any tube amp gurus ???

[awilson40]

I have an old MusicMan HD130, solidstate preamp, 12ax7 phase inverter and 6ca7 power tubes.

The amp seems to clip sooner than it should and as I was tracing a tone with a scope I see that the positive going
wave clips, while the neg doesnt at the plate of the 1st stage of the PI tube.
I swapped tubes, same thing. I checked supply voltages and resistor values, all OK.
I havent replaced the coupling caps, perhaps one is leaking.

heres a pic of the wave at the grid


http://home.earthlink.net/~awilson40/pin_7_grid.jpg

and at the plate

http://home.earthlink.net/~awilson40/pin_6_plate.jpg

any ideas ??

the plate load resistor is 100k, cathode res is 1.5k. I tried going up and down a bit on the cathode res
and it made no diff.  Would changing the plate resistor better center the clipping ?

Thanks

[SteveB]

Maybe someone here can help you out, & you could check out these sites, too.

You'll have to register to access, but you'll find lots of info, & hopefully some help here.
http://76.162.4.197/Forum/yabb2/nph-YaBB.pl?board=TubeAmpBuilding

Another one.
http://music-electronics-forum.com/forumdisplay.php?f=22

Steve

[jrem]

looks like you're hitting the rail.  Check the bias on the grid, there should be a resistor holding the grid to a reference voltage, probably in parallel with a cap.  I can't google the schematic, but post it if you have it and I'll take a look. 

You say it's clipping on the postive side so it isn't the power supply crapping out.  Check that it's really positive, if it's inverted then maybe you're dragging the power supply down.  Change the frequency of the input signal and see if it changes things.

Regards, John.

[R.G.]

Probably the big question is whether the clipping there ever makes it to the output tube plates. 6CA7 plates are biased at about -40 to -50V. Any voltage bigger than the bias voltage either positive or negative is going to either be trying to turn the tube more off than it already is or more on than it can do.

Can you provide the static DC voltages (no signal) on those pins, as well as the magnitude of the signals in and out when it is just beginning to clip? Even better, do you have a schematic or a partial schematic of those sections as well?

[awilson40]

I left the amp at church but will bring it back when I have some more time and ideas to troubleshoot.
Here is a schematic, just ignore the black "Clapton Mod" changes.

http://home.earthlink.net/~awilson40/Clapton_HD130.pdf

The clipping does make it to the output tubes as I get the same signal when I scope the speakers.
I'm pretty sure its positve and not inverting. This amp has a high power mode and low power
where the power section voltage is cut. I get the same "early clipping" in the postive direction
in high power mode, but I do have more head room before the clipping begins.

The clipping is being generated from the 1st stage of the 12AX7

In HP mode , the plate voltage to the top of R50 is 350v, R50 measures 100K
R48, cathode measures 1.5K

Thanks Guys

[awilson40]

looks like you're hitting the rail.  Check the bias on the grid, there should be a resistor holding the grid to a reference voltage, probably in parallel with a cap.  I can't google the schematic, but post it if you have it and I'll take a look. 

You say it's clipping on the postive side so it isn't the power supply crapping out.  Check that it's really positive, if it's inverted then maybe you're dragging the power supply down.  Change the frequency of the input signal and see if it changes things.
Regards, John.

I never checked that, I'll look there. My tube theory is really rusty, havent messed with tube stuff sence the 70's when I was in the Marines .
The Grid has a 10k resistor to ground but no cap. Would altering the resistor value be a good idea ?

[jrem]

Would altering the resistor value be a good idea ?[/quote]

I think your right, the plate is in phase, the cathode is out of phase.

I ain't no RG, so I would start with whatever he suggests.

But if I didn't have his advice then I would check R51-R56 to make sure they all supposed to be what the schematic states.  I would lift C36 & C37 and see how it acts.  And get the voltage on the grid, plate, and cathode at idle.

[brett]

Hi
It might be that the grid resistor (grid to ground) has come lose.  That makes the grid voltage too negative (it accumulates electrons) and the tube cuts off.  You'll detect this by measuring the grid voltage with a high reststance DMM.  (you'll get something like -3 volts instead of -1 or 2V on the grid).

In any case, check the cathode and grid voltages, coa there's something screwy in there.  Be safe with those volts.   All 350 want to zap you.
cheers

PS I've got a MusicMan 65W (2x10).  Nice clean amp.  My daughter plays her iPod through it, mostly.

[awilson40]

thanks Brett,
I pulled , checked and resoldered the cathode, plate and grid resistors so they are OK. The plate resistor
was loose and fixing this solved another distortion problem I was having.
When I get it back home,I'll check the dc voltages.

