Usual Shottky diodes are well known with their low forward drop (and low max reverse voltage) and Cree SIC (Silicone Carbide) ones are known as high reverse voltage and high forward drop ones.
The presence of these 50Hz harmonics indicates the two half waves are not equally rectified.
Is the secondary winding center tap ?
When the two legs of a center tap are perfectly equal, there is only harmonics of 100Hz, if not, the two windings or the two diodes are not identical. The wiring the tracks a poor solder joint could make an imbalance too.
Is the secondary winding center tap ?
When the two legs of a center tap are perfectly equal, there is only harmonics of 100Hz, if not, the two windings or the two diodes are not identical. The wiring the tracks a poor solder joint could make an imbalance too.
They are not center tap but 2 x 18 VAC. It is Toroidy transformers and they are usually exact at their output voltages and if I measure this also indicates this.
On one mono block the harmonics are much worse than the other. Just below -100 dB on one and below -110 dB on the other. So one diode bridge could be a little off compared to the other? .....it is these standard square 35A bridges with a hole in the middle for mounting. 1.1V forward drop.
I did not see this when I used the high voltage Cree SIC diodes. Transformers are the same.
Think I will try a Vishay Schottky type in TO-247 with 3-legs.....double diode type where the two diodes are run in parallel so 4 devices are used in a bridge (8 diodes).
If I buy a number I may measure them up front and try to find identical ones.
I found a 40V type (0.49V forward drop)......but seems better to go for a 45V type for safety margin?
On one mono block the harmonics are much worse than the other. Just below -100 dB on one and below -110 dB on the other. So one diode bridge could be a little off compared to the other? .....it is these standard square 35A bridges with a hole in the middle for mounting. 1.1V forward drop.
I did not see this when I used the high voltage Cree SIC diodes. Transformers are the same.
Think I will try a Vishay Schottky type in TO-247 with 3-legs.....double diode type where the two diodes are run in parallel so 4 devices are used in a bridge (8 diodes).
If I buy a number I may measure them up front and try to find identical ones.
I found a 40V type (0.49V forward drop)......but seems better to go for a 45V type for safety margin?
I may go with this type:
https://docs.rs-online.com/f271/0900766b80a38b04.pdf
But I can understand that the SiC types has even lower switching noise but higher forward drop that change "a lot" with temperature. That was my experience. The above may give me the best of both worlds. Relative low noise compared to standard Si-diodes and stable low voltage forward drop. Then I can match the diodes for low uneven harmonics.
https://docs.rs-online.com/f271/0900766b80a38b04.pdf
But I can understand that the SiC types has even lower switching noise but higher forward drop that change "a lot" with temperature. That was my experience. The above may give me the best of both worlds. Relative low noise compared to standard Si-diodes and stable low voltage forward drop. Then I can match the diodes for low uneven harmonics.
If you use one bridge for 2x18 VAC it mean 36VAC and so you want about 100V rated diodes or higher. If you use two bridges - one for each 18VAC winding - then 45-60 VDC rated diodes have to be ok.
Last edited:
I understand your two 18V secondary winding are floating. The two windings are not connected to implement a center tap.
This is a way I like to implement a -V 0 +V regulated PSU using two identical positive regulators.
Your trouble is likely from non identical diodes.
Is it really bad, significant compared to the 100Hz harmonics ?
May be, what you see is within diodes tolerances on Vf.
Edit: Looking at first posts, I see it is a very little effect, I would not care to try to understand exactly what is going on.
This is a way I like to implement a -V 0 +V regulated PSU using two identical positive regulators.
Your trouble is likely from non identical diodes.
Is it really bad, significant compared to the 100Hz harmonics ?
May be, what you see is within diodes tolerances on Vf.
Edit: Looking at first posts, I see it is a very little effect, I would not care to try to understand exactly what is going on.
Last edited:
If diodes in bridge is a bit off (not 100% identical) this will create a bit of uneven harmonics (150/250 Hz etc.). Will this also create a bit of DC in toroid transformer that theoretical can create a bit of mechanical hum in the transformer?
Think I will try using the low voltage drop Schottky diodes and then try to match them as close as possible and make 4 quads. Now with lower voltage drop the difference in dV may be smaller.
