Adding a Bleeder Resistor

[a90s2cs]

I couldn't find definitive answers to my questions after a few hours of reading search results for "Bleeder Resistor".

1a) If I want to add a bleeder resistor is there a formula or a rule of thumb for the value of resistor to use? 1b) Is it related to B+ Voltage?  1c) Is it more dependant on what types of caps you're using?

2a) Does the bleeder resistor value vary depending on where you put it?  2b) In general where is a good place to put it? 

3) Is the resistor value only relevant in terms of how long it will take to drain the caps?  For example If a 220k drains the caps in one minute would a 2M resistor drain them in ten minutes?

4) Specifically, I'm looking at adding one to a JCM800 2204 build.  The simplest, most effective place to put it seems to be on the first cap can.  The two positive lugs on this can are jumpered, so a 220k-470k resistor from either of the positive lugs to the ground lug seems like it would just be too easy.

5) I'm guessing what I want to do in question 4 would lower B+ voltage but not enough to warrant making other changes to compensate.  is this a safe assumption?

[2deaf]

Do what you said in the second #3) with a 1M 1/2W resistor.

[a90s2cs]

I fixed the numbering...  So the Idea in #4 with 1M resistor will do the trick?

[92Volts]

You need to take into account the resistor's power dissipation and voltage ratings.

Power = Voltage * Current
because Current = Voltage / Resistance this can be simplified to:
Power = Voltage^2 / Resistance

1M ohm resistors are big enough they don't burn much power even at high-ish voltage but you'll need at least a 1W resistor and possibly larger to get a voltage rating greater than 300v... if you need to buy, for example, a 3W resistor to handle the voltage this high power rating would enable you to go smaller (such as 220k) which will drain the caps quicker.

If you want you can research and use the RC Time Constant to figure out how long it takes to drain the caps to a safe level, and choose a resistor for those reasons. The speed it drains is influenced by the size of the caps. I wouldn't worry about it though. The idea is to drain the caps to a safe level in (much) less than a day instead of a few weeks with no bleeder. Whether that takes a minute vs. an hour arguably isn't as important as just having something.

[2deaf]

Mouser's $0.15 .5W 1M is rated at 3.5KV.

[Ledzepp007]

I use a 220k/3W on the second filter cap. Don’t know if that’s the best way, but it seems to work.

[2deaf]

1a) If I want to add a bleeder resistor is there a formula or a rule of thumb for the value of resistor to use? 1b) Is it related to B+ Voltage?  1c) Is it more dependant on what types of caps you're using?

2a) Does the bleeder resistor value vary depending on where you put it?  2b) In general where is a good place to put it? 

3) Is the resistor value only relevant in terms of how long it will take to drain the caps?  For example If a 220k drains the caps in one minute would a 2M resistor drain them in ten minutes?

4) Specifically, I'm looking at adding one to a JCM800 2204 build.  The simplest, most effective place to put it seems to be on the first cap can.  The two positive lugs on this can are jumpered, so a 220k-470k resistor from either of the positive lugs to the ground lug seems like it would just be too easy.

5) I'm guessing what I want to do in question 4 would lower B+ voltage but not enough to warrant making other changes to compensate.  is this a safe assumption?

1a,b,c)  The current going through the typical bleeder resistor is very small compared to the current going through the tubes.  The tubes in a 2204 use around 86mA at idle and a 1M bleeder resistor uses 0.5mA while a 220K bleeder uses 2.1mA.  Under these circumstances, about the only considerations are how fast you want the discharge and what resistors you have on hand.  However, see 2b) below.  The 1M/0.5W bleeder resistor will discharge the capacitors to a safe level by the time it takes to get the chassis out of the cabinet.  A 220K bleeder might discharge the capacitors to a safe level by the time it takes to flick the standby switch, remove a tube, and to insert a knitting needle into the socket.

2a,b)  The best place for a bleeder resistor on a 2204 is at the first capacitor after the rectifier.  The bleeder resistors don't draw much current, but they still cause a voltage drop when they are in series with your 10K dropping resistors.  If you put a bleeder at the furthest node from the rectifier, there will be three 10K dropping resistors in series with it.  A 1M bleeder will drop the voltage at that point around 10V more than if there were no bleeder there.  A 220K bleeder will drop the voltage around 42V at that point.

