The VOX AC30
What makes the sound of a Vox AC30 so unique? Is it even all that unique? Let’s investigate and arrive at our own verdict. A sample question and answer session will be followed a more detailed explanation.
|Q: Does the ‘fact’ that the AC30 is a Class A amplifier make it sound so unique?||A: No. The amplifier is not an excellent example of Class A design. When played at maximum volumes, the amplifier is far more Class AB than Class A.|
|Q: What about the tube rectifier, the GZ34? I’ll bet that contributes to the sound of an AC30!||A: Not if you think about it for one minute. I’ll explain later.|
|Q: Well, do the EL84 output tubes give the AC30 ‘that sound’?||A: Not entirely. Many small Gibson amplifiers used EL84/6BQ5 output tubes in a Class A configuration, and yet sound nothing like a Vox amplifier.|
Q: Do the Celestion speakers at least contribute to that unique sound of my AC30?
A: Yes, they certainly do. Those low powered speakers driven by an amplifier having preamp and output distortion yet no negative feedback make a tasty recipe!
Q: The absence of negative feedback makes that much of a difference?
A: Hell, Yes! It’s now time to study the Vox AC30!
The Vox AC30 has always been considered a ‘pure’ Class A design, and is the first example offered when someone inquires about what a Class A amplifier ‘sounds like’. Unfortunately it is not a complete Class A amplifier, and the confusion stems from the fact that it is using cathode biasing on the output tubes. By definition, any Class A amplifier draws’ maximum’ plate current at all times, and it should not deviate between idle (quiescent) and full output power. Another way of saying this is that the ‘maximum’ plate current is flowing at all times. The Class AB amplifier will have its plate current increase approximately 25% between idle and full output, while each tube takes turns ‘not conducting’.
OK, the secrets of ‘Tone’ have gotta be in there somewhere, don’t they?
The same plate current is said to be flowing between ‘appreciably more than half and less then the entire electrical cycle’.1 Studying a schematic and taking voltage and current measurements when you have the opportunity will help you understand things a little better. Lastly, by looking at a tube manual for the 6BQ5, we see that a Class AB amplifier needs a common 130-ohm resistor to achieve an output of 17 watts (4% THD), with the other circuit parameters taken into account, including a quiescent plate current of 72mA and maximum plate current of 92mA (both tubes). Do the math and check out that the AC30 has a common 50-ohm resistor biasing four EL84’s. Even the schematic states that the voltage measurement at the cathodes should be 10VDC at quiescent and 12.5VDC at full output. That works out to approximately 200mA at idle through four EL84’s and 250mA at full output. By their own (not very) silent admission, the AC30 is not even close to a true Class A amplifier.
If there is no ‘sag’ in a Class A amplifier, why do you need a tube rectifier?
Moving along to the power supply, many people feel that the tube rectifier contributes to the tone of an AC30. For an interesting conversation starter, ask your local tube guru the following question. “If the AC30 is a Class A amplifier, could the tube rectifier offer any power supply sag?” Should the amount of sag available in any amplifier be significant, a GZ34 has very little to offer, as compared to a 5U4/GZ31. Sure, our AC30 is more like a Class AB amplifier, but our trusty tube manual lists the common ‘sag’ voltage of a 5U4 at approximately 50VDC, compared to a 5AR4 with an approximate 17VDC voltage drop during high current demands. There just simply isn’t much sag to the power supply in an AC30. The amplifier should be drawing maximum current at all times in any event, and therefore no increase in current demands at high playing volumes or transient peaks can force the power supply to ‘dip into its reserves’. Does a Fender Champ have ‘sag’, with its’ 5Y3 and single-ended 6V6 output? The tube rectifier is still handy, nonetheless, to ensure a ‘soft start’ to the powering up of the amplifier. Seeing as the amplifier has no standby switch, the delayed warm up time of the indirectly heated cathode on the GZ34 helps avoid certain potential problems at the output stage, real or imagined (read on). Many people enquire about adding a standby switch to any amplifier that does not have one, including the AC30. The truth is that any amplifier with a GZ34/5AR4 rectifier does not need a stand by switch. Lastly, with regards to the power supply (and a slight return to the output stage). Most Vox AC30 amplifiers have at least some hum, which is actually normal. Part of the reason is the output transformer is not a Class A transformer, meaning it is perfectly balanced and rated for a 100% duty-cycle. Any imbalance in the transformer primary and we get unbalanced DC through the transformer core. Plus we get more 2nd order harmonic distortion (and therefore more THD), and the common mode (hum, noise) rejection benefits are reduced. Add to this the 16uF filter capacitors found in AC30 amplifiers, and you end up with hum in AC30 amplifiers. Most players learn to live with it for the nostalgic vibe and to maintain the vintage value of their amplifier. However, you can increase the amount of filtering and reduce the hum considerably, although not entirely. Be careful not to increase the amount of filtering to a level that will stress the rectifier tube. Remember, at turn-on the capacitors are essentially a ‘dead short’ until they charge.
