Bottom line is that you can use 0,22uF before and after the choke. That value is not very specific..does the job..almost always according to Morgan Jones.
Theory..
Because the voltage from the diode at some times drops to zero..the choke wants to maintain current and will develop a voltage.
The snubber in other words filters out the voltage spikes..
Hope this helps..
Theory..
Because the voltage from the diode at some times drops to zero..the choke wants to maintain current and will develop a voltage.
The snubber in other words filters out the voltage spikes..
Hope this helps..
It's funny how some trivial questions cause endless reams of glib replies, while a little tricky one recieves few.
I think the concept of snubbing inductors is a little confused.
There are 2 moments in time when a CR series network connected in parallel with the choke can be useful in a choke filter circuit:
1) To reduce the initial voltage peak that can be generated from the choke at turn on.
This is to reduce the chance of first capacitor, choke insulation or rectifier breakdown.
2) In a choke input circuit to reduce the peak voltage across the choke on every cycle.
This is to reduce the chance of choke insulation or rectifier breakdown. And can reduce vibration noise from the choke, though most people just add a small value cap from the choke input to ground, even though this causes a voltage rise.
Of course the correct type and rating of choke is presumed.
Bas's "rule of thumb" value of 0.22uF sounds about right from my experience. Make sure the voltage rating is at least 2x the max in circuit.
As for the resistor: You must use one. It's where the energy is going! So, consider a dissapation of several watts. And the value? I can't remember; it obviously depends on a number of factors, but since we're into "rule of thumb", I'd start with 10K and work downwards.
I think the concept of snubbing inductors is a little confused.
There are 2 moments in time when a CR series network connected in parallel with the choke can be useful in a choke filter circuit:
1) To reduce the initial voltage peak that can be generated from the choke at turn on.
This is to reduce the chance of first capacitor, choke insulation or rectifier breakdown.
2) In a choke input circuit to reduce the peak voltage across the choke on every cycle.
This is to reduce the chance of choke insulation or rectifier breakdown. And can reduce vibration noise from the choke, though most people just add a small value cap from the choke input to ground, even though this causes a voltage rise.
Of course the correct type and rating of choke is presumed.
Bas's "rule of thumb" value of 0.22uF sounds about right from my experience. Make sure the voltage rating is at least 2x the max in circuit.
As for the resistor: You must use one. It's where the energy is going! So, consider a dissapation of several watts. And the value? I can't remember; it obviously depends on a number of factors, but since we're into "rule of thumb", I'd start with 10K and work downwards.
I have now bought some 0,1 uF capacitors with a voltage rating of 2000V DC.
But do you have an idea of how I can measure the voltage peaks produced by the choke with an oscilloscope?
I have simulated the power supply with Duncan's PSU Designer and the first peak was about 3000V!!!
While the others were a bit less than 2000V peak.
In my design I'm using a mainstransformer with two secondaries with 2*260V, these windings are coupled in series giving a voltage of 520V, and then we have a brigde consisting of 12 UF4007(fast 1N4007). The filter is a LC filter with 10H/32ohm choke and two 1000uF capacitors giving 500uF.
But do you have an idea of how I can measure the voltage peaks produced by the choke with an oscilloscope?
I have simulated the power supply with Duncan's PSU Designer and the first peak was about 3000V!!!
While the others were a bit less than 2000V peak.
In my design I'm using a mainstransformer with two secondaries with 2*260V, these windings are coupled in series giving a voltage of 520V, and then we have a brigde consisting of 12 UF4007(fast 1N4007). The filter is a LC filter with 10H/32ohm choke and two 1000uF capacitors giving 500uF.
Gunderz,
Disclaimer:
Please only do this if you are responsible and comfortable with high voltages.
Ensure capacitors are fully discharged before touching anything.
The Biz:
The real-world voltages will be slightly less, but still very high.
This is because the simulator doesn't take into account all of the chokes properties - just the important ones.
To measure the peaks, you will need to make a 100:1 scope probe.
If you already have a 10:1 probe, you can modify it:
You will need 10 x 1Mohm resistors.*
Connect the resistors in series between the high voltage test-point and ground.
Connect your 10:1 probe across the resistor nearest to ground.
There will be a slight HF loss, and a 5% measurement error, but it shouldn't matter in this case.
Remember to multiply your scopes reading by 100 to get the result.
