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

[bob p]

Has anyone else noticed that the new versions of the SVT tend to chew up and spit out the 12AU7 that they're using as drivers?

I'm lucky enough to have a pair of SVT-2 Pro amps right now, one an SLM/USA product from around 1993 and the other a LOUD/Korea product from shortly thereafter. Both are sporting a well-matched set of Sovtek 6550WE in their output sections, well-matched Sovtek 12AX7WA, and well-matched EH 12AU7. Both bias up well using the LED idiot lights and both sound strong like an ox... but the SLM/USA amp seems to get louder sooner, while the LOUD/Korean amp seems to need a little more twist of the Gain and Master to reach the same bone-crushing output levels. This might not actually be meaningful, as both amps are louder than I can stand to run full-tilt. It might take gigging with these beasts to determine if the difference in volume related to the knob positions are of any consequence.

What is of consequence though, is that the 12AU7 tubes consistently test differently from one amp to the next. I pulled the tubes from both of them and ran them through my tester. Much to my surprise, all of the 6550WE in both amps were very tightly matched and tested very strong. All of the 12AX7 tested strong as well, with the only exception being V1 (first preamp tube) on the LOUD amp -- it had a triode section that measured low-output / reject on the tester, so I replaced it with a strong, well-matched 12AX7. That didn't solve the problem.

Interestingly, the 12AU7 EH tested differently on both amps. The SLM amp had it's 12AU7 EH all test in the high-normal range, with tight matching between triode sections within each tube, indicating that the tubes were very strong and well-matched. The LOUD amp had it's 12AU7 EH all test in the low-normal range, with decent but not as tight matching between triode sections. Although the 12AU7 EH all tested as "normal" in both amps, the 12AU7 in the SLM amp all tested at the high end of the normal range, with tight grouping, while the 12AU7 in the LOUD amp all tested at the low end of the normal range, again with tight grouping.

This has me thinking that even though the tube tester rates all of the 12AU7 with passing grades, well into the "normal" range, it seems that the amp that has the tubes testing in the lower half of the normal range isn't performing as well as the amp that has the 12AU7 ranking in theupper half of the normal range. This brings to mind the old adage that "the circuit you're using is the ultimate tube tester." This may be a case where my mutual transconductance tube tester just isn't as sensitive at identifying "normal" tubes that turn out to be too weak to perform well in an SVT circuit.

continued in next message...

[bob p]

The obvious next step is to swap the 12AU7 from one amp to the other, and to order some new 12AU7 EH for screening and matching, but this brings up some important questions:

First, I know that the performance in the SVT is highly dependent on drivers, and that if the driver tubes are weak, then the output section can't be driven to maximum performance. With that in mind, the LOUD amp does bias up quite well using the idiot lights on the back, and listening to it on it's own, it seems like the amp runs just fine. What strikes me as odd though, is that the SLM amp (the one with the stronger-testing 12AU7) just sounds like it has more balls with less of a twist of the knobs.

Are these amps known to eat the 12AU7 that are used for drivers? What's interesting is that in the less ballsy amp, it's not only the two drivers that test as low-normal. The 12AU7 in the preamp circuit tests low-normal as well. This makes me wonder if these amps are just prone to use-up and destroy their 12AU7.

Second, why is it that the SVT-2, SVT-2 Pro, and the later SVT reissues use 12AU7 tubes in the driver positions instead of the old-school 12BH7? I'm thinking that this decision had to be made based upon widespread availability of the 12AU7 and the lesser availability of good 12BH7 in today's marketplace. Comparing the tubes, I think that the 12AU7 is probably being pushed to it's limits in this application, and that this could relate to the tubes having a shorter lifespan. I'd like to ask if anyone has experience in this regard.

Third, the natural question that follows is whether any of the new production 12BH7 are any good (such as the 12BH7 EH or the TAD 12BH7A), and if the amp would benefit from modification to deploy a more robust tube in the driver positions to replace the 12AU7.

Any insights would be greatly appreciated.

[bob p]

FYI, Enzo was kind enough to post the schematics for the SVT-2 Pro in posts 6, 7 and 8 of this thread:

http://music-electronics-forum.com/t33604/

Thanks, Enzo!