Hints and ‘Tricks; on Reducing AC30 Hum
Here are a few ‘tricks of the trade’ aimed at helping you reduce the ‘hum’ in your AC30 to a much lower level. Each amplifier is different, so treat each ‘patient’ as an individual. That said, these ‘tricks’ do work well separately or in tandem.
- Determine if the hum is a filament winding imbalance. If so, install two 100-ohm/1-watt resistors from each filament tap to ground. The most convenient spot is at the AC mains indicator.
- A ‘shield’ directly below the ‘Cut’ control often helps. The coupling capacitors from the phase-inverter are right here, and are susceptible to stray (induced) magnetic fields. I have seen a ‘bracket’ that had been fabricated to fit over the ‘Cut’ control (thus eliminating the need to drill holes and ‘mount’ the shield); this is simple, and does work.
- Try different output tubes. Here, matching (or ‘unmatching’ just enough) can make the difference. This isn’t a great ‘cure’, since your tubes will age ‘differently’, and you’ll eventually have to go through the whole process again, but it sure helps in the recording studio. On a live stage, most players seem to ‘put up with’ a little bit of hum.
- If the amplifier has a grounding problem, you can try to ‘find and fix’ the offending ground connection. A ‘quick fix’ until you do is to have the ‘Normal’ volume control at about the ‘1 O’clock’ position.
The output tubes have a moderately important role in determining the overall sound of an AC30 (or any other amplifier, for that matter). The EL84/6BQ5 has always been considered a 9-pin version of the 6V6. The truth is that the 6V6 is a Beam-Power Tetrode tube while the EL84 is a Power Pentode tube. Pentodes have a poorer damping factor compared to Tetrodes in similar circuits, which means we can get more output tube distortion. Of course, with a bit of feedback in the amplifier, we’d be hard pressed to differentiate between the two sonically, but the fact remains that the pentode output stage is run without any negative feedback, making clean headroom take a back seat. The only ‘drawback’ to a real Pentode versus a Beam Power Tetrode is the sad fact that Pentode tubes are more susceptible to a potentially damaging Screen current rise at maximum signal than Power Tetrodes. This is because the beam forming plates in a Beam Power Tetrode drastically cut down on secondary emission, which will help keep the Screen Grid current increase at full throttle to a minimum. 6BQ5’s are at least designed a little better than an EL34 (another Power Pentode), in that the 6BQ5 has its’ plate connection well separated from the heater connections. When you blow up an EL34, it is always between pin #2 (heater) and pin #3 (plate) that you develop a carbon track build-up. On the 6BQ5, the heater connections are pin #4 and pin #5. Pin #6 has an internal connection, and pin #7 is the plate. A well though-out tube. This partially explains why you see some amplifiers with the same 6BQ5’s in them years later, while some Marshall amplifiers can’t make anything other than a vintage EL34 last more than a week.
All guitar amplifiers played at 50% of full power are operating in Class A.
Having no negative feedback in an amplifier is a unique sound, and the AC30 is certainly no exception. As the amplifier is only rated for 30 watts of output power, most players tend to turn up the preamplifier volume. Without any feedback to keep a leash on the preamplifier, there is copious amounts of distortion present at the output. The smaller output transformer will also add a little distortion of its own.