*(1.5M, or 2.2M will do, but the measurement error will be greater)
Hope this helps
Disclaimer:
Please only do this if you are responsible and comfortable with high voltages.
Ensure capacitors are fully discharged before touching anything.
The Biz:
The real-world voltages will be slightly less, but still very high.
This is because the simulator doesn't take into account all of the chokes properties - just the important ones.
To measure the peaks, you will need to make a 100:1 scope probe.
If you already have a 10:1 probe, you can modify it:
You will need 10 x 1Mohm resistors.*
Connect the resistors in series between the high voltage test-point and ground.
Connect your 10:1 probe across the resistor nearest to ground.
There will be a slight HF loss, and a 5% measurement error, but it shouldn't matter in this case.
Remember to multiply your scopes reading by 100 to get the result.
*(1.5M, or 2.2M will do, but the measurement error will be greater)
Hope this helps
To decrease the risk of hurting yourself, I would also add some bleeder resistors of say 300k across the caps. So that if you switch off the capacitors are slowly discharged. Don't feel safer though, because it is this feeling of safety that gets you into trouble.
Electrolytic capacitors especially tend to keep their charge quite long in my experience.
I have a 540v choke input psu..and the caps are 630volts rated and they have been fine so far..(3 months)
Electrolytic capacitors especially tend to keep their charge quite long in my experience.
I have a 540v choke input psu..and the caps are 630volts rated and they have been fine so far..(3 months)
Thanks....
Sadly I don't have a 100:1 probe, but I have a lot 1M resistors.
What's the reason for the 5% error?
I'm currently using 1% resistors.
Are there more things I shuld be aware of measuring these peaks, I don't want to detroy my scope, I have just bought an digital scope(Tektronix TDS2000) so I'm a bit anxios about "destroing" it,but I have an old one too(Tektronix 2235)
I'm aware of about high voltages and the thing about discharging capacitors.
In fact I went to a high school couple of years ago where it was an accident with charged capacitors.
It was a 2200pF capacitor holding a voltage of 10k-50kV, and this capcitor discharged through a person holding the capacitor in one hand and the other hand at a metalcabinet.
Sadly I don't have a 100:1 probe, but I have a lot 1M resistors.
What's the reason for the 5% error?
I'm currently using 1% resistors.
Are there more things I shuld be aware of measuring these peaks, I don't want to detroy my scope, I have just bought an digital scope(Tektronix TDS2000) so I'm a bit anxios about "destroing" it,but I have an old one too(Tektronix 2235)
I'm aware of about high voltages and the thing about discharging capacitors.
In fact I went to a high school couple of years ago where it was an accident with charged capacitors.
It was a 2200pF capacitor holding a voltage of 10k-50kV, and this capcitor discharged through a person holding the capacitor in one hand and the other hand at a metalcabinet.
Gunderz said:I use resistors(470k) to ensure half the supply voltage over each capacitor, but maybe I should have a bleeder too?
I'm currently using two Rifa PEH200 with a voltage rating of 350V.
And for my own security I uses gloves while working, and... one hand at my back while the other is working
No..I don't think you would need a bleeder in that case...
I just had an idea....
Bas Horneman said that the reason for the voltage peaks is because none of the diodes is conducting.
But I had this idea about SMPS, in SMPS we often find LC-filters at their ouptus. And we also have a transistor as a switch, open-closed.
But during the time the transistor is open the choke got a current path through a switchdiode, if the switchdiode had been removed we wouldn't have a current path for the choke and large voltage peaks had been devoloped by the choke.
But couldn't this be transfered to the LC filter used in my power supply, to provide a current path for the choke while none of the diodes is conducting I could use a switchdiode between the input at the choke and earth, for example UF4007.
Bas Horneman said that the reason for the voltage peaks is because none of the diodes is conducting.
But I had this idea about SMPS, in SMPS we often find LC-filters at their ouptus. And we also have a transistor as a switch, open-closed.
But during the time the transistor is open the choke got a current path through a switchdiode, if the switchdiode had been removed we wouldn't have a current path for the choke and large voltage peaks had been devoloped by the choke.
But couldn't this be transfered to the LC filter used in my power supply, to provide a current path for the choke while none of the diodes is conducting I could use a switchdiode between the input at the choke and earth, for example UF4007.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.