The Hundred-Buck Amp Challenge

I wonder if you can modulate a sharp cutoff tube using the screen?
George will know, from his screen-drive experiments...

Varying the screen grid (bias) voltage definitely changed the gain in my 6JW8 preamp stage, but I didn't try experimenting with feeding it a control signal for compression. I was just trying to find a good DC operating point.

I expect it will take a lot more modulation voltage at G2 compared to G1. Maybe that's why Nachbaur used the scheme he did?

-Gnobuddy
 
I would have thought modulating G1 with DC would cause all sorts of thumping and bumping noises at the anode,

There's a pile of "vibrolux" type circuits that do pretty much the same, but usually by hitting the cathode - e.g. this Supro (EL34 world)

I've also seen compressors (e.g. EF86 based by Bruce Vicknair ) with Vcontrol on G1 and the input on G2

Ken Stone did a Russian "rod" pencil pentode based unit using G2 for Vcontrol.

Using remote cutoff pentodes as VGAs is a technique as old as the hills: if you can think of it, someone's probably tried it.

More than one way to skin a cat.
 
There's a pile of "vibrolux" type circuits that do pretty much the same
I've seen "bias wiggle tremolo" before. In this case, there is at least the advantage that the tremolo waveform is more or less a sinusoidal one, and its frequency is well below the lowest guitar frequency, so you have some chance at high-pass filtering it out via small coupling caps. (Though I have still read a lot of complaints about thumping sounds when tremolo is turned on.)

I would think a compressor would be more of a headache, considering that you might want fairly short attack times - which, by Fourier transform, equate to correspondingly higher frequencies in the control signal.

(This might be a good time to mention that there was a period in my twenties when I failed miserably at multiple attempts to design a well-behaved guitar compressor. The least-bad ones were optical, using an LDR as the gain control element. Since the LDR is just a resistor, and can be fed a purely AC signal, there are no DC changes to "thump"!)

I have a feeling Nachbaur played electric guitar rather differently than many of us. He seems to have enjoyed playing classical guitar pieces by Sor and Carcassi et al, on his electric guitar, using extremely clean tone. Perhaps this type of playing (no sharp pick attack transients) helped cover up some compressor thumping?

More than one way to skin a cat.
Indeed, but the good audio compressor circuits I've run across all have some sort of circuit topology designed to keep the compressor control signal itself from showing up at the same circuit node as the audio output (using balanced bridges, etc). Pretty much the exact opposite of putting the control signal on G1, along with incoming audio!

-Gnobuddy
 
George will know, from his screen-drive experiments

Changing the voltage on the screen grid of a pentode will affect the current through the tube, and thus vary the Gm. This will change the tube's gain without having to disturb the DC voltage on G1. The amount of gain change depends on the tube's characteristics, but I doubt that a very large change can be affected by this method alone. Multiple stages will be needed.

There are a few "dual control" pentodes around that are probably more suited for something like this. They look like a regular pentode, but a second control grid is substituted for G3. There are several of these but the only number that pops into my head right now is the 6AS6. I have a box full of these and have been tinkering with them an a vacuum tube version of a Moog style ladder VCF. These have been successfully used in audio compressors.

Using remote cutoff pentodes as VGAs is a technique as old as the hills

A common sharp cutoff pentode has a control grid wound as a spiral wire of constant pitch. A remote cutoff pentode has the grid wound with a variable pitch, usually tightly wound near each end, slowly changing to a wider spacing in the middle. This way as the voltage on G1 is increased from cutoff, only the center of the tube is being used. The fine pitch part remains cut off. As the grid voltage is brought less negative more of the cathode area is active, and less is cutoff. These tubes were intended for variable gain RF and IF amplifiers in radio and TV sets, but do find themselves in some audio applications, including compressors.

The original Fender Champ 5C1 used a 6SJ7 sharp cutoff pentode for the input tube, and grid bias is developed by the grid leak method. Driving the snot out of a 5C1 with my trusty germanium fuzz box would send the grid voltage negative enough to prove that blocking distortion can indeed happen in an input stage. This was in the 1960's.

