Troubleshooting Selmer T'n'B Mk III

[EFK]

I've got this old Selmer head that I've been trying to kick start. When I got it, all of the main filter cans were replaced and an adjustable bias pot was added. One power tube socket is a replacement along with filament wiring to power tubes only. Standby switch (stdby-off-on) added. New rectifier diodes (16A 1200 V hexfreds). All of the old mustard caps (save one) are still in place and much of the wiring etc. is original. I had some problems biasing, initially, but have managed to get it biased. With NO input cable and both volume pots all the way down, there is still a fairly annoying hum which comes through the speakers. When I bias via OT shunt, with black test lead on center tap and hot lead on the respective plate, reading direct mA, the hum completely vanishes. Amp works - bass channel (2 inputs) seem fine, but when I plug into treble channel inputs the hum gets louder and sound is kind of weak. Also, when plugged into treble inputs, if I work the volume knob on any guitar it is very scratchy (oddly enough the amp knobs do not seem scratchy). Some of the hum may be AC filaments, but what's up with the treble input? Anyone have any ideas where I should start? Input jacks/resistors are all original but look ok, pots are almost all original but also seem to work fine. I enjoy working on these (carefully!) but am not particularly knowledgeable when it comes to troubleshooting so I'm a bit at a loss as to where to start. Anyone have any ideas? Thanks!

[melvin]

Do you have a schematic?

First of all, there should be no sound coming out of the amp when using the shunt method to bias the amp. So that's normal.

As far as the hum... Are the new heater wires connected each to the same heater pins on the output tubes? (pin2 to pin2) Is the bias supply cap new as well? Did the power supply resistors get changed with the cans? Are any preamp tubes noisy?

Most likely though, it sounds like the input jack and connections in the treble channel needs to be resoldered. And if that is so, there may be other solder joints that are adding to your woes....(ground connections?)

[Enzo]

My first thought is the amp is unstable and oscillating at high freq, above audio. The result is usually hum along with a weak signal. A scope to the output would detect that.

You may have replaced all the B+ filters, but did you replace the bias supply filters? If the old B+ caps were in a can, did you ground hte new caps to the same place? As opposed to grounding them here and there?

[EFK]

Thank you both very much for the initial thoughts! I will check the filament wiring and ensure whoever rewired it kept 2 on 2 and 7 on 7. Should be simple as they're two different colored wires. The filter caps and bias cap all look pretty new. I didn't replace them so I don't know how long ago they were replaced but they sure look new. The power resistors in this amp are all 27K and they look like they are the original resistors with some new solder on them. I will take a harder look at the inputs and the treble channel soldering. All the input jacks are original and I have to admit that when plugging in they feel kind of loose - however, it didn't seem to make any difference if I wiggled the cable input around. The scratchy guitar volume pot was interesting, though - is that DC leaking back into the guitar wiring?

Enzo I didn't do the cap job so I don't know if the grounding scheme is different than when new. I'll attach a schematic I found online which does seem to match up as best as I can determine with what is in the amp. I think what I would like to do first and foremost is to check all the grounds I can find - some look recently resoldered but some are definitely original soldered grounds. Anything that involves a scope is going to not involve me because I don't have one! If I can't figure out the problem myself with a meter and soldering iron, I'm going to have to send it out to someone or drive an hour + to hit civilization.

Is there a safe and/or effective way to check all the coupling caps for leakage while they are in-circuit? I'd like to do that too but I don't like to just helter skelter start poking around with a meter in there.

[Steve Conner]

Have you tried swapping the preamp tubes around? A scratchy guitar volume often means a bad first preamp tube. Try swapping the tubes between the normal and bass channels, and see if the fault follows the tube.

The OT shunt biasing method shorts out the OT, so it's hardly surprising that the hum disappears. Are the tubes matched? (ie, similar bias reading for both sides?) If they're not, the hum will be worse. These amps have quite a well filtered power supply, so they should only hum a little.

I had a T'n'B 50 SV, they're great amps. I ended up rebuilding it completely and turning it into a combo.