[Enzo]

I have mentioned this often in various threads. Yes, the 12AU7 is I think the weakest link in these amps. replacing the 12AU7s is THE one most common repair I do to them. I suggest swapping them out whenever the complaint is can't get the bias lights to settle.

I replace them with new production tubes, and have not had one come back so far. SO it has not occurred to me to try NOS or good used pulls and compare. Nor have I characterized any in a tester. In a tester are you talking about emissions? COuld be the currents in real world use are greater than the tester demands. Might make a difference.

If I had to guess what is killing them, my first thought might be cathode to heater voltage. The heaters are at ground, the bias voltage is derived from -180v. That is controlled down to about -45 by the 12AU7 triodes. But when the amp is fired up, I believe that full -180v appears on the cathode until the tube heats up and conducts. It might cause some damage. Or it may be something totally different from that. Not to mention that when cold, there is 560v across that tube, potentially.


As to why didn't they us 12BH7 or other alternative, I think you are exactly right - availability. They need production quantities of the tubes on a reliable basis. 12AU7 is a very common tube. We use them, hifi guys use them. 12BH7 much less so. I imagine they look ahead and think it will be a lot easier to meet production demand with 12AU7 than anything else.

This would be a good place for a couple of Steve's MOSFETs, but then they could no longer claim all-tube power amp.

I have no idea if todays 12BH7s are just repackaged 12AU7s or something.


One thing I feel the need to point out is the difference between common failure and reliability. Knowing that WHEN I see bias problems, the 12AU7 is most likely to be at fault, is not at all the same thing as 12AU7s fail all the time in these amps. MOST do not, only WHEN the amp fails, they are the prime suspect.

[Enzo]

Just thinking, it looks like less than 3ma per triode through them.

[bob p]

I have mentioned this often in various threads. Yes, the 12AU7 is I think the weakest link in these amps. replacing the 12AU7s is THE one most common repair I do to them. I suggest swapping them out whenever the complaint is can't get the bias lights to settle.

It makes sense that the two 12AU7 that are being used as output stage drivers would be critical to performance, that they would be likely to cause bias problems, and that they would early-wear items. It's interesting that on the SVT-2 Pro, it looks like all THREE of the 12AU7 are wearing out, while only two of them are seeing a lot of current as output stage drivers. Those two drivers (V2 and V3 on the amp board) sit at the extreme rear of the amp, between the 6550 output tubes and the back panel bias controls, right along side of V1 (the 12AX7 phase inverter). It certainly makes sense that V2 and V3 on the power amp board would go bad from sustained high current, and that that would be a weak link in the amp.

What strikes me as odd though, is that I'm having no problems with V1 on the amp board (12AX7), while the other 12AU7 in the preamp also tests just as weak as the two 12AU7 drivers. That other 12AU7 is V4 on the preamp board, which is a CF that drives the EQ. I'm failing to understand why that tube should be wearing out as fast as the power amp drivers, yet the 12AX7 phase inverter (V1 on the amp board) escapes unscathed.

Can you explain that?

[bob p]

I replace them with new production tubes, and have not had one come back so far. SO it has not occurred to me to try NOS or good used pulls and compare. Nor have I characterized any in a tester. In a tester are you talking about emissions? COuld be the currents in real world use are greater than the tester demands. Might make a difference.

Yes, the tester is basically a "mutual conductance" emissions tester; it has to be using a low-current supply that pales in comparison to the actual SVT circuit. In addition to the supply current being lower than the actual SVT circuit, I'm certain that the RMS voltage being applied to the tubes also pales in comparison to what the tubes actually see in the SVT circuit. I think the take home point from this is that on my tester, the tubes need to test in the upper half of the range that's defined as "good" to really perform well in the SVT circuit. It certainly appears that what the testers calls "good" in the lower half of the range seems to be less than optimal in the SVT circuit. As far as the tester is concerned, I think that I'm going to have to seek out tubes that test in the high upper range of normal for these amps.

If I had to guess what is killing them, my first thought might be cathode to heater voltage. The heaters are at ground, the bias voltage is derived from -180v. That is controlled down to about -45 by the 12AU7 triodes. But when the amp is fired up, I believe that full -180v appears on the cathode until the tube heats up and conducts. It might cause some damage. Or it may be something totally different from that. Not to mention that when cold, there is 560v across that tube, potentially.

