#1
So Hi. I builded an amp from a kit (ampmaker se-5a) which has a pre-amp with 2 stages.
So I wanted a bit more distortion. so I decided to add one more tube (3rd stage + cathode follower). I designed it in order to have a 150k resistor in my anode but ended up doing with a 100k.

So I have a huge problem. It distorts, the sound is great up to the point that when I turn the "2nd" gain more than 50% I start having a high pitch sound.One of those really really annoying that gives you headaches, and at the max position it's like "game over" from a 80's video game(not kidding that's what I heard).

What could it be..a bad connection?(I tested all new solder points and they all have perfect contacts.) or something else..please ideas.

I have bypass in order to function with only 2 stages. I don't have an option to shut down the cathode follower for now.

The schematic is this one, the bottom right part is the part I added, the rest was there before.
#3
you need to properly attenuate the input to that stage. you're also missing a gridstopper.
Prs se Holcomb is the answer
#4
Parasitic oscillation. You just added a LOT of gain to the signal chain and as Min above stated, you have no grid stoppers anywhere to help reject/mitigate oscillations.

Try adding a 33k grid stopper to your input stage, a 100k grid stopper to the 2nd stage, and a 100k to the third stage. Put the resistors as close to the grid pins as possible. See if that helps.
#5
Thanks ! Actually I think the sound is fully distorted right before the oscillation starts. I'll add the resistors tyvm
#6
Quote by CECamps
Parasitic oscillation. You just added a LOT of gain to the signal chain and as Min above stated, you have no grid stoppers anywhere to help reject/mitigate oscillations.

Try adding a 33k grid stopper to your input stage, a 100k grid stopper to the 2nd stage, and a 100k to the third stage. Put the resistors as close to the grid pins as possible. See if that helps.


Between the 1st and 2nd stage there's a 1M resistor(R4 in the schematic).
Ty I'll add some resistors around
#8
R4 is not a grid stopper though. It is part of a voltage divider that is attenuating the signal before the next stage.

An example of a 3 stage plus CF is the JCM 800.
#9
http://www.ampmaker.com/images/ak01kit/ak01sc2.jpg This is the original schematic.
The R4 is doing the same thing as the 470k in the jcm800 no?

http://www.kbapps.com/audio/schematics/tubeamps/marshall/images/JCM800mastervolumeM2203-100W.gif
I honestly can't see the diference between JCM800 and R4, obviously besides the value.

I have a potentiometer for gain in the 3rd stage, R26 is representing it(the schem in 1st post). You mean I have to put 2 resistors on in the signal path and other between the grid and ground? LIke the ax84 high octane? (that's what my R26-1MegPot will be for no?

Thanks for the help.
Last edited by forevertheuni at Nov 24, 2011,
#10
Yes, the 470k resistor you're looking at in the JCM800 is part of a voltage divider just like R4 is in your schematic. However, the fact remains that it is not a grid stopper and does not perform the function of a grid stopper.

One of the main problems with your modified design is that your gain control is at the end of your preamp chain. It doesn't really make any sense to do that. The gain control should come after the very first gain stage, particularly since you took out the master volume control and are running a single ended output tube. Personally, I would put the master volume back in if I were you.

Once that is addressed, you'll likely still need to run grid stoppers. The grid stoppers are placed between the grid leak resistor and the preamp tube grid. An example is R22 in the pic below. It needs to physically be placed as close as you can put it to the grid pin on the tube socket (preferably soldered right to the pin).

#11
Quote by CECamps
Yes, the 470k resistor you're looking at in the JCM800 is part of a voltage divider just like R4 is in your schematic. However, the fact remains that it is not a grid stopper and does not perform the function of a grid stopper.

One of the main problems with your modified design is that your gain control is at the end of your preamp chain. It doesn't really make any sense to do that. The gain control should come after the very first gain stage, particularly since you took out the master volume control and are running a single ended output tube. Personally, I would put the master volume back in if I were you.

Once that is addressed, you'll likely still need to run grid stoppers. The grid stoppers are placed between the grid leak resistor and the preamp tube grid. An example is R22 in the pic below. It needs to physically be placed as close as you can put it to the grid pin on the tube socket (preferably soldered right to the pin).



I still have the master. I didn't take it out.

