best way to quickly find cause of distortion, 1955 twin

[centervolume]

Greetings -

I have an old tweed twin I've been cautiously restoring (re-coned speakers, recapped). All voltages look good but it goes into distortion too quickly, lacks clean headroom and overall power as a result. Given the age and value, I really want to be non-invasive as possible and locate the problem. I have heard a couple of these amps and they've been much louder before breaking up. Should I trace the signal with a scope to find if the signal starts to degrade at x location? one clue I have is that I think it's running through tubes at V3 where the inputs are summed before the tone controls. Any recommended approaches appreciated:



http://www.kbapps.com/audio/schematics/tubeamps/fender/twin5e8a.php

[sluckey]

Change tubes first. Then check voltages at all B+ nodes and ***ALL*** tube pins, even the pins that measure zero volts. Compare to voltages on this schematic/layout. Post voltages here if you need a hand. Signal tracing with a sig gen and a scope is a good idea too. Inject a 200mV 500Hz signal into the input jack and chase it through the amp.

[centervolume]

Change tubes first. Then check voltages at all B+ nodes and ***ALL*** tube pins, even the pins that measure zero volts. Compare to voltages on this schematic/layout. Post voltages here if you need a hand. Signal tracing with a sig gen and a scope is a good idea too. Inject a 200mV 500Hz signal into the input jack and chase it through the amp.

I'm happy to have a hand for external verification if nothing else. All tubes replaced with known good as per tube chart except using 6L6GC rather than 6L6G. I'm also using a home made bucking transformer to knock the wall down to about 106VAC. Following voltages reflect this attenuated wall voltage. I see the grid of triode 2 of the phase inverter is significantly off (hopefully in red below)

V1 (12AY7)
 plate 1: 125     plate 2: 123
 grid  1     0      grid  2     0
 cathode   2      cathode   2

V2 (12AY7)
 plate 1: 129     plate 2: 126
 grid  1     0      grid  2     0
 cathode  .001   cathode   .001

V3 (12AY7)
 plate 1: 111     plate 2: 226
 grid  1    .01     grid  2  111
 cathode  1.6     cathode 113.7

V4 (12AX7)
 plate 1: 197     plate 2: 248.3
 grid  1    .03    grid  2   18.3
cathode  1.3      cathode  46

V5, V6 (6L6GC)
Plates:  391
Grid 1:  393
Grid 2:  -36
Cath     -.001

V7, V8 (5U4GB): +394 on pins 8 and 2 (pins 1, 3, 5 not connected show .001 to .004; pins 4, 6 show "OL" DCV)

HEATERS
 5.8 VAC
 0   VDC except V4 (pins 4,5) and on V6 (pin 7) which goes between 0 and .001

*NOTE: there does appear to be a slight variation from the schematic in that the center tap of the OT has been moved upstream of the choke, the 6L6 suppressor grid voltage remains downstream

 

[Sonny ReVerb]

There is a bit of positive feedback through the 1M grid leak in the phase splitter. The resistors are original, I assume? If the grid leak resistor drifted lower in value it may cause early distortion.

[sluckey]

V2 cathodes should be same as V1 cathodes (2v) because they are connected together. Probably a measuring or writing error?

V4 grid 2 is OK. It's nearly impossible to get an accurate reading there because the impedance of that bootstrap biased circuit is very high and your meter will load it down considerably. Different meters will show a wide difference in measured voltage. A better way to measure that grid voltage would be to measure the voltage at the bottom of that 1 meg grid resistor and just assume that same voltage is really on the grid.

V5 and V6 grid 1 and grid 2 voltages are reversed. Probably a writing error? Otherwise, voltages look OK.

All tube voltages look good except the errors I mentioned. No clues in the voltage readings.

I'm also using a home made bucking transformer to knock the wall down to about 106VAC.

I think your bucking transformer is dropping the line voltage too much. I'd plug the amp straight into the wall and reevaluate. How much AC voltage do you measure across the filaments (not filament to ground) when using the bucking transformer?

there does appear to be a slight variation from the schematic in that the center tap of the OT has been moved upstream of the choke

Any idea who did it and why? The 5F8A was wired like this but it had four output tubes and only had one rectifier tube. It's possible that Fender did this for some reason (like maybe didn't have a proper choke on the bench). If your choke is the right one I'd connect the OT as per the schematic.

