucd will produce noise alot, maybe less but will have noise and will follow with your volume control, because this system work like RF.

And the heart of this amp is R1 on your schema is R10, you can use trimpot for this, if u wanna try, first take 1k trimpot and tune it at maximum 1k resistance. turn on the amp and start lower the 1k resistor until u get the sound at output. Dont make it lower than that, because u just got proper value for the R10 when the amp start osc and u get sound at output, and u will found mosfet will cool even u play hard. i`ll post here with your setup but i`ll not include op-amp yet okay, i`ll do put op-amp later and simulate will post result also.

here is mine first. Input 2vrms freq 100Hz, power rails +/- 48vdc , 100Hz freq output +/- 24 vac at 6 ampere at 4 ohm load, 470n filter cap, 30uH inductance.


Starting source stepping with srcstepmethod=0
Source Step = 3.0303%
Source Step = 33.3333%
Source Step = 63.6364%
Source Step = 93.9394%
Source stepping succeeded in finding the operating point.

Fourier components of V(out)
DC component:0.0000834

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+02 2.262e+01 1.000e+00 -124.68° 0.00°
2 2.000e+02 9.930e-04 4.391e-05 -35.90° 88.78°
3 3.000e+02 2.565e-02 1.134e-03 -168.04° -43.36°
4 4.000e+02 5.921e-03 2.618e-04 142.70° 267.38°
5 5.000e+02 8.691e-03 3.843e-04 -62.36° 62.32°
6 6.000e+02 1.849e-03 8.177e-05 -120.58° 4.09°
7 7.000e+02 4.359e-03 1.927e-04 46.69° 171.36°
8 8.000e+02 1.655e-03 7.318e-05 -17.55° 107.12°
9 9.000e+02 2.364e-03 1.045e-04 157.75° 282.43°
Total Harmonic Distortion: 0.025101%


Fourier components of V(in)
DC component:6.85373e-012

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+02 2.000e+00 1.000e+00 0.00° 0.00°
2 2.000e+02 7.652e-11 3.826e-11 -41.81° -41.81°
3 3.000e+02 7.637e-11 3.819e-11 150.80° 150.80°
4 4.000e+02 5.838e-11 2.919e-11 -148.85° -148.85°
5 5.000e+02 1.304e-10 6.519e-11 -34.88° -34.88°
6 6.000e+02 6.466e-11 3.233e-11 121.59° 121.59°
7 7.000e+02 9.702e-11 4.851e-11 -171.13° -171.13°
8 8.000e+02 1.236e-10 6.180e-11 23.47° 23.47°
9 9.000e+02 1.144e-10 5.720e-11 65.58° 65.58°
Total Harmonic Distortion: 0.000000%



Date: Sun Feb 26 09:16:53 2012
Total elapsed time: 896.354 seconds.

tnom = 27
temp = 27
method = modified trap
totiter = 13922616
traniter = 13919918
tranpoints = 3697359
accept = 2594680
rejected = 1102679
matrix size = 95
fillins = 206
solver = Normal
Matrix Compiler1: 23.5 KB object code size 8.8/5.9/[2.5]
Matrix Compiler2: off [2.7]/3.2/17.0
 

norazmi

diysmps Senior Member
here is the graph. mosfet use is IRFP250N.
 

Attachments

  • input.jpg
    input.jpg
    36.5 KB · Views: 66
  • Load.jpg
    Load.jpg
    42.1 KB · Views: 60
  • inductance load.jpg
    inductance load.jpg
    46 KB · Views: 70
  • wave.jpg
    wave.jpg
    41.5 KB · Views: 65

norazmi

diysmps Senior Member
Stewin,

all parts use with your schema.

Done simulate running with +/- 65 vdc, note that if u want go more high voltage, the 5401 and 5551 must change to mpsa 42 and mpsa 92 which have good voltage at 300 vdc.

simulate 100Hz, mosfet irfp250N, gate 22 ohm, input at 2vrms. load 4 ohm.


Starting source stepping with srcstepmethod=0
Source Step = 3.0303%
Source Step = 33.3333%
Source Step = 63.6364%
Source Step = 93.9394%
Source stepping succeeded in finding the operating point.

Fourier components of V(out)
DC component:0.0557483

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+02 1.964e+01 1.000e+00 -133.12° 0.00°
2 2.000e+02 1.336e-01 6.803e-03 33.11° 166.23°
3 3.000e+02 7.249e-01 3.691e-02 158.50° 291.62°
4 4.000e+02 6.938e-02 3.533e-03 106.74° 239.86°
5 5.000e+02 1.047e-01 5.329e-03 -120.00° 13.12°
6 6.000e+02 6.021e-02 3.066e-03 -166.71° -33.59°
7 7.000e+02 1.254e-01 6.385e-03 151.94° 285.06°
8 8.000e+02 4.795e-02 2.442e-03 -80.11° 53.00°
9 9.000e+02 2.058e-01 1.048e-02 -118.20° 14.92°
Total Harmonic Distortion: 4.019748%


Fourier components of V(in)
DC component:-1.59857e-010

Harmonic Frequency Fourier Normalized Phase Normalized
Number [Hz] Component Component [degree] Phase [deg]
1 1.000e+02 2.000e+00 1.000e+00 -0.00° 0.00°
2 2.000e+02 2.431e-10 1.216e-10 130.67° 130.67°
3 3.000e+02 3.124e-10 1.562e-10 113.37° 113.37°
4 4.000e+02 8.134e-11 4.067e-11 63.90° 63.90°
5 5.000e+02 1.796e-10 8.981e-11 19.93° 19.93°
6 6.000e+02 1.649e-10 8.247e-11 -46.62° -46.62°
7 7.000e+02 7.209e-11 3.604e-11 -81.00° -81.00°
8 8.000e+02 4.815e-11 2.407e-11 98.47° 98.47°
9 9.000e+02 1.340e-10 6.700e-11 -0.05° -0.05°
Total Harmonic Distortion: 0.000000%



Date: Sun Feb 26 11:47:17 2012
Total elapsed time: 773.724 seconds.

tnom = 27
temp = 27
method = modified trap
totiter = 10492943
traniter = 10490527
tranpoints = 2731542
accept = 1939583
rejected = 791959
matrix size = 94
fillins = 226
solver = Normal
Matrix Compiler1: 28.1 KB object code size 10.4/6.7/[2.8]
Matrix Compiler2: off [2.5]/3.1/16.2
 

stewin

Member
thanks thanks thanks . funny it is true at high volume it produces distortion very audible and irritating but when i reduce the bass it sort of reduces . thanks man i was puzzled i thought i had done a poor layout of my pcb . my ucd class d is based on the ucd 2.5 class d from r u but i guess it is because he uses smd components and cool fets . based on your simulation are you saying if i change the resistors to be like on your schema it will have power and the distortion will disappear?
 

norazmi

diysmps Senior Member
ur welcome, no, dont change like mine, i`ll change it and test first with simulation, then u can fo change with what i`ll suggest for ya, then tell me how the sound produce. i`ll make test your schema tonite, and will tell u the value, at the mean time u can do with R10, try use trimpot 1k ohm, do it like what i have explain above, tell us here :D
 

stewin

Member
hi nozrami if you are going to use your schematic please change d15 to uf4007 or mur120 because 1n4148 fails if you use over +/-35 . And also i would omit d5 and do without it because in all ucd discreet i have made d5 always fail after some time and if i do without it it works fine. do d16,11,17,18 help performance in any way?
 
Top