Major noise issue with AB763 build

[Apexelectric]

I built a single channel AB763 based on both the fender and Hoffman designs but with a few mods.  It's layed out Fender style but with a 4 section cap can.  I changed the mid pot to a raw pot, Added a master vol, eliminated the tremolo section and tube, added a selectable NFB, bright switch, and a bypass cap selector on V1a. Used 4 and 8 ohm output jacks with the Nfb attached to the 8 ohm.

Fired it up for the first time and had a loud hum which was controlled by the master volume. Arcing on the standby switch when powering down. Checked all my wiring, connections, etc and realized I needed a 47k resistor to ground between the 0.1 cap leaving V3 and the .001 cap entering V4? Now there's a loud motor boating. I pulled the pre amp tubes one by one and with V3 removed it disappears.

Not sure where I went wrong but I think I might be missing something when I removed the tremolo section?

Any help would be appreciated.

[shooter]

start here

http://el34world.com/Forum/index.php?topic=17701.0
work slow and methodical, repeat

[Apexelectric]

I've triple checked all of my connections, jumpers and values to see if they correspond to the schematic but since its not a stock design, I'm thinking theres an issue with the design that Im missing.

[sluckey]

Show us your schematic. Make it big. Also post a bigger version of your layout.

[Leevi]

The motor boating tells about oscillation somewhere, try to move wires during operation and listen how it affects.


If there is continuous background hum one source can be the negative bias voltage if there is ripple due to bad filtering.


You can try to localize the the problem by shortcutting the PI input to ground. If the hum still continues then it's coming from the power amp.


/Leevi

[Apexelectric]

I'll look into it some more over the next few days but it seems that when I corrected the missing 47k resistor problem it created another issue because it lit up my current limiter when I fired it up. I'll try to see if I damaged something in the power section before I take it any further. Voltages in the preamp seemed a little high but in the ballpark but I did notice an inequality on the the PI plate voltages.

Wondering if I damaged the multi section cap can in this process?

No blown fuses to this point.

More to come when I have some definitive info...

[John]

You really need to post an as-built schematic. For instance, I'm trying to figure why there'd be a 47k resistor going to ground between 2 (coupling??) caps, and without a schem I have no idea. Sluckey probably has the AB763 schem memorized, but you've made changes which means it's not really that circuit anymore.


Schematics rule, dude. 

[Apexelectric]

Pardon the fact that I'm still trying to get up to speed with this process. I have not created a schematic as of yet. Still looking into all the software options out there and avoiding having to learn to navigate one or two more platforms.

Here is a repro of the Hoffman Single Channel with the resistor that I missed the first go around.

[Apexelectric]

I was also curios to know if I should adjust a dropping resistor value to adjust for the missing load of the tremelo tube. The overall voltages seem to be about 10% higher than expected. I'll reserve my judgement till the issue is resolved but based on what I've noticed so far in the preamp and the output section, the voltages are off a little.

[labb]

I wouldn't worry a lot about the voltages at this point. Most of the Fender schematic say voltages are +/- 20%. Most likely you have a wiring problem unless you have really modified the standard circuit to the point where it just will not work.

[shooter]

I'll try to see if I damaged something in the power section before I take it any further.

are you still on the current limiter?
do you have voltage readings at each PS tap?
Both DC and AC would be helpful

[Apexelectric]

Here is an edited schematic reflecting the current design. I didn't bother to edit minor variations in values such as 47uf vs. 50uf or 20uf vs. 22uf but shared cathodes have been edited and all the mods have been included.

Let me know if anything jumps out?

I'm hesitant to test voltages with the standby switch on for fear of damaging anything but all off the rectified B+ voltages were around 470-490. Preamp was anywhere from 240-300. Heater voltages were right on.

These were not on the current limiter.

The phase inverter plate voltages were 25-30V different from one another. One was right where it should be but the other too high.

I can pull V3 and check voltages without worrying about the issue but not sure how much good that would do.

Thanks!

[John]

Thanks for the schematics! Your motorboating *might* be as simple as having the brown and blue wires in the wrong phase, since you're using NFB. As I understand it, that can show up in several different ways/noises.


And I feel like a dummy, because that 47K to ground makes no sense to me. It seems like it should be a 470K? But, I know very very little about the hows and whys of Fender designs... I'm sure someone will jump in who knows.

[sluckey]

that 47K to ground makes no sense to me

It simulates the missing 50K INTENSITY pot. The amp will work without it but the signal feeding the PI will be much larger and the amp will sound more gainy, ie, will lose some of that clean AB763 sound.

[Apexelectric]

I'll try swapping the OT leads but that probably wouldn't explain why my current limiter is lit up?