I just realized that I forgot to post a schematic/layout with voltages listed. Here it is...

     http://el34world.com/charts/Schematics/files/fender/Fender_TWIN_5E8A.pdf

[Willabe]

....... the 6L6 suppressor grid voltage remains downstream.

That's the screen grid.

The suppressor grid is not shown in the schemo, it is tied internally to the K.(K=cathode)
 

[centervolume]

thanks for correction Willabe, my error on that.

Regarding V1, V2 cathodes. I noticed there was grime covering parts of those connections (strapped over the socket) so I was getting insulation from that. I scraped that off and now showing about 2.04VDC on all V1, 2 cathodes.

V4 grid 2: on the layout schematic, pin 1 (plate) of V4 shows 210 VDC going to a .02/600VDC (coupling?) capacitor. On the other side of that cap, there is the 18+ volts that is connected to / showing up on grid 2 (pin 7, V4). I was not sure if that 18 volts was there because of leaking issues with that .02/600VDC cap.

As for the VAC across the filaments: the 5.8VAC was measured across the filaments with bucking transformer. Without that voltage reduction, the VAC across the filaments is : 6.7.

V6, 7 (power tubes): perhaps I reversed name designation. Pin 3 is plate (now 458 without bucking transformer and after bias adjust for new power tubes); pin 4 is 460 and pin 5 is -45VDC.

Not sure about the OT center tap re-route upstream of the choke. Almost everything on the amp is stock but it does appear that this solder joint has been molested (shiny). I checked and all iron is original so I can restore this to factory.

I'll scrap the bucking transformer end remeasure even though I was happy to see the voltages were pretty close to published schematic with that reduction of the wall voltage provided by the bucking transformer. Lastly, not sure about the differences between the 5E8 and the (5E8-A) update, but this amp shows 5E8 on the tube chart. 

[sluckey]

On the other side of that cap, there is the 18+ volts that is connected to / showing up on grid 2 (pin 7, V4). I was not sure if that 18 volts was there because of leaking issues with that .02/600VDC cap.

It's possible that the cap is leaking. The only way to know for sure is to disconnect the side of the cap that connects to the grid and then measure the voltage on that dangling cap lead.

But there has to be a positive voltage on that grid and that voltage will be 1 or 2 volts less than the voltage measured on the cathode. Your meter will always measure that grid voltage wrong in any bootstrapped biased stage.

So, how does the amp sound without the bucking transformer? Any difference?

[PRR]

> The suppressor grid is not shown in the schemo, it is tied internally

RCA would claim that the 6L6 has no G3 suppressor grid. If it did it would infringe the European patent on the Power Pentode. Sarnoff was not about to pay that license.

There are two bits of metal (not grids!) along side of where a suppressor grid would go. These are indeed tied to cathode, like G3 on many-many pentodes. These and some clever proportions give a suppressor-like effect. But not a suppressor grid!! RCA would not steal a good idea (if they had to pay royalties).

And yes, the Screen (G2) Grid goes near B+ choke somewhere.

And yes, 106V is kinda low even for a 1955 amplifier. 110-115V was normal. {EDIT} indeed your 6.3V would be spot-on for 115V at the wall. Use a 6V buck-down, not 16V.

[centervolume]

thanks for additional info all

So, how does the amp sound without the bucking transformer? Any difference?

Not sure whether it's from new tubes or operation at 120 wall voltage (also spent some time cleaning connections after I noticed how pervasive the grime issue was) but the "early" distortion seems to be less bothersome, possibly normal since I really dont have a clear idea of what this amp is supposed to sound like, my only other tweed amp is a 60 deluxe. 

My concern now is really with gain, it doesn't have that aggressive Fender volume, feels like there is a -20dB pad on the amp or something; almost like a wide panel 53-54 type circuit with that more muffled sound. I guess I need to bust out the scope and see if the AC signal sine wave is losing amplitude. My problem there is that I have approximately zero hours logged as a scope user, but no time like the present to jump in.

With the DC voltages being ok (with possible exception of 18 volts on PI grid 2), I'm inclined to focus on what happens to the AC signal as it moves through the circuit. Does this sound logical?

