70s SVT Bench Test Results

[coreyfyfe]

All 6550 equipped svt's I've had on my bench, over 10 now, have put out right around 270 watts at the point of any clipping seen on my oscilloscope. I hold line voltage at 120 with my variac so I can compare tests from different days etc...

My 6146 equipped 70 does about 340 or so Before clipping. Just to share my findings.

Were they 6146 converts or newer SVTs? I think the plate voltage on the 6146 was lower then the 70s 6550 amps. I have a few different 6550 labeled schems that show 695V or 660V, but I think the 6146 shems show 630V so a converted amp will not be as powerful (edit - doesn't look like they do, plus B-o-T's comments below). The SVT-CL and VR schematics I have show 660V as well. There's an old 70s manual for the SVT that says 300 watts less than 3% THD. Never benched one myself but I wonder how much variance there was over the years.

 

[beans-on-toast]

Were they 6146 converts or newer SVTs? I think the plate voltage on the 6146 was lower then the 70s 6550 amps. I have a few different 6550 labeled schems that show 695V or 660V, but the 6146 shems show 630V so a converted amp will not be as powerful. The SVT-CL and VR schematics I have show 660V as well. There's an old 70s manual for the SVT that says 300 watts less than 3% THD. Never benched one myself but I wonder how much variance there was over the years.

Bill Hughes has said that the 695v was a schematic error that carried over through some revisions. The early 6550 units should have 660V printed on the schematics.

Here is the refernce: My "super SVT" and a polarity switch question...

 

[coreyfyfe]

Bill Hughes has said that the 695v was a schematic error that carried over through some revisions. The early 6550 units should have 660V printed on the schematics.

That clears up that question at least!

 

[beans-on-toast]

That clears up that question at least!

Plate voltage will vary with the power supply health and wall voltage. So it isn't unusual to see differences when taking readings.

Measured power output can also be subjective, depending on the procedure used. With an audio analyzer output power can be measured over a 20-20K sweep; THD can also be swept. That allows you to conclude what the output power is at a specified THD or less. It can provide a better idea as to what is going on. Often people will only look at one frequency such as 1K. Adopting different approaches can provide different results. If you do the same thing on different amps, it gives you a measure of how they perform relative to one another. That's a good thing.

I always like to finish by testing with my ears.

 

[agedhorse]

If there were any revisions to the output transformer that altered the plate load that the tubes saw reflected back from the load, this too can alter the output power (provided the plate impedance and power supply capacity are adequate). I don't know the revision history of the part, but it's not uncommon to revise if there are any changes to tube compliment or plate voltage.

 

[Mark Reccord]

Plate voltage will vary with the power supply health and wall voltage. So it isn't unusual to see differences when taking readings.

Measured power output can also be subjective, depending on the procedure used. With an audio analyzer output power can be measured over a 20-20K sweep; THD can also be swept. That allows you to conclude what the output power is at a specified THD or less. It can provide a better idea as to what is going on. Often people will only look at one frequency such as 1K. Adopting different approaches can provide different results. If you do the same thing on different amps, it gives you a measure of how they perform relative to one another. That's a good thing.

I always like to finish by testing with my ears.

I was in a bit of a rush on Sunday but I should clarify that when I tested the output power of mine the other night I didn't use my variac so the line voltage could well have been below nominal (didn't check that either) and I measured it with a 100Hz sinusoidal signal. It was a quick and dirty test and may not be all that meaningful. I may redo with the variac some other time.

Still working on getting my hands on another one to run the sweeps on as well.

 

[seamonkey]

0.25V rms * 1.41 = 0.35V peak
I'd expect some THD around 5% with generic diodes anaway.

Besides clipping, diodes have capacitance to.
It was probably some designer listening, and tried it with and without diodes and personally liked it better with. The capacitance rolls off some highs that people don't want for bass anyhow. As one pushes the amp the clipping diodes soft clip and let the player know their reaching the upper limit. Ampeg's new line of tube amps use these clipping diodes also. The diodes seem to be part of the Ampeg "Signature sound". It's just not "all tube" as they advertise. Doesn't need to be if people didn't think solid-state was some kind of poison, and just took the amp at what it sounds like.

