Toast Rack SE

[Steve Conner]

Hi all

I've started modding the "Toaster" amp that I built back in 1999, and I heard that some people here like Ray Ivers had shown interest in the design, so I thought I'd post what I'm doing.

The only real annoyances I had with the amp in the 7 years I used it were that it sounded a little dark and muddy at high volume, could only drive a 16 ohm speaker, that the "OMG" channel was quieter than the "Dirty" channel, and that if you turned it up really loud it would switch itself off, due to the geeky protection circuit I installed that shut the B+ off in case of screen overdissipation or loss of bias. This last thing was what motivated me to mod it: it never bothered me when I used it for practice and recording, but now I'm gigging with it, I don't want it crapping out on me, or worse, someone else.

So this weekend I pulled it apart and started poking around in there. I tried modding the screen protection circuit so it would shut off for a second and come back on, but I just couldn't get it to work. I'd used a Rube Goldberg arrangement of transistors, lost the schematic, and forgotten how it worked. I also noticed that there were violent RF oscillations happening when the protection tripped. In the end I "fixed" all that by ripping the protection board out. After all, Hendrix and Noel Redding got by fine without screen protection circuits ;-)

I designed the regulator under the assumption that the protection would shut it off and prevent sustained overloads, so I'm not sure I trust it in the absence of protection. I'm tempted to rip out the regulator too, or relegate it to running the screen voltage and preamp tubes, but then I guess it wouldn't be the same amp any more :-(

Now that it couldn't turn itself off and spoil my fun, I tried fitting KT88s and loading the 16 ohm output with 8 ohms. The regulator circuit ran out of headroom, but even so I measured over 100W out, bonus!

BTW: I found a mistake in the regulator schematic I posted on my site, that would stop the foldback current limiting from working.

[Ray Ivers]

Steve,

Does it sound a lot better without the screen-protection circuit? I've heard so many stories over the years about protection circuits doing strange things, far afield from their original intended duties. I think the main problem must be that during normal high-power amp operation, voltages and currents routinely run into the 'fault-condition' range briefly - and installing a time-delay to prevent false trips results in protection really not much different from a fast-blow fuse.

Your original regulator circuit worked well for me in simulation, but I tweaked it a little (removed R13/R14, replaced Q1 with an ECG103A Ge transistor, made R11 adjustable, etc.) just to fine-tune for my needs; it's still 90+% your circuit, though. The simulator indicates a very smooth transition from rock-solid regulation to gradual current-limiting, which sounds like a plan to me.

So now you've got a two-tube, 100W KT88 output section - same as me. I've always liked two-tube PA's, KT88's, and high power - seems like a natural to combine 'em all.

Ray

[Steve Conner]

Hi Ray,

The screen protection circuit couldn't modulate the screen voltage in any way. All it could do was pull the plug on the amp completely (by shorting across the TL431 with a transistor) if it measured excess screen current for too long. Sagging of the regulated supply would also trigger it due to the half-assed way I chose to measure screen current.

So I don't see the sound being any different now it's gone, except that I can now hear different sounds made by cranking the power amp to places it's never been before. I'm hoping to take it to a rehearsal space and try it with a full stack of 4x12"s

I originally put the protection circuit in because I discovered that the main consequence of cranking an amp with the speaker disconnected is massive screen overcurrent that makes the screen wires melt and "ka-blam". I tuned the protection circuit using dummy loads, but the impedance of speakers varies widely with frequency, and driving the speaker at its bass resonance showed a high impedance load that would draw heavy screen current and trip the protection.

I guess it's not too hard to just check the speakers are hooked up and working before going up to 11...

[Ray Ivers]

Hi Steve,

The screen protection circuit couldn't modulate the screen voltage in any way.

Oh, OK. I figured it might have been "micro-cycling" on signal peaks or some such, but evidently not.

...I can now hear different sounds made by cranking the power amp to places it's never been before.

I'll bet they're pretty cool places, too.