Where are you? These were British made amps, and it would be interesting to see where they ended up.

[EFK]

Steve it was tubeless when I acquired it so I added all brand spanking new tubes. The (2) EL34s are biased within about 1 mA of each other so they seem pretty well-matched. I'm in northcentral PA in the middle nowhere, meaning it has definitely wandered quite a distance from home!

Is it normal for the preamp voltages to be quite low on these? The plates on V1 were pulling only @ 114/116 V, V2 133V, V3 132/119V and V4 209/199V. Maybe I don't know what I'm talking about but those seem low for a 50W amp that supposedly does not dirty-up very fast.

[EFK]

Well so far I've found a bad 1M resistor on the treble inputs. Also, the power resistors which are supposed to be 27K have drifted (apparently) upwards to about 32-33K by now.

Take a look at the attachment below: what is the purpose of the 1M resistor in-line with the signal *after* the two 470K resistors?

[Gtr_tech]

Check for DC on any grid in the weak/noisy channel. If you have a leaky coupling cap it'll cause exactly what you describe.

[melvin]

What is in the circuit at that spot?

[EFK]

Despite some mods to the bias (added pot) this circuit is essentially intact; it matches the schematic. Where I indicated the 1M resistor above, there is a 1M resistor. Of the few other amps I've worked on, Gibson and Marshalls, there is nothing following the mix resistors so I'm wondering why this 1M resistor was put in the signal path b/t the two channels and the next gain stage (via the echo loop). ???

[EFK]

I have the amp biased and the inputs sorted-out, everything meters out fine there, and all of the grounds have good continuity. But, there is still a godawful loud hum as soon as I take it from standby to power. I would REALLY like to check caps for leakage but I'm not sure how - do I need to desolder all of these? Can I check for DC somehow on the grid side of any cap without sparks flying? Frankly I'd just replace them - there's really not that many - but part of me hates to mess with the old hallowed "mustard" caps and I'd like to keep them there if possible.

A stupid question: if the circuit specifies a 1M pot, and I take a resistance reading across the outer legs while the pot is in-circuit tied to tone pots, should it still read about 1M? The pots all seem to work but I get some REALLY funky readings off them.

Any HELP!!!!!! very much appreciated!

[melvin]

The tone pots should give you strange readings in circuit. Is the hum just momentary or is the amp always humming? If not the latter, it could be the standby switch or any number of things.

From the schematic, the 1M resistor should be safe to shunt. It may have been necessary for the echo in/out but I don't see why. I would think the amp will sound a lot better without it.

[Wakculloch]

As the bias circuit is the area where mods have been done, i would be tempted to check it over and make sure the smoothing cap is OK and there is no ac there.

[Steve Conner]

To answer the earlier question: The 1M forms a potential divider with the 47k, to lower the signal level to something more suitable for an outboard echo unit. Or you could say, to lower it to a level that won't overdrive the loop recovery stage.

The designer could have achieved the same by changing the two 470k's upstream of it to 2.5M, or changing the 47k to 10k. I'm not sure why he added this extra resistor, unless he didn't want to buy 2.5M just for that one position. IIRC, my T'n'B had the same circuit and I thought it was weird too. One of the first mods I did to mine was to change it to a parallel FX loop, with higher level to drive a digital reverb, and a cathode follower (there was a spare triode in there) driving the send.

How does the amp perform if you go straight into the "echo in" jack, disabling the preamp? Is the hum and weak signal gone then?

Does it still hum if you pull the PI tube? If so, we're probably looking at badly mismatched power tubes, bad filter caps, shorted choke, and/or a problem with the bias circuit.

