NFB making me friggin nuts....

[Icetech]

I built this ODS from norms layout here.. like 8 months ago now.. and am finally close to considering it done.. BUT this one prob has been hounding me bad since i built this thing.. and i keep thinking i have it fixed but it's always there.. taunting me..

heres the layout
http://ampgarage.com/forum/viewtopic.php?t=892

The problem is... a crackling sound.. at the begining of the note, that fades out fairly fast.. with the amp very low it doesn't happen, soon as you get up around 5 it starts..

This is with bassman iron, i have tried anther OT, rerouted wires.. the only thing that stops this from happening is if i disconnect the NFB totally.. then the amp is literally twice as loud as with the NFB attacked and just wild.. early distortion..

I have used from 3.3k to 25k resistors on the NFB to see if that would change it.. as long as the NFB is connected in anyway it crackles at the front of the notes..

Any ideas on what to try, im open for.. this amp can't play clean at any decent volume at all without this happening..

Also should the NFB drop the volume of the amp that big? for example, with the NFB, amp on 10 i can sit in front of the 4x12 and its not loud enough to bother me much.. with the NFB off.. its like OMG i want outa this room loud...

BTW, i have tried this with 6V6's 6l6's el34's prob is exactly the same..

I would post a pic, but i don't think it would help much, i have ran the NFB wire over the board.. up in the air 3" above the board.. below the board... all the same effect..

[Normster]

It could be a bad 1uF cap or that the cap is backwards. I've been trying to get a grip on the whole NFB concept using Aiken's formulas. It looks like Dumble reduces gain by -8dB for both the 100watt and the 50watt. The 100watt, as we know, uses a 4k7/390 voltage divider to the PI input. Gil mentioned that the 50watt uses an 8k/1k voltage divider. Oddly enough, they both calculate out to about -8dB but I assume the cathode voltage would be slightly higher on the 50watt PI. (Just guessing, I haven't check it.)

I have since changed the way I wire the Presence control. I now run the 390R (or 1k) from the input to the ground of the pot and use a 1uF tantalum between the wiper and the ground. You can try that and see if it solves the problem. Just make sure that you wire the ground side of the cap to ground if using an electrolytic.

BTW, DO post a pic of your Presence pot and wiring. It may help.

[heisthl]

Can't tell from the layout but I'm making the assumption that your using the 4ohm tap as the NFB source. I would troubleshoot this initially by using some test clipleads and connecting the transformer, output tube and Phase Inverter grounds together. If no change connect to the 4way resistor juntion on the PI cathode/grid chain and individually check the R values. If no problem found and a scope isn't telling you anything, sub in replacement caps one at a time in the NFB/PI area. If at this point you haven't found anything you would have to make the assumption that the noise exists even without the NFB attached but is being masked by the loud volumes and could be caused by anything from PI plate resistors to P/S noise to poor grounding elsewhere. I'm hoping for your sake its a Homer Simpson DOH problem.

[pamaz67]

Hi Icetech.
this , to me, sounds like a self oscillation that starts just on certain output levels and/or particular frequency.
The only way to verify such a possibility is with a scope and a signal generator or a variable freq sine wave generator ( i have a cd with sweep test tones that is good for such things) .
If you have access to a scope, put all your tone controls clockwise, turn pab and midboost on and place a powerful dummy load at the output of your amp. Then inject the sine wave ( something like 1.5v pp, quite close to a peak to peak strong guitar output) starting from low freq ( 30 hz) to higher ones ( over 20khz) and verify if at the output you have a decent reproduction of the waveform, or you see something that does not resemble a sine wave in any way.
If you are in this second situation, you are probably looking at an ugly self oscillation of the amp.
There is no fixed rule to solve such an issue, but first of all , go and buy some screened cable. In some places it helps........



