This is far too long and does not comply with actual regulations. The idea is that after unplugging the power plug from the wall next you touch the connector pins. Thus a discharge timeconstant in the ballpark of 1sec (1megOhm||1uF) is required. I am talking about the X2-caps that are used for EMC filtering. Maybe you considered the bulk caps beyond the rectifiers - these are "untouchable" from outside so there are no special requirements for my knowledge.
About 7 minutes to discharge those to around 5 volts give or take using a 1k. Worst case dissipation would be over 3 watts. Not a great idea really.
Does the amp itself not pull the voltage down quickly once you power off?
Does the amp itself not pull the voltage down quickly once you power off?
it takes really long time. It's a 3W 10K resistor (two of them connected to each cap). It pulls down but very slowly.
Attachments
Last edited:
The best way to auto-discharge a bulk cap battery is paralleling some PTC-Resistor. This will dissipate between 0.5 ~ 1.0Watts permanently. After shutdown it cools down rapidly, its resistance drops and thus discharges quickly the caps.
So centre tapped. Four caps per rail but using a 10k. That would take nearly 40 minutes to reach 5 volts.
Interesting idea 👍 if you can get a suitable thermistor.
Another thought would be a DPDT relay (so just a standard relay) held open when the power is on and that closes at power off. You can get them I think with 48 volt coils but if not its only a single series resistor needed. Super simple. A 100 ohm would pull it down in as little as 30 seconds. The relay would have to fed from a separately derived rail of course, otherwise it would still take for ever for the relay to drop out. Still super simple.
Thank you Mooly, can't I use only resistor(no relays) ?
What value and wattage should I use to discharge caps in less time without any I'll effect to audio quality or unnecessary power dissipation when in operation, otherwise I can live with 40min discharge.
What value and wattage should I use to discharge caps in less time without any I'll effect to audio quality or unnecessary power dissipation when in operation, otherwise I can live with 40min discharge.
Why do you insist using this resistor? You may omit it as well, saving unnessecary power waste all the time. Only before doing some internal works, you discharge manually using one 100Ohm/4W wire wound resistor.
Not really. To discharge that amount of capacitance in a short time takes a low value resistor and that means it dissipates lots of power when running normally.
Thank you Mooly and Bucks bunny
I'll be adding back LED indicator, keep existing arrangement and keep 100Ohm resistor to discharge manually.
I'll be adding back LED indicator, keep existing arrangement and keep 100Ohm resistor to discharge manually.
I was about to say, use the LED as an indicator but LEDs do bust so don’t rely on it to be the only discharge route putting a bleed resistor in parallel is advisable if used ie (LED+Rled)||Rbleed
Also what are the safety caps as a circuit with? Mains AC primary side (ie part of a filter) or on the DC side (ie secondary side)?
If it’s the primary side they should be seeing AC, but the switch off may leave them charged at a non-zero voltage. They should be relatively small however careful the transformer when switching off as it may act like an inductor.
For the AC primary side I wouldn’t worry about it - I tend to leave the to an IEC inlet filter manufacturer to save time. The DC rectification is a different beast and I would add a bleed.
I like to use a MOV to limit any voltage spike when switching off.
Also what are the safety caps as a circuit with? Mains AC primary side (ie part of a filter) or on the DC side (ie secondary side)?
If it’s the primary side they should be seeing AC, but the switch off may leave them charged at a non-zero voltage. They should be relatively small however careful the transformer when switching off as it may act like an inductor.
For the AC primary side I wouldn’t worry about it - I tend to leave the to an IEC inlet filter manufacturer to save time. The DC rectification is a different beast and I would add a bleed.
I like to use a MOV to limit any voltage spike when switching off.
Last edited:
Thank you
I already have a bleeder resistor 10K 3W, that is what I meant by keeping existing arrangement, it's just that it takes time and I am ok with that after discussion.
Long time back, I learned this the hard way when I touched a laptop charger capacitor charged to 220V. Though I knew abt caps, it's just I wasn't paying attention when opening the charger. Now when I go near electricity I check for capacitors, thermistors, oil tra sformers 😀
I already have a bleeder resistor 10K 3W, that is what I meant by keeping existing arrangement, it's just that it takes time and I am ok with that after discussion.
Long time back, I learned this the hard way when I touched a laptop charger capacitor charged to 220V. Though I knew abt caps, it's just I wasn't paying attention when opening the charger. Now when I go near electricity I check for capacitors, thermistors, oil tra sformers 😀
Last edited:
I'd anticipate the OP was referring to a secondary side ss amp filter given the LED and series 15k ohm comment in post #9.
Some safety regs called for <60Vdc within 1 minute, and whilst 60Vdc still imposes a safety risk I still use that as a guideline for valve related B+ supplies. I also don't rely on the valves to drain B+ caps, so a bleed is always added, as some maintenance/repair activities can have the valves pulled out, and safety is about mitigation whenever practical. Also with valve amps I would use a bleed to also provide a 100:1 divider to allow external measurement/monitoring (along with output stage bias voltage levels) as a convenient check, and to avoid having to up-end and probe for that information.
The risk for ss amps with large bulk capacitance is not per se the voltage as it reduces below say some 'safe' level of 12-24Vdc, but rather the risk related to the residual energy within an arc should one start probing around.
For ss amps, it is easy to obtain 72Vdc rated PTC that could support connection across a main rail (via a series current limiting resistor). Unfortunately larger ss amps have higher than 72Vdc rails so some extra effort is needed there. However like a LED, a PTC part is not as robust as a power resistor operating within its limits, so there would seem to be a reasonable argument for redundancy even if one was well trained to always measure before starting any repair or maintenance.
Some safety regs called for <60Vdc within 1 minute, and whilst 60Vdc still imposes a safety risk I still use that as a guideline for valve related B+ supplies. I also don't rely on the valves to drain B+ caps, so a bleed is always added, as some maintenance/repair activities can have the valves pulled out, and safety is about mitigation whenever practical. Also with valve amps I would use a bleed to also provide a 100:1 divider to allow external measurement/monitoring (along with output stage bias voltage levels) as a convenient check, and to avoid having to up-end and probe for that information.
The risk for ss amps with large bulk capacitance is not per se the voltage as it reduces below say some 'safe' level of 12-24Vdc, but rather the risk related to the residual energy within an arc should one start probing around.
For ss amps, it is easy to obtain 72Vdc rated PTC that could support connection across a main rail (via a series current limiting resistor). Unfortunately larger ss amps have higher than 72Vdc rails so some extra effort is needed there. However like a LED, a PTC part is not as robust as a power resistor operating within its limits, so there would seem to be a reasonable argument for redundancy even if one was well trained to always measure before starting any repair or maintenance.