 

[beans-on-toast]

I was in a bit of a rush on Sunday but I should clarify that when I tested the output power of mine the other night I didn't use my variac so the line voltage could well have been below nominal (didn't check that either) and I measured it with a 100Hz sinusoidal signal. It was a quick and dirty test and may not be all that meaningful. I may redo with the variac some other time.

Still working on getting my hands on another one to run the sweeps on as well.

I'm lucky to have an amp friendly 117VAC power line service. The only problem is, it can vary a bit. It climbs during the day, in the morning it might be 116.5, in the afternoon 117.5. Well within spec but enough to affect readings over a session. I need to check the line voltage when I start, then again during the process. Over a couple of hours, it will drift a bit. It depends on how accurate you want or need to be.

I had one of these ferroresonant units at one time: [Invalid or Expired Link Removed]. It hummed and gave off heat, it had to be isolated in a small room. There's better technology available these days.

 

[timv]

Yeah I check my voltage and will do most tests @ 100hz. I'll usually check at 250 and 500 also but don't see much difference.

 

[cerrem]

THE SVT Output transformer has always remained the same plate load from 69 on up to Early 80's.... I have measured and analyzed them as well as wind them..
They did have slightly lower inductance in some of them due to the E-I laminations were stacked differently in either 1x1 or 2x2 alternate groupings, thus changinging the effective air gap..
The plate load is slighlty different between the 4 Ohm and the 2 Ohm tap...
As for the Power transformer... All the 6146B and 6550A amps made around 700V on the plates give or take...when healthy and set-up properly...regardless of the schematic.. Never seen 660V in any amp during the Blue or Black line era, unless a filter cap was very leaky...
The Screen voltage for 6146B amps were around 230V, while the 6550A amps were around 350V on the screens..
When you convert a 6146B amp to 6550 tubes....the amp will usually put out 275W at best...due to the screen supply being to low...
In those amps typically put a voltage doubler in the screens, this ends up at around 400V DC on the screens, which makes for a punchy loud SVT getting up to 400W output...
If you have strong transconductance 6550 tubes..you should make good power..typically get 360 W before clipping at 400 Hz into 8 Ohms ....
I would not use a VARIAC during measure power output from amp...unless you compensate for the droop the VARIAC will make...so adjust for 120V measured at output of VARIAC when the amp is at full power output..
I am interested in putting the SVT on my HP 3577A Network Anylazer... I know that the low frequency respeonse will change depending on the output level, due to the OT ...

 

[ThisBass]

Besides clipping, diodes have capacitance to.
It was probably some designer listening, and tried it with and without diodes and personally liked it better with. The capacitance rolls off some highs that people don't want for bass anyhow. As one pushes the amp the clipping diodes soft clip and let the player know their reaching the upper limit. Ampeg's new line of tube amps use these clipping diodes also. The diodes seem to be part of the Ampeg "Signature sound". It's just not "all tube" as they advertise. Doesn't need to be if people didn't think solid-state was some kind of poison, and just took the amp at what it sounds like.

Personally I think the diodes did originally protect the power amp banks A/B from glowing red.
At the Rolling Stones tour it sometimes happend that bank A or bank B randomly started to glow red so the engineers had to implement a "protection" on the fly just to fix the issue. The power amp driver stage was redesigned anyway after the tour ended but "very likely" the diodes remained, may be for "additional" protection purpose or may be for sound reason as well?

IME the svt needs little more voltage than stated at the schematics to push 300 watt at the output terminal.
And I think it's lacking in practical relevance to bench mark a MI bass amplifier with narrow-minded THD numbers anyway.
I'm pretty sure the svt would hit pretty good THD numbers if there were no diodes present at the input stage of the power amp but, I see no reason to act like a nit picker on probably the most famous bass guitar amplifier in the world.
(hope nobody feels offended due to some direct words)

The TE V8 has got similar diodes but in this case the diodes can't impact the signal like it it happens with the svt. At the other hand side the TE V8 is known for it's sudden coming saturation.