...I discovered that the main consequence of cranking an amp with the speaker disconnected is massive screen overcurrent that makes the screen wires melt and "ka-blam"

Yeah - with the plates slamming against insanely hi/lo stops and just a low-value resistor from the screens to B+ to limit current a bit, I don't doubt the screens would take an extra-heavy beating, above and beyond the normal heavy beating they take from us guitar folks anyway.

I guess it's not too hard to just check the speakers are hooked up and working before going up to 11...

Famous last words but I agree, start out at low volume and make sure everything's OK before cranking up, and 99 times out of 100 everything will be fine.

I worked up a speaker-protection circuit a while back, maybe I mentioned it already - two 43V, 5W back-to-back Zeners in series with a 24 ohm, 3W resistor, all across the speaker output. It works about the same as a 200 ohm resistor across the output with no load connected, but steals no power or 'speaker bounce' during normal operation. It looks like average screen current jumps from about 25mA to 35mA or so (in simulation) with no speaker load.

Ray

[Steve Conner]

Hey, I like your speaker protection idea Ray! I can build something like that up and screw it to the heatsink inside the chassis no problem. Then I can fit a thermal cutout to the heatsink too.

For my next round of tests I hooked the amp up to a big 14 ohm wirewound rheostat as a dummy load, and hooked a hi-fi speaker between the wiper and ground so I could hear an attenuated version of the output. I put a switchable 15 ohm resistor in parallel with the output so I could change the loading between "16" and "8".

I put it on the clean channel and played slap bass through it with the master volume cranked up high enough to drive the power tubes into heavy distortion. I noticed the cathode current meter on the front panel (300mA fsd) would slam off the end stop in time to the beat. The regulator didn't seem to sag much when driving 8 ohms, I could see a good 40V peak output on a scope, and I didn't hear any unpleasant clipping, just good dirty distortion, so I think I'll leave it in place. I think that was a pretty harsh test, since the tweeter in a hi-fi speaker always seems to show up nasty things.

I preferred the sound of the overdriven power tubes when driving 16 ohms, though. The bass sounded fatter and heavier. I know the OT in it saturates somewhat on loud bass notes, and I think it saturates more with 16 ohms and adds extra farty goodness.

I guess I should mention I'm using the KT88s (a pair of GECs that I pulled from scrap lab equipment at the uni) with the screens connected to the same supply as the plates through a 1k resistor.

[Steve Conner]

P.S. yay for KT88s! I also increased the volume on the OMG channel, added a footswitch that "turns" both gain knobs up to max (by shorting across them with H11F1 optoisolators) and changed the old "Treble" knob so it adjusts the Q of the second EQ band instead. It ought to be quite a naughty beast for either bass or guitar now. I just need to take it to a jam for testing.

The original Toaster was described at http://scopeboy.com/toaster.html

[Steve Conner]

The test was successful! I took it to a practice session of our new band, which seemed to end up being a fight club for ex-metalheads who play in indie pop bands and need to let off some aggression now and again. The rehearsal rooms had Marshall Valvestate 4x12" cabs so I plugged it into one of those (ick)

To cut a long story short, it worked great and never gave any trouble in spite of having the master volume more or less dimed into an 8 ohm load the whole time. I think I almost felt one of my pant legs flapping, and my ears were ringing for 2 days afterwards, though I blame the drummer for that

I did have some problems: the power transformer was too hot to touch afterwards, and though it seems louder than it was with EL34s, it's obviously still no BAGA

[Steve Conner]

The next part of the plan is a nice faceplate made of laser cut and engraved plexiglass with LED lighting. I got to know some guys who run a laser cutter so I thought why not!

I used a black background and amber LEDs, so it's kind of the inverse of a plexi fronted Marshall.

I had some trouble getting the nuts on that hold the pots in place, due to the plexi being pretty thick. (I asked the laser guy for the thinnest he had, and he gave me 3mm.) But I think most of them should just about hang on.

[Steve Conner]

Update: I measured the power wrongly (using a wirewound load resistor with too much inductance) The true output into 8 ohms is just under 70 watts.