tube amp problem - need help....

[gusfinley]

I just finshed my rebuid of a Gibson GA-5 Les paul Jr amp...

It is working good except that when I turn the voulme knob past a certain point I get this really BAD distortion, like the speaker is cutting in an out- we're talking some SERIOUS clipping....

I tested it earlier with an oscilloscope and the problem lies in the preamp section... The wierd thing is that when I had a dummy load connected to the amp and turned the volume up the tube started ringing - mechanically!! Te point that the tube starts toring is the point where the speaker starts to distort...

also, when no input is connected to the amp, and the volume knob goes past that magic point, it looks like two electrodes on the output tube short together - there is a blue glow between them.

This is really an odd problem, and I have not been able to find any help yet with it....

any toughts?

 

[ez_willis]

Did you swap out the volume pot? If you soldered onto it you may have overheated it, maybe.

 

[boingoman]

Sounds like a it could be an oscillation. That means something in the amp is generating a signal, at a very high frequency beyond what you can hear. This sucks up a ton of headroom, and can cause the symptoms you describe. When the amp gets to that "magic" point you talk about, it runs up to full output power. An oscillating amp can make no sound, squeals, have low output, make ghost notes on top of what you play, make it sound like the speaker is blown, and have a bunch of other sounds.

Oscillations can be a pain to track down, but you are working from a known good design, so that helps. And you have a ringing preamp tube, so that may help locate the problem. But it may not.

The problem is that it could be almost anywhere. It could be a bad coupling cap, bad filter cap, bad solder joint, bad circuit board, bad ground, long resistor lead, two wires too close together, anything. I'd resolder the preamp section, and check all the coupling caps as I did. I'd try some simpler checks first, though.

If you changed the layout at all, even wire layout in the chassis, it could be the problem. Where the wires run in a tube amp can make a huge difference in stability. Did you orient all the tubes and transformers the same way as the original? And run all the wires just like they were in the original?

The easy things to check are wire layout and grounds. Check all ground connections with a meter, especially transformer center taps and grid resistors. No connection to ground should be more than an ohm or two. Make sure all wires and component leads are as short as possible, while following the original layout. The problem may be as easy as moving a wire 1/2 an inch somewhere.

Check all solder joints and coupling caps. Check everything, but starting with the easiest thing is best.

If the amp has a negative feedback loop, disconnect it. If the problem goes away, your transformer is wired backwards. Switch the leads on either the primary or secondary.

 

[gusfinley]

sweet!!

Sounds like a it could be an oscillation. That means something in the amp is generating a signal, at a very high frequency beyond what you can hear. This sucks up a ton of headroom, and can cause the symptoms you describe. When the amp gets to that "magic" point you talk about, it runs up to full output power. An oscillating amp can make no sound, squeals, have low output, make ghost notes on top of what you play, make it sound like the speaker is blown, and have a bunch of other sounds.

Oscillations can be a pain to track down, but you are working from a known good design, so that helps. And you have a ringing preamp tube, so that may help locate the problem. But it may not.

The problem is that it could be almost anywhere. It could be a bad coupling cap, bad filter cap, bad solder joint, bad circuit board, bad ground, long resistor lead, two wires too close together, anything. I'd resolder the preamp section, and check all the coupling caps as I did. I'd try some simpler checks first, though.

If you changed the layout at all, even wire layout in the chassis, it could be the problem. Where the wires run in a tube amp can make a huge difference in stability. Did you orient all the tubes and transformers the same way as the original? And run all the wires just like they were in the original?

The easy things to check are wire layout and grounds. Check all ground connections with a meter, especially transformer center taps and grid resistors. No connection to ground should be more than an ohm or two. Make sure all wires and component leads are as short as possible, while following the original layout. The problem may be as easy as moving a wire 1/2 an inch somewhere.

Check all solder joints and coupling caps. Check everything, but starting with the easiest thing is best.

If the amp has a negative feedback loop, disconnect it. If the problem goes away, your transformer is wired backwards. Switch the leads on either the primary or secondary.

Finally, An answer that makes sense!! Actually, I didn't really have a layout to work from, so I just did one myself. I do have a lot of wires crossing over eachother....

I already thought that the transformer may be the problem so I changed the connections on the primary and it seemed to sound better...

How would I check the coupling/filter caps? My mulitmeter has a capacitance setting, but is this enough to see if a cap is working alright?

I believe I have checked all the grounds, because I originally had thought it was a grounding problem... they all read less than 1 ohm....