[EFK]

I like a fairly straightforward circuit so I cut out the echo loop and did away w/ the 1M resistor. The less in the signal path, the better. Chased the hum around today, actually I'm not sure if I should describe it as a hum or maybe motorboating. Pulling all the preamp tubes, the amp is dead silent. I mean, REALLY quiet. Putting them back in one at a time, the very loud noise kicks in when I plug in V2. If I pull V2 but leave all the other pre tubes in place, the amp is dead silent. If V2 is left in place, as soon as I kick it from standby to full power, the noise starts low and quickly - about 3 to 4 seconds - gets VERY loud. Like really loud TV loud, through the speaker. It actually vibrates the amp chassis and speaker. V2 uses only half the triode, and the solder joints all look good. I cleaned the socket and tried different tubes but it makes no difference. It kind of sounds like a hummy motorboat. Not a slow 'put put put' though, a pretty fast one. This is with nothing plugged into input. The cathode cap on V2 is new and oddly enough the only coupling cap in the entire amp which has been replaced is this one - the .022 uF (see above schematic) off the V2 plate is some pink russian looking thing - K40Y or something like that. Looks very new. So where do I go from here? The preamp voltages seem slightly low to me, but I don;t know what they are supposed to be so maybe they are normal. Voltage on V2 pin 1 is 133V and it stays stable. So where do I go from here? I've never had this problem before!

[Alex R]

One of the things I'd try (in my reckless way) is grounding the grids of V2 - not on the pins though as there's a big danger of grounding the plate voltage and flash bang! Go back up the grid wires and ground the other end. Use a coupling cap as a probe and a croc lead to ground to be safe from accidentally grounding HT. This sounds like an oscillation problem and if you can stop it by grounding the first input grid then we have isolated the problem to one small area.

I am wondering whether you have lost a ground reference somewhere in the input area, allowing a grid voltage to drift upwards. Those 1 meg resistors on the inputs? I seem to remember some of these amps don't ground out the input contacts when the jacks are out - so if you lost the ground ref resistor too that might cause such a problem.

Agree with Steve that these are great amps. I have a couple of pro customers who use them in preference to anything they can get via endorsements etc, on sound quality and reliability.

[EFK]

I grounded out the input grid of V2 with an electrolytic cap and the amp went absolutely silent. I mean dead silent. What does this mean? Only half of V2 is used and that half is the final gain stage for the signal from both V1 (bass channel) and V3 (treble channel). If I leave V2 in place and listen to the loud noise, and then ground out pin 2 (from inputs) on V1 or V3, there is no change. If I ground out pin 7 on either V1 or V3 at the volume pots, the sound veeeeeerrrry slightly gets quieter, but barely. Hardly noticeable. But ground input to V2 and...sweet silence. I replaced the grid wire to V2 with shielded wire and it made no difference at all. Now what? Unfortunately I can conduct these tests easily and safely, but I do not know what these results mean!!!! Anyone? What next? Thanks VERY MUCH for all the help so far.

[Gtr_tech]

Look for DC on that grid.

[EFK]

So using a meter, do I just hook black lead to chassis ground and start checking the grid wiring and grid side of coupling caps with the red lead while the amp is running? It should all be negative voltage, correct? Any positive DC = bad? Is it absolutely necessary to desolder coupling caps to do this?

[Gtr_tech]

Originally Posted by EFK

So using a meter, do I just hook black lead to chassis ground and start checking the grid wiring and grid side of coupling caps with the red lead while the amp is running? It should all be negative voltage, correct? Any positive DC = bad? Is it absolutely necessary to desolder coupling caps to do this?

Gnd lead of the meter goes to chassis gnd. Amp on and off standby (play). You should see no dc on gain stage grids....they should be at or *very* near ground potential. Even a couple volts is enough to throw off the operating point of a small triode. Don't disconnect anything when doing this check. Beware the meter probe will cause a fairly loud pop in the speaker when you contact a grid connection. Expect it and it won't startle you as much.

[Alex R]

You've now isolated the problem to the general area of the echo send/return circuit. Since you made some changes there, particularly that 1M resistor, it makes sense to check that area out, and try undoing the changes.

That is, unless the noise stops if you leave V2 in and pull either V1 or V3...

[EFK]

I will test as described above. Thanks!