I don't know about the other reissue SVT amps, but when the SVT-2 Pro amp powers up, there is a time-delay between the time that the heaters come on, and the time that the B+ circuit is actually activated. Assuming that you leave the standby switch in the B+ ON state at all times, when you turn on the amp, the process of flipping on the power switch only turns on power to the filament transformer. There is a time delay relay in series with the user's STBY switch that waits a minimum of 20 seconds before passing AC to the B+ transformer. Because the TDR and the STBY switch are wired in series, both have to be activated before the B+ circuit is energized. This forces a minimum wait period for B+ in spite of what the user may be trying to do.

This suggests that at a bare minimum, there's no way that the cathode-to-heater voltage should ever be at -180V on a cold tube. At least as I understand the circuit, there should be some minimal warm up before the B+ gets turned on and the cathode-heater voltage is blasted to -180. Of course, this assumes that the TDR circuit works as it should, and that 20 sec is enough time for warm-up. Personally, with an amp that has CF drive with negative cathode-heater voltages like this, I'm the kind of guy to leave the standby off for a couple of minutes before flipping it on, just to provide a little extra margin of safety. Of course, I have no idea what the previous owner(s) may have done with the amp. They may not have been as kind.

[bob p]

As to why didn't they us 12BH7 or other alternative, I think you are exactly right - availability. They need production quantities of the tubes on a reliable basis. 12AU7 is a very common tube. We use them, hifi guys use them. 12BH7 much less so. I imagine they look ahead and think it will be a lot easier to meet production demand with 12AU7 than anything else.

So it sounds like your solution is to just replace the 12AU7 with new production tubes whenever they seem to be causing a biasing problem. Do you mind if I ask which new production 12AU7 you've been using? I don't really have any experience with 12AU7 in this sort of extreme application, so I thought I'd ask if there are any new production 12AU7 that might work better in an SVT than any others. Any experience that you have in this regard would be appreciated.


One thing that I find interesting about this amp is that I'm NOT having any symptoms of biasing problems, and I'm not having any real symptoms of weakness in the amp. The bias lights work just fine, and the range of GREEN-ON before RED-ON is similar for both amps, suggesting that there's no biasing problem. Or another way to say it is that the bias lights are failing to provide an early warning indication that the 12AU7 seem to be going bad.

The amp's performance isn't really weak either -- even with the 12AU7 that test "low-normal" on my tester, the amp still has lots of balls, and you might never think there's a problem with it... that is, until you put it side by side up against an SVT with really strong 12AU7. Then it becomes obvious that the tube tester is telling you about a subtle problem that's evolving before most of the other methods of detection are capable of identifying it. I have to admit, if I didn't have to of these amps sitting side by side, I'd have never bothered to test the tubes and I would have never figured out the tube-wear problem.

This would be a good place for a couple of Steve's MOSFETs, but then they could no longer claim all-tube power amp.

Which Steve? Connor? Bench?

I have no idea if todays 12BH7s are just repackaged 12AU7s or something.

I'm lucky enough to have a tube caddy that's full of NOS RCA 12BH7, so naturally I'm thinking about whether NOS or new production 12BH7 might be worth putting into the amp. But I have to admit that I'm not sufficiently familiar with the differences between the 12BH7 and the 12AU7, and the demands of the circuit, to determine if this would be a good idea or a bad idea.

Anyone?

[bob p]

One thing I feel the need to point out is the difference between common failure and reliability. Knowing that WHEN I see bias problems, the 12AU7 is most likely to be at fault, is not at all the same thing as 12AU7s fail all the time in these amps. MOST do not, only WHEN the amp fails, they are the prime suspect.

I think my experience with the tube tester compliments your observations, and also shines a bit more light on the problem.

