Squealing/positive feedback - explanation of what is happening

[sleipnir]

Guys - I wrote a pretty long post last night with links to photos - I think it got routed for some sort of moderator review (maybe due to the number o links I included?). If it doesn't show up by this afternoon, I will give it another shot!

[Chuck H]

A lot of things can go wrong with long posts here. Not sure why you think it may be under some kind of review (unless you saw a notification or got an email to that effect). I haven't seen that in my twenty years here. More likely that post is just gone. Probably better just to go with a couple of photos you have and just post them without much fanfare. More photos and discussion can follow.

[sleipnir]

Thanks Chuck - I did get some sort of note stating that the post had to be reviewed, but based on your experience, I will assume it's just a technical glitch. I will break my post up into a couple of smaller ones...the good news is that I had copied the post before submitting it so all is not lost.

And for the record, I am not at all worried about you guys making fun of my wiring, soldering skills, and various melted insulation! I will take it all as constructive feedback :-)

First, please see my layout as it might be helpful. Note that I have now pretty much disconnected all of the wiring related to the mod switches located on the front panel. There are some switches which I have not disconnected: The switch between the rectifier tube V6 and power tube V5 allows me to switch in 5 zener diodes to lower my B+ (my PT is 380-0-380 but when I switch in the zeners, I get down close to the voltages found on the original 5F6-A schematic). The amp squeals in both high and low power modes and I have been running it primarily in low voltage mode lately while debugging. The two switches located between the power tubes are for pentode/triode mode and to change between 1K and 470 Ohm screen resistors (the latter might have been somewhat pointless to include as a switchable option - at one point I was going to use it do 470 Ohm or no resistors, but I have since read that might not be a good idea even though the JMP50 amps supposedly did not have screen resistors). Here's the layout:
http://i270.photobucket.com/albums/j...0mods%20v8.jpg

Here's what the amp looked like wired as per the layout above:
http://i270.photobucket.com/albums/j...202015%206.jpg

And after replacing capacitors, adding shielded wires to grids and presence pot, twisting OT primaries and secondaries, disconnecting the front panel switches, unsoldering and re-soldering various wires and components too many times while splattering solder and flux all over, this is what it looks like now:
http://i270.photobucket.com/albums/j...02/Chassis.jpg

[sleipnir]

You can see a few more close-up photos here:
Tweed Bassman 2 Slideshow by osingiusa | Photobucket

The bias circuit will look a little odd with its 3 50k pots - this is what I came up with in order to allow me to bias the amp in tube rectifier and low voltage (with zeners) mode, and also maintain proper bias in high voltage (no zener diodes) tube rectifier and SS rectifier modes. I have not touched this while debugging but don't how this could affect the squeal (maybe I am wrong?)

I was able to place the chassis upright so that it resembled how it would be relative to the guitar while mounted inside the amp cabinet. The squeal occurs when I approach the chassis from either the back or front with the guitar positioned as it would be strapped around my neck. From the front, when I get too near the OT, a much louder and more obnoxious squeal takes over, which I assume is normal and due to magnetic coupling taking place. I can record and post a video if this would be helpful, but the squeal I am trying to remove is the one shown in the video with the link pasted in an earlier post.

RG - I have tried a couple of different cables, and today I connected two 20ft cables using a simple junction box and found that it improved the situation a little bit. Presumably, the added cable capacitance reduced the highs enough to force me to move even closer to the amp to get the squeal.

RG - I think I can rule out acoustic feedback as the speakers are several feet away. I am curious about electrical field feedback and magnetic field feedback. The problem still exists when I shield the open chassis with a piece of sheet metal and aluminum tape (as in the video) and also have tube shields on the preamp and PI tubes, so I am not sure where the electrical field feedback is coming from. However, I experienced squeal with the current OT as well as with one wired to the amp but located outside the chassis on the bench and at one point I concluded that the feedback was coming from some sort of coupling between the speaker wire and the guitar pickups, which I guess would be electrical field feedback? Is there anything definite I can do to determine whether this is electrical or magnetic feedback?

I had made a video when I had the second OT connected and positioned outside the chassis, but I was testing with just and instrument cable connected and no guitar, so I don't know if this is relevant:
http://vid270.photobucket.com/albums...A%20squeal.mp4

I truly appreciate you guys taking the time to look at some of this and would love any further suggestions as what to do next. I would be happy to take photos from other angles as needed as I realize it is difficult to see everything given the messy wiring right now.

[g1]

Sometimes you get logged out while writing a long post. When you try to reply again, look for a "restore auto saved" button in the lower left area. You may get lucky and it may have been saved. Sometimes not.
When doing long posts, it pays to copy it before you push the post button. Then if it gets lost you can just paste it back in. (I see you did that!)
Or if you get the error message, and can still see your post, be sure to copy it before you log back in, just in case you need it.
Also, when logging in, check the "remember me" box.