Should I see zero DC on the grid with no input

[R.G.]

Should I see zero DC on the grid with no input

Yes.

For that circuit, you should have 0V on the grid, about 1.0-1.5V on the cathode and about 230-260V on the plate with no signal.

It might be that the grid resistor (grid to ground) has come lose.

Good one, Brett. Especially since they use a 10K (!) grid resistor there.

The signal on the plate of a tube is inverted in phase compared to the grid. The cathode follows the grid, the plate is pulled down. Signal going flat on the high side indicates that either the input is clipping or the plate voltage is too high, and it bumps into the power supply and can't turn any more off than off.

[awilson40]

Good info, Thanks
I'll bring it home on Wed and check some voltages.

Would my best bet be to first alter the grid resistor to get zero at the grid,
then alter the plate R if required to stop the positive clipping  ?
Would I want a larger or smaller value on the plate ?
Thanks again

[R.G.]

Actually, the best bet is to find out what is keeping it from working as designed.

The first thing to do is to measure the DC voltages and find out what its bias point is exactly. This circuit has to provide about 50-60V of swing on the plate. The phase inverter is a simple split-load affair, with a gain of one to the cathode output and a gain of minus one to the plate. So all of the grid swing has to be generated in this triode stage.

To get 50V of positive signal swing on the following grid, your plate has to be more than 50V below the power supply at DC. The power supply you've measured as 350V from an earlier note. You need the plate of the triode to be sitting at well under 300V with no signal so it has enough room to turn off.

For the values shown in the schematic, the plate should be sitting at 250V plus or minus a few if the parts have the values shown. If it's not, then something is wrong elsewhere in the circuit, and correcting the flaw with changes in the grid resistor or plate resistor is just pasting over the symptom, not curing the patient.

This is an odd circuit for a tube setup, in that it's driven from the output of an opamp with a low-value grid resistor. In one position of the bass-boost switch, the opamp output is directly coupled to the grid. So if the opamp has an erroneously large minus voltage on it, even half a volt, it would be pushing the grid way more negative than it should be, and could cause this symptom. So if the grid is not sitting at 0Vdc, then the problem could be in the opamp circuits. Changing the value of the grid resistor is not going to change the bias point at all.

Another way this could happen is that the power supply is sagging under positive swings. You could find that out by putting your scope probes one on the plate and one on the power supply. The power supply should remain constant on all signal swings on the plate. If the power supply sags down to meet the plate, the problem is somewhere in the power supply.

If the cathode resistor were unusually high, (it's not, you measured it) then it could also be choking off the tube by making too high a reverse bias.

Here's a thought - are your tube socket pins dirty? Bad contacts on the tube pins looks like a big resistor in series with the tube pin.

In all these cases. measuring the voltage on the tube pins tells the tale. Or at least points the way.

[awilson40]

That makes good sense. Thanks, I'll check .

Funny you should mention the tube socket. It was dirty, I would wiggle the PI tube and get all kinds of cracks and pops.
I sprayed it witn contact cleaner and it seemed better.
I will ohm the connections and see if there is any resistance.

[awilson40]

Here are some further checks and observations.



I cleaned , wiggled,and bumped the  PI tube to see if it would change the signal. It didnt.
Power supply voltage is not changing when a signal is applied. Its holding steady.

here are the voltages with no signal/
High power : Voltage at point F  343v
                    at pin 6 (plate) 219v
                    pin 7 (grid)   0v
                    pin 8 (cathode)  1.6v

Low Power    point F  233v
                    pin 6   160v
                    pin 7  0v
                    pin 8   .95v

here are some test settings and observations.
test signal at pin 7(grid) is a 750hz wave, 5v peak to peak
output signal at pin 6 (cathode) is 120 v p2p with no clipping
signal at cathode is 2v p2p, no clipping

grid DC level at this point is -1vdc

If I increase the signal at the grid to 10v p2p, the output signal has bad positive clipping and
is at 140v p2p.
At this time the grid is at -.8v.
signal at cathode is at 3v p2p with neg clipping

As I increase the signal strength to the grid, clipping at the output gets worse and the grid DC goes further negative.

This neg voltage at the grid is not coming from the preamp as its dc level remains zero. If I throw the 'Deep switch"
removing the coupling cap and directly connecting the preamp to the PI, then the grid DC voltage stays at zero
but the output still clips.

Any other ideas, opinions ?? Or is my input signal just driving to hard.

Thanks

[R.G.]

