1kW smps project (based on MicrosiM design)

behrad

Member
thanks mr silvio
1:bottom pic is correct about winding?
pv.jpg
2:what is point 1 and 2 in bottom pic?
3:i did not found c17(1uf 825v).can i use 1uf 630v polyster instead it?
zzz.jpg
4:can i use 7812 reg ic instead 7815?
5:can i use inductors in output rails?(like mr silvio 1kw smps)
 
thanks mr silvio
1:bottom pic is correct about winding?
View attachment 6415
2:what is point 1 and 2 in bottom pic?
3:i did not found c17(1uf 825v).can i use 1uf 630v polyster instead it?
View attachment 6416
4:can i use 7812 reg ic instead 7815?
5:can i use inductors in output rails?(like mr silvio 1kw smps)

Here's the winding order :
1/2 primary
1/2 secondary
1/2 secondary
1/2 primary
pv.jpg
You can wind it ether Clockwise or Counter Clockwise but you have to do all your winding with only one of them


point 1 is a jumper that will pass through your Current core ( this is for Protection board and if you don't want to use protection than you can just Jumper it)
123.jpg
Point 2 is were your Current core winding will go
 
Last edited:

behrad

Member
i want to build protection unit and use it
what is ct toroid size?and How many rounds of wire should I turn it on?
i have yellow iron powder toroid from computer psu?is it proper?
secondary winding on toroid go to point 2???
 
i want to build protection unit and use it
what is ct toroid size?and How many rounds of wire should I turn it on?
i have yellow iron powder toroid from computer psu?is it proper?
secondary winding on toroid go to point 2???

I use this one also taken out from a PC power supply ( I put it next to an IRFP460 for you to get it's size )
222222.jpg

Start with 22 turns and yes, just connect it's wending to point 2 that goes to protection board
 

Silvio

Well-known member
thanks mr silvio
1:bottom pic is correct about winding?
View attachment 6415
2:what is point 1 and 2 in bottom pic?
3:i did not found c17(1uf 825v).can i use 1uf 630v polyster instead it?
View attachment 6416
4:can i use 7812 reg ic instead 7815?
5:can i use inductors in output rails?(like mr silvio 1kw smps)

1) Yes winding is correct
2 point 1 is where you pass through the current trafo ( the small ring toroid Badboy mentioned
3 point 2 is where you go with the toroid secondary windings (22 turns)
4) yes you can but if you can use 15v it will be better, You can add a small zener of 3.3v in series with the ground of the 7812 and it will lift up the voltage to 15v.
5) Yes you can use output inductors but you will experience some voltage drop on the output with load. These are there because the trafo will see a dead short during charge up of the secondary caps. with the inductors there will be some current limit. I used 1.5uH and current will limit around 10 to 15 amps. You can leave some space for them in case you need them. You can always bridge the traces in case you don''t use them. ( I had 4000uF and 90v peak on each rail)
6) You can use the capacitor with a lower voltage rating of 630v. This takes all the current through it and has to be a robust one. That is why the higher voltage rating.

Regards Silvio
 
Hello dear Silvio
I test my New 400W amp with this SMPS and in output I get a bit fuzzy signal
gjfg.jpg
But with old 50Hz transformer I get a completely clean output waves
IMG_20171024_195021.jpg

Is this normal???
If not, is there any way to fix this ????
 

Silvio

Well-known member
@Badboy
It could be that some EMI is getting through to the amplifier. I suggest that:-
1) you make a couple of 1uf film capacitors between + and ground and - and ground on the output of the smps. This will filter HF noises on the supply rails
2) when the smps is fitted in the amp box make a metal screen between smps and amplifier this will screen a lot of interference generated by the smps. I think you still got a lot of wires going round and these all serve as antennas. I also suggest you add an input filter at the mains. Try also to decouple the center ground with a small ceramic cap to chassis.

Is that fuzzy signal audible? Does it increase with more power? Did you try a radio near the smps ? (about a foot away) you will hear some noise for sure and the more load the more the noise. This will tell you what you are dealing with. Try to screen the cables to the input of the amp and also the supply wires from smps.
EMI is not so easy to get rid of and good filtering is needed.

Silvio
 
Last edited:
@Badboy
It could be that some EMI is getting through to the amplifier. I suggest that:-
1) you make a couple of 1uf film capacitors between + and ground and - and ground on the output of the smps. This will filter HF noises on the supply rails
2) when the smps is fitted in the amp box make a metal screen between smps and amplifier this will screen a lot of interference generated by the smps. I think you still got a lot of wires going round and these all serve as antennas. I also suggest you add an input filter at the mains. Try also to decouple the center ground with a small ceramic cap to chassis.

Is that fuzzy signal audible? Does it increase with more power? Did you try a radio near the smps ? (about a foot away) you will hear some noise for sure and the more load the more the noise. This will tell you what you are dealing with. Try to screen the cables to the input of the amp and also the supply wires from smps.
EMI is not so easy to get rid of and good filtering is needed.

