It seems that the new SVT like to eat their 12AU7 Drivers

[g1]

The obvious next step is to swap the 12AU7 from one amp to the other

For me the definitive test would be to swap tubes between the two amps and see if the "balls" follow them.

Did swapping confirm it was the tubes and not the amp?

[bob p]

I guess it is all in your expectations. I see J25, J26 and the legend "to standby switch." That to me means those points go to the switch. They don't show it because it is not on the board, it is just indicated by that legend.

I'm the kind of guy who says that if there's a switch in the amp, then it needs to be somewhere on the schematic. But it isn't anywhere on the schematic, there's just a reference to it on a pin header. A reader has no way of knowing what kind of switch is present in the amp by looking at the schematic, and I think that's just wrong. The user has to pop the hood on the amp to determine that it's an SPST switch. that's something that you might be able to infer from the schematic, but when the amp documents don't show a switch on a schematic and they don't show a switch on a parts list, i think that's a shitty schematic and a shitty BofM. IMO it would be better if they were to put the switch on the schematic and put a dotted box around it that says "on chassis" or "not on board". Schematics that don't show all of the parts of the circuit aren't very good, IMO.

I've given up reading the modern Marshall schematics. I've decided that I don't care about the amps enough to spend the time making sense of their crappy drawings.

Your delay relay control legend says from power amp. And on the power amp drawing, bottom center I see relay control to AC terminal board. I find that easy enough to follow without too much stress. For me I think the interconnections between pages are clear.

Help me out here -- the AC board legend says that the "relay control from power amp" is on J35. The power amp board's legend says that the "relay control to AC term board" is J12. Looking at the AC term board, J12 is the line input to the power switch. There is no translation table that tells you that J12 on one board is equivalent to J35 on the other, and there is not even a diagram that shows a connection between the terminals that bear a different numbering system. This schematic is clear as mud.

I guess you're supposed to figure out on your own that J12 on the amp board goes to the AC terminal board, but it doesn't go to J12 on the terminal board, because J12 on one board actually connects to J35 on the other. There is no logic to the bus numbering system. There is no conservation of bus channel numbering from one page to the next, there is no diagram that shows that one thing connects to the other, and there is no translation table to tell you how the numbering system changes at one end of the cable vs. the other. It's up to the reader to figure out where the connection NEEDS to be, because the schematic doesn't show you where the connections are actually made.

This is SHITTY work on the part of the guy who drew the schematic.

How are you supposed to figure this stuff out?

I don't buy that you don't have any problem with seeing these things as you jump from page to page, because even though you're "comfortable" reading these crappy drawings, you made the mistake in not seeing the time delay STBY circuit that comes from charging C13 on the amp board. To me, if Enzo can't make see the time delay standby when he looks at the schematic then.that's prima facie evidence the schematic is bad.

[Steve Conner]

This is a classic case of a schematic that was drafted for the purpose of PCB layout, so it doesn't include any parts that aren't on PCBs, such as the standby switch and the inter-board wiring loom.

[bob p]

You're probably right -- somebody drew it in a package like Eagle or ORCad or an equivalent, and drew what they had to draw on the schematic to get the PCB layout function to work, and that's as far as they bothered to go. They never bothered to complete the schematic, they only did as much as was needed to print out the Gerber file for the PCB.

Around here we call that a shitty schematic that was drawn by a lazy ass.

[g1]

Sorry to go off on a tangent from your tangent (agree about the state of schematics these days), but how 'bout that tube swap? Do the bad tubes make the good amp sound weaker (and vice versa)?

[Enzo]

Expectations. I see the relay transistor base getting a control signal from the power amp. I wasn't looking at the J35, as in troubleshooting, I will be looking for a signal there at the transistor, and if not present, I'll go first to the source of the signal, which is on the other page. However, I look to the left and see the layout, and see J35 is a push-on pin, so I know it will be a single wire. I then look at the power amp board and spot the relay control to AC board signal line. Again, not really looking for the wire, as I am more likely to look at source and destination, not the wires between, until I find the source has something that the destination never receives. But I see the J12, and it is a push-on over near a five pin header.

I note the drawing convention on this schematic is for multi=pin connectors to be drawn in a rectangular block, while individual wire push-ons are shown as arrowheads. With multipin connectors, there will be a mate, so perforce the pins will have the same numbers. But the connectors themselves, like all the other parts on each board, are numbered for that board. So every board will have a J1 a J2 a J3, etc. J11 may be a push-on on the AC board, but a 16 pin inline is J11on the power amp board.

SO yes, I figure out on my own that J12 on one board connects to J35 on another, by dint of the fact that those two signals are the same and would be connected together.

