• Please review our updated Terms and Rules here

Time to wait for power supply caps to self-discharge?

...Right back at you Hugo... Right back at you...
I'm glad we agree. I'm not opposed to discharging capacitors with current limiting resistors, where it is required. I'm just opposed to direct shorts.
By far and away the best tool to discharge a CRT anode connection, if a person wants to rush under the anode cap immediately after the set is switched off, is the EHT probe as it has a suitably proportioned resistor to do it.
 
Ouch !

I have electrocuted myself a few times too but always avoided the big one.

In a HV flashover, its often breathing in the copper vapour cloud that does for the unfortunate person. We used to rack HV breakers in without any arc flash PPE, but since that youtube video of the flashover on racking an ACB into service, we now wear full Nomex suits with face protection.

Still not sure it would save us though.
 
Hugo has a lot of good info there! However, TV's can absolutely hurt you... I can attest to working on a TV in mid-1970 and my arm HAIR came close to the flyback diode lead (of course, and the energy stored behind it) and the sudden discharge literally blew me out of the big wooden cabinet, ripping a deep gash down my arm as it scraped on a sheet of metal inside the TV cabinet. The TV owner did NOT unplug the TV 2 days earlier as I had asked but only just before I arrived to diagnose [and later fix it]. My bad, I never experienced a live supply before while working on other folks TVs and so did not test (VOM) the TV before reaching in to it. A lesson learned the (very) hard way!

So, I shall share a few thoughts:

I strongly disagree with anyone who would direct-short a cap (AC or DC!) or lay a PCB on a "damp" cloth... I'm sorry, what are you guys thinking?

Last, first: A damp cloth? Anyone worth a grain of salt in this tech world of ours KNOWS moisture and electricity are not good play-buddies!! Especially in this day an age of electronics, laying a PCB (stuffed of course!) on a damp ANYTHING after removal from it's mounting (and especially if recently powered up) can cause all manner of shorts and can actually harm transistors/FETs and even kill pins on IC's, via over-voltages and especially if back-biased, but certainly from instantaneous discharges of energy. Consider a lightning strike as a mental picture.

NOW, direct shorting can shorten the life span of a cap and can cause all manner of damage (most often INTERNALLY) to the capacitor (but often times some damage can be observed EXTERNALLY), especially if the cap was large valued, and as Hugo pointed out: direct shorting can harm YOU - and BTW: anything "sensitive" in close proximity to the discharge path. I've been an electrical/electronic design engineer for many decades and OMG the "ideas" novice or even some well-meaning people come up with on all sorts of electronic related topics. It is amazing.

PLEASE understand not all caps self-discharge quickly, and many which do self-discharge are very SLOW dischargers - like edlc/super/ultra caps, and many aluminum electrolytics, and almost ALWAYS: film and poly caps. Use a resistor to bleed a cap down (alligator leads or clips work just fine to hold the resistor in place); measure the voltage before removing the resistor; measure it AFTER removing the resistor (use appropriate VOM/DVM/DMM scales). Yes you may find many caps have a "memory" as Hugo spoke to, so you may see some residual voltage... just get into the practice of handling a "discharged" cap as if it were FULLY CHARGED. BTW: a small value small wattage resistor can detonate on you due to the joule-related discharge power... so the wise will do some due diligence and size the value and especially the wattage of a resistor, before using it...

Also come to appreciate: (CV^2)/2... equals JOULES... ergo: stored ENERGY! Consider this associated relationship: Volts x Amps (which = Watts), x time (the amount of time the watts are applied) = J (Joules)... and as relates to 1 Joule: reflects the "work" required to produce: 1 watt of power for 1 second, otherwise known as: one watt second. Thus: 1 amp passing through a 1 ohm resistor for 1 second, equates to 1J. For clarification: 1W (Watt) = 1J (Joule)!

So, capacitor-value times applied-voltage (use Vpeak if AC) will get you a number... ex: 47000uF at 50VDC equates to 58.75J; 3300uF at 220VAC equates to 159.72J [remember to convert to Peak... VRMS*(sqrt 2) = 311.126V]. Be aware that current as low as 1 milliamp (0.001 amp) can cause fibrillation or even kill under the right circumstances, so when dealing with AMPs, "careful caution" is the thought of the day. In an appropriate and appropriately ISOLATED hospital set up, anywhere from roughly 100J can completely stop a VERY healthy persons heart in a nano-beat. Outside in the "other" real world, where many other factors can come into play, cardiac arrest (heart stoppage) can occur with significantly less energy. Understand a roughly 10mA DC current running up your arm can cause severe muscle contractions... perhaps you can now appreciate why when people grab a live wire (120VAC of many tens of amps) most folks can not let go... due to muscle freeze. They generally die.