Silvio

Thank you for your reply
since I'm using this app for driving a subwoofer, i can't hear any noise on my subwoofer (like it's not even connected to amp) and also this happens as the output power on amp increases (when smps goes under load)
I also used the input main filtering from an Pc psu and that did't effect anything
I also made a new pcb for my amp and now there is much less wires involved
I just connect the amp to 50HZ transformer and put the amp next to the smps as it was under load (about 600W)
there was no distortions and very clean signal in amp output
I made this test with and without input EMI filters on smps and there was definetely no differences

So I guess this problem has nothing to do with magnetic interference or SMPS input EMI filtering

I test the radio thing and yes, I definetely could hear some noise in 1 meter radius while smps was under 600W load

I think this problem comes from SMPS output rails

I will try that 1uf idea and let you know the result
 
Last edited:

leland331

New member
Hey,

I would like to build this supply to run a Pass Aleph amp.

I have two questions about it.

1: is there any way to make it work on both 120v 60hz, as well as 240v 60hz? I will mainly use 240v, but sometimes only 120v will be available here.

2: This is a big single ended class A amp. it'll draw a steady current. Is it worthwhile/possible to make it regulated?

Thank you
 

Silvio

Well-known member
@Leland
Yes Ludo smps can work on both voltages by using the voltage doubler for the 120v mode. For you to make it regulated it is also possible but you have to build a separate board for the feedback circuitry. You also have to include the output inductors and make arrangements to the pcb to take them.

For a steady current load the smps cannot give 1000 watts continuous as the copper wire in the trafo has to be somewhat thicker to cater for the continuous current load. One other option is to use a bigger core or make some kind of forced cooling on the trafo to cool it down but this will make some noise. If you can tell us what is the power consumed by your amplifier I can see things better. I will also want to know what is the regulated voltage, the minimum and maximum current absorbed.

Silvio.
 

leland331

New member
Thank you for the reply.

By the looks of it, the amp will draw 6-800w steady. It won’t increase or decrease.

The final voltage will be +45 -45
 

Silvio

Well-known member
@Leland
Ok about the current consumption of the smps so for 800w at 90v your current will be approximately 9 amps You can add a small pcb with the feedback loop back to pin 1 and ground. I don't know if you are familiar with smps and feedback circuitry but if the amp will draw a steady current then you will have a more or less stable voltage at load will little fluctuations. If the amp can tolerate a little fluctuation in the input voltage say around 10v or so then peak voltage will be around 55v and sags with load to around 45v or so. Like this there is no need for stabilization, this will make things simpler for you. Chopping the waveform during stabilization is more prone to noise in the output voltage rails and more filtering is needed.

Regards Silvio
 

Silvio

Well-known member
Hi Badboy, I cannot really tell if it is ringing or not. Take more scope shots of gate to source and drain to source of each transistor.
Please say if it is with load or without load.

I guess that the scope shot you sent in your last post is with the probes across the primary correct?

That is rather an odd wave shape. One half cycle looks good while the other looks like it is having a double trigger. Ringing does not look like that usually.
 
Hi Badboy, I cannot really tell if it is ringing or not. Take more scope shots of gate to source and drain to source of each transistor.
Please say if it is with load or without load.

I guess that the scope shot you sent in your last post is with the probes across the primary correct?

That is rather an odd wave shape. One half cycle looks good while the other looks like it is having a double trigger. Ringing does not look like that usually.

hello dear Silvio and thanks for your kind help
yes the picture are from the primary of ETD59

and it was ringing because by using snubbers on my IGBTs the problem go away and now everything is good

I split my SMPS into 2 pcbs as it was too big to fit in just one
the first one that contains input bridge, caps, protection board, Controller, 15V voltage driver And the IGBTs is finished
and right now I'm working on the second one that will contain ETD59, Output diodes, output caps
I will try to finish it by tomorrow and will post the pictures of it
It will handle 1.5KW continuos power in output without any problem.
I'll be back soon
 

Silvio

Well-known member
Hi Badboy, It is not always a must to use snubbers at the input. Sometimes they make things go worse. snubber values can vary and there will be a difference if using a soft snubber or a hard one. This is usually used if ringing is rather large. We that we have no output inductors will not usually be needed. At times output inductors create ringing themselves which reflect in back to the primary. This I can prove as it happened to me.
I am curious why do you need this much power. I know that sub woofers are quite demanding but other frequencies do consume around 30 to 60% depending on type of music played.

Regards Silvio
 
Hi Badboy, It is not always a must to use snubbers at the input. Sometimes they make things go worse. snubber values can vary and there will be a difference if using a soft snubber or a hard one. This is usually used if ringing is rather large. We that we have no output inductors will not usually be needed. At times output inductors create ringing themselves which reflect in back to the primary. This I can prove as it happened to me.
I am curious why do you need this much power. I know that sub woofers are quite demanding but other frequencies do consume around 30 to 60% depending on type of music played.

Regards Silvio

I agree because in my prototype smps I didn't use any snubber and there was no problem at all

I go for this much power as I'm going for more powerful amps in the future
 

MicrosiM

Administrator
Staff member
Snubbers needed if

You have nasty EMI being injected somehow at load or no load.

or

You have Spikes at the MOSFET or the IGBT.

or

Unstable operation.
 
Top