The very efficient Celestion ‘Bulldog’ speakers have a great impact on the tone of every AC30. With the lower power handling capability comes wonderful things like a smaller plate gap, higher efficiency, and thinner cone paper. These speakers break up quite nicely for us at moderate volumes, and we tend to need ‘at least’ moderate volumes to use a 30-watt amplifier on stage in the modern playing environment. After all of this ‘studying’, we now come to a very interesting note regarding the AC30; it does not follow verbatim the ubiquitous Bassman circuit! It is a long and arduous task to understand the circuit completely, and rationalize why this circuit was chosen, so we won’t bother. We can, however, study a portion or two and look at how these circuit variances may contribute to a sound unlike other amplifiers we are familiar with. Keep in mind there was a circuit ‘evolution’ to the AC30; more so than just whether or not the amplifier had a tube rectifier. Below is a partial schematic of the ‘Bright’ and ‘Normal’ inputs common to many variations.
At first glance, the input section seems like most other amplifiers you inspect on your bench. There is not a lot of gain to the ‘Normal’ channel. If you trace it through (highlighted in ‘Blue’), there is only one gain stage, a Volume control, and then the phase inverter! This ‘channel’ does not seem to be the most popular to use, and I can see why. Even with a 220K plate-load resistor, the gain is very modest. It is still enough to drive EL84’s (somewhat), but not enough to drive the output stage into clipping as much as the ‘modern’ guitar player would appreciate. Following the ‘Bright’ channel (seen above highlighted in ‘Orange’) shows more of what you’d expect. The Volume control appears before the tone control circuit, although you can ‘move it over’ for a little variation in tone and drive characteristics. I like to have the Volume control fed from the Treble control, and I always make sure to have some type of grid resistance in place of the original Volume control after this modification. If I do this modification, I tie a 1Meg resistor between the grid and ground. If the 100pF capacitor makes the ‘Bright’ channel seem a little too ‘bright’ you are free to alter that value. I like using a 250pF. The ‘Bright’ channel has a Bassman/Marshall ‘smell’ to it, but there is the odd difference. The tone circuit isn’t what you may be used to, but it does have a good ‘feel’ to it. The phase inverter also takes its cue from a vintage Bassman phase inverter (up to a point). However, if you look closely, you may note that the ‘tail’ resistor is 47K. This ‘neuters’ the overall gain and drive voltage, but it does help balance the two outputs a little better. You may recall Fender usually used a pairing of 82K/100K for plate-load resistors in their popular phase inverter, but the ‘tail’ resistor could be as low as 6800-ohms. This yields a lot more gain, but the balance to the phase inverter is compromised. I never worry about that detail, as no phase inverter is perfectly balanced, nor is the output transformer. Even though it may seem Vox is ‘wasting’ a lot gain, keep in mind EL84’s do not require a lot of ‘drive’. All we need is 28.3VAC grid-to-grid to achieve all the output power possible. You are free to alter the circuit, but you’ll just be overdriving the output tubes more, and taking away from the characteristic Vox tone. However, if you feel you ‘need’ to constantly use distortion pedals with your AC30 to get a tone you are happy with, I have given you a few ideas to explore. The real ‘secret’ to the Vox AC30 tone is concealed in the phase inverter. While the partial schematic above only hinted at it, I purposely kept it out of full view until now. Below is just the phase inverter tube, and the associated circuitry.