I discovered a pseudo cure could be obtained by plugging a 6SK7 remote cutoff pentode into the input tube's socket. I would soon learn that the real cure was to eliminate the grid leak bias by using a cathode resistor and a bypass cap. I still liked the sound of that amp with a 6SK7 when pounded with my DIY fuzz tone.

There are plenty of common tube types that exist in both sharp cutoff and remote cutoff versions. Look at the IF strip in any radio or TV set. The most common pair is the 6EJ7 sharp cutoff, and 6EH7 remote cutoff. Going back in time, the 6AU6 sharp cutoff, 6BA6 remote cutoff......6SJ7, 6SK7.......6J7, 6K7.

Often as variants of existing triodes (e.g. 6DJ8 vs 6ES8 )

Yes, this same technique was applied to triodes, usually for TV tuner duty. The dual triode was usually configured as a cascode amplifier with signal applied to the lower grid and AGC voltage applied to the upper grid, although both could be applied to the lower grid as well. Again these found their way into audio compressors.

The most famous is the Fairchild 670 which now sells for STUPID money. It used the 6386 tube which was originally designed for military aircraft radios. NOS versions of this tube sell for $200 and up.....way up! I have a few well used ones.

You can find the Fairchild schematics online. The key to using any of these variable gain tubes in audio is not only the thump, pump, or other control voltage effects, but the nonlinearity of the tubes themselves. When you build a tube that has a variable gain, it is inherently nonlinear. Some tubes are worse than others.....much worse. The good ones require a rather large change in control voltage to effect a change in gain, thus the audio signal will be a smaller percentage of the total voltage across the tube. The good tubes tend to have a constant radius to their "bent" transfer curve, thus creating only even order harmonics, predominantly 2nd. The key here is to use a totally balanced design to cancel out those harmonics. The Fairchild is balanced and transformer coupled throughout, and multiple stages are used to get the required control range.

The VGA cell is an integral component of any compressor. It can be a simple LDR, or a complex vacuum tube or FET circuit. I will eventually get to tinkering with the VGA, after I conquer the VCF, both for music synth reasons. I have several tubes to test and some 6386's to compare them with. I need to get some reasonably priced magnetics, probably Edcor, to try out some Fairchild type circuits. Either way, it's not going to happen soon.

BTW, the LDR makes for a wicked resistive element in a Moog ladder filter. Weird delayed release though.
 
There are a few "dual control" pentodes around that are probably more suited for something like this. They look like a regular pentode, but a second control grid is substituted for G3. There are several of these but the only number that pops into my head right now is the 6AS6.
I ran across this across this article on radiomuseum.org by Joe Sousa: "Dual Control Pentodes (Dual control Pentodes). I thought I'd mention it here in case others are interested in experimenting with these.

In the article, Joe mentions the following list of pentodes designed for dual control applications:

A list of pentodes designed for dual control​
I leave out the various heater voltage variants that a number of these tubes have.

6AS6 was constructed similarly to a 6AK5 but with dual control. It has high frequency capability and was used in high end communications receivers.

6DT6, 6HZ6, 6GY6, 6GX6 are all dual control single pentodes used mostly in FM detectors of American TV sets.

6F33 is the only dual control European pentode I know of. Jacob Roschy shared this info.

6BU8 two pentodes share a control grid G1 and Screen G2, but have separate suppressor grids G3 and Plates.

PF86 for Transitron oscillator applications

EF50 designed as a conventional RF pentode, but can be biased for dual control.

6BN6, 6KS6 gated beam detectors were specialized dual control beam pentodes with very sharp (high mu) input characteristics for improved limiting action. It was often used as a limiter in FM tuners.

The following tubes are all G2-Compactron types with 12 pins. Each tube includes a dual control pentode for FM detection and an audio power amplifier for service as the audio section in American TV sets.