[92Volts]

It's worth cleaning the pots with DeOxit or similar product. Usually you'd notice crackling and stuff rather than just loosing signal, but who knows. After 60 years I'd be more surprised if they were working properly.

Even if your DC measurements looked good, 5.8v on the heaters is low enough it was likely part of the problem.
If you really want to reduce the voltage I'd use 115v, which is probably what this amp was built for. That should get you real close to 6.3v on the heaters, which arguably matters the most. Whatever remaining differences there are from the schematic were likely there from the factory.

[PRR]

> exception of 18 volts on PI grid 2

Forget that! The 18V is only when you poke it with the meter. When not poking it is surely a couple volts lower than the 44V on Cathode; otherwise you would not have 44V at cathode.

This type bias gives a VERY large impedance at the grid. On one hand, it reduces loading on the stage before. On that other hand, the self-resistance of a meter will cause a large drop of DC level. If P and K are OK, G probably is fine.

[shooter]

zero hours logged as a scope user

start by "analyzing" your sig-gen signal

Inject a 200mV 500Hz signal

I typically use "AC coupled", there is no DC reference this way, I use my fluke for DC.
happy hunting, signal chasing is a great hobby

[Sonny ReVerb]

> exception of 18 volts on PI grid 2

Forget that! The 18V is only when you poke it with the meter. When not poking it is surely a couple volts lower than the 44V on Cathode; otherwise you would not have 44V at cathode.

This type bias gives a VERY large impedance at the grid. On one hand, it reduces loading on the stage before. On that other hand, the self-resistance of a meter will cause a large drop of DC level. If P and K are OK, G probably is fine.

You could use a VTVM for a more accurate reading if you really want to know and you have one handy.

[sluckey]

You could use a VTVM for a more accurate reading if you really want to know and you have one handy

Even that meter with it's 21MΩ input resistance will load that circuit and pull the voltage down, although not as much as a 10MΩ digital meter. The best way to accurately measure that grid voltage is to simply measure the voltage present at the bottom end of that 1MΩ grid resistor. Since there is no path for dc current through that 1MΩ resistor, there will be no voltage dropped across it, so the dc voltage at the top of that resistor will be exactly the same as the voltage at the bottom of the resistor.

[centervolume]

ok got the scope hooked (leader 508b). I'm able to see the amplification factor between the grid and plates of V1 and V2 are all uniform (v1 input 1 and 2 go respectively to a triode of V1 and same for ch. 2 => V2 triodes). Visualizing the amplification is pretty cool, waves amplitude goes up the same amount on all inputs from respective grids to plates. So I'm inclined to think things are normal at this point (for example, ruling out input problems (jacks, input resistors etc) as potential cause for the 'quiet' performance.

By the time I get to V3 I can adjust the height / amplitude with the volume control. So while I have the scope working (good news) I'm not sure exactly what to look for. Clipping should set in about 7-8 on the volume control right? Not sure how far toward output to trace it down, guessing all the way is fine just concerned about increasing DC voltage impact on scope, probe. (The input max volts is 600 so I should be safe there, but the probe allows for 2 settings on the handle 1x v 10x which appear to have different sensitivity. On the preamp (lower volts) I used 1x setting but I'm uncertain about going past the phase inverter into the output section - whether I need to change to the 10x setting due to the higher voltage.

I also noticed what appears to the a crack in the baffle board. Im not sure if it goes all the way through (I dont hear rattling on low notes), but I will have to check that. I'm wondering in the worst case scenario (that the baffle is split in two) if that would result in this quieter performance. I'll hook it up to another 2-12 4 ohm baffle currently mounted in another amp to see if its a remarkable difference.

[shooter]

I'll hook it up to another 2-12 4 ohm baffle

you replace a 80db speaker with one around a 100 and it should blow your hair back

whether I need to change to the 10x setting due to the higher voltage.

I typically break out the 100X if I get on the output plates, typically I just go from grid to output side of OT tranny, then if something looks "bad" i'll go to PA plates

Clipping should set in about 7-8 on the volume control right?

depends on design, I usually design for clean signal into the PA section, but large enough to spank the pants off the PA tubes.
you may notice more asymmetry than clipping, look at the + and - parts of the sine wave, one might only be 2/3 the other, tubes tend to squish more than sharp cut-off