Why does my amp sound worse than an AC30CC?!

[Gaz]

I've been working on this 30 watt amp for a while now, and I was pretty happy with it, but A/B'd it with a Vox AC30CC (supposed to be pretty close to a JMI Vox), and the Vox was definitely a superior amp - not by a ton, but enough to drive me nuts!

The Vox had more punch, and was more articulate, which really surprised me. I'm not sure if the voltage on the power tubes is just higher, or it's because of the smaller 100R screen resistors, or what?! The Vox also has a little more bottom than my amp, but in a way that's not muddy.

My design has a little more gain on tap, and is more controlled sounding when overdriven, which some may prefer. It sounds closest to the Vox with the "Voice" switch OFF in the middle position, which just means that there is just the 47k resistor to ground off the bottom of the tonestack, hence a lot more mids.

I'd appreciate any ideas of where I can go with the amp, like where I could add a little more bass without muddying it up (preamp or poweramp?), and if maybe switching to the Vox stack could help.

It sucks when a Chinese amp sounds better than your own!

[Cliff Schecht]

What speakers are you using?

Also your phase inverter-->power tube coupling caps could be larger, maybe 0.1 uF or so and your grid stoppers on the EL84's are quite large compared to the typical 1k-1.5k for most amps. Why so large?

[Gaz]

Tried both through the stock ceramic Celestions in the Vox, and with a Vintage 30 loaded oversized Mesa cab. the Vox had the same qualities through both cab, but I liked them both better through the Mesa cab.

[Cliff Schecht]

Edited my post, read above.

[Gaz]

Yeah, Cliff, I thought maybe increasing the coupling caps to the power tubes would be a smart place to add bass.

The grid stops are relatively large to inhibit blocking distortion, but maybe I'm losing a little character there. The amp is still nice and bright. IIRC, 10k is the recommended (schmecommended) on the datasheet. Thanks for the input.

[guitarmike2107]

Yeah it is crap when that happens, I don’t think you can build a decent amp without having some baseline to check them against.

Why have you got such a big plate bypass capacitor on the tone stack driving stage?

Have you got a schematic for the AC30, is the tone stack is cathode follower driven? What’s the power supply like?

Muddiness is often too much low Mids, The Vox tone stack is very scooped You can cut the low mids slighty by increasing your mid cap to 0.047uf , this will allow you to turn up the bass a bit more and help your bass sound a bit clearer and punchy.. that’s probably why Fender used that value in his plate driven tone stacks.

You are comparing two different amps though, handmade and boutique parts don’t always equate to awesome tone… unfortunately

[Fischerman]

The circuit looks different enough to consider it not an AC30 Top Boost (or Rocket) type circuit. AFAIK, the AC30 Top Boost and Rocket have 2 gain stages then a cathode follower driven tone stack and the tone stack is different (1M pots and a 50pF treble cap among other differences).

Maybe play with Duncan's Tone Stack calculator and see if you can tweek that stack to be similar to the Vox stack response (just remember yours is plate driven and the Vox is CF driven). Maybe try increasing the value of the cathode bypass caps, the Vox uses 25uF on the first stage and no cap on the second but you'll probably need a bypass cap on your second stage to make up for the gain you're losing in the stack (your plate driven stack arrangement is more lossy than the CF driven stack, but CFs have no gain).

[Zippy]

OP:

What are you using for iron?

Where did this schematic come from? Why the differences from an AC30 if that is your goal?

[Andy Le Blanc]

Play it out... break it in....
leave it on in the in the shop for couple days...
play it out again and see how it does...

you could use more critical bypass cap values in the power side
those 220uf are huge, didn't vox have one cathode resistor?

look at your power side first, thats where the amp responds.

[Gaz]

Thanks so much for all the responses. The amp design is bits and pieces taken from various EL84 amps. I wasn't trying to build an AC30 clone, of course, and I didn't think I wanted one, but the AC30 I tried definitely had some characteristics that I wanted to have in my own design, which weren't quite there - the punchiness and clarity.

I set up a test rig to A/B the Fender/Marshall stack in the schem with the Vox, and I think it really helped. The Vox is definitely more scooped, which helped the clean tone a lot. The treble control has less high mids, which makes the treble response sound smoother and less brash when turned up - more chime. The Vox stack also has greater bass emphasis, which helps me understand why such a small coupling cap is used in the Top Boost circuit, which is by no means thin sounding. But man, the Vox stack is soooo interactive, but hey if it sounds better, I can get down with it.

BTW, I'm using the Heyboer HTS-8093-1 (like Dynaco 470 with no UL taps) and the Heyboer HTS 9144 with the 250-0-250 taps.

I should also mention that the AC30CC was using a Weber can rectifier deal in place of the tube recto, which is probably bringing up the voltages and eating the tubes...which is probably why is spitting out such good tone!

I'll have to try it against the Vox again to see if it shares any of those qualities it was lacking before. More suggestions are welcome.

[Gaz]

BTW, the 470p across the plate resistor driving the stack is just there to cut hiss, and I can't hear affecting anything negatively.

[Darkbluemurder]

I'd appreciate any ideas of where I can go with the amp, like where I could add a little more bass without muddying it up (preamp or poweramp?), and if maybe switching to the Vox stack could help.

More bass - increase the output coupling caps after the PI to at least 0.047uf, better 0.1uf.

More punch - change the slope resistor to 56k and the treble cap to taste (150pf to 180pf). I would not necessarily go to the full Vox tone stack as it is a bit quirky in that it creates a huge low mid dip when the bass control is turned up full.

To get more sparkle I would change that 470pf snubber cap over the 2nd stages plate resistor to 220pf. That should still work for the hiss reduction.

Cheers Stephan

[Darkbluemurder]

The treble control has less high mids, which makes the treble response sound smoother and less brash when turned up - more chime.

That is due to the 50pf/1MA combination for the treble control.

The Vox stack also has greater bass emphasis, which helps me understand why such a small coupling cap is used in the Top Boost circuit, which is by no means thin sounding.

I believe Vox reduced the coupling cap in order to keep the bass coming from the first stage in check - the cathode bypass cap in the first stage is so large that all frequencies in the guitar passband are fully amplified. Furthermore there is no shelving filter between the coupling cap and the gain control like e.g. the 470k II 470pf filter in the 2203/04 Marshalls which would further reduce bass. Matchless used 1000pf in the Lightning, Bad Cat used 1200pf in the normal channel of the Hot Cat but that is not a Top Boost circuit. But I don't think the bass reduction has to do anything with the tone stack.

Cheers Stephan

[Gaz]

I was just saying that the Vox stack has more bass emphasis (or less treble, I suppose), and small coupling cap after the first stage keeps things balanced. With the other tone stack I was using there was less bass, and I needed a larger cap to get the same amount of bass. Hope that makes sense.

I think I've found a good combo with the Fender stack - 100pf/.1uf/.047uf or 150pf/.1uf/.033uf - that mimics the Vox stack well without the quirks.

[Gaz]

I did some more experimenting today and found that a 47pf/.1uf/.047uf with the Fender pot values got me the closest to the Vox stack. This is with both controls set at noon. I can't mimic other settings of course because of the Vox's unique interactive behavior. I thought the values were worth mentioning because the frequency plots on Duncan's TSC show other values as being closer to the Vox. I guess I can't count on it for that kind of accuracy when comparing two stacks next to one another...