If you study the schematic closely, you may note the following. Rather than have the second half of the ECC83 at AC ground (as seen in every Fender clone), the input is actually used. It is fed from the Tremolo channel, and the coupling features a very clever ‘notch’ filter. Voiced in the midrange of the guitar frequencies, this gives the AC30 its characteristic ‘honk’. Some guitar players like to use a wah-wah pedal as a notch filter, and the principal is similar. It isn’t a circuit you see every day, and helps explain why no other amplifier has an AC30 tone. Keep in mind it is only one portion of the recipe, albeit an important one often overlooked. It has been brought to my attention, by more than one reader, that the ‘notch’ is also (or is it primarily?) used to take care of any low frequency ‘thump’ courtesy of the Tremolo oscillator. Is this so? I cannot say for certain, but here is a thought. The typical coupling network of 1Meg and .022uF will have a cutoff of about 7Hz. Fair enough. However, changing that to 470K and .01uF will now have the low-end cut off at 33Hz; the same as 1Meg and .0047uF would. So why go through all the trouble of a multi-section filter? Of course your filter has a steeper roll-off, but do you need it for any special purpose? The typical Tremolo oscillator will operate between 1Hz and 10Hz. Why is the notch centered around 33Hz and not 10Hz? I may never know. Also, it has been duly pointed out that the grid of the aforementioned ECC83 is at AC ground via the .047uF capacitor and the Volume control. As long as the Volume control is set to ‘0’, the grid is at AC ground. Fair enough. Also mentioned is that having the Volume on ‘0’ removes the notch filter from the circuit, and thus it cannot have any effect on the ‘Normal’ or ‘Brite’ channels. This is absolutely true, but I mentioned it solely on the observation that no one I know uses the ‘Normal’ or ‘Brite’ channel. Lastly, I’d like to address an issue regarding Vox AC30 amplifiers, as seen in print. In the excellent tome ‘The Ultimate Tone’ writer Kevin O’Connor states “All Vox products – but especially the AC30 – are a serviceman’s nightmare…. wired with short, brittle leads…. One of the weakest points in the AC30 assembly is the lead dress for the output transformer.” Mr. O’Connor then goes on to explain that the one extended transformer lead is still ‘too short’ and sits tightly against the chassis. This is combined with the fact that Vox uses ‘only’ 600-volt rated heat shrink to extend the transformer leads. It seems that peak signal can drive the voltage on the leads to 800-volts and this will cause the insulation to become brittle, and in conjunction with time and exposure to high ambient temperatures, the insulation eventually wears through. These accusations may be well and true for Mr. O’Conner. However, I can state that while the very first AC30 I worked on was a slightly confusing puzzle, from the second one on it wasn’t that big of a deal. Even the first AC30 I worked on seemed less formidable after an hour or so. I wasn’t impressed on the design of the chassis and speaker cabinet, but overall everything worked out just fine. And I have never seen an AC30 with ‘worn out’ insulation on the transformer leads. In fact, the only problem I have seen (other than routine servicing and attempting to reduce ‘ghost notes’ and/or hum) was the single instance of shorted diodes! It blew a hole right through the ‘board’. The results were messy, and it took an extra effort to restore the amplifier to playing condition, but that was one single instance. I spoke with other experienced service technicians, and no one complained about how awful AC30’s are to service. I would never let anyone’s opinions interfere with what guitar/amplifier I bought during the last 30 years, nor should you. If the AC30 has the tone you like, buy one. Don’t be afraid of someone else’s misgivings. This reminds me of the story of when I was a teenager, and went shopping for a ‘new’ amplifier. Marshall was still the new kid on the block, and the guitar magazine photos of Led Zeppelin and Deep Purple looked interesting. I asked if the store carried Marshall amplifiers, but the salesman steered me instead towards a Fender (which he carried). “Don’t buy an imported amplifier, kid.” The salesman lamented. “Can you imagine the wait you’ll have for parts when you need to get the thing fixed?” So I went home with a Silverface Bandmaster Reverb head and 2X12″ cabinet. Not quite the same, and every picture of Jimi Hendrix I saw still had these ‘imported’ amplifiers. It would take me years to get over that fear, and buy my first Marshall. Your homework assignment for this ‘Lesson’ is to study a Vox AC30 schematic, and compare it to a Gibson GA-79RVT. They aren’t close to being similar, but look at the power supply and the output section. Both use four 6BQ5’s in a cathode-biased output scheme. Keep in mind the Gibson amplifier was designed to be a stereophonic amplifier (a separate ‘Left’ and ‘Right’ channel), but can be played in either ‘mono’ or ‘stereo’. We just have to think a little to achieve that goal.
- RCA Receiving Tube Manual, RC-30, page 25.