I added a 470k resistor right before the grid of the 3rd stage. Will add another one in the previous stages. ty again.oh and also added a 33k between the input and the grid(should I put a bigger value in here?
#12
I wouldn't have put a 470k grid stopper in. You are going to dull the tone with that high of a value. Go with my original suggestion of 100k for the second and third stages. 33k is fine for the first stage.
#13
if you copy the jcm800 you'll see that there's also a divider network before the third state/cathode follower.

that should be enough to stabilize the amp as long as your lead dress and everything else is good.

the jca20h has a different 3 stage layout.

tons of amps have 3 stages.

also switch where you take b+ for the preamp stages. the 3rd stage has to be taken before the input stages.
Prs se Holcomb is the answer
Last edited by AcousticMirror at Nov 24, 2011,
#14
Quote by AcousticMirror
if you copy the jcm800 you'll see that there's also a divider network before the third state/cathode follower.

that should be enough to stabilize the amp as long as your lead dress and everything else is good.

the jca20h has a different 3 stage layout.

tons of amps have 3 stages.

also switch where you take b+ for the preamp stages. the 3rd stage has to be taken before the input stages.

Lol Asked that question afraid of the same thing almost everyone said..oohh it's the same thing.(will do it anyways).

Will those 0.00047 capacitors in the jcm800 make a big difference?
So I added a lot of grid resistors(470k each) before every stage It should be something like this now:

I added R20(470k) R21(470k) R19(33k) R17(470k). Btw when I say(replacing in the schematic It's for me to know it's a VR and not a static resistor)

It still screams. So if I reduce the second gain to +/- half if stops, and if I put the second gain to max and the 1st gain to max it still screams a lot. Besides the B+ is there something else?

Thanks a lot for bearing with a noob
Last edited by forevertheuni at Nov 24, 2011,
#15
In order to understand what you added, we'll need to see a schematic representation. The ref des means nothing.

Can you post an updated schematic that shows the circuit exactly as you have it now?
#17
OK, there are a few things that need to be addressed here. First off, as I stated before, 470k resistors are WAY too high in value to be used as grid stoppers in this amp. Again, I would highly recommend changing R17 to 100k.

Next, there is still no grid stopper on your 3rd stage. You DEFINITELY need one there. Put a 100k grid stopper there.

I'm not understanding VR26 in your schematic. You have two values listed: 1M and 400k. Is that a static resistor or a pot? What's the value? If it is a pot, it is connected incorrectly in the circuit and serving no useful purpose. Personally, I would remove that completely unless you want to run two gain controls--in which case you need to connect the pot correctly (pin 1 of the pot to C20, pin 2 to R20, and pin 3 to ground).

I'm not understanding why you connected VR1 the way you did either. It is serving no useful purpose as it sits, just acting as a static 500k resistor. If you run a gain control, it really needs to be in VR1's position. But it also needs to be connected correctly to work. Connect pin 1 of the pot to R4, pin 2 to R17, and pin 3 to ground. If you only want one gain control, it needs to be positioned where VR1 is.

Lastly, there is a TON of gain going on here. You have 3 cascading stages and all of them are bypassed resulting in full gain across the useable guitar frequency range. With your VR1 configured like it is, a normal guitar input voltage is resulting in about 4 volts at the grid of the 2nd stage and you have no way to attenuate that. That's going to be beginning to overdrive that stage. Further, given this same hypothetical input voltage, you'll have about a 100 volt input signal at the grid of the third stage! And with no gain control to attenuate it at all and no grid stopper to prevent oscillations. It's no wonder it's squealing.

So perhaps incorporate the changes I've listed above and also think about removing C3 to bring the gain down. You may even want to remove C13. Just depends on what you're going for with the amp.
#18
Quote by CECamps
OK, there are a few things that need to be addressed here. First off, as I stated before, 470k resistors are WAY too high in value to be used as grid stoppers in this amp. Again, I would highly recommend changing R17 to 100k.

Next, there is still no grid stopper on your 3rd stage. You DEFINITELY need one there. Put a 100k grid stopper there.
I had a 470k in there as a grid stopper.Only afterwards, and after still having problems, I tried to add the extra resistance R21. So is it better to have smaller values? I though a bigger one would decrease in a higher extent the gain in order to reduce my problem.

Quote by CECamps


I'm not understanding VR26 in your schematic. You have two values listed: 1M and 400k. Is that a static resistor or a pot? What's the value? If it is a pot, it is connected incorrectly in the circuit and serving no useful purpose. Personally, I would remove that completely unless you want to run two gain controls--in which case you need to connect the pot correctly (pin 1 of the pot to C20, pin 2 to R20, and pin 3 to ground).