3)  Yes.  The math isn't like that, but the smaller the resistance of the bleeder the faster the discharge.  Wattage rating, cost, heat, and power supply effects become concerns if the bleeder resistance becomes too small.

4)  It does seem to be just too easy and that makes me suspicious, too.  Still, a 220K/2W, 470K/1W or a 1M/0.5W will almost undoubtedly work fine.  I have all of these resistors on hand and I still use 1M/0.5W in this situation.

5)  I think this is covered above.

Other)  It has been stated that a large resistance bleeder is better than no bleeder at all because electrolytic capacitors have a charge memory and will develop a charge over time even when they are disconnected.  They won't do that with a bleeder resistor across them, though.   

[Joel]

I build in a heater elevation circuit into all of my builds (for at least 30V elevation) off the first or second power cap.  Bear with me... Whats a heater elevation circuit got to do with a bleeder circuit?  Because they also happily work as a bleeder circuit!

This circuit is simply a voltage divider made up of a 220k and, say, a 22k resistor (tailored to suit) both 3W.  I also chuck a 10uF or so cap across the lower resistor as a bit of filtering for the elevation voltage to the heaters.  Two birds.  One stone.

[jukeboxman]

http://www.angelfire.com/electronic/funwithtubes/3_Simple_Power.html

[dude]

Been using a 220K, 3 watt, bleeding resistor for years, never a problem and the voltage it's gone in a few minutes every time.
But always check with meter, you might have a volt or two hanging on for five or ten minutes but never got zapped from 1 volt.
Put the resistor on the first filter cap to ground.


al

[92Volts]

Been using a 220K, 3 watt, bleeding resistor for years, never a problem and the voltage it's gone in a few minutes every time.
But always check with meter, you might have a volt or two hanging on for five or ten minutes but never got zapped from 1 volt.
Put the resistor on the first filter cap to ground.


al

Once it's down to 20v or so I sometimes short to the chassis with a screwdriver to bring it to 0v. "Technically" it never reaches 0 on its own-- as voltage decreases, so does the rate of discharge. You can either short the supply to reach 0v quickly or not worry about it, as a couple volts usually isn't a problem (definitely not a shock hazard).

[pompeiisneaks]

Also electrolytics have that recharge and it will often climb back up to a few volts like 2-5VDC but that's not enough to do more than tickle :D

~Phil

[a90s2cs]

Thanks for all the information... very helpful. 

So far I've been aligator clipping a 10k 3w from ground to where the rectifier diode and first cap can meet, that drains it down pretty quick without any arcing.  I leave it there for a few seconds then aligator clip that point to the chassis and leave it there like a ground strap while I'm tinkering.   With this build I've had a lot of kinks to work out and I'm starting to get paranoid that I'll forget one time and kill myself.  At this point I'm looking to add some safety redundancy while still ritualistically draining the caps manually and double checking with the meter before touching stuff. I also have a stupid cat that apparently likes the smell of charged capacitors and hot soldering irons.  I'd hate to leave it on the bench without remembering to drain the caps and have that varmint get zapped.

The other consideration I have is that I've got 125v coming in from the wall and the PT is putting out 670v instead of 650v so I wouldn't mind seeing the B+ drop a little, trying a 470k 1w or even a 220k 3w might help a little with the high voltages I've been getting across the board.   

That brings up different but related question.  If I'm seeing high voltages (+10-15%) across the board because of the "out of spec" AC input stated above would putting resistors on all 3 cap cans be an acceptable means of compensating while simultaneously providing safety redundancy?  Seems like a win win to me...  starting with something like a 1M on the first can and then 2.2M on each lug of the 2 remaining cans?

[dude]

If the voltage is a little high, use at least a 5 watt, cement resistor, say 5k. 470k and 220k are way too high. You could also use zeners in a string to ground 0n the PT’s center tap, say 5 watt, 10 volts, in series off CT to ground. Three would reduce the B+ 30 volts

[sluckey]

High ohm bleeder resistors ain't gonna reduce your B+. Maybe consider this. See page 2...

     http://sluckeyamps.com/misc/Amp_Scrapbook.pdf

[2deaf]

I'm seeing JCM800/50W PT replacements with 690 to 720V secondaries.  You guys sure 670V is a problem?