Other amplifiers (without additional driver stage) hit lots of THD due to take-over THD at high levels but, amplifiers like svt and TE V8 don't do so.
In my opnion MI amplifiers shouldn't be THD bench marked like audio amplifiers.
I think it's more of interest which kind of THD and spectrum MI amplifiers can push at around 50% of rated power.

 

[beans-on-toast]

That clears up that question at least!

Included the refernce to Bill's comment in my post above.

 

[beans-on-toast]

Given the SVT power amp prototype was based on the 811A with a 1500V plate voltage, not a surprise that the first SVT run had another transmitter tube, the 6146B.

 

[ThisBass]

If you have strong transconductance 6550 tubes..you should make good power..typically get 360 W before clipping at 400 Hz into 8 Ohms ....

360 watt before clipping that's a number I can validate.
In my case at 100Hz into 4 Ohm

Besides quality of tubes there is probably more severity to rail voltage and power supply losses and losses due to class A/B design and preamp power demand including heaters that just limits output power to around 360 watt just before clipping.
Not sure the amp did ever any less than 360 watt.

edit,
some tube sellers may mention that tube brand x may bring the benefit of a (significant) plus of watt.
With those oldish led sleads like an svt the accessible output power is restricted by the design itself which means VA of the supply and rail voltage which also regards to the design of the whole amplifier.
Class A/B 6550 amplifiers waste way to much of thermal losses inside the amplifier than any high quality 6550 pentode may lead to believe there was actually some "plus" of output power with brand x of 6550 at the output terminal available.

 

[tubedude]

I'm not too concerned with plate voltage.changes, or screen either. I can adjust bias to keep them from exceeding dissapation . The real issue is the 6.3V on the cathode. I want that as close as possible.

 

[agedhorse]

The real issue is the 6.3V on the cathode. I want that as close as possible.

Why are you so concerned? Even +/- 10% isn't going to be an issue as that is how an amp is specified per international safety standards anyway.

The cathode needs to be hot enough to boil off enough electrons, any more isn't of any practical advantage, though a possible disadvantage might be a little shorter tube life at high line of +10%.

 

[bobyoung53]

All 6550 equipped svt's I've had on my bench, over 10 now, have put out right around 270 watts at the point of any clipping seen on my oscilloscope. I hold line voltage at 120 with my variac so I can compare tests from different days etc...

My 6146 equipped 70 does about 340 or so Before clipping. Just to share my findings.

I have heard that before, that the 6146 SVT's put out more power, I might just try to grab one.

 

[beans-on-toast]

It was reported that there was a problem with the 6146B's ability to dissipate the heat when pushed hard. That is one reason why they switched to the 6550.

 

[timv]

Also talking about how much power it puts out really is just numbers because anything in this power range is definitely loud enough for most situations.

As a quick comparison that I have a lot of experience with. A Mesa/boogie bass 400+ only tests out power wise the way I do it at about 240 watts clean before any waveform clipping/distortion but nobody is ever going to say it isn't as loud as an svt!

It has a lot more to do with how that clipping/distortion happens and what it does after that point, that's one of the points of a tube amp.

 

[beans-on-toast]

I'm not too concerned with plate voltage.changes, or screen either. I can adjust bias to keep them from exceeding dissapation . The real issue is the 6.3V on the cathode. I want that as close as possible.

The cathodes on the SVT power tubes are grounded through a pair of one ohm resistors which are used for setting the bias. I think that you are referring to the 6.3 heater voltage which Andy addressed.

It is important to note that variation in the wall voltage will affect the performance of most amps. It will affect the bias, the plate voltage and power supply voltages in general. The amps are designed to work that way. Some touring bands have power supplies that keep their line voltage stable. Others will rebias the amps at each venue to keep the performance consistent. Then again, not everyone is overly concerned. There are other factors that can affect the sound on any given night.