I think that the occurrence of bias problems is a late-warning indicator that there is a problem with the 12AU7, because by the time that bias problems occur, the tube is already weak and is entering a failure mode. In contrast, my tube tester experience suggests that I'm watching 12AU7 being degraded and consumed in these amps, as the tube tester results change from high-normal to low-normal readings. This suggests that the 12AU7 probably do fail all the time in these amps, that the process of failure takes some time, and that the process of failure can be monitored with something as simple as a mutual conductance tube tester. I'm thinking that once the 12AU7 start to test in the lower-half of the normal range, that's an early predictor of impending failure and it's time to replace them BEFORE you start to have the bias problems.

The real question is whether I'd be better off just re-populating those 3 positions with new production 12AU7, or if I'd be better off using a more robust tube that can handle the job better, like a NOS or a new production 12BH7.

I'm just dying to hear more about this from someone who knows more than I do.

[Enzo]

Beware correlations, they are not necessarily causation. We have failing driver tubes, and we see low readings on the tester, then we see the same readings on the preamp tube. It may not be that weak emissions are the failure mode. We might just have a batch of low gain tubes. SO the fast the preamp tube measures the same as the failing drier tubes is not necessarily due to them "wearing out" in use. Unless we know that those preamp tubes were higher gain when installed and only NOW have gotten weak, we cannot assume the correlation is causal.

This suggests that at a bare minimum, there's no way that the cathode-to-heater voltage should ever be at -180V on a cold tube. At least as I understand the circuit, there should be some minimal warm up before the B+ gets turned on and the cathode-heater voltage is blasted to -180.

Look at the driver board schematic, page 2. The -180v supply is off the filament transformer. It comes on immediately. B+ is dependent upon the standby, but not the bias supply. -180 will be on those cathodes the moment power is turned on.

That aside, I don't see a B+ delay. There is a relay in the mains lead of the B+ transformer, yes, but I see no delay. If I missed it, point it out please. What I see is an enable relay powered by the heater voltage. But the transistor controlling the relay is powered right off the rectified section of the heater string. So in other words the B+ transformer cannot come on unless there is heater supply present. But I see no delay involved. If we left the standby switch ON, then turned on the power, B+ would come up immediately. And then we'd have not only a moment of -180 on the cathodes, but also +380 on the plates. 560v across that tube. Then again, most folks with SVTs use the standby.


I have not been selective with 12AU7s, so they would likely be Sovtek or JJ tubes in my shop.

[bob p]

I find the SVT-2 Pro schematics to be quite arcane. I have difficulty following schematics that hop all over 20 pages in 6 different PDF files and use jumper cables to jump from page to page. It makes understanding the circuit all that much harder. What makes it even harder to follow is that the schematics that I received from LOUD are not the same schematics that I got from you. Some things appear to be different, perhaps due to ECOs. LOUD told me that the schematics that I received included the most recent ECOs.

Here is an excerpt from the Owner's Manual that describes the automatic standby function:







more to follow...

[bob p]

Here's a snippet from the AC board that shows the relay that's in series with the standby switch. Just to make things as clear as mud, the STBY switch isn't even shown on this schematic, or on any other page in the schematic library. You're supposed to figure out on your own that J25-J26 go to a switch that isn't shown anywhere.



It's likely that there isn't a strict time delay countdown, but that the time delay occurs while the control voltage ramps-up gradually from 0V to 15V. This takes about 20-30 seconds with a cold amp, less time with a hot amp. So I guess it's really only a 20sec delay when powering on a cold amp, and the instruction manual is somewhat inaccurate about the 20-second delay when the amp is first turned on. The delay is less than 20 seconds if the hot amp is shut off and turned back on again.



PS -- I see what you mean about the -180V not being delayed.

[Steve Conner]

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.

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.

[Enzo]

You are absolutely correct, I overlooked the timer control on the relay transistor. It is on my files too., I just overlooked it. The -180 still comes right up, just not B+.

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 don't find these hard to follow. In general I always prefer a system schematic to drawings for each assembly, but this way they reuse the same assemblies in other models without drafting a whole new document. So I deal with this. If you want impossible to follow, go look at the Marshall TSL100 drawing set.


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.

The 15v rails come right up. The delay I see comes from IC2b, bottom center power amp board page, just upper right from that is C13. it takes it a moment to charge up, pulling the + pin high on the IC, which toggles the output positive.

[Enzo]

OH, and when I was thinking MOSFETs, it was Mr. Connor came to mind.