[a90s2cs]

I don’t know if it hinders the amp’s functionality but it does make it difficult for a newb like me to troubleshoot when voltage readings are off by 20-50v across the board. At first it had me thinking that I used the wrong diodes in the rectifier then wrong/bad resistors.  Spent days quadruple checking everything only to settle on the PT ‘running hot’.  It’s a bit frustrating.

[2deaf]

Where did you get the target voltages?

[a90s2cs]

Where did you get the target voltages?

[2deaf]

Mark doesn't give us the measured AC output of the PT secondary.  The 320-0-320 is probably what the transformer is supposed to be.  He is showing a B+ of 468V, so the minimum that the secondary could be would be 662V_rms (468/1.414 = 331 x 2 = 662 or 331-0-331) presuming a sine wave.  The secondary is undoubtedly a little higher than that because you just don't tend to get the theoretical values in the real world.

It seems like you should be getting a B+ that is fairly close but lower than Mark's example.  Perhaps you could post your B+ voltage as well as the DC voltage on each node.

[a90s2cs]

I'll list the voltage from the schematic and the actual reading.  Schematic / Actual

right to left across the top of the schematic

468 / 512
467 / 510
390 / 420
348 / 389
334 / 378
 
Pin 4 on both V4 & V5 = 511

So they're not crazy high, and they are pretty close to "tolerance" (if that is +/- 10%)....  And they're relatively consistant.

[2deaf]

Are all three dropping resistors 10K?
What kind of power tubes?
Did you build the bias circuit like Mark shows?
Do you know what the idle current is through the power tubes?

[a90s2cs]

Are all three dropping resistors 10K?
What kind of power tubes?
Did you build the bias circuit like Mark shows?
Do you know what the idle current is through the power tubes?

They are all 10k

A set of TAD el34 and a set of JJ’s. Both yielded the same #s.

Sort of...  I used a 50k trim pot and a 56k resistor.

I do not know the idle current, I haven’t gotten that far yet.

[PRR]

What is your Wall Voltage today?  (Do not get killed!!)

[2deaf]

What is the bias voltage you are using (pin 5)?
With the amp on standby, what is the range for the bias?

[a90s2cs]

What is your Wall Voltage today?  (Do not get killed!!)

It’s fluctuating from 125 to 128.  I live 1/4 mile from a transforming station and 1/2 mile from a hospital, 124 is the lowest I’ve ever seen it, but the upside is the power never goes out for more than 1/2 an hour.

[a90s2cs]

What is the bias voltage you are using (pin 5)?
With the amp on standby, what is the range for the bias?

It’s been set at -50v but it has a range from -35v to -70

[2deaf]

It’s been set at -50v but it has a range from -35v to -70

That seems like it would idle the tubes kind of cold.  If the tubes are idled a little hotter, it will pull the B+ voltage down.  If you don't have one-Ohm resistors between the power tube cathodes and ground, you might want to consider installing them so that you can easily set the bias.  Once the bias is properly set and the B+ at that point is known, we can adjust the resistors between the nodes to get whatever voltages you want.

After looking at your node voltages, it appears as though V3 is drawing way too much current and V2 is not drawing enough current.  If this is correct, how come?

[PRR]

127V today and 117V typical when the amp was new suggests everything "should" be 8%-9% high.

Your voltages are 9.5% high of the "documented" values, which seems reasonable.

It's not precision work. Generally tubes can be -20% to +20% on voltage and be happy. True, when the original amp design PUSHed the voltages, we like to be not-too-high. But I would not panic. I did not read the whole thread but it sounds like you have some specific problem(s) other than voltage.

BTW: I get my power from 10 miles up the road. (There's a closer substation but my side of town does not connect.) My voltage at the street is *solid* 125V all the time, except *one* summer day it hit 124V. Now, my line from the street is long and skinny so in the house I get 125V down to 108V when dryer and heater run. But the street voltage is constant. Most utilities have tap-changers at street-level. (OTOH we have way too many loose trees and crazy jeeps. Power does go out, and sometimes for days.)