[ac427v]

I'm wondering if it is related to the NFB arrangement. Usually Fender used a 47 ohm resistor to ground when NFB was connected to the 8 ohm tap. Using a 100 ohm multiplies the amount of NFB. You can test this theory by switching to the NFB 1500 ohm  position or the off position.
--Craig

[sluckey]

...it seems that when I corrected the missing 47k resistor problem it created another issue because it lit up my current limiter when I fired it up.

Installing that resistor would not cause the current limiter to light up. The bright limiter bulb has to be a coincidence.

[Apexelectric]

None of the switches change the issue. Removing the resistor to ground off the reverb jack changes the low loud motorboating to a higher pitched squeel. Noises will be almost completely silenced by the volume controls. 

[shooter]

as an experiment try giving V3B it's own cathode R to ground, and turn your verb knobs to 0.

Also take everything off V1A's cathode except the 1.5k.

better/ worse?

[Apexelectric]

Pulled all the preamp tubes, still an issue on the current limiter. Pulled the power tubes too and only left the rectifier, lamp burns half bright on the limiter, full bright with the power tubes in. This is all with the standby switch off. Center tap on the filament coil an issue?

[labb]

Pull the rectifier tube and see if the limiter still lights up. If it does then look close at your pilot light for a short.

[Apexelectric]

Just a dim glow with the rectifier removed

[PRR]

> dim glow with the rectifier removed

What size lamp?

This is a big amplifier. I'd use at least 100W _incandescent_ (or 87W-90W halogen). A 25W sure will glow with the idle current of this beast.

[Apexelectric]

It's a 75w halogen.

Tore into it this morning and isolated the different power supplies and the only one causing the current limiter to light up is the heater circuit. First disconnected the high voltage tap, lamp still glows, then the bias, then the rectifier, same result. Only when I finally disconnected the 6.3V heater supply did the lamp not light up. Then I removed the tubes. Preamp tubes not causing an issue with the limiter just the output tubes. Checked the output tubes again and they tested good. Heater winding reads around 6 ohms. No bad connections on the tube socket to the heater terminals. Pulled and reattached the heater leads on the output tubes no changes.

[Apexelectric]

Any chance there's an issue with the heater tap windings? Small load on it with the pre amp tubes and the problem doesn't surface. With the output tube heater loads, the problem shows itself? Never blew a fuse this entire time.

[shooter]

Never blew a fuse this entire time.

If your fuse is the correct value, I suspect not.  Did your hum/noise change after you re-did stuff?
with everything loaded and no limiter, does the PT get warm fast?  Are all the PS rail taps stable?

[Apexelectric]

Nothing seems to have changed. After turning it on this time the noise started out more like a 120hz hum for 10 seconds then progressed into the motorboating.

[Ed_Chambley]

I do not recognize your Power Trans. What is the orange wire, why ground.?


Just looking at the red and blue I guess bias tap.


The primary side have different taps?

[Apexelectric]

Orange wire apparently has to do with an anti static band. Spec says to ground it. Red/Green is the bias voltage wire to the bias board from the PT then exits on the red wire to the eyelet board. Primary has 4 different tap options. 120V option uses two different pairs of wires and leaves two wires unused.

[Apexelectric]

PT is a ClassicTone 40-18029 if you're interested.

[Apexelectric]

Well, finally got it sorted out.  After reconfiguring the ground scheme I noticed there was a missing ground from the reverb pot tab. Got the one on the DWELL but missed the one on the MIX. Sorry for the oversight. My bad.

I'm going to tweak a little more on it to get it dialed in but a couple things I found since I got it up and running.

My PT is kicking out about 10% more voltage on all taps than I would like. It's rated at 120V and I have 122V at the plug so I'm assuming due to the reduction in tubes in the design that it is now under loaded and kicking out more voltage? My voltage after the rectifier is now around 500VDC instead of the 460VDC I was looking for. The filament voltages are also high by the 10%. I tried the lower voltage B+ tap on the PT, which drops the voltage by 100V, but it's not giving me the tone I'm looking for.

I'm assuming I should not worry too much about the filament voltages and change the dropping resistor values on the filter caps but I'm looking for an idea on how much of a change I should start with? They are the stock values now. Or would changing the rectifier tube be better?

I also got a high pitched squeel out of the reverb when I turned up the mix and dwell. Not sure which one was the cause. The noise was too offending to want to reproduce it again.

I'm going to get back on in it a bit. Waiting for everyone to wake up before I start making noise.

Thanks for all the help. There's definitely a steep learning curve here. I apologize for being the newbie but I'm learning from my mistakes.

[labb]

Just got to ask, is the 500 volt reading with all of the tubes installed?

[Apexelectric]

Yeah. Fully loaded.

[labb]

How much negative voltage do you have for the bias?

[Apexelectric]

Just remembered that the 1/2w resistors are only rated at 350V. It did seem quite a bit quieter when I was running it on the lower voltage B+ tap. I'm guessing that a 1W resistor would be a better choice when the voltages get up a bit higher.