Here below is a linkto the picture of the layout of the terminal board.........
http://cc.usu.edu/~davbradshaw/imghost/diyamp.jpg

 

[boingoman]

Finally, An answer that makes sense!! Actually, I didn't really have a layout to work from, so I just did one myself. I do have a lot of wires crossing over eachother....

I already thought that the transformer may be the problem so I changed the connections on the primary and it seemed to sound better...

How would I check the coupling/filter caps? My mulitmeter has a capacitance setting, but is this enough to see if a cap is working alright?

I believe I have checked all the grounds, because I originally had thought it was a grounding problem... they all read less than 1 ohm....

Here below is a linkto the picture of the layout of the terminal board.........
http://cc.usu.edu/~davbradshaw/imghost/diyamp.jpg

Cool. Not much to that thing, is there? Is that turret board? That stuff is cool. I have only seen it in Hiwatts.

I'll be the first to admit, I'm no expert on this layout stuff. Most of my experience is from fixing amps that are just broken and already laid out, or people have done wrong things when making mods. As far as design, I'm pretty clueless.

Your meter should be able to tell you if the cap is working- also check for DC leakage in them. There will always be some, apparently, and the mfr. should be able to tell you how much is acceptable. That is probably not the problem, though, as you are using all new stuff and haven't played with the layout. It's up to you which you want to check first. Just realize you have to unsolder one end of the cap to check it.

The transformers are at 90 degrees to each other, right? Since you knew to check the trannie leads, I am gonna assume you did that.

A bunch of this stuff comes down to just poking around in there with a wooden stick, nudging wires and components until you (hopefully) find the problem.

I think one of the first things I would do is to route the AC away from the circuit board, as much as possible, and do the same with the AC line that feeds the pilot light. That mostly will take care of hum, I think, but any little bit has to help, right? It looks like you have some slack to work with. Aside from that, I am pretty clueless, except for maybe tippiing those orange drops a little away from the filter caps and general poking around. Also maybe shorten those leads from the pots to the board, and get that resistor on the tube on the left snugged up closer to the tube. Those green wires going to the tube on the right, twist them tighter into a pair.



Looking at a layout from an amp with a similar configuration might help some.
Here is one from a bassman. It has the power switch and pilot light in the front, like yours.
http://www.drtube.com/schematics/fender/bassman-5f6-a-layout.gif

One big thing to notice is that the AC is all on one side. The other thing about layout I do know is that you want to keep all the similar stuff together as much as possible, and things should be laid out in a manner that mimics the amp's function. AC on one side, input going to preamp tubes directly as possible, then to output tubes, output trannie, and to the speaker. Looking at this layout, you can see the signal moves front to back, then right to left through the preamp to the power amp. The layout of the front panel controls is even linked to this concept. The tone controls, you notice, are placed so they are close to the tone circuits.

It's not always possible to strictly follow it, but using that as a guiding principle for laying out your wires or troubleshooting might help.

I noticed on yours that the AC comes in on the opposite side from the power trannie, right? That may not be easy to change, or at least not fun. If it's not like that, ignore me.

If so, is that a line out opposite the AC on the back? You might consider switching them and running longer lines to the pilot light and switch, or at least routing all the AC around the inside edge of the chassis.

Changing the AC stuff is probably nit-picky, except for routing the AC away from everything else.

Something else to try as a desperation move, if you can figure out a way to do it safely, is to onhook the whole board and move it around in a random way. It might help. (I can't believe I'm saying this)

Hope some of this helps.

 

[gusfinley]

yup........

Cool. Not much to that thing, is there? Is that turret board? That stuff is cool. I have only seen it in Hiwatts.

----------------------------------------------------------------
Yup, its a simple amp, good for a first build, and this turret board stuff is pretty cool....
-----------------------------------------------------------------

Your meter should be able to tell you if the cap is working- also check for DC leakage in them. There will always be some, apparently, and the mfr. should be able to tell you how much is acceptable. That is probably not the problem, though, as you are using all new stuff and haven't played with the layout. It's up to you which you want to check first. Just realize you have to unsolder one end of the cap to check it.

---------------------------------------
How do I check for DC leakage?
---------------------------------------

The transformers are at 90 degrees to each other, right? Since you knew to check the trannie leads, I am gonna assume you did that.

A bunch of this stuff comes down to just poking around in there with a wooden stick, nudging wires and components until you (hopefully) find the problem.