[a90s2cs]

It’s been set at -50v but it has a range from -35v to -70

That seems like it would idle the tubes kind of cold.  If the tubes are idled a little hotter, it will pull the B+ voltage down.  If you don't have one-Ohm resistors between the power tube cathodes and ground, you might want to consider installing them so that you can easily set the bias.  Once the bias is properly set and the B+ at that point is known, we can adjust the resistors between the nodes to get whatever voltages you want.

After looking at your node voltages, it appears as though V3 is drawing way too much current and V2 is not drawing enough current.  If this is correct, how come?

I have no idea why V2 and V3 are like that.  Like I said, I quadruple checked all the resistors, I took those readings with a set of JJ ECC83S tubes but I've also had Sovtek and Tung Sol 12AX7's in it and pretty much had the same results. 

I did put 1-ohm resistors between the cathodes and ground.

I just got it to the point of firing it up with tubes in it for the first time.  Unfortunately, I'm not able to devote time to it every day so its been a slow process of an hour or two here and there and playing a bit of catch up when I can get back to it.  I probably won't have a chance to do much with it until next week, I'm in the process of remodeling my home myself and this is what I do to take my mind off of that and my day job.

So far I've only fired it up 7 times in one single session since it has been 'complete', obviously I'm not counting the preliminary starts for checking the PT and related wiring.

First time... no tubes, on the bulb limiter, no issues

Second... with tubes, on the limiter, no issues

3rd... no tubes, no limiter, took readings. 

4th... no limiter, with tubes, took readings.  Got sound but the volume was really low and through examining the voltages that I was getting concluded that I put a 470k where a 470 should have been. Put the correct resistor in it then checked every component value several times.

5th & 6th and 7th times I just experimented with swapping tubes and taking readings and pretty much got the same results with all of them.  It had full volume now but sounded like hell, squealing, crackling, farting... as far as I could tell there is something not right with both the input and tone stack, the low input jack does not work at all and turning down the bass pot completely kills the signal like a volume pot.  I also swapped the OT primaries on the tubes which improved the farting and squealing but did not eliminate it.

That last series of tests is as far as I've gotten. That is also when I posed the original question here, because I knew at that point there was going to be a lot more of...  turn on - take readings - poke at it with a stick - turn off - change something - turn back on - take more readings - repeat.  I wasn't looking to fix a specific problem yet so much as prepare for massive diagnostics and general trouble shooting.

So that's where I'm at with it.... I'm not frustrated with it or in a huge hurry to make progress, this is a new hobby to me so just tinkering with this thing and problem solving it is all part of the fun (stress relief) for me.  Who knows, maybe subconsciously I did some things wrong just to drag out the project.

My next step is to re-read all about how to set the bias.  I also think before I turn it on again I need to clean it up a little, quintuple check the resistors,  reflow some ugly solder joints, re-work the shielded wires (do you ground the shielding?, I'm not clear on that), do a better job on all ground points, and re-examine my work on the tone stack and input jacks.  Then thoroughly blow it out with compressed air to remove any solder droplets, loose wire braids, trimmed component legs, etc.

Although slightly disappointed that it didn't turn out better at initial start up I am quite pleased that it didn't billow smoke, catch on fire, blow fuses or arc anywhere... and that there were good signs of life. I wasn't exactly ready to post any "how do I fix this problem" questions yet but my curiosity did lead me to jump the gun a little on sorting the B+.

I love how you guys jump right in with trying to figure out what is wrong and trying to help, this is a great community.  It's a good feeling knowing there's so many people willing to give their time, knowledge and experience to help me out with this...  Thanks!

That said, I feel like I should give you a heads up that it may be several days before I can get a chance to work on it and that I am in no way blowing off the help you are offering.  To me it seems a little rude for somebody to offer you help and then not respond to them for days.

[PRR]

> sounded like hell, squealing, crackling, farting...
My next step is to re-read all about how to set the bias.

WHY??

That's like trimming the idle-speed on a car that won't go, won't stop, the doors are flapping from engine vibration....

Stop farting around- stop the farting etc. If the power tubes are not stone-cold and not glowing-red, they will work "OK". Power tube bias is for when the amp "works".

> rude for somebody to offer you help and then not respond to them for days.

Go away. Come back when you have news. We all have Real Lives on the side, we know how it is. We don't generally hang on the edge of our chair waiting for reply.