6AD10A gm1=2.5mS + Audio amp
6AL11 gm1=2.5mS + Audio amp
6BF11 gm1=1mS + Audio amp
6T10 gm1=1mS + Audio amp
12AE10 gm1=1ms + Audio amp
13V10 gm1=1mS + Audio amp
18AJ1 gm1=2.4mS + Audio amp

The next three tubes include a gated beam beam pentode that is similar to the 6BN6.

17AB10 gm1=0.7mS gated beam + Audio amp
6Z10 gm1=0.4mS gated beam + Audio amp
6J10 gm1=0.4mS gated beam + Audio amp

These two tubes were used in American TV sets for sweep synchronization and synchronous color demodulation.

6BA11 gm1=1.8mS twin-plate pentode + triode

6BV11 gm1=3.7mS dual pentode gm3=2m7S
 
how about this one? a 5F1 wannbe:D

110612a-016.jpg


i got the 2 x 12LW6 tubes for a $1 each at rogalski's...
the 12AX7 cost P350 or about $7.90

the irons, a power traffo, a single ended opt, 4k:8ohm, and a choke for about P2500 or about $56....

easily less than a hundred green ones....:D

Wow, Im jealous
 
Two years ago in post #2105 I stated:

The VGA cell is an integral component of any compressor. It can be a simple LDR, or a complex vacuum tube or FET circuit. I will eventually get to tinkering with the VGA, after I conquer the VCF, both for music synth reasons. I have several tubes to test and some 6386's to compare them with. I need to get some reasonably priced magnetics, probably Edcor, to try out some Fairchild type circuits. Either way, it's not going to happen soon.


Those experiments have begun. I wired up a close replica of the Fairchild 670 gain cell with ONE 6386 tube and a pair of $12 Edcor transformers. It works very well at 1KHz with about 60 dB of dynamic range The THD at 10 V RMS output is less than half a percent at 1KHz. It rises at each end of the audio spectrum, but that's not the real problem. The frequency response sucks to put it mildly. Loss at the low end due to transformer saturation was expected. Severe amplitude loss at each end of the spectrum was not.

The -3dB points are 240 Hz and 3.3 KHz....not even good enough for guitar or synthesizer use, for get HiFi audio. WTF?????

Multiple experiments ensued, including digging up an old UTC transformer. It seems that Fairchild used some pricey transformers that were maybe a wee a bit better than my $10 Edcors. Better Edcors are on order.

The UTC worked better, but still falls short......Fairchild uses 4 X 6386 tubes in parallel. Could this be needed? I wired up a second 6386, and yes the frequency response got better. Four in parallel would probably get me in the ball park. OK, so to clone a Fairchild 670 I would need 4 X $120 tubes and probably the same $$$ worth of transformers....for ONE channel. I guess that's why old 670's fetch upwards of $20K!!!!

Some creative engineering will obviously be needed here to make a low buck compressor, but for now I need to focus on the variable gain cell, or VGA. armed with a full characterization of my single 6386 gain cell, I set out to do better.

I put together a list of possible tubes similar to the list in post #2106. Several of the tubes in that list were the same as mine, and I went off is several other directions as well. The dual control pentodes, variable Mu triodes like the 6386, remote and semi remote pentodes, beam deflection tubes, sheet beam tubes...... I'm about half way through the testing of the tubes on my list.

There have been some surprises along the way, but I'm not quite ready to build a Fairchild clone yet. Another brick in the vacuum tube modular wall, maybe!

This thread is about cheap guitar amps....and I will have more of those in the future. To avoid adding some VGA stuff to this thread, All future VGA and synthesizer related stuff will go here....until there is the need for a compressor thread.

Modular Synth build
 
Here's my contribution to the thread. I designed it about 4 years ago and had some PCBs made up about six months later. Life got in the way and I ended up finishing the first one about a year ago. Unfortunately I gave it away, so I'm going to have to track down some pictures. I built a second one as a hi-fi amp with a few modifications to lower gain, and it works great. A third is on the way (about 95% done, just waiting on a guitar speaker, cabinet, and some jacks/misc).