I think I read schematics pretty well. I read relay logic trees, I read TTL logic diagrams, I read these amp drawings. The fact I overlooked a control line had nothing to do with the style of schematic. I can read the damned Marshall ones too. yes, I'd like to see system schematics, but I have no trouble reading these. A mistake reading is not the same as a defect in the reading material. I sure don't blame Ampeg for my oversight. If you had asked me about a delay back when we were discussing how the bias circuits worked, I would not have overlooked it, because it was in the heart of that IC circuit on the other page. What I did overlook was where it CLEARLY stated TIME DELAY on the base lead of that control resistor. That is not their fault, that was mine.

[bob p]

haven't swapped tubes yet. those tests are being deferred until I have ordered replacements.

as far as schematics goes, it used to be common courtesy to provide a header numbering translation diagram when you didn't use a ribbon cable, and when you had single wires going from location to location. here's an example of a schematic that has a separate chassis assembly, PSU board, amp board, display board & front panel controls, and numbered header pins that connect to different numbered header pins in other locations. this diagram is the equivalent of a translation table, a Rosetta Stone, so to speak.



a wiring diagram like this one makes figuring things out very easy. The SVT-2 Pro schematic is much harder to work with.

[bob p]

For me the definitive test would be to swap tubes between the two amps and see if the "balls" follow them.

Also measure the power output, frequency response and distortion of both amps to see if you can get a correlation with the subjective sensation of balls. If you find that they both do 300 watts before clipping, then maybe it is just a difference in preamp gain that has nothing to do with the driver tubes.

Thanks for looking at this, Steve. I think I remember you saying that you've played through a 2 Pro before, so I was hoping you'd notice the thread and respond with your experiences with the amp.

FWIW I did compare both amps with the tone controls flat and with the same speaker load. One amp seemed to get loud much quicker than the other with respect to dial positions. I guess that another thing to look at would be to see if the GAIN, DRIVE and MASTER pots might have different values or even a different taper.

I've got the gear to sweep the amp for frequency response. That's also on the to-do list, as it's possible that dried out caps might be effecting one amp vs. the other. The main impediment to measuring individual caps for C and ESR is the fact that disassembling these amps is a royal PITA. I don't know if you've ever done this, but the 2 Pro assembly is a lot tighter than a regular SVT. They have 2 stacked PCB in the amplifier section, and a 3rd sheet metal board for mounting the 6550 sockets to the chassis. The amp requires a lot of disassembly for service and the amp can't be run while it's apart.

I hadn't measured power output at clipping because I don't have a suitable load to test 300W RMS at 4 or 2 ohms. I've got a resistive test load that's suitable for that at 8 ohms, but it de-rates in power as I tap it at lower resistances, so I'll have to fab a new load box. Looking in my box of resistors, I did find 10 of the 10R/100W Ohmite 270, which I could use to make a 2R/500W load or a 4R/1kW load. I'll just need to mount them in a permanent box before I can do the tests. (Wow, these amps are monsters.)

For longer life (and extra geek points) you could try replacing the driver 12AU7 with a JJ ECC99. It's just a cathode follower, so changing to a different tube type won't affect the bias point too badly.

I'm all for earning g33k points, but I have to confess I'm not familiar with the ECC99, so I'll have to look at the data sheet. What is it about the ECC99 that makes you think of using it instead of a 12BH7? I think I mentioned earlier that I had considered a 12BH7, but I hadn't gotten around to comparing the original/reissue SVT circuits to see if any changes were made to accommodate the 12AU7 vs the 12BH7. I'd obviously be in the same boat if I had to calculate load lines and parts replacements to use the ECC99. Before going that route I'd like to ask as many questions as possible, so if you know anything more I'd be glad to hear it! By any chance do you think the 12BH7 or the ECC99 might be drop-in replacements? Re-biasing the amp is not at all difficult.

Thanks!

[Steve Conner]

Hi Bob, yes I like the SVT2 Pro. I think my local rehearsal studio are about to junk their fleet of the things in favour of Aguilar solid-state amps. The guy told me he was "tired of them blowing up". I think he probably meant that they would wear out their power tubes quickly under rehearsal room duty. The Aguilars do sound great.

The ECC99/12BH7 would be a drop-in replacement if only they hadn't used half of the tube as a gain stage, which I didn't notice before. It might still work, but the bias point (of the gain stage, as well as the power tubes) will need checking.

Finally, all the schematics I draw at work are full of those mystery headers. I always label them on the schem, though. "To RF board J1" or the like.

[J M Fahey]

Hi Bob, yes I like the SVT2 Pro. I think my local rehearsal studio are about to junk their fleet of the things in favour of Aguilar solid-state amps. The guy told me he was "tired of them blowing up". I think he probably meant that they would wear out their power tubes quickly under rehearsal room duty. The Aguilars do sound great.

FWIW that same fact has helped me sell a lot of my SS 300W Bass amps to former SVT owners (who keep the 8x10" of course).

That definitely helped raise the perceived value of my stuff

[R.G.]