VR26 is a 1M pot with an integrated switch,the 400k in the picture is just a way to simulate a midvalue in LTSpice, should have replaced it when made the picture.
So when the pot is at 0 the sound is bypassed and 3rd stage and follower never get AC on the grid, when the pot is turned on the switch makes the sound pass through them, that's why I have two gains.(while acting as a potentiometer).
It is connected exactly like you suggested.
Quote by CECamps

I'm not understanding why you connected VR1 the way you did either. It is serving no useful purpose as it sits, just acting as a static 500k resistor. If you run a gain control, it really needs to be in VR1's position. But it also needs to be connected correctly to work. Connect pin 1 of the pot to R4, pin 2 to R17, and pin 3 to ground. If you only want one gain control, it needs to be positioned where VR1 is.

Again, the same applies, VR1 is a 1MVR (1st gain) the 500k indication is a midvalue for LTSpice.
It is also connected like that.
I basically have the two gains because I can bypass the signal to only have 2 stages.

Quote by CECamps

Lastly, there is a TON of gain going on here. You have 3 cascading stages and all of them are bypassed resulting in full gain across the useable guitar frequency range. With your VR1 configured like it is, a normal guitar input voltage is resulting in about 4 volts at the grid of the 2nd stage and you have no way to attenuate that. That's going to be beginning to overdrive that stage. Further, given this same hypothetical input voltage, you'll have about a 100 volt input signal at the grid of the third stage! And with no gain control to attenuate it at all and no grid stopper to prevent oscillations. It's no wonder it's squealing.

So perhaps incorporate the changes I've listed above and also think about removing C3 to bring the gain down. You may even want to remove C13. Just depends on what you're going for with the amp.


The 1st two stages I would prefer to let them as is (for a clean sound when bypassed).
So with this. How can I decrease the gain to the 3rd stage when the sound goes through it? Adding a big resistor between C20 and VR26?
Again, I tried to use big grid stoppers to decrease the gain to a maximum (doesn't work like that?).
The sound is like I want it to be, until I increase VR26 more than half or reduce gain 1 to half and full at VR26, and there is no squealling up to that point only after that it starts to squeal like a pig.
Basically I modified the 2stage amp following a setup somehow like the AX84 hi-octane...
So if I was not to have VR26 and only added a resistor, what would be the value so that gain would be at maximum?

Once again, THANK YOU very much for your time spent helping me.
#19
you should really just think about copying the jcm800 to start with and then changing things after you get that stabilized.

3 stages shouldn't be squeeling that bad.
Prs se Holcomb is the answer
#20
Quote by AcousticMirror
you should really just think about copying the jcm800 to start with and then changing things after you get that stabilized.

3 stages shouldn't be squeeling that bad.

The thing is, the JCM800 only really has 2stages+cathode follower with two independent inputs that do +/- the same thing (or I really need to learn how to read the schematics, or have been looking the wrong ones).
The sound in the amp is really cool, goes from clean to lead with no background noise at all. So I though in, adding a lot of resistance somewhere in the path(right before the 3rd stage?).

I still have to switch the B+.
#22
Hmm no capacitors in almost all cathodes.damn it.I don't get it then. This schematic isn't that much different than mine. :S

hmm wait the's a 10k resistor in the 2nd stage...isn't that really high?
Last edited by forevertheuni at Nov 25, 2011,
#23
So I changed the B+ order and added a 470k resistor between the 3rd pin of 2nd gain poti and ground. I could almost have no squealing with both gains almost at the max. However I should check the voltages between anode and cathode and intensity right? So that the tubes don't blow really fast in order to be in the perfect curve.
#24
Yo. Well,came back becaause I wanted to say thanks for the help to AcousticMirror and CECamps.
It's almost finished. and the noise is almost gone. So in the 1st stage I replaced the 1M for a 370k. And in the 3rd stage(before the poti) I added a 1M and I also added a voltage splitter to ground with 100k(with a resistor following to act as grid stopper 100k), this allowed a big dissipation of the signal I guess.

I think I should add a capacitor in parallel to the 1M resistor. I noticed this is still guitar dependent. When my tone knob was set to 0 I didn't get these problems(if my memory serves well it's a 47pF cap in it).
I think I'll copy it from the JCM800. Again thank you for all your help.