[Justin Thomas]

^^^^^^^^^ WHAT G1 SAID!

From personal experience...

Justin

[Chuck H]

sleipnir,

Ok... I'll be the first to jump in. That's not a 5f6a layout. And it's actually hard to tell what's going on with some detail. The circuit is hard to trace because of how it is wired now. It does look like lead dress could use a lot of work. Those caps floating high are just more added grid lead length. You've used a lot of shielded lead in places that shouldn't need it. Not a big issue except that you've daisy chained grounds in many places so your shields can amount to larger static conductors at worst. And those daisy chained grounds can be especially problematic with gain. I know you see it on PCB designs, but those are actually thought out carefully and prototyped sometimes many times to get to a satisfactory working design. That is, not just any convenient path to ground is used and the order in which things are grounded when chained can be very important to stability. That's why I use individual ground leads for every circuit in my builds. I hate to say this, but I would probably strip and rebuild it with caps tight to the board, shorter leads in many places, shielding to the input grid only, individual grounds and make sure all OT leads approach from the PA side of the board, including the presence control/nfb lead. Distinctly as far from the preamp as possible. A cleaner build makes for easier troubleshooting because it's easier to eliminate the little things, Not to mention being able to trace the circuits. Floating leads and connections and using the junction of multiple shields as building posts are really icing on the cake WRT difficulties tracing the circuits and grounds. If the trouble you're having is just one thing I still think it would be very hard to pinpoint in there. Try not to take this too offensively, but... I'll liken to walking into a really messy kitchen and there's an "odor". You take out the trash but it's still there. So you scrape the old food off the top dishes but it's still there. You know then that you're going to need to clean the whole kitchen before the odor is gone.

[sleipnir]

Chuck - I definitely do not plan to leave the wiring the way it is now, the floating wires and all of those shielded grid wires will be replaced. I am not sure if you had a chance to look at the picture of the amp as it was wired before I started replacing stuff left and right, but the caps were pretty close to the board and the wiring was at least more orderly than where I am at right now. When it was wired that way, I got the squeal, and since then I have tried a lot of different things. As I went along, I did not undo changes which is why the amp looks the way it does right now. Perhaps that approach is bad in that so many things are now different and could be preventing me from actually finding a fix (if there is one). Maybe I will start by getting the 5F6-A part of the circuit back in order while still leaving the mod switches out.

If you look at my layout, do you see any issues with the grounding scheme there, or only with the current state of the amp? I have isolated pots and jacks on the front panel, except for one input jack which serves as the preamp ground and to which the rail on the back of the pots is grounded. Only the input grids are shielded, although that might not be clear from the layout. Are you doing more than a preamp and a power amp ground in your builds?

I will see if I have enough lead length to move the OT secondaries to the other side (would require a new hole, but I can maybe remove the speaker wires for now while testing). However, I don't believe this is an issue as I had squealing even with an external OT and the OT secondaries and speaker jack outside the chassis. It is worth a try, though! Since the squealing is even worse with no negative feedback, I am not sure the feedback wire itself is a problem, but again I cannot say that with 100% confidence.

Thanks!

[Tom Phillips]

Please attach your photos to the forum posting. That way the whole thread is preserved for future reference instead of dead offline links when someone reads this thread later (maybe much later after your photobucket account is long gone).

Regulars - I know that there was a loss of attachments at MEF a while back but I think that overall that loss was less than the quantity of dead links in the historical MEF record due to dead photo sharing accounts.

Cheers,
Tom

[Chuck H]

Well I just looked at the "before" shot. Nice and tidy One of the mod switches does look like it has the circuit crossing back over itself, but the switches aren't there now so that's moot. As are some of your try to fix add ons like the shields. So we are indeed looking for something that would result from the initial build even without the mod switches.

So I only have two suspects right now. I'll look over the grounds as best I can see and if anything looks potentially problematic. Other than that I'd like you to test the PS filters to make sure all nodes are in out of phase pairs. Nothing about your skills, but mistakes happen and those connections aren't all apparent in the pics.

[Chuck H]

Looking over the pics and layout more I just don't see anything about the grounding that looks so bad. Daisy chains look to be limited to cathodes really. I don't prefer to ground the presence control with the preamp, but that's not your problem. It is very hard to see much in the pics with black wires in the shadows and a lot of leads under the board. The layout shows a single 8uf filter for all four preamp triodes. I probably wouldn't do that either, but the way the circuit is arranged I wouldn't expect it to cause your problem. There's a switch that controls the "speaker in" circuit. I'm not even sure what that circuit is because I can't read the switch writing and again, the leads aren't visible. It looks to be on a rotary switch with impedances marked that is other than the output impedance switch. I'm confused. But I can tell you this. If those leads have anything to do with the OT secondary then you need to get them entirely away from the preamp. As in that switch cant be in the middle of the tone stack with leads running across or under the preamp. Bad stability juju. But that's all I have right now.