It's very straightforward. Two 6GV8 tubes (or ECL85 or 6F5P) are the basis of the amp. Power comes from a 12v switching psu. This powers the heaters, power LED, a fan, and a switching 230v inverter. The inverter goes into a doubler (not really, since it's a square wave, it's more of a 1.4x boost) and CLC filter. The PCB has a voltage regulator on board that gives 250v so the triode section of the 6GV8 doesn't exceed it's maximums and I can use smaller/cheaper 250v rated capacitors instead of 400v or 630v.

Schematic is attached.
6GV8_guitar.gif

The design is simple. A standard grounded cathode gain stage feeds into a source follower. The triode is biased by a LED for maximum gain. The source follower feeds a James style tone stack. That feeds into a volume control (R12/R13). After that is another grounded cathode gain stage biased by a LED that feeds into a concertina Source-ode-dyne. Capacitor C10 restores the full AC swing that the R23/R24 voltage divider halves, but is optional if you need more gain. The output tubes are standard grounded cathode and use a garter bias with capacitor bypass. The output transformer is an Antek toroid (10VA, 7v).

The first one was built with the outputs operating in pentode mode and has huge gain and distortion. Even with a cheap 6" speaker, it's way too loud, so I added a L-pad to the speaker to get all the distortion but at practice volume. From now on I'm building them triode mode, which should give about 5 watts out. The values for the tone stack in the schematic are optimized for the "hi-fi" amp I recently built and would need to be adjusted for guitar use.
 
Last edited:
Also a picture of the 12v to 230v PSU...
Nice project, and a handsome PCB! :)

Can you share any more info on the PSU? Using an SMPS instead of old-school 60-Hz iron used to only be an option for large corporations (Peavey, etc), but now it's starting to be viable for one-off DIY projects.

On another thread, someone shared this: High Voltage DC-DC 150V-420V Converter NIXIE & Tube HV Power Supply 1.25-12V adj 699904388163 | eBay

If the voltage ratings are to be believed, there's enough there to power any number of small "classic" guitar amps (push-pull 6V6, etc.) What doesn't seem to be too clear is how many watts this PSU can actually handle. The ad says "High voltage output current: 50-100mA adjust", which isn't terribly illuminating. Is that 50 mA at 420V (21 W)? Or is it only 100 mA at 150V (15 W)? Or something else entirely?


-Gnobuddy
 
My text seems to disappear. Let's try this. A cheaper version of the inverter can be found searching 'DC-DC 10-32V to 45-390V High Voltage Boost Converter Step-up Booster Module', won't take my link.


I have an amp project with the same architecture, of a dual triode with a Mosfet follower and splitter. I probably will do cathode biased on the output tubes though.
 
Last edited:
The converter linked by Gnobuddy is a single ended boost converter based on an old TL494 chip. It's efficiency goes down as the difference between the input and output voltages go up. It is also a relatively low frequency device. The data sheet examples run at 10 and 20 KHz. I wouldn't expect more than a few mA st 300 volts or so before parts start to get hot.

The converter I see in pictures from Printer 2's description only has one power device, so it's probably single ended as well. It does use a transformer for step up, so it may do better in the higher voltage ranges.

I haven't tried either.

Many of the cheap inverters for powering small electronics in the car turn 12 volts DC into a high DC voltage, then chop it up into a quasi sine wave with a mosfet H bridge. I got some cheap ones made for European 240 volt 100 watt electronics on Amazon and tapped into the DC voltage inside. It's about 270 volts and good for 50+ watts of power continuously. As I found, they blow up if accidentally shorted. This was a few years ago and the devices I bought are no longer there, but I'm sure there are new ones.
 