Alex, I cut out the echo loop in an attempt to find the hum/motorboat - it was already making the loud noise. Didn't change anything. I cut out the loop just before the 1M resistor that follows the 470K mix resistors and tied it back in just before the .022uF cap that leads into the input of V2. I do want to test that .022 cap in-circuit although I tested it for capacitance with an LCR meter when I had it out and it was almost dead-on .022uF, and I also tested it with a battery to see if it would pass DC which it did not.

(BTW, isn't that .022 a little odd or redundant, especially considering that you have caps just before the 470K mix resistors anyway?)

[EFK]

Well now I think it's not really V2 at all. When I pull V2 the hum is largely gone but still faintly there. When I pull PI it is silent. So I tested all the grids on V1, V2 and V3 and have no DC voltage on any of them. But, I hit the grids on the PI and I have 7.58 VDC on pin 2 and 6.97 VDC on pin 7. When grounding pin 2 it is almost quiet (still faint) but when I ground pin 7 is is absolutely dead silent. So which caps are suspect here? Should I suspect the two .047 caps to the output grids, or should I suspect one or both of the .1 caps down on PI pin 7 (can't see from the schematic how they would tie into pin 2???). Getting closer I think! Have my fingers crossed.....

[EFK]

BTW, does anyone have a photo of the board in one of these? Specifically, I'd like to see a good detailed photo of the PI section, caps and long tail pair etc. Thanks!

[Steve Conner]

Here are a couple of pictures of my SV, back when it was relatively unmodded. I may have some clearer ones, but they show a lot of mods that may confuse things.

They're full resolution as they came from the camera.

[EFK]

Steve thank you for taking the time to post those. Those SV amps are considerably different, apparently. Chassis layout is definitely changed-up quite a bit. I found some internal pics on the vintagehofner site (which has a lot of very good information regarding Selmers) but they're not the most detailed quality. Still, I can see that it appears mine is wired correctly. I'm a bit at a loss to explain how 6 to 7 volts DC is getting on my PI grids but whatever is causing it would seemingly be the problem. I replaced one of the .1 caps - the top cap in the schematic into pin 2 - and it made no difference whatsoever. The .022 into pin 7 looks brand new (for a mustard) and tested good (out of the amp w/ a 9 volt battery). So where is this DC coming from? I guess I'll try replacing the other .1 to ground and see if that makes any difference. Anyone have any more ideas or systematic approach to tracing back this DC I'm all ears!!

[EFK]

Take a look at this poorly-drawn sketch (attached). Which diagram matches up with the schematic, and does it make any difference either way? I found two blurry pictures of the board on a Mk II and Mk III, and they appear to be soldered up as per diagram #1. However, when I acquired this amp, mine is soldered up as per #2 (which seems easier given the size of the .01 caps).

Opinions?

[Steve Conner]

The two diagrams are electrically identical.

Am I the only one who thinks that 6-7V DC is nowhere near enough for a LTPI? The grids should have something like 20V on them. The earlier remark about checking for DC on grids doesn't apply to the LTPI, the DC is a vital part of its operation.

The SV and MkIII are almost identical, but the schematic for the SV has voltages marked, and the LTPI cathodes say 20V, so the grids should be around 18V.

So now we have two symptoms, lack of voltage and hum. Both could be explained if the electrolytic caps happened to be bad, and since they're known to go bad with age, this is a possibility you need to check out.

Selmer TnB 50w SV Amplifier Schematic - SV
Selmer Tn'B 50 MkIII Schematic - MkIII

So, I'd replace all of the 32uF caps as a matter of course. The ones labelled "6x 32uF" on the Vintage Hofner schematic. Can you post some pics so we can see if they've already been done? 32uF is hard to get, but 47 will do.

Did we check whether the amp plays properly with no hum, if the "treble" channel preamp tube is pulled?

Alex R: If you're reading this, the channel I kept was the "Bass" one with the full size coupling cap. Is that a bad thing, would the "normal" one be considered better sounding?