[sleipnir]

Chuck - thanks for looking at this further! Regarding the power supply filters - can you maybe elaborate on what I should be looking at with regards to the nodes being in out of phase pairs? Just to explain what I did - I differ from the 5F6-A filtering scheme in that for the first node (where the OT center tap and one side of the choke are attached) I use two 100uF 350V caps in series with 220k balancing resistors in parallel to the caps to increase the voltage rating to 700V (my power supply without the zener diodes is 380-0-380 resulting in B+ voltage near the voltage ratings of single capacitors). The resulting 50uF capacitance is a little higher than the 40uF in the original 5F6-A with two 20uF in parallel. For the second node, which is right after the choke and powers the output tubes, I use two 47uF 500V caps in series again to get a higher voltage rating. The resulting capacitance of 23.5uF is close to the original design which uses another 20uF cap. For the third node, which powers the PI tube, I have a 22uF cap similar to the 20uF used in the original design. Then, as you pointed out, I have an 8uF cap for the final node which goes to the two preamp tubes. This is exactly how the original 5F6-A amp design from what I could find. Can issues with the powering perhaps be leading to oscillation? This is not something I have explored much, and I have also not made any efforts to move the under the board high voltage wires. I could disconnect them and run temporary wires above the board if necessary.

Regarding the speaker switches. The one closest to the PI is the impedance selector and there are wires coming out of the bottom of the chassis right near it and also relatively close to the PI tube and its circuitry. I am using the same hole for the secondaries as the original 5F6-A, although my PI circuit is slightly shifted to the left compared to the original amp (due to my adding some additional switchable caps and resistors earlier in the circuit). There is a chance that this shift in the PI circuit (and thus tone circuit) is an issue due to its relative position to the OT secondaries or other wires, but since I had problems even when I used an external OT and had moved the speaker switch out of the chassis entirely, my gut tells me this may not be a problem. The wire that runs from the impedance selector under the board in the schematic is the NFB wire - it goes to a switch allowing me to disconnect the NFB altogether or select between two NFB resistors. However, the way I have it wired now, this under the board wire and the switch are out of the circuit and only the 27k NFB resistor is in circuit. I was always a little concerned about all the wires in my design running across the board, which is one reason I disconnected those first assuming they were perhaps the culprits.

The second speaker switch, the one closer to the V4 output tube, is currently out of the circuit. When in use, this switch allows me to choose between the speakers in 2 Ohm or 8 Ohm mode. Essentially, I wanted the ability to use this amp as a typical 2 Ohm tweed Bassman amp using the 4 speakers in the cabinet. However, I also wanted to be able to run the amp to an external speaker cabinet (hence the multiple tap OT) or use the 4 speakers in the Bassman cab as an external cab for another amp. That is why you will also see the two speaker jacks. The speaker in is where you connect a speaker cab, and the speaker out is where you would connect another amp in order to use this cabinet's speakers. When using the amp as intended, I just run a small speaker cable from one jack to the other.

If I have some time tomorrow, I will reverse some of the "fixes" I had attempted (i.e. get rid of shielded wires, put back the previous caps, shorter wires, etc. while still leaving out the mod switches. I can then take some new photos and post them directly here (assuming I don't end up with image size issues). It might be easier for you to see what's going on with a cleaner setup.

Thanks again!

[Tom Phillips]

...I can then take some new photos and post them directly here (assuming I don't end up with image size issues)...

If you have questions about attaching files to your posts there is a FAQ section on the subject at Music Electronics Forum FAQ That link will take you directly to the section about managing attachments.
Tom

[Steve A.]

sleipnir:

In one of your earlier posts you mentioned not wanting to add pF caps to preamp tube sockets- that is one of tools we use to reduce or eliminate oscillations. I recommend putting together an assortment of 1kv pF ceramic caps, like between 12pF and 150pF for tube sockets.

Someone mentioned that your layout is not a 5F6A (I didn't look.) Fender's layouts can be critical like getting the transformers oriented properly. Every time you add an extra wire or switch you increase the chances of getting oscillations or other gremlins. Copying the original grounding scheme can be important.

Moving leads around with chopsticks is one trick to improve lead dress.

Extreme settings of controls can contribute to problems, like turning the volume and treble up to 10. I would also guess that you are using a 12AX7 for V1, and probably a high-gain one at that.

Just some random thoughts..

Steve Ahola

P.S. You might want to check out the tube amp debug page at R.G.'s site

http://geofex.com

EDIT I just looked at some of your pix. Although I love Orange Drop caps they can act as transmitters and receivers of signals within your chassis. Sometimes I have to wrap insulated wire around the body and connect it to ground or another circuit point to quiet things down.