I know my problem is not huge and I have very little noise as is......but interesting to try it out.
Think I will try using the low voltage drop Schottky diodes and then try to match them as close as possible and make 4 quads. Now with lower voltage drop the difference in dV may be smaller.
I know my problem is not huge and I have very little noise as is......but interesting to try it out.
Yes, for Vf and it will probably be best at the operation point but that will be difficult as it is pulsing with high current spikes.
A diode tester with very low current will probably not give a useful reading. It could be a starting point to find the best matches. Else I have to mount the diodes on a heat sink and then maybe use a lab supply with current limiter and set it to maybe 2 amps and measure Vf? .....and then maybe 0.5 A and 4 A? .....then I have to wait until temp has stabilized.
A diode tester with very low current will probably not give a useful reading. It could be a starting point to find the best matches. Else I have to mount the diodes on a heat sink and then maybe use a lab supply with current limiter and set it to maybe 2 amps and measure Vf? .....and then maybe 0.5 A and 4 A? .....then I have to wait until temp has stabilized.
I think you should use a bridge test jig:
2 diodes 2 resistors
High current feed and measurement of the voltage between the two legs to get Vl1 - Vl2.
You could use one of the two diodes position for a reference diode and use the other position to test all the other diodes.
Diode temperature is likely a trouble maker as well as resistor temperature coefficient.
2 diodes 2 resistors
High current feed and measurement of the voltage between the two legs to get Vl1 - Vl2.
You could use one of the two diodes position for a reference diode and use the other position to test all the other diodes.
Diode temperature is likely a trouble maker as well as resistor temperature coefficient.
What you could do is using the final system to figure out the diodes that better match.
On a 4 diodes bridge, let's name them 1 2 3 4.
Two of them conduct during each half wave, alternatively 1&3, then 2&4.
Matching that matters is: Vf1+Vf3 against Vf2+ Vf4.
Measure and/or listen what you get, then swap diodes 3 and 4, to compare these two alternative configurations.
With 4 diodes, there are only two ways that matter, to make a bridge about matching diodes.
With a lot of diodes to pick from, there are many more configurations and procedures not yet invented to find the optimum matching.
On a 4 diodes bridge, let's name them 1 2 3 4.
Two of them conduct during each half wave, alternatively 1&3, then 2&4.
Matching that matters is: Vf1+Vf3 against Vf2+ Vf4.
Measure and/or listen what you get, then swap diodes 3 and 4, to compare these two alternative configurations.
With 4 diodes, there are only two ways that matter, to make a bridge about matching diodes.
With a lot of diodes to pick from, there are many more configurations and procedures not yet invented to find the optimum matching.
Last edited:
I will get a tube with 30 T0-247 double diodes and I need 16 TO-247 for my two mono blocks. So I have a few to select from. It helps that I can make combinations of two to add up to same sum. I have also ordered heatsinks for each TO-247 so I will mount 16 and try to measure them as good ad possible and then make the best double-pairs. If not possible to make nice double-pairs I can select from the 14 spare TO-247 devices.
The PCB I use has room for a snubber (R + C) and a C. Normally 10R + 100nF and a 100nF is better than nothing I guess?
The C rated at 63 VAC should be sufficient? ....do we expect some voltage spikes so a higher voltage rated capacitor is needed across the AC?
The PCB I use has room for a snubber (R + C) and a C. Normally 10R + 100nF and a 100nF is better than nothing I guess?
The C rated at 63 VAC should be sufficient? ....do we expect some voltage spikes so a higher voltage rated capacitor is needed across the AC?
4 quads.
Then you say you will use 16 pairs.
I do not quite understand.
4 bridges need 16 diodes.
A bridge could be made with a pair of common cathode diode plus a pair of common anode diodes. ( A bridge cannot be made with two pairs of the same kind ).
Does that mean you make a bridge with four pairs where you parallel the diodes in each pair ?
Why do you use TO247 case diodes for 18 VAC secondary voltage? Is there such a large current?
...
Oh, I see now, you have 1.5 A load. It is just fine for TO220 case diodes (for example, 10A rated or more, if you want a large margin). They need a heatsink of cause.
For example, VS-10TQ045PBF or cheap MBR1645 (MBR1045).
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.