Resistive Bleeder Question

[Randy Knutson]

Hi,
I'm designing a +/- 50 Volt power supply that has about 20,000 uF of bulk capacitance on the output. Does anyone have any suggested rules of thumb as to how many time constants this amount of charge should discharge through a bleeder resistors in?

Thanks,
Randy Knutson
Mankato, MN

[benny]

T = RC gives you the time constant, so you say have 0.02F of capacitance, and say you use a 50k bleeder then your time constant is 1000 secs... although I guess you already knew that.

So after 1000 secs in this example, you're down to 36.8% of your initial charge, after the next 1000 secs you're down to 13.5%, then the next you're down to 5%, then 2%, then 1%.

After 5 time constants you're only loosing a tiny % of you're initial charge, and it never quite reaches 0%.

Does this help, or is this not what you wanted?

[Sch3mat1c]

Under a minute is a good figure. Say, 30-60 seconds = 5RC (99% discharged). 330 ohms will have to be rated a good 10 or 15W to handle the continuous power; if you connect it automatically on shutdown, 1 or 2W will do (it has to absorb 25J, mostly in the first few seconds).

Tim

[leadbelly]

Originally posted by Sch3mat1c
if you connect it automatically on shutdown, 1 or 2W will do

I've never seen this technique. So all you need is a double throw switch for power that will tie the rail to ground through the resistor in the "off" position?

[benny]

Originally posted by leadbelly

I've never seen this technique. So all you need is a double throw switch for power that will tie the rail to ground through the resistor in the "off" position?

If you want to discharge quickly as schematic sugested, then to have the bleeder permanantly conected into the circuit is not a good idea... 330 ohms with 50V across it draws a little over 150mA... not only will this create extra work for the filter caps, it will also waste 15W (7.5W per side) of power... considering nasty effects of excess current draw on transformers such as sag, it's best to conserve as much power as you can in audio designs.

A DPDT switch will work, A 3PDT might be even better... yes, all you want to do is tie the resistor to ground on power off.

By the way, I suggest you use a little more than 1 or 2W. Considering resistors are dead cheap, you're best to use at least 5W. Those few extra cents just gives you that extra peace of mind that you have even less chances of something going wrong... I always believe that you should use the maximum rating that you can afford... within reason of course, I mean, there's no need to go using 50W here is there!

Cheers

[Sch3mat1c]

Originally posted by benny
not only will this create extra work for the filter caps, it will also waste 15W (7.5W per side) of power... considering nasty effects of excess current draw on transformers such as sag, it's best to conserve as much power as you can in audio designs.

No.

For one thing, if it were best to conserve power, we wouldn't be in this conversation in the first place because everyone would be building class D amps!

The general rule of thumb is 10% of rated maximum current, so unless you're at only 1.5A max., 330 ohms is quite conservative. AT 20mF capacitance, 0.5 to 1A bleed would be more realistic.

A DPDT switch will work, A 3PDT might be even better... yes, all you want to do is tie the resistor to ground on power off.

Can also use a relay, especially if you have one in use already for say, speaker protection.

... I always believe that you should use the maximum rating that you can afford...

Some of the resistofreaks here should consider a group buy from these guys:

Hum, can't find the link. Well, I was thinking of these spiral-wound stainless or nichrome resistors - usually used for breaking on diesel-electric locomotives.

Tim

[benny]

Tim, you take too many things the wrong way (not just in this post, other posts I have observed your responses in).

Just for your information, buying the maximum reasonable rating you can afford does not mean being sucked in by snake oil salesmen.

If there is only a few cents difference between the price of 2W resistors and 5W resistors (which there is, atleast where I buy my suplies from), then what is the harm in this? It does not mean buying some titanium coated, ferrofluid cooled resistor that's going to cost you an arm and a leg because it theoretically contains some special properties that really don't exist for the component to posess anyway.

Also, just because not everyone is building class D amps, does that mean that it is best to waste power? Does that mean that we're more interested with vacuum tube amps because we like to waste as much power as possible? I'm sure we'd all love it if tube amps were a little more efficient... when you design a power amp isn't one of your design goals to get the maximum efficiency from the tube you're using? (of course, there are others we need to consider such as THD and what not, but efficiency is one).

I don't think we all necessarily look for the most efficient design (operating class, tube/ss, etc), but we do seek to find a certain amount of efficiency within the design we're working with.

So of course it's advisable to conserve as much power within the confines of the design. Even if you can waste it, why waste it when you can quite easily avoid it? Maybe a little silly tacked on the end here, but a penny saved is a penny earned.

Have a nice day,
Ben