Generalized snubber
Mark Johnson came up with a generalized CRC snubber in his Ringnot PSU for transformer of about 200VA.
• Cx = 3.3 nF, metallized polyester film, stacked, rated 50VDC and up
• Rs = 22 ohms, 1/4 watt or higher
• Cs >= 680 nF, metallized polyester film, stacked, rated 50VDC or higher
See here.
John Swenson believes that an overdamped RC snubber is good enough and is happy with 330ohm + .022uF for transformers 10VA to 1000VA.
See here.
Hope that helps.
Denis
Here are a couple of thoughts for snubber values:
Mark Johnson came up with a generalized CRC snubber in his Ringnot PSU for transformer of about 200VA.
• Cx = 3.3 nF, metallized polyester film, stacked, rated 50VDC and up
• Rs = 22 ohms, 1/4 watt or higher
• Cs >= 680 nF, metallized polyester film, stacked, rated 50VDC or higher
See here.
John Swenson believes that an overdamped RC snubber is good enough and is happy with 330ohm + .022uF for transformers 10VA to 1000VA.
See here.
Hope that helps.
Denis
Yet it does. This is great.
i went through what i have and found 1micro farad panasonic 250v film cap and 3.3 nano farad film cap also added to my order both 22ohm and 30 ohm yageo 5% 1/2 watts. so i should be all set now. thanks to all.
i went through what i have and found 1micro farad panasonic 250v film cap and 3.3 nano farad film cap also added to my order both 22ohm and 30 ohm yageo 5% 1/2 watts. so i should be all set now. thanks to all.
Last edited:
You should also do a before and after listening test and see if your ears agree with the oscilloscope reading.
I'm building another power supply for an AJ, first was for an F5—anyone see any issues with putting the snubbers on the bottom of the board? So if I want to later I can hook it up to a different transformer? and resnub. Not that I have any plans...but changing them after the diodes are installed is a PITA.
As long as the PCB is mounted o standoffs long enough to clear the components, there is no issue at all.
That's what I thought. Thanks Jim!
And.... because there's always MORE questions...
In the build guide you used a 2.2K resistor for the bleeder? I didn't catch this until just now—I used a 22K in the first PSU because that was the part number in the BOM—and I didn't know enough at the time to ask the questions, but the range listed is 4.7K - 22K (you used 2.2K)(am I not understanding the values correctly?). I've read what I can find in this thread about bleeders... and I get that there's a range with different lengths of time related to how much power at the rails etc... My question is, am I bleeding faster or slower with 22K? slower? More resistance?
The bleeder is there to drain the PSU caps after it's powered down. Noting more complicated.
A higher resistance will be slower, a smaller resistance will be faster - if you have too low a resistance, say a short circuit, they will discharge very, very quickly. 🙂 You want the bleeder to be slow enough to not present any load to the PSU under normal operation, which in our case would be measured in hundreds of ohms, so if it's higher, no big deal. I have a box full of 2.2K and that value works nicely.
Once this resistor value is above, say, 1K, the actual value is not critical at all.
A higher resistance will be slower, a smaller resistance will be faster - if you have too low a resistance, say a short circuit, they will discharge very, very quickly. 🙂 You want the bleeder to be slow enough to not present any load to the PSU under normal operation, which in our case would be measured in hundreds of ohms, so if it's higher, no big deal. I have a box full of 2.2K and that value works nicely.
Once this resistor value is above, say, 1K, the actual value is not critical at all.
If you choose Cs >= (15 * Cx) for the input snubbers, then the same capacitor values can be used for both different transformers. Only the resistor value Rx changes, and so only the resistor Rx needs to be mounted on the underside of the PCB.
Most people here tend to choose Cx=0.01uF and Cs=0.15uF (a ratio of fifteen to one), which means these capacitors can be mounted topside and not changed when swapping transformers.
It would be wise to purchase an extra Cx and an extra Cs when ordering parts. Then throw the extra ones into your Quasimodo project box, to be used at some future date when you swap transformers. It feels good to know you've got the blood brothers of the parts on your PCB, there on your Quasimodo tester.
Most people here tend to choose Cx=0.01uF and Cs=0.15uF (a ratio of fifteen to one), which means these capacitors can be mounted topside and not changed when swapping transformers.
It would be wise to purchase an extra Cx and an extra Cs when ordering parts. Then throw the extra ones into your Quasimodo project box, to be used at some future date when you swap transformers. It feels good to know you've got the blood brothers of the parts on your PCB, there on your Quasimodo tester.
Mark, that is exactly what I did (purchase extras). And what my plan was with undermounting the resistor, and not the caps. Thanks for confirming!
i bought film wima 0.01 and 0.15 but the 0.15 is 63v is that enough? the 0.01 is 250v lol
for bleeder resistors, the guide says, 4.7k to 22k. i did buy some 20k but the wrong power rating. i just foiund a pair of dale 5 watts 800 ohm resistors. smaller resistor shorter discharge time. whats the problem? does it affect the normal performace of the power supply?
The voltage rating of Cs should be at least 110% of the transformer secondary voltage you'll be using.
(The extra 10% is just for safety as your mains supply can vary by that much. Mind you, the capacitor voltage rating usually has a safety margin too, but I still like to add a bit more.)
(The extra 10% is just for safety as your mains supply can vary by that much. Mind you, the capacitor voltage rating usually has a safety margin too, but I still like to add a bit more.)
800 ohms is going to be burning a decent amount of power while the amp is running, generating heat and putting more load on the transformer.
I = V/R. Put in your voltage and resistance and calculate the current. That should give you a good idea of whether or not it's too much.
I = V/R. Put in your voltage and resistance and calculate the current. That should give you a good idea of whether or not it's too much.
in a moment of brain fart ... it is 1/2 watt. I uinderstand the point about the bleeder consuming energy and turning it into heat during normal operation so i wont be doing that. also the dales 5 watts are too big i will have to fit them vertically so ill be ordering 20k or 22k for the bleeder i guess.
Totally fine. 24v into 20000 ohms is 0.0012A , which is dissipating 0.02W
The bleeders drain the PSU capacitors to empty after power down, that’s it. Use the 20K and they will just drain more slowly, which is completely inconsequential.
Stuff them in the PCB and carry on. 😀
The bleeders drain the PSU capacitors to empty after power down, that’s it. Use the 20K and they will just drain more slowly, which is completely inconsequential.
Stuff them in the PCB and carry on. 😀