That's enough to do some guessing on.

here are the voltages with no signal/
High power : Voltage at point F  343v
                    at pin 6 (plate) 219v
                    pin 7 (grid)   0v
                    pin 8 (cathode)  1.6v

Low Power    point F  233v
                    pin 6   160v
                    pin 7  0v
                    pin 8   .95v

Those are dead-normal for a 12AX7 in that situation.

test signal at pin 7(grid) is a 750hz wave, 5v peak to peak
output signal at pin 6 (cathode) is 2 v p2p with no clipping
grid DC level at this point is -1vdc

ACK! 5V p-p on a 12AX7 grid??? The only saving grace here is that the cathode is unbypassed, so the cathode is coming up by a volt to add to its normal volt or so of reverse bias and the grid is ... marginally... below the cathode. Any more signal than that and you go into grid conduction.

If I increase the signal at the grid to 10v p2p, the output signal has bad negative clipping and
is at 3 v p2p. At this time the grid is at -.8v. As I increase the signal strength to the grid, clipping at the output gets worse and the grid DC goes further negative.

That's because the grid starts conducting on positive peaks, and starts Hoovering in electrons from the stream passing through it on the way to the grid. Grid conduction is building up electrons on the coupling cap. You're into grid blocking. That's what it will do. Again, normal.

This neg voltage at the grid is not coming from the preamp as its dc level remains zero. If I throw the 'Deep switch"
removing the coupling cap and directly connecting the preamp to the PI, then the grid DC voltage stays at zero
but the output still clips.

That's because the opamp output can eat the electrons the grid puts out and still hold the grid at 0V.

So far, what you describe is absolutely, positively normal for a typical 12AX7 triode.

I think there is a different problem than what you're trying to solve.  I did a sim run on this and the results match your observed results. However, the triode in question is putting out nearly 200V p-p signal on its plate before it clips, and that produces a neat 100V or more of grid drive on the output tubes. The bias voltage on the output tubes is no more than -48V by the schematic, so you have about twice as much grid drive available as the output grids need to drive them fully on and fully off.

The question ought to be - how come those output tubes aren't producing more power and less distortion with that much grid drive.

I think you may have an output stage problem, not a predriver problem.

[awilson40]

Thanks,
BTW... I edited my above post as I was at first observing the signal at the cathode... late night
I revised showing the correct voltages and this was all taken at the Low power setting.
High power yields similar results but at a higher level.

The way I came up with the values for the grid input signal was watching my guitar signal at hard pick attack
and setting the reference sig at a similar level. 
Any ideas / pointers for me to begin checking in the power tube section.

Also, the amp is very loud and clean, as long as I stay below 3 on the pre amp volume. I have just thought something was wrong as I cant turn the preamp vol above 3 before the
PI tube clips. Might just be due to the fact that I have hotter pickups so I cant run the preamp vol higher.

Thanks again.

[R.G.]

Also, the amp is very loud and clean, as long as I stay below 3 on the pre amp volume. I have just thought something was wrong as I cant turn the preamp vol above 3 before the PI tube clips. Might just be due to the fact that I have hotter pickups so I cant run the preamp vol higher.

If it's very loud and clean, then there may be nothing wrong. I may have misread your posting "clips sooner than it should" for "clips at low volume".

A preamp and volume knob will only produce as much power as the power stage can put out, no matter if that happens at 10 on the volume knob or at 2.

[JHS]

The MM are the only amps running in class-B mode (or the 725V would kill the tubes).
EC's amp is running on full bias w the pot at max (got this info from his amp tech and my MM is running on full BIAS too w/o any problems).

1.) Remove the spikekiller diodes, mayby this helps already to get a more sym. signal.

If not then

2.) Put in a new matched quartett of Sylvania 6CA7 power tubes (no El34 or 6CA7/EL34 and no SOVTEK , they will burn out after a short time in HI-power mode) .

3.) Replace the coupling caps C36 and C37 with 1% caps and replace R61 with a 2W 0,1% metalfim R.

This should help a bit to get a sym. signal.

4.) Trim the output signal for sym. by altering the value of R53 or R54 (a high ohm R in parallel of one of the resitors should work fine).

(a trimpot is fine and would allow easier adjustment but could loose the adjustment due to vibrations)

BTW: If the output signal is total sym. the amp can sound somewhat "HiFi", sterile or artificial. Old Tweed amps never ever have a total sym. signal at the output.

JHS

[Gus]

I did not see if the tubes were changed in any of the posts

Did you try a different set of tubes?

The reason I post this
is it is easy and
R.G. posted it might be an output stage problem

[awilson40]

thanks a lot for the additional info, I'll be checking further.
Yes, new JJ 'matched' tubes from Euro Tubes. I tried the factory original tubes also, and
they made no difference.
The spike killer diodes are coming out tonight, I have read some stuff on the MusicMan forum
about what happens if  these short