I think one of the first things I would do is to route the AC away from the circuit board, as much as possible, and do the same with the AC line that feeds the pilot light. That mostly will take care of hum, I think, but any little bit has to help, right? It looks like you have some slack to work with. Aside from that, I am pretty clueless, except for maybe tippiing those orange drops a little away from the filter caps and general poking around. Also maybe shorten those leads from the pots to the board, and get that resistor on the tube on the left snugged up closer to the tube. Those green wires going to the tube on the right, twist them tighter into a pair.

-----------------------------------------
I did tidy up those AC wires a bit, they do stay more to the edges of the amp now...

I was pretty careful with my layout, All of the AC is on the right of that pic including the power transformer. The input and output are on the left side along with the output transformer.

I ran into a problem when I had most of the parts in the amp. I was planning on putting a lot of components on some terminal strips in between the powet tube, but the small chassis wouldn.t allow that The cramped section on the top left of the pic I posted shows those....... I'll try moving those components around a bit and see what happens.....

--------------------------------------------------------------

Other forumers at a different board have suggested shieliding the leads that cross things like the output transformer leads, and ground points... Could this elimiately the problem with the oscillation?

The amp is quiet until the "strangeness" occurs, so I am not too worried about doing little things like twisting the filament leads (they weren't long enoughand it make them difficult to solder in the cramped space.....

oh, and by the way - I added a few resitors to the filament leads to ground referance them - the picture hasn't been updated to reflect that yet.....

 

[boingoman]

How do I check for DC leakage?.....

Sorry- DC leakage will just be continuity. Just strap your meter across it, out of the circuit, set on ohms, and see what you get. If you get overload, or some ridiculous high number of ohms, you are OK.

I did tidy up those AC wires a bit, they do stay more to the edges of the amp now...

I was pretty careful with my layout, All of the AC is on the right of that pic including the power transformer. The input and output are on the left side along with the output transformer.

Awesome. Sounds pretty good to me.

I ran into a problem when I had most of the parts in the amp. I was planning on putting a lot of components on some terminal strips in between the powet tube, but the small chassis wouldn.t allow that The cramped section on the top left of the pic I posted shows those....... I'll try moving those components around a bit and see what happens..........

That sure might help.

Other forumers at a different board have suggested shieliding the leads that cross things like the output transformer leads, and ground points... Could this elimiately the problem with the oscillation?

oh, and by the way - I added a few resitors to the filament leads to ground referance them - the picture hasn't been updated to reflect that yet.....

Yes, shielding those wires could help. I'm not sure if the resistors in the filament circuit make a difference, but I assume you've checked it out. At some point, you may want to twist those filament wires.

Oscillation comes from capacitances and reactances in the amp. They could really come from anywhere. Oh, and try grounding the speaker frame to the chassis.

Very nice project, gus.

 

[gusfinley]

let's see,,,,,,,

Gus, Is this the schematic for your amp? It looks like you've got something other than a 6V6 for an output tube and you've got a diode bridge for the power supply. A while ago if you remember seeing it, I had done a rebuild of an old tube tape recorder into a GA-5 design.

That is pretty close... I am using an EL84 in the output like the new Reissue Les Paul Jr GA-5 Goldtone amp. I have a few EL84's laying around....

I also added a tone knob using the circuitry from a Kalamazoo model one. I was orignally going to build the kalamazoo, and then wanted to build the GA-5 Reissue and so I combined three different schematics to come up with what I wanted. Then I found the real schematic for the reissue and I was only off by a few resitor values (it also used a different tone network.)

Maybe you've already done this, but have you done a point to point check of the wiring compared to the schematic just to ensure things are going where they should.

Yes, I have done this MANY times!!

I tried to follow some of the wiring on your turret board and the resistor (220k?) that I think goes to pin 6 of the 12AX7 looks as if it may have 2 caps tied to it or I can't see the second 220k which should go to pin 1 and have a coupling cap tied to it. It also appears in the design in yours the triode halves are swapped in the pin numbering compared to the schematic I had which doesn't make a whole lot of difference.

I'll post a pic with the wiring so you can see what is goin on under the board.... Yes, the triodes are swtiched - it was going to work better for my layout.....

One thing you might try, is taking a capacitor (.01uf or so, higher than PS voltage) and connecting one side to ground, use the other side as a test point to prod different parts of the circuit. The idea of this is the cap should suppress the oscillation by grounding it. Use caution as the cap could hold a charge. Try the 12AX7 grids, cathodes, and plates. Also after the 22k resistor on the last filter cap. It may or may not give a clue where the problem lies.

