first meeting with X-over distortion - Ampeg SVT-3 Pro

[pjd3]

Hey everyone,

While I"ve successfully built two all tube guitar amps I have never contended with an amp with a solid state output section as come in the Ampeg SVT-3 Pro bass amp.

My bass player started having issues with this amp, one of them being notable distortion no matter what settings he used on his amp, which he had come to know pretty well. (Although, he did just get an active bass, but, claims he used the -15db switch which from my analysis does drop a sine wave from very clipped to nearly pure-ish sine wave).

Up till this point I've never witnessed or directly contended with crossover distortion on my new scope until this amp. Looking at 12ax7 plate signals I was looking at very pure amplified sinewaves for the preamp section but across the 8 ohm resistor dummy load at 1Khz, the crossover notches were visually apparent. For the heck of it, I cautiously made adjustment to a trim pot on a board close to the rows of power mosfets - looks like its a push-pull mosfet output, 4 and 4. Right away I was able to bring the crossover notch into line with the otherwise non-interrupted purity of the sinewave.

This post is certainly meant to be of a general nature since I have not yet developed instincts regarding the visual nature of crossover distortion on a scope and how that translates into the quality and quantity of distortion.

Until I develop a more definitive assessment of this, are there any assertions you have learned and developed relating what you see and what you hear? In other words, how bad does it have to look before it is making a notable audible difference.
I know the rubber will truly meet the road when I haul a worthy speaker up here to hear the difference in real time but, was really interested in your experience with this particular facet of solid state amp outputting.

Thanks, and appreciate your patience with my green-ness in this area. I'm a boob toob head so far. (not complaining, Slucky's 6V6 plexi is giving me alot of mileage and compliments over lots of gigging).

Best,
Phil Donovan

[Stevem]

First off I am not sure exactly what you mean here.

Crossover distortion and output stage clipping are two very different things.
If that amp IS producing crossover distortion then it’s coming from the first part of the SS section of the driver / PI/output stage.

When the amp starts to go into clipping how much has the voltage on the power supply rails dropped down to?

How much wattage is it outputting into that 8 ohm load when it starts to clip and also what exactly is the resistance of that load resistor when tested on a meter ?

It might be time to replace some or all of the 4 5600uf@80 volt filters.

I like to also check the resistance of my load resistor(s) when they get up to 130 degrees f as shown on my thermal gun.
A .6 ohm difference in a 8 ohm load resistor is 3.5 watts on a 50 watt amp.
A .4 ohm difference in a 8 ohm load resistor testing out a 250 watt amp is near 9 watts.

By the way, this amp uses a 12Au7 as the last tube stage that drives the SS output section.

This is chosen for the 12AU7s current output, if you have a 12AX7 in that socket the amp will never work as intended and it will have greatly reduced output wattage.

[LOUDthud]

It's been my observation that amps using IRF Fets tend to have more Xover distortion at higher temperatures. Check that you have current through all eight of those 0.47 Ohm 5 Watt resistors.

A word of warning: I'm looking at a version of the schematic dated 6/25/98 (rev 15). The way the bias pot AP1 is wired, it will fry those Fets or blow a fuse if the pot is turned to the position that shorts B-E of Q2. That is a really poor design decision to save a 1 cent resistor.

Schematic Link: https://el34world.com/charts/Schematics ... _3_pro.pdf

[pjd3]

Thanks for your fast responses here!

The clipping was more about me purposely causing clipping at the input section in order to see how well the -15db input pad lowered the signal going into the first stage of V1 which in fact, it did take a notably clipped signal down to a normal looking sine wave at the output. That was just to confirm that my bass player wasn't overloading the input with a hot, hot signal that would cause distortion right at the start. He claimed he was certain he used the pad with his active bass. I'm pretty much on board with the difference between crossover distortion at the output section and clipping in the preamp stages. At this point, I can't be sure what my bass player was hearing for distortion, how much, or where it was actually coming from, just covering all the tracks.

However, I am not versed in observing crossover distortion on a scope, and translating that into any distortion that would be heard audibly. Something I wasn't very comfortable with was that after I trimmed out the crossover distortion that I was seeing at the output, I noticed that there was barely any rotation left on the trimpot, that it was nearly already very close to the end of its travel. Usually on a built in adjustment, it is designed to be on average somewhere in the middle of the pots travel, not 90-95% at one end or the other. But, this is what it is. I was just really curious over how much crossover distortion it takes to be a distortion that notably corrupts a signal and makes it arguable unusable. My bass player generally gets a very deep and clear, robust tone so, I don't imagine it takes much of anything to bother him. So, I'll need to get in touch with how this power section crossover factor plays into the sound. That's for me to figure out ultimately but was surely interested to hear about it from anyone else that has a developed perspective on it. For all I know, the amount of crossover distortion I saw on my scope wouldn't amount to much of anything. But Im sure my bass player was hearing something of a distortion that made his amp unusable for him. I'm trying to see if I can source the problem and if I'm fortunate, offer a remedy.