[EFK]

Steve and everyone - I can;t thank you enough for bearing with me. I'm sure it's tough to try to diagnose from a distance, especially when I don't really know what I'm talking about.

This has three cap cans, 2 which are 2X32uF and the mains which are 2X50uf (someone must have upped the values). I replaced both 32uF cans today, although they were F&T and did not look very old, as well as installed new monster 27K 5W resistors off the cans, and it made NO difference at all. I did not replace the main 2X50uf "Ruby" can as I swear it looks like it was installed yesterday, it's that new. For giggles, I also used a continuity tester on the preamp pins, one end on the respective grid pins then touching all the others in turn, just to reassure myself that none of the pins were shorting out with each other. Still VERY loud hum, 6 to 7 volts Dc on PI grids. There is absolutely no DC on any of the earlier preamp grids, V1 through V3. But, I'm getting that 6 to 7 VDC coming into the PI off the .022 cap which feeds the PI from V2 (last gain stage). I replaced that cap - no difference at all. (How is the DC getting through that cap?) I think the hum might be coming from either V2 or somewhere earlier than V2 (V1 or V3) and the added gain on V2 is amplifying it. IF I touch any of the grids on V1 or V3 with the meter, no voltage and nothing changes. If I touch the V2 grid (one half is used only), the hum gets even louder, almost a double tone hum, and I can feel the vibration right up the probe. But, there is no DC voltage there. I checked for DC coming through any of the coupling caps on V1 or V3 while the amp was running and got absolutely nothing. To reiterate, all of my grounds in the amp check perfect.

This is driving me crazy!!!!! Help!!!

Is there any way something in the pots for either channel could be causing this? I've got replacements lined up. Honestly, I've got enough replacement parts to rebuild the entire preamp but I would REALLY like to figure this out rather than just blindly replace things.

[Steve Conner]

Oh, well if they were F&T, someone has already replaced them. F&T haven't been in business long enough for any of their caps to need replacing yet, AFAIK. How about the bias filter capacitor? has that been done?

Another few things to check: This was a British amp, so it would have been set up for 240V line voltage and a three-prong power cord. (We've always had three prongs over this side of the pond.)

Now, if the voltage selector were set for 240V, that would explain the low voltages on the PI grids, and if an American 2-prong cord had been unthinkingly grafted onto it, that would explain the hum, since the amp would be completely ungrounded.

So check the voltage selector is set to 120V. (should be a little plastic dial on top of the PT if it's like mine?) And check that it has a three-prong power cord with the green wire connected solidly to the chassis. Also check that all of your outlets on your workbench have functioning grounds.

Back on the component front, check all of the components and connections around V2, and try swapping V2 with one of the other tubes, in case it's bad.

[EFK]

Steve I feel like I'm running up against a wall. I did check the PT as soon as I got the amp and it is set to 120. Today I replaced the main 2X50uf can; now it has three brand-new cap cans, and just for the hell of it I replaced the 100uF/100V bias cap. Also rechecked all the ground straps and resoldered a few just to satisfy myself. Can you guess? (Drum roll please.....) - no difference. If I ground out the PI grids or if I ground out the grid side of the .022 cap feeding the PI, sweet silence. Otherwise, I have a beautiful vintage boat motor. I also replaced both volume pots but I've noticed that if I fiddle with them while the amp is on and singing its lovely humming song, there is crackly static here and there. Both channels - so there is DC getting through to the pots? I replaced those tiny 220pf and 500 pf caps, but I guess I'm going to have to say goodbye to those lovely mustard coupling caps too. I can't think of what else to do.

If there was a grounding problem in one of the LTPI resistors, could it cause this? And if so, which resistor would be the likely culprit? One of the 1M?

BTW - anyone know how this power transformer is supposed to 'read?' It's a lay-down, and the terminals are numbered 1 to 10. What would happen if the rectifier diodes were installed incorrectly? As I mentioned somewhere previously I think, someone replaced the BY128 called for on the schematic w/ two hexfreds. The two negative sides of the diodes are joined and soldered to the red wire which runs to one of the + terminals on the 50uF can, and then the two positive ends are respectively soldered to PT terminals #5 and #3. Aside from the hum everything seems to work properly so I assume they are installed correctly, but just thought I'd doublecheck in the event that someone might be familar w/ this transformer.

