5W Soldano SLO
Soldano SLO needs no special introduction. When it came out in the late 80s, it redefined high gain hot-rodded Marshall sound and served as an inspiration to countless modern amps (looking at you Randall Smith!). It adds a cold biased clipping gain stage that has that refined, smooth, yet punchy distortion character. I always wanted to have one, but didn’t really need all that power and couldn’t afford to buy one, so why not build a scaled down version that has all that awesomeness in a beadroom-friendly package?
I started with the official SLOCLONE schematic and analyzed the circuit, trying to figure out what would need to be changed to accomplish the goal. I figured that output power of around 5W would be plenty and looked around for tubes that could provide that kind of output in push-pull class AB operation with fixed bias, just like the original SLO. In cases like this, many people opt for a simpler solution that uses one 6V6 or EL84 in single ended class A configuration, or use a double triode in push/pull configuration, but I wanted to keep as much of the character of the original SLO, so I really wanted to preserve the topology of the power amplifier and use two power pentodes in push-pull class AB operation..
The smallest current production power pentodes I know of are EL84 but they deliver around 20W which is still too much for me, so I dug deeper into the world of NOS tubes and found some 7-pin small tubes from the 50s and 60s that were designed for much less power. On paper, EL91 sounded like a winner, so I bought a few from eBay and started playing the datasheet. EL91 are small 7-pin power pentodes, probably the smallest ones in the EL family and can deliver 1.7W in single ended and around 5W in push/pull configuration.
– The Power Supply
The power amplifier section had to be modified to use EL91 tubes instead of the 6L6 used by the original. Looking at the EL91 tube datasheet, the little El91 can take only 250V on the plates which is half of the voltage used in the original SLO to feed the big tubes. 500V would fry them within seconds for sure. I could reduce the voltage of the whole amplifier to 250V and run the preamp tubes at around 230-240V, but I didn’t want to risk changing the character of the preamp, so I went with a more complicated option that involves cloning the power supply, so we have two independent power supply branches that have separate secondaries on the power transformer and share only the ground reference. One branch delivers ~350V to the preamp and the other branch delivers 250V to the poweramp. Both branches are very similar to the choke-less power supply used in AX84 projects. The negative voltage bias supply is also taken from AX84 projects.
For the power transformers I’m using a custom toroidal transformer, wound by a local company Trafomatic. They can wind even a single custom transformer with any given number of primaries/secondaries. All that at a very good price (much cheaper than EI style Hammond transformer with similar power). For this particular transformer, they wound it according to these specs (voltages are under load): 220VAC primary, 300-0-300V @ 40mA, 200-0-200V @ 100mA and 3.15-0-3.15V @ 3A secondaries.
One more thing I did differently than the original SLO is to use the same secondary for powering the heaters and to power the switching LDRs.
– Output Stage Changes
The small EL91 tubes have less headroom than the big tubes, so I was concerned about overdriving the output stage too heavily. To compensate for smaller tubes, I put voltage dividers instead the two plate resistors (R36: 82K and R37: 100K) of the phase-splitter stage, so I can tap into the output of the phase-splitter and take only a portion of the voltage swing it can produce. Normally, we take the (maximum) output from the plate directly, but we can tap the plate resistor and take potentially less than that. The 82K resistor is replaced with 27K+56K and the 100K resistor is replaced with 33K+68K where the smaller of the two resistor is facing the plate. That way, we are effectively reducing the output of the phase-splitter by about a third.
Another way to achieve very similar results would be to use a lesser gain tube for the phase-splitter, like 12AT7, 12AY7 or even 12AU7, but I was anal about it and wanted to keep the same tube, bias it the same way as the original, just reduce the output.
– Other Mods
- Replaced 12AX7/ECC83 preamp tubes with Russian military 6N2P-EV tubes. These are about the same as 12AX7 with a tad less gain and slightly different heater wiring that operates only in 6.3V mode. Pin 9 is used as internal shield between triodes that should help reduce noise and it should be grounded.
- Added Depth control in the negative feedback loop (NFB). It’s possible to get this as an optional feature when buying a SLO from Soldano, so I thought – why not. It gives extra control on the bass making it easier to nail the tone that’s fat, but not boomy. In the SLOCLONE schematic it’s marked as VR9.
- Reduce fx loop “Send” level to be more pedal friendly. Some say that SLO has a relatively hot FX loop that works well with professional effects that can take line level signal without clipping, but could drive regular stompboxes into clipping. To reduce the level, I added a 1K (R50) resistor in parallel with existing 2.2K resistor on the cathode resistor of the last triode before the “Send” jack. That reduces the output level. To compensate for the lower signal level we send out of the amp, the triode in the recovery gain stage after the “Receive” jack cathode resistor is bypassed with a 1uF capacitor (C13). It boosts all frequencies above ~70Hz for 6db.
- Switchable treble bleed capacitor on the OD channel gain pot (a.k.a Warren Haynes mod) to change the high end response when gain pot is not turned all the way up. In the SLOCLONE schematic, the switch is S3 and it takes C6 in and out of the circuit.
- A single 8ohm output jack installed. No impedance switch.
- Added a switch to choose between 2ohm and 8ohm OT taps for the NFB loop. The 2ohm tap sends less signal to the NFB and it makes amp sound a bit more aggressive.
For the output transformer I got a Hammond 125C which is capable of taking up to 8W of power, so it’s more than capable of handling the output of EL91. The secondary is wired to lugs 2 and 4 which give about 22.5K primary impedance, according to the Hammond 125-series datasheet. Two EL91 in push-pull like to “see” 20K which is close enough.
With only ~5W of power, there’s much less current, so B+ line needs less capacitance to filter out AC ripple. I got JJ 100uF+100uF 500V can capacitor, so that each power supply uses one 100uF section as first filter cap. The rest of the filter caps are mounted on the main board, so there’s no need for a dedicated power supply board.
I should note that photos show ELON and Tung electrolytic capacitors that I bought for the project but they started causing problems pretty much since day one and I finally realized that they are fake Chinese capacitors that have old capacitors pulled from who knows what old device with completely wrong specs. I took them out as soon as I found out and replaced them with F+T caps. It served as a lesson never to buy cheap electrolytic capacitors again. Since then I have switched to using brand name capacitors bought from reliable shops.
Click on a thumbnail to play the video on YouTube.
Click on a photo to see more details