I did one Point-to-point construction (a prototype)
was a BC337 with altered polarities.
Like filter, can use nucleus of filter of SMPS ATX, which works very well.
Give 30 turns of thread in him that goes to work
YouTube - Amplificador classe D de tecnologia UCD, montado em circuito poto-a-ponto
was a BC337 with altered polarities.
Like filter, can use nucleus of filter of SMPS ATX, which works very well.
Give 30 turns of thread in him that goes to work
YouTube - Amplificador classe D de tecnologia UCD, montado em circuito poto-a-ponto
Last edited:
A sort of class d amp config
Have been looking at class D amps for a couple of weeks now. Started by wanting to know what a class t-amp was n obviously ended up here, as it turns out they are simply a single chip implementation of a class d amp.
I have looked with interest at your simple discrete self oscillating circuits. The basic concepts of this amp reminded be of a "sort of class - d" output stage of an aircraft analogue interface circuit that i worked on some 25-30 years ago. Remember (or for you young ones - be informed 🙂) back then an 8-bit Z80 uP clocked a 4MHz was state of the art! So, to produce 8 high voltage modest current analogue outputs the uP would write to an ADC (very expensive beasties in those days) and an 8 bit analogue mux. Each new value would be used, momentarily, to drive the +ve input of a comparator the -ve input being driven by a class D o/p attached to a gurt big cap. Once the comparator changed state the drive to both high and low side o/p trannies were removed (Big Bipolar trannies - power MOSFETs hadn't been invented in those days). The idea obviously, was the cap was charged up to the desired voltage and maintained it through the "dead interval". I remember this cct well because being a wet behind the ears engineer at the time, I thought this is all wrong ( i was arrogant in those days) I thought why not simply store the charge on a small cap and use a class b o/p stage to provide the drive? I very soon learnt the notion of efficiency!
Any way, this whole class d amp idea got me thinking and I came up with the attached as a possible scheme - its very raw. My thinking is that shoot through currents wont happen. Start up isn't a problem and the amplitude of the high freq noise is significantly reduced. An obvious disadvantage is the size of the storage cap required - lots of uF 🙁 As for distortion your guess is as good as mine 😕
I'd love to bread board the concept just to check that my memory isn't playing tricks on me and to see if such a scheme would work. Unfortunately I no longer have access to the necessary test equipment.
I thought I'd share it with you anyway just in case you find it an interesting notion and not just the mindless ramblings of a old codger.
Have been looking at class D amps for a couple of weeks now. Started by wanting to know what a class t-amp was n obviously ended up here, as it turns out they are simply a single chip implementation of a class d amp.
I have looked with interest at your simple discrete self oscillating circuits. The basic concepts of this amp reminded be of a "sort of class - d" output stage of an aircraft analogue interface circuit that i worked on some 25-30 years ago. Remember (or for you young ones - be informed 🙂) back then an 8-bit Z80 uP clocked a 4MHz was state of the art! So, to produce 8 high voltage modest current analogue outputs the uP would write to an ADC (very expensive beasties in those days) and an 8 bit analogue mux. Each new value would be used, momentarily, to drive the +ve input of a comparator the -ve input being driven by a class D o/p attached to a gurt big cap. Once the comparator changed state the drive to both high and low side o/p trannies were removed (Big Bipolar trannies - power MOSFETs hadn't been invented in those days). The idea obviously, was the cap was charged up to the desired voltage and maintained it through the "dead interval". I remember this cct well because being a wet behind the ears engineer at the time, I thought this is all wrong ( i was arrogant in those days) I thought why not simply store the charge on a small cap and use a class b o/p stage to provide the drive? I very soon learnt the notion of efficiency!
Any way, this whole class d amp idea got me thinking and I came up with the attached as a possible scheme - its very raw. My thinking is that shoot through currents wont happen. Start up isn't a problem and the amplitude of the high freq noise is significantly reduced. An obvious disadvantage is the size of the storage cap required - lots of uF 🙁 As for distortion your guess is as good as mine 😕
I'd love to bread board the concept just to check that my memory isn't playing tricks on me and to see if such a scheme would work. Unfortunately I no longer have access to the necessary test equipment.
I thought I'd share it with you anyway just in case you find it an interesting notion and not just the mindless ramblings of a old codger.
Hi, thanks for your question.
The C in question is acting as a tank capacitor providing current once you have turned off the high/low drivers it is not part of the low pass filter. It will be quite large - rough calcs suggest it will be in the order of 10s of uF for 200kHz switching frequency.