BTW the two leads on film and foil caps like Orange Drops are not interchangeable. For optimum results the lead that goes to the outer wrap of foil should go to the circuit point closer to ground. (I'm not sure about the smaller caps but on the 600vdc caps I use a visual inspection will reveal the outer foil lead.)

If it was me I'd probably revert circuit to that in the original Fender layout and add in your mods one at a time, correcting problems as they arise.

[Chuck H]

Chuck - thanks for looking at this further! Regarding the power supply filters - can you maybe elaborate on what I should be looking at with regards to the nodes being in out of phase pairs? Just to explain what I did - I differ from the 5F6-A filtering scheme in that for the first node (where the OT center tap and one side of the choke are attached) I use two 100uF 350V caps in series with 220k balancing resistors in parallel to the caps to increase the voltage rating to 700V (my power supply without the zener diodes is 380-0-380 resulting in B+ voltage near the voltage ratings of single capacitors). The resulting 50uF capacitance is a little higher than the 40uF in the original 5F6-A with two 20uF in parallel. For the second node, which is right after the choke and powers the output tubes, I use two 47uF 500V caps in series again to get a higher voltage rating. The resulting capacitance of 23.5uF is close to the original design which uses another 20uF cap. For the third node, which powers the PI tube, I have a 22uF cap similar to the 20uF used in the original design. Then, as you pointed out, I have an 8uF cap for the final node which goes to the two preamp tubes. This is exactly how the original 5F6-A amp design from what I could find. Can issues with the powering perhaps be leading to oscillation? This is not something I have explored much, and I have also not made any efforts to move the under the board high voltage wires. I could disconnect them and run temporary wires above the board if necessary.

Regarding the speaker switches. The one closest to the PI is the impedance selector and there are wires coming out of the bottom of the chassis right near it and also relatively close to the PI tube and its circuitry. I am using the same hole for the secondaries as the original 5F6-A, although my PI circuit is slightly shifted to the left compared to the original amp (due to my adding some additional switchable caps and resistors earlier in the circuit). There is a chance that this shift in the PI circuit (and thus tone circuit) is an issue due to its relative position to the OT secondaries or other wires, but since I had problems even when I used an external OT and had moved the speaker switch out of the chassis entirely, my gut tells me this may not be a problem. The wire that runs from the impedance selector under the board in the schematic is the NFB wire - it goes to a switch allowing me to disconnect the NFB altogether or select between two NFB resistors. However, the way I have it wired now, this under the board wire and the switch are out of the circuit and only the 27k NFB resistor is in circuit. I was always a little concerned about all the wires in my design running across the board, which is one reason I disconnected those first assuming they were perhaps the culprits.

The second speaker switch, the one closer to the V4 output tube, is currently out of the circuit. When in use, this switch allows me to choose between the speakers in 2 Ohm or 8 Ohm mode. Essentially, I wanted the ability to use this amp as a typical 2 Ohm tweed Bassman amp using the 4 speakers in the cabinet. However, I also wanted to be able to run the amp to an external speaker cabinet (hence the multiple tap OT) or use the 4 speakers in the Bassman cab as an external cab for another amp. That is why you will also see the two speaker jacks. The speaker in is where you connect a speaker cab, and the speaker out is where you would connect another amp in order to use this cabinet's speakers. When using the amp as intended, I just run a small speaker cable from one jack to the other.

If I have some time tomorrow, I will reverse some of the "fixes" I had attempted (i.e. get rid of shielded wires, put back the previous caps, shorter wires, etc. while still leaving out the mod switches. I can then take some new photos and post them directly here (assuming I don't end up with image size issues). It might be easier for you to see what's going on with a cleaner setup.

Thanks again!

I attached this incredibly long (but understandably necessary) post so it would be clear what I was responding to ...

So those are indeed OT secondary's across the preamp to enable the "speaker in" circuit. Even though the switch is gone, do any of the wires remain? OT secondary's are exactly what you need to keep away from any preamp leads. This is so critical that many builders have trouble even with careless routing of the NFB/presence lead from the usual impedance selector switch. Super important that anything about the OT isn't close to the preamp. Some more contemporary designs do have the OT on the preamp end of the chassis. But the design is such that only NFB (rather than PFB) is the result. This is a careful and tricky business. Much more typical of invading the preamp with OT leads is instability. Unless you have the time and budget to design multiple prototypes to this end it is always best to simply keep the OT leads entirely away from the preamp. So do check for any leads that are still present near the preamp. Regardless of whether the switch is connected. If the leads are there it could be a problem.

Unless you are running the triodes of the first preamp tube in series for a cascade gain stage the power supply as it is wired shouldn't be a stability problem.