More math: What is the remaining voltage on a given capacitor after time (t)? ref: Vcapacitor = Vapplied*(e^−(t/RC)) See the example below:

NOTE: "t" refers to time and is called: TAU.
NOTE: referring to the SI Unit of the Coulomb, a single proton has a charge of about 1.6 x 10-19 Coulomb. (e=1.60217662×10^−19C) This is called an elementary charge, abbreviated by the letter e. On your technical calculator, the key marked: e to the x (ie: e^x) is what you will use unless you enjoy longhand.

Using the 3300uF (<-- internal resistance ignored) at 311V, as an example, we then apply a 1000 ohm resistor to bleed the stored voltage toward zero. In 1 second it has only bled off around 81 volts! It will take around 15 seconds to bleed down to just 3.3V... H O W E V E R : recall the J value is approx: 159J!... 159 Watts!! Using a 1/10th watt resistor will vaporize as soon as the leads are connected. You NEED to size the bleed resistor wattage to handle (in this case ONLY) slightly longer than 1 second of discharge current (as heat). A half watt package will suffice, but if you use (say) a 10 watt package, you'll be safe for a broad range of voltages and stored energy values. (Again, this is only ONE example and is not intended to be an across-the-board solution! Different stored voltages require serious and generally different resistor value and power option considerations!)

That said, imagine the bleed resistor was only 1 ohm in value... in 15 milli-seconds (0.015S) the residual voltage would be about 3.3 volts! THATs FAST (like lightning) thus you'd want to use at least a 5W resistor so your fingers don't get heated up if you happen to be holding the resistor body! Get this: the power dissipated in the resistor (if continuously applied) would be 159 watts... so in THAT case you'd want to use at least a 600W resistor! Quite a difference - YES?! Most would choose a 750W device to try to keep thermal damage/wear/impact to a minimum OR use a really good chunk of aluminum as a heatsink. A general minimum rule of thumb for selecting a decent power rating is 4 times the expected max dissipated wattage. In some situations up to 10 times might be used!

So... this is clearly not an electronics class forum, so ultimately, your job *if you want to do things safely* is to do the due diligence, do the research, do the math, use the right set-up, tools and equipment, and please be careful. As concerns electricity whether it be AC or DC: if you are ever unsure about what you are considering doing: DON'T - let someone who does, do the deed.
 
Generally, the at risk current for electrocution is thought to be >30mA. This is why RCD's (ELCB's) are designed to trip at that current. In medical facilities, operating rooms etc the RCD's normally trip at 10mA.

The at risk voltage is thought to be anything over about 70V because it has to be at least that high to establish a current of >30mA via human contact, meaning the resistance of the pathway for a 70V high risk shock would need to be in the 2000 to 3000 Ohms vicinity. In a sense, it is another way of saying that you require at least a minimum of 2 watts of continuous energy flow and at least 30mA for an electrocution event. Though the risk is altered by the timing of the voltage application compared to the cardiac cycle and whether it is AC or DC.

Generally, especially for small TV's VDU's the EHT supply, derived from the LOPT during the resonant phase of the operating cycle (flyback time) is designed to supply currents in the 10's to 100 of uA region, to support the CRT's beam current. Overload effectively shorts out the transformer (LOPT) and damps the resonance. In essence these supplies have a relatively high internal resistance. Though the ones in >26" color TVs were getting more risky in this respect and could reach a dangerous threshold. By this time in history though of the color TV, the EHT cap insulation and wiring was pretty good and hard to make accidental contact. Also as noted the energy storage in the CRT bulb is relatively low for small monochrome VDU's. This is why, over many decades of TV servicing, including many TV's in the early days that didn't even bother to insulate the EHT anode caps there were few if any electrocution events associated with contact with the EHT terminal. In fact I have never heard of one from the EHT power source in the last 50 years. If there was an electrocution from the TV set , those reported would have likely been from the line supply side or the B+ high voltage supplies.

This is one of the advantages of flyback supplies, in general, the intrinsic current limiting on overload. So much so you can buy things on ebay like Jacob's ladder kits etc with exposed wires, based on flyback supplies and the sellers are not too worried about electrocution events. Yet the terminal voltage appears "scary" being in the high kV region. Here is one using a TV LOPT and a electronic driver circuit, though they might be pushing the envelope, it looks like a color TV LOPT, it would be worth knowing what the short circuit current was or the current into a 2500 Ohm load:


Here is another kit using a TV LOPT:



However, it is a completely different cup of tea with the other voltage sources inside a TV or VDU, early Tube TV's they were typically in the 400V region, the power supplies could source over 50 to 100mA in many cases, potentially lethal.
Even more modern color VDU's the B+ voltages can be over 100V B+ supplies and source high currents and are potentially lethal, to say nothing of the high voltage and the capacity storage on the line side of the SMPS, also potentially lethal, again because it is a high voltage supply, >70V with a very low internal resistance.

So it is a good question the OP had about how long to wait. Most sensible designs put bleeder resistors across the high voltage filter caps (these can sometimes go open circuit). Due to the electrolytic cap being intrinsically leaky though, most will be well discharged by the following day, even without bleeder resistors.
 
Last edited:
Back
Top