Thanks,
Phil

[pjd3]

I figured it out.

When I tried the amp, I used the left most speaker output, It sounded fine, the sine wave out looked like the sinewave in, unless you hit the input hard, which my bass player never does.

The other night, I plugged the actual speaker into the other speaker output jack and all sonic hell broke loose. First thing was to physically exercise the jack and there it was - an intermittent contact of Jack to pcb. Took out the power circuit board and found a few fractured solder joints, reflowed them and all was beautiful. I had to laugh that I was using the good jack to get "good results" and my bass player was using the bad jack for his "bad results".

Since I have never played with crossover adjustment, or perhaps output balance adjustment, I adjusted the trimmer while listening and looking at the scope, and was surprised to hear a notable harmonic pop out even from a little bit of crossover distortion. So, that was adjusted just until the harmonic disappeared, and I was left with a nice pure tone and pure looking sine wave.

Thanks for all your input and comments on this amp problem. I'm happy that I can give the amp back to my bass player in good repaired and adjusted form. I know he's looking forward to getting it back. He's got a few good ones but this is his favorite gig amp!

Best
PJD3

[wpaulvogel]

I’ve never actually heard crossover distortion in a tube amp but I’ve observed it many times while watching the output waveform. On a tube amp it’s usually subtle and right at the point of clipping and never gets to be more than a slight turn in the center point of the sine wave until clipping is heavy and dominant. Solid state devices are more hard at their limits than the softer nature of tubes. The amp you’re working with is to my knowledge a class AB design and will eventually present crossover distortion at the class B threshold but as you have mentioned the bias threshold is adjustable. Definitely be careful to follow the manufacturer’s instructions on setting the bias of the output. I do know that with high voltage tube output sections that crossover notch sometimes cannot be completely eliminated sometimes as full clean output is reached without setting the bias dissipation past the recommended 70% of maximum dissipation. This is aggravating but due to reliability issues it’s a necessary condition to extend the life of the components (tubes). The SVT amp is a high power output design and it may just be a juggling act to get the expected clean output. Lowering the mains voltage can allow the output section to remain crossover distortion free but will lower output power before clipping as a result. Finding the sweet spot with the volume control is easier than tweaking voltages.

[R.G.]

I was just really curious over how much crossover distortion it takes to be a distortion that notably corrupts a signal and makes it arguable unusable. My bass player generally gets a very deep and clear, robust tone so, I don't imagine it takes much of anything to bother him. So, I'll need to get in touch with how this power section crossover factor plays into the sound.

Crossover is funny. For all distortions, the more abrupt the change into/out of the distorted section, the more noticeable and unpleasant the distortion is to most "average listeners". Crossover is a bit unique as the human ear seems to be more offended by kinks and discontinuities near the zero crossing than to flat spots on bigger signals.
The very best advice here is to take advantage of having an amp that you can adjust into and out of visible crossover with a pot and hang a speaker on that amp. Put in a test signal and listen while you adjust that pot. You'll have the start of an understanding of how much visible crossover correlates to audible distortion.

As a guide, the more horizontal the crossover region looks, the more prominent the distortion. Tube crossovers tend to make the crossover regions slanted, which is a lot less objectionable. SS output stages (which I think you're describing) can have very flat little regions between the + and - signal polarities. This sounds ... horrible, to me at least. In SS amps, the biasing for minimizing crossover is much trickier than for P-P tube output stages. This is one of the things that was probably behind the 1970s "transistors are harsh and ugly sounding" mantra that persists to this day. MOSFET stages are gentler than bipolar in general, although they have other faults.

So - the sharper the corners and the flatter the crossover region, the worse it sounds.
To me at least.

[pjd3]

Thanks for stopping by RG.

While I didn't listen to an actual bass guitar/speaker while adjusting the crossover trimpot, the sine wave sure did highlight the visual tuning in and out of the crossover distortion. Looking back I wish I did play through the amp at varying degrees of the visual X-over on the scope.

But it was an eye opener and certainly educational to hear and see what and how much that crossover takes a sine wave out of its "purity". This has inspired me to take the guitar tube amps I've built and gig with, and see what its distortion waveforms looks like at various settings, pre vs power section and so forth.

Thanks you everyone, my bass player has his amp back and says it sounds as good as it ever has.

Best,
Phil