I can't tell directly from the schemos presented here, but is one section of the 12AU7 used as a follower to drive the output tube grids?

I think that running with continuous high currents may be exhausting the cathode oxides early.

Has anyone tried subbing a power MOSFET follower for the CF if so? Lots of transconductance there, and immunity to high current problems if so. If the schemo permits this, then SVTs of this ilk might make a GREAT thing to pick up cheap and improve.

This is all speculation, of course.

[bob p]

I can't tell directly from the schemos presented here, but is one section of the 12AU7 used as a follower to drive the output tube grids?

Now *THAT* is funny. These schematics are so bad that a professional amp designer can't even make sense of them!

Here are two pictures that answer the question: V2 and V3 in the following diagram are 12AU7 cathode followers that go to pins 13 and 15 on the connector cable:



The next page shows that pins 13 and 15 on the ribbon cable do indeed drive the output tube grids:



There is also a CF in the preamp stage, where a 12AU7 CF drives the send to the EQ.

I think that running with continuous high currents may be exhausting the cathode oxides early.

Interestingly, the 12BH7 draws twice the amount of heater current as the 12AU7.

Has anyone tried subbing a power MOSFET follower for the CF if so? Lots of transconductance there, and immunity to high current problems if so. If the schemo permits this, then SVTs of this ilk might make a GREAT thing to pick up cheap and improve.

I've picked up two of these amps because they are absolute power monsters, and really good sounding amps, that aren't given that much love by bass players. Conventional wisdom by non-tech users has led to sticky internet fables about lightweight transformers that are grossly inaccurate. With those things in mind, people who are Ampeg snobs tend to scoff at SVT from the SLM era and even moreso at the LOUD products. In this economy it's possible to find good deals. Looking at the construction of these beasts, they are extremely well made amps. I can't see any difference between the USA/SLM amp and the Korea/LOUD amp.

I'm willing to experiment to make these amps more reliable. I can make sense of the 12BH7 substitution, but I don't know jack about mosfets.

[Steve Conner]

The funny thing is that in this circuit, both halves of the 12AU7 run well within their ratings. At idle, the cathode current is about 3.5mA for both stages, and the plate dissipation is less than 1W for the gain stage, and about 1.5W for the follower. (The rated dissipation is something like 2.75W per plate.) The circuit doesn't allow the followers to try to source big grid currents into the power tubes, either. The grid resistors are largish, and the clipping diodes on the input prevent really heavy overdrive.

Maybe they just don't make 12AU7s like they used to?

PS: The version of the story I heard was that Ampeg made the output transformer a bit smaller in the SVT2-Pro than the original SVT, because they felt that a little transformer saturation improved the sound. It's still hardly "lightweight".

[R.G.]

Here are two pictures that answer the question: V2 and V3 in the following diagram are 12AU7 cathode followers that go to pins 13 and 15 on the connector cable:

Got it. Yep, I think a power MOSFET could be subbed in. I'll have to tinker a bit to ensure I understand what they do, but a source follower could probably live in those positions, and would never be eaten up by the current loading.

I'm willing to experiment to make these amps more reliable. I can make sense of the 12BH7 substitution, but I don't know jack about mosfets.

I'll see what I can suggest. The only real trick will be making sure the bias adjustment pots can maintain the right range. A 12AU7 has a grid voltage that's 10-20V below the cathode, but a power MOSFET needs a gate voltage that's about 5-6V above the source. It may be as simple as subbing in the MOSFET and readjusting the bias pot, but I need to look to be sure that's not taking away an important part of the range for bias.

I'm not a professional amp designer, necessarily, although I did design a tube amp for the commercial market. Well, several, only two of which were taken to market. But I do make my living designing electronics and have been a practicing EE since 1973. I just have this weakness for musical electronics.

[bob p]

The funny thing is that in this circuit, both halves of the 12AU7 run well within their ratings. At idle, the cathode current is about 3.5mA for both stages, and the plate dissipation is less than 1W for the gain stage, and about 1.5W for the follower. (The rated dissipation is something like 2.75W per plate.) The circuit doesn't allow the followers to try to source big grid currents into the power tubes, either. The grid resistors are largish, and the clipping diodes on the input prevent really heavy overdrive.

Maybe they just don't make 12AU7s like they used to?

I can't say that I see an obvious reason for driver failure either. As you've said, the cathode current and plate dissipations are within spec for the AU7, and ~15K of grid resistors should do a good job of limiting grid current. I'm not sure I follow why those 4007 are there. Those are what you were referring to as the "clipping diodes", right?

Hearing how these amps typically fail due to driver failures, it really makes me wonder if the less ballsy nature of one amp is just a batch of weaker than normal tubes or a just compliment of tired tubes. I guess the other possibility is that the amps really do eat them for reasons that aren't readily apparent. It's frustrating when you do the math and you can't figure out why there's a problem.