Good luck

Paolo

[pamaz67]

Hi Icetech.
this , to me, sounds like a self oscillation that starts just on certain output levels and/or particular frequency.
The only way to verify such a possibility is with a scope and a signal generator or a variable freq sine wave generator ( i have a cd with sweep test tones that is good for such things) .
If you have access to a scope, put all your tone controls clockwise, turn pab and midboost on and place a powerful dummy load at the output of your amp. Then inject the sine wave ( something like 1.5v pp, quite close to a peak to peak strong guitar output) starting from low freq ( 30 hz) to higher ones ( over 20khz) and verify if at the output you have a decent reproduction of the waveform, or you see something that does not resemble a sine wave in any way.
If you are in this second situation, you are probably looking at an ugly self oscillation of the amp.
There is no fixed rule to solve such an issue, but first of all , go and buy some screened cable. In some places it helps........



Good luck

Paolo

[Icetech]

Paolo, yeah it is at certain freq's i had the scope hooked to it yesterday and did what you said (and not having a dummy load made it painful for me) and i could sweep the sine generator to tones that would make it constantly do the crackle.. and change the sine on the scope..

[tonelab2]

Since the crackling is only there with the NFB connected I would shotgun the whole NFB circuit. Also have you tried using a signal trace (with dummy load) into a reference amp tracing signal right from input through NFB section to spot a faulty component.
My amp is wilder (feedsback easier) but only slightly louder without NFB, are you 100% sure everything is wired correct.
Once had a amp that had intermittent (spelling) crackle, turned out to be a resistor wrapped around turret but not soldered.

[Icetech]

Tone, im gonna double check every part and solder points asap.. im just frustrated.. and yeah i had one amp i did make me REAL nuts... everything looked perfect and seemed right cape i had put a like 800K in place of a 800ohm.. or whatever value it was:)

Im sure i will figure it out..

[pamaz67]

Ciao Icetech.
Just a question have you got a dumblelator / effects loop wired in the amp or outside it?.
Let me know
Paolo

[Icetech]

no. it's setup jsut like that layout.. i ordered some parts from mouser today to swap out.. will let ya know whats up later in the week...

[pamaz67]

Just some more ideas to trace the problem.
-Have you wired the grid resistors directly on the sockets? ( if not , do it)
-Have you got a 1 Meg master volume? ( i had a self oscillation using a 250 k one)
-By using chopsticks have you tried to move the wiring and to check on the scope is something is changing while you move the wires?
- Are you using some Huge ground wire ( remember that what's important is not the resistance but inductance at the high frequency of self oscillations)
- Have you tried to bypass with a cap ( let's say 100p to 330 p) the anode resistance on the clean channel? ( The values suggested could kill the tone of the amp, but probably you'll be able to identify where the oscillation raise)
- if you have a ground connection on your tube socket( i mean) the one in the center of the socket, connect it to ground. Sometimes it helps.
- try to diconnect the pream from the power amp , and give a look if they individually are working fine , or if they already oscillate when left apart.

Probably youalready have tried everything, but I'm just trying to give you ideas for searching the solution.
Ciao and good luck

Paolo

[pamaz67]

some more hints that come to mind now.
Just talking of the preamp side.
try to understand what the oscillation frequency is ( on a scope you'll be able for sure)
Now you know that you have to kill all the gain the amp have over that frequency.
Therefore, you'll be able to calculate the value of the capacitor to place across the anode resistor. That cap will bypass the resistor over the desired freq, and therefore will kill the gain of the preamp at that freq.
i'm quite confident that the oscillation will be over 100khz, therefore if you are able to get a suitable cap across the anode resistor ( think like 47 pf if you have a 220k anode resistor, or 100 pf if you have a 100k resistor) , that would be not really annoying for the tone of the amp, considering a 15khz bandwidht to be more than enough for a guitar.
ciao

Paolo

[Icetech]

Thanks paolo, i will check on that stuff when i get home from work..

All the wiring is 22ga teflon..

On the scope the oscilation starts in the driver section... preamp stays nice and clean..

Moving wires and cleaning up dress was the first thing i did... which turned out good anyway, i had a very small hum i got rid of:)