Does this theory make sense?

[Tony Bones]

I doubt there's more than 350V across any of the resistors, except maybe briefly at power up.

But, generally speaking, higher power resistors are quieter than lower power of the same construction.

[Apexelectric]

There's definitely 300+ at the plates of V1. I think I'll go back to the lower B+ tap to compare the noise level again.

Tubes are biased cold at the moment at -55v from the bias pot. About 30MA of current with plate voltage at around 485-490V.

[labb]

Don't over think the 350 volt max. rating.    Lots of amps running with 1/2 watt resistors that have 400+ volts on the plates. Power rating (watts) is a different story. As Mr. Fender says on the schematics, all resistors are 1/2 watt except as noted.

[labb]

Would be interesting to see what the plate voltage went to if you raised the current to about 40 ma. I't trying to understand why the data sheet for the 40-18029 is so far off. Looks like you should be at about 460/465 vdc.

[Apexelectric]

Ok. I won't.

Been through the whole amp looking for bad connections and chopsticked the heck out of it with no result that improved the background hiss. Tried putting up metal barriers between transformers and other areas to try to isolate an area that might be inducing the noise. No luck.

Also checked the power source, fluorescent lights, cell phone, etc. 

I used a blank chassis that's more like an 18w size so I thought things might be too crowded and inducing noise into areas. The wires to and from the grids on V1 are shielded. Also changed to shielded wire on the heater supplies.

[Apexelectric]

Yeah, I know. Thought that the lesser load due to the elimination of the tremelo tube might have driven the voltage up a tad.

Should I change to a higher value dropping resistor at the filter caps or add on in Series to the choke?

[labb]

zener diodes work pretty good but you have to take care of the heat. You are using a 200 amp xformer and probably only pulling about 90 amps at the most if you are at 60 ma for the power tubes. 

[Apexelectric]

So what rating on the diode would be recommended? in series with the rectifier output?

[labb]

Most are in the PT ground. Do a search. Should be a lot of discussion about zeners used to drop B+ voltage.

[ac427v]

You could drop B+ by using a 5U4GB rectifier or by switching back to the lower voltage secondaries and using Doug's solid state plug in rectifier.
--Craig

[Tony Bones]

> There's definitely 300+ at the plates of V1.

What's the voltage across the resistor, from end to end? At idle that would be 300V - plate voltage. Of course, at full tilt the plate voltage swings below quiescent by a couple dozen volts, but the resistor never has even 300V from end to end, never mind > 350V

[Apexelectric]

I'll check the voltage across the resistor later when I'm back home. Im assuming it should drop quite a bit. Just a bit unsure on how the noise relates to the resistor voltage ratings. From what I'm hearing it sounds like it should be something other than the 1/2 W resistors.

Was also wondering if I would have been better off with a cap board instead of a 4 section cap can. I ran out of real estate on the back of the chassis.

Thanks for all the advice guys!

[pompeiisneaks]

I'll check the voltage across the resistor later when I'm back home. Im assuming it should drop quite a bit. Just a bit unsure on how the noise relates to the resistor voltage ratings. From what I'm hearing it sounds like it should be something other than the 1/2 W resistors.

Was also wondering if I would have been better off with a cap board instead of a 4 section cap can. I ran out of real estate on the back of the chassis.

Thanks for all the advice guys!

I've used a cap can with no problems on a trainwreck liverpool, I've heard someone once on FB say that using them in both preamp and power amps is bad because of the shared ground but I don't notice it (they claimed it was one of the biggest problems in most marshalls for noise).  So if you wanted to be extra careful, you could use one can for the power and PI side and another can for any preamp stuff to be 'cleaner' but I don't know how much weight i put on that...   Otherwise they're great from what I understand.

~Phil

[Apexelectric]

I might try to separate the pre amp from the power amp filtering to isolate them into two different grounds. Never been that much of an issue in the past having all four share on ground lead. I might be able to squeeze in one more can on the back. Otherwise I might have to stack the filter caps vertically on the back of the chassis.

[Apexelectric]

Well, I've made some progress in silencing some of the background noise. It does run a bit quieter at the lower B+ tap voltage but I've got it too the point where it's almost acceptable. The reverb pot seems to be the one thing that injects the most noise into the circuit than anything at this point. The dwell pot position seems to not cause an issue, just the mix pot.

I still have a 20v difference on the phase inverter plates (V4) 265vdc vs 245vdc as well as the same readings on the plates of V3. Imbalance in the triodes? Resistors seem to be within proper tolerances. Output tubes are at 29.3 and 30.8 ma at 479V on the plates. Bias voltage is at -53.

I'm guessing I shouldn't be seeing that much of a voltage difference between the plates of  the two triodes of V3 and V4? Could that be adding to the background noise?