This amp has the biggest replacement 3-prong power cord I have ever seen in my life. 14 ga wire was used. Trans-atlantic cable anyone? Whoever did it got the colors wrong but it is just a piece of cable with a replacement screw-together plug as well, and I checked the wire routing and things work out to where they are supposed to be so the colors don't really matter. A standby switch was installed also and it all works properly.

[melvin]

I think you should go back to the preamp section. I still haven't read whether or not you have tried pulling only V1 or V3 to see if the hum goes away. And it would help if we knew whether it was hum (60 or 120Hz) or if it was "motor-boating" which usually gets worse as you turn up the amp and is some kind of parasitic oscillation.

If it's a parasitic, and it involves something before the PI, you could try shortening the pre-amp grid wires, but more importantly, grounding the last filter caps before each preamp section near the cathodes of their respective gain stages. You might also try replacing the 27K resistors they are attached to, assuming you haven't already done so...

But tell me how it goes when you pull V1 and V3.

[Steve Conner]

Also, can you measure your main B+ (at that red wire going from the diodes to the cap can) and heater voltage, and get back to us with the results? Should be about 420V and 6.3V. We still haven't determined why the PI grid voltages are half of what they should be.

I'm beginning to suspect a grounding SNAFU. When this replacement power cord was installed, where was the ground wire connected?

And when the 2x50uF cap can was replaced, what about the negative wire of it? You should find that a wire runs from the negative directly to the PT center tap, maybe via the HT fuse holder. And another wire should run from the cap negative to the rest of the amp's ground system. The wire from the PT center tap should not share any other grounding function. This keeps charging current out of the ground system.

If the rectifiers were wired wrongly, it would most likely blow fuses or destroy the PT.

[EFK]

OK guys this is a lot so forgive me for being terse: (but not with the THANKS for hanging in there with me!!!! Thank you!)

Off the main 2X50uF can negative terminal:

(1) Black to HT fuse and jumper from black to a chassis ground. Purple off tail of HT fuse holder to terminal 4 on Power Trans.

(2) Another black wire to terminal 2 on PT, which jumps over to terminal 7 (which is between 6 and 8 which are filament taps), then jumper from terminal 7 to a joined ground with both 32uF can negative terminals.

(3) Ground strap to brass plate (across front behind pots) which has all the preamp ground straps ties to it (the brass plate). Positive end of bias cap is tied to this ground strap.

(4) Bias resistor off the adjustable bias pot is also tied to negative term. on main 50uF can.

I have 485 B+ directly off red wire from 50uF can to PT diodes. 6.1 filament voltage on all socket pins.

Ground wire on 3-wire cord is attached to chassis w/ it's own ground bolt and soldered. Tested good for continuity.

Negative feedback wire runs from 7.5 ohm jack to tag on board, then 100K resistor goes on from there. Should there be a ground continuity between the chassis and the junction of this wire and the resistor? There is.

Filament taps and chassis have ground continuity.

Merlin will let you know about the tube pulling shortly - I already did this but unfortunately have forgotten what happened. I have swapped in about 10 diff. tubes and no difference.

[MWJB]

If wired stock ALL your filter caps garound to the samp point. Separate the PI & preamp filter caps and run ground wires to the buss wire at the input jacks. Only the main & screen caps should be grounded to the chassis where the caps are physically situated.

If you are familiar with Fenders the filter caps may be grounded like them - e.g. PI cap is grounded to buss wire about halfway down the chassis, preamp cap goes to bus @ input jacks...but those Selmer buss wires are quite busy, so both caps could go to the jack end, try both ways...

Alternatively, keep both volumes turned up at all times.

This hum is a normal "feature" of Selmers that I have seen.