This sounds like some good advice - I was thinking that a capacitors placed somewhere in the cicruit to block the oscillation would be a way to fix it . I will be seeing my professor next week about the probelm to see what advice he might have to offer.....

 

[gusfinley]

oh yeah......

One odd thing that that this problem seems only to have to do with the position of the potentiometer. I can run a booster pedal and get a lot more volume out of the amp, and the volume level going into the amp doesn't seem to matter at all, but when the volume knob is turned past about 15K or so is when it starts to go crazy. I am missing a LOT of headroom, I can turn it up to about 1 or maybe 1.5 (if I had numbers on it)

 

[gusfinley]

some pics

updated pics to show the layout and all wiring.... Other than AC and filiment


http://cc.usu.edu/~davbradshaw/imghost/turretboard.jpg
http://cc.usu.edu/~davbradshaw/imghost/chasissfront.jpg
http://cc.usu.edu/~davbradshaw/imghost/chassisback.jpg

 

[gusfinley]

cab pics.....

For interests sake - here is a drawing or what the cab is going to look like.......

and a link to a bigger pic.......
http://cc.usu.edu/~davbradshaw/imghost/cab_big.jpg

 

[gusfinley]

alrighty........

The only part I couldn't quite make out was the terminal strip at the top. From pin 3 of the EL84(cathode), it goes to the one side of the parallel 25uf/210R on the outer lugs of the terminal strip. The other side of that appears to go to pin 7(?) plate of the EL84 via a small black wire. Something may be hidden or obstructing what I can see, but the low side of the 25uf/210R should return to ground/PS(-).

Yes, that does return to ground - it appears to go to pin 7 because I also used the ground point on that terminal strip for the ground reference for the two 100ohm resistors in the filament supply. Those are the brown blob that you can see near the filament pins on the EL84. There is no connection between the ground point and pin 7 of the EL84

Do you have any give to be able to lift the turret board up a bit? Some of the wiring in the output stage loops back under parts of the preamp gain stages. Noticably pin 2 (grid) of the EL84 (violet wire) and the leads of the output transformer. Possibly if you can relocate the EL84 pin 2 wire and putting a twist in the output transformer leads may help. Also not sure if you incorporated negative feedback in the circuit-couldn't see where it was if there.

I may be able to get a shot of of the wiring underneath. I may not have enough slack in the OT leads to twist them. I'll try to get a pic of the mess I made under there. Like I said, I originally had a pretty good layout but I overestimated the size of the chassis.

I don't have negative feedback in the design... from the kind of "distortion" that I get when the amp starts to freak out - well, its not the kind of thing that negative feedback will fix. I seems to be distortion about the distorted right around the biaspoint and about the time axis, rather than at the voltage peaks of the wave

Where did you get the transformers for the project? I've been thinking about building something from scratch, rather than rework old stuff.

I got my transformers from Angela insturments (www.angela.com) THey had some good prices - best I have seen on the net - and they have a huge selection of them!! while you are at it you can pick up tube sockets caps and a lot of other compnents....

I really wish that there a was a place that I could get everything for an amp - as it is now I thing I have 5 or 6 suppliers for the parts that I've been using.

 

[gusfinley]

alternate layout.........

I conjured up this alternate layout with photoshop.....
http://cc.usu.edu/~davbradshaw/imghost/layout.jpg

Here are some things that it would do...

- get first gain stage away from the speaker jack
- get the feed to the grid of the EL84 away from the OT primary wires
- get the leads from the input jack and volume pot above the turret board and away from the OT primary wires

I could also get those 2 220K resitors a little further from one another.....


and if possible, I'll have to check it out, I could move the OT behind the Power tranformer and eliminate the long OT primary lead running all the way across the board...... but where would I put the speaker Jack?

How much do you think any of these steps would help elimate the problems I am having?

THanks for your input guys....... I hope I can get this figured out soon. I really don't want to get a new chassis for the amp. I liked the smaller one so could put it in the cab that I posted a pic of........

 

[gusfinley]

uh........

come on guys, any feedback on this alternate layout? What are the chances that it will fix my problem?

 

[SBax]

I would try posting in the homebrew troubleshooting section of Ampage or maybe Weber's.

http://www.firebottle.com/fireforum/fireBB.cgi?cfg=ga&select=ganews&first_days_old=


http://www.webervst.com/amps/