The C in question is acting as a tank capacitor providing current once you have turned off the high/low drivers it is not part of the low pass filter. It will be quite large - rough calcs suggest it will be in the order of 10s of uF for 200kHz switching frequency.
Hi Luka
Once again thanks for showing an interest. I'll say now - it is a long time (decades) since I was employed professionally as an electronics engineer I very quickly ended up moving into systems engineering. So I am very rusty on the finer points here, but i'm working on the premise that at this node you cannot think of C in analogue terms.
This is the way i am looking at it:
If Fs is ~ 200kHz, T~5uS
T is divided between High/Low on time: Ton, and High/Low off time: Toff
Then, considering the supply current Isup and the load Current Iload and using V=(I x T)/C you get:
At the end of Ton, the voltage on the cap will go up/down by:
Ton x (Isup - Iload)/C
and at the end of Toff the the voltage on the cap will go down/up by:
Toff x Iload/C.
Obviously, at peak signal voltages the amount of voltage decay on the cap will be much different to that at lower voltages - a source of non linearity possibly, however, I'm thinking that if Fs and C can be set to values to minimise this the harmonic distortion at higher output voltages could be minimised.
Hope this makes sense and hope i've not overlooked anything obvious regarding interactions between C and the low pass filter stage.
Once again thanks for showing an interest. I'll say now - it is a long time (decades) since I was employed professionally as an electronics engineer I very quickly ended up moving into systems engineering. So I am very rusty on the finer points here, but i'm working on the premise that at this node you cannot think of C in analogue terms.
This is the way i am looking at it:
If Fs is ~ 200kHz, T~5uS
T is divided between High/Low on time: Ton, and High/Low off time: Toff
Then, considering the supply current Isup and the load Current Iload and using V=(I x T)/C you get:
At the end of Ton, the voltage on the cap will go up/down by:
Ton x (Isup - Iload)/C
and at the end of Toff the the voltage on the cap will go down/up by:
Toff x Iload/C.
Obviously, at peak signal voltages the amount of voltage decay on the cap will be much different to that at lower voltages - a source of non linearity possibly, however, I'm thinking that if Fs and C can be set to values to minimise this the harmonic distortion at higher output voltages could be minimised.
Hope this makes sense and hope i've not overlooked anything obvious regarding interactions between C and the low pass filter stage.
An externally hosted image should be here but it was not working when we last tested it.
My amplifier it has power supply 40V symmetrically.
He has not good profit, in the output only it produces 17V in 60Hz of sign the sound.
What I can to do improve the profit and to produce pelomenos 30V sound in the output?
Hi,
Driver system not have efficiency, you can improve performance on C4-C5, change to 1uF not polarity.
Regards
Driver system not have efficiency, you can improve performance on C4-C5, change to 1uF not polarity.
Regards
Good to see your amp is working gr8. I too got my PCB I shall assemble it today night.
Well, I could not get exact meaning of "Profit" here. but I will suggest you to download simulation file uploaded by M-tech. using this you can find exactly what is the Vp input for distortionless output.and at the given supply voltage what could be the maximum output.
Well tomorrow I will post my updates. just keeping my fingers crossed. Hope my PCB work too.
After you instructed me to insert capacitance of 1uF now it was so:
power supply = 40V
Voltage of sound = 25V / 60Hz in 4 OHMS
is it possivel to improve the profit?
power supply = 40V
Voltage of sound = 25V / 60Hz in 4 OHMS
is it possivel to improve the profit?
Hi,
this amp have various defect (thd,slow response..etc) but if you like this project, (maybe is simple) just demostrate that problem is "driver scheme",
it is slow,non-linear,with low current for gate mosfet,infact, after change to 1uF, now increment 17 to 25v out, you can change res 2K (on out comp) to 1K, also comparator,( out type), is not suitable perfect for driving totem-pole direct (current is insufficient). with this 1K change, you can gain 3-4 V at output....no other help for this🙂
Regards
You have changed 2K to 1K res? you test it?
Out of 393 is open collector, 2K produces small current for driving totem-pole bjet.
change comp with totem-pole outputs or add an bjet .
Hey, I assembled my the PCB. some changes had to be made. I had to use 7812 and 7912 for 12 volts regulator. I am working on transistors heating and Coil heating ( actually only core heating. )
will update you. your suggestions on transistor heating and core heating are welcome.
I am posting my current diagram.
will update you. your suggestions on transistor heating and core heating are welcome.
I am posting my current diagram.
Hey guys, after proper coil was replaced, heating of transistor and coil stopped now both are perfectly cool. but there is small noise at the output. want to remove it. do you all face same problem of noise ?
- Status
- Not open for further replies.