If you search ebay for "DC-AC Converter 12V to 110V 220V AC 150W " you will find one like the one I'm using. It has a ST SG3525a PWM chip in it. It runs fixed frequency/duty cycle as far as I can tell. It outputs a square wave at 230v and I think 20KHz. Pretty easy to filter. You can even use a voltage doubler to get about 350v. The inverter transformer is probably good for at least 50w at 20KHz, 150w seems...optimistic. You can probably run it at 19v instead of 12v if you change the main capacitor to get 50% higher output voltage. They're cheap enough to run one for each channel.
 
An Ebay search with those terms reveals 4 different things. A square blue board, a square green board, and a long skinny rectangular board that looks like yours. It also brings up a red and black 12 volt to 220 volt car inverter.

The red and black car inverter looks like the one I got from Amazon. It took me 10 minutes to blow it up. I still have a silver Philips branded 70 watt inverter that I got from Amazon at the same time. It did a good job powering my laptop in the car, so I didn't rip it apart.

I suspect that the other three are the same design. I have the square blue version. It does have a ST branded SG3525 chip in a socket! It is a push pull design and capable of lighting up a 40 watt bulb without smoking. It produces an AC voltage so you must use your own rectifier and filter. Do not use ordinary 1N4007's, they start smoking in a minute with no load. UF4007's work fine. The board does not have any provisions for feedback, so the regulation sucks, but it's probably OK for a guitar amp. DC feedback could probably be added from the rectified output, but I never got that far.

Mine is still sitting a box trying to get my attention, but my ADHD powered brain wandered off in a different direction again. It's still alive though, I didn't manage to blow it up.
 

Attachments

  • P3320727_x.jpg
    P3320727_x.jpg
    621.3 KB · Views: 381
Last edited:
The square blue board and the the skinny blue board are probably the same thing just in different form factor, right down to the same components and positioning of parts around the socketed SG3525. The square one might be a bit better - there are heatsinks on your board and mounting holes in the corners.

I use SiC diodes without problem. I tried running feedback and it didn't work, but it was probably because I found a cold solder joint or two on my feedback circuit after I took it apart. For a class-A or guitar amp it should work fine running all-out.
 
I got a chance tonight to take pictures of the first prototype I built. The primary difference with this one is that it runs at 230v, the preamp sections are regulated using a voltage divider/RC instead of zeners, and the output tubes are run as pentodes.

The cab is built with 3/4" pine and uses a piece of L shaped aluminum for the controls. The speaker is a Parts Express buyout 6.5" car speaker.
The back has a 80mm fan (Enermax Marathon, silent), a L-pad to reduce volume while keeping amp distortion, the 12v input jack, and a 1/4" jack wired to the second secondary winding to power either an external cab or as an output to a PA system (it's a little hot but works well with the input turned down).

The guitarist I gave it to says he really likes the sag and distortion at higher volumes, probably due to the pentode operation. He also likes the high gain of the amp so he doesn't need the gain of any of his effects boxes.

P.S. Sorry about the rotation of the photos, the forum software apparently doesn't like portrait orientation and rotates them.
 

Attachments

  • combo_back1.jpg
    combo_back1.jpg
    654.8 KB · Views: 360
  • combo_front1.jpg
    combo_front1.jpg
    665.7 KB · Views: 345
Last edited:
I guess this fits the bill. I had a bunch of leftover stuff: poplar for the cabinet, tolex from an earlier project, a bucket of line transformers, screws from an earlier project.

Tubes were given to me but it uses 955 and 6gw8. I pulled capacitors, knobs, resistors, pots from my parts bin.

Speakers were pulled from an old tv I was getting rid of.

Power transformer came from a hamfest for $2

The panel was made by me using some 2" aluminum bar stock I had.

I purchased some cabinet corners and handle.

Out of pocket cost was $21. If I'm being fair, prices for other parts I had are....

955 tube: $7
6gw8: $20
Line transformer: $5
Pots, caps, resistors: $15
Wood: $12
Tolex: $7

Adding the $21 I spent for this project leaves it well below $100.

It sounds great too. Makes 2 watts. Goes to 11.


PXL_20210108_212445215.jpg PXL_20210108_212456007